diff --git a/.gitattributes b/.gitattributes
index 1ef325f1b111266a6b26e0196871bd78baa8c2f3..45d6d87109900a19ab930a9c7690a1eff70a6175 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -57,3 +57,13 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+checkpoint-1096/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-137/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-274/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-411/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-548/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-685/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-822/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-84/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-959/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
diff --git a/README.md b/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ca517555e82393af6ad676793bdd73ff18edd99c
--- /dev/null
+++ b/README.md
@@ -0,0 +1,61 @@
+---
+library_name: peft
+license: other
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+model-index:
+- name: dpo
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# dpo
+
+This model is a fine-tuned version of [/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/](https://huggingface.co//raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/) on the 60k dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 9e-06
+- train_batch_size: 2
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- total_train_batch_size: 16
+- total_eval_batch_size: 64
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine_with_min_lr
+- lr_scheduler_warmup_steps: 10
+- num_epochs: 3.0
+
+### Training results
+
+
+
+### Framework versions
+
+- PEFT 0.12.0
+- Transformers 4.47.1
+- Pytorch 2.5.1+cu124
+- Datasets 3.1.0
+- Tokenizers 0.21.0
\ No newline at end of file
diff --git a/adapter_config.json b/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..154797ebe6197329495f8f587101cda9b9d46ce8
--- /dev/null
+++ b/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "o_proj",
+    "k_proj",
+    "up_proj",
+    "gate_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/adapter_model.safetensors b/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..ec1793e4177220cb6fc3396a023e5d5827d31b36
--- /dev/null
+++ b/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55162de64ad006819ed48c4133cb0d41cd73c8588ff99e0437eeb5cd44b911e7
+size 828527688
diff --git a/all_results.json b/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..f79cbd52b02b920694c001d8856154bc13f2f095
--- /dev/null
+++ b/all_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 3.0,
+    "total_flos": 216736934658048.0,
+    "train_loss": 0.47654047750291373,
+    "train_runtime": 2544.5527,
+    "train_samples_per_second": 0.519,
+    "train_steps_per_second": 0.033
+}
\ No newline at end of file
diff --git a/checkpoint-1096/README.md b/checkpoint-1096/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-1096/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-1096/adapter_config.json b/checkpoint-1096/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-1096/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-1096/adapter_model.safetensors b/checkpoint-1096/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..8515d41bb50e4fc9904079c7f56a2c2fab224bcf
--- /dev/null
+++ b/checkpoint-1096/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20fbc6f817582b0f3fdbffb2f67fa458c05ab67b75fd66caa2f92b47061233f5
+size 207244392
diff --git a/checkpoint-1096/latest b/checkpoint-1096/latest
new file mode 100644
index 0000000000000000000000000000000000000000..23f5d8c383f7b2a1bf8ad6c0618e5f14c8d36497
--- /dev/null
+++ b/checkpoint-1096/latest
@@ -0,0 +1 @@
+global_step1095
\ No newline at end of file
diff --git a/checkpoint-1096/rng_state_0.pth b/checkpoint-1096/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b346349ce12dd5a17d4b91ed2a5722bb52550950
--- /dev/null
+++ b/checkpoint-1096/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
+size 15984
diff --git a/checkpoint-1096/rng_state_1.pth b/checkpoint-1096/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..68f3c6994456cb8d0592a5375d99503c8924b1c4
--- /dev/null
+++ b/checkpoint-1096/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73
+size 15984
diff --git a/checkpoint-1096/rng_state_2.pth b/checkpoint-1096/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..be044f6ceeed587d30e80c2f72d5aa19fdc9947b
--- /dev/null
+++ b/checkpoint-1096/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b
+size 15984
diff --git a/checkpoint-1096/rng_state_3.pth b/checkpoint-1096/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..fc825249656a9b858782542bd3f4386250f1dfe0
--- /dev/null
+++ b/checkpoint-1096/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc
+size 15984
diff --git a/checkpoint-1096/rng_state_4.pth b/checkpoint-1096/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d30f52a44be563c152ae09db6ae934da6da0d3ed
--- /dev/null
+++ b/checkpoint-1096/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972
+size 15984
diff --git a/checkpoint-1096/rng_state_5.pth b/checkpoint-1096/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..c8715d27ab23ae545d58039cf949cc44ecc1da5e
--- /dev/null
+++ b/checkpoint-1096/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991
+size 15984
diff --git a/checkpoint-1096/rng_state_6.pth b/checkpoint-1096/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1ed791b6ef76eadf0b0c55a5733411771e2ae027
--- /dev/null
+++ b/checkpoint-1096/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa
+size 15984
diff --git a/checkpoint-1096/rng_state_7.pth b/checkpoint-1096/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..800c3bbbc5edf7db01a8316069d439c5fb8d8c30
--- /dev/null
+++ b/checkpoint-1096/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773
+size 15984
diff --git a/checkpoint-1096/special_tokens_map.json b/checkpoint-1096/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-1096/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-1096/tokenizer.json b/checkpoint-1096/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-1096/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-1096/tokenizer_config.json b/checkpoint-1096/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-1096/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-1096/trainer_state.json b/checkpoint-1096/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a701074d34e61ea9c94fc535c64938423c21eed0
--- /dev/null
+++ b/checkpoint-1096/trainer_state.json
@@ -0,0 +1,16473 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9900090826521344,
+  "eval_steps": 500,
+  "global_step": 1096,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    },
+    {
+      "epoch": 0.49954586739327883,
+      "grad_norm": 2.1769776344299316,
+      "learning_rate": 9.65745789630079e-06,
+      "logits/chosen": 0.11112834513187408,
+      "logits/rejected": 0.10816515237092972,
+      "logps/chosen": -83.5405044555664,
+      "logps/rejected": -83.06329345703125,
+      "loss": 3.0939,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.142024040222168,
+      "rewards/margins": 0.1820124089717865,
+      "rewards/rejected": -2.3240363597869873,
+      "step": 275
+    },
+    {
+      "epoch": 0.5013623978201635,
+      "grad_norm": 1.7329221963882446,
+      "learning_rate": 9.654994540260396e-06,
+      "logits/chosen": 0.0653618574142456,
+      "logits/rejected": 0.08004368096590042,
+      "logps/chosen": -80.83209228515625,
+      "logps/rejected": -82.97142028808594,
+      "loss": 2.7161,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.851841688156128,
+      "rewards/margins": 0.2948168218135834,
+      "rewards/rejected": -2.146658420562744,
+      "step": 276
+    },
+    {
+      "epoch": 0.5031789282470481,
+      "grad_norm": 1.592657208442688,
+      "learning_rate": 9.65252271146268e-06,
+      "logits/chosen": 0.09880789369344711,
+      "logits/rejected": 0.14229761064052582,
+      "logps/chosen": -67.1727294921875,
+      "logps/rejected": -75.03417205810547,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8714643716812134,
+      "rewards/margins": 0.37937116622924805,
+      "rewards/rejected": -2.250835657119751,
+      "step": 277
+    },
+    {
+      "epoch": 0.5049954586739328,
+      "grad_norm": 1.6487712860107422,
+      "learning_rate": 9.650042414948133e-06,
+      "logits/chosen": 0.13465353846549988,
+      "logits/rejected": 0.12865689396858215,
+      "logps/chosen": -76.4417724609375,
+      "logps/rejected": -78.6947021484375,
+      "loss": 2.6085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.875291347503662,
+      "rewards/margins": 0.35175251960754395,
+      "rewards/rejected": -2.227043867111206,
+      "step": 278
+    },
+    {
+      "epoch": 0.5068119891008175,
+      "grad_norm": 1.6523009538650513,
+      "learning_rate": 9.64755365577451e-06,
+      "logits/chosen": 0.04238567873835564,
+      "logits/rejected": 0.07994347810745239,
+      "logps/chosen": -76.92097473144531,
+      "logps/rejected": -83.20886993408203,
+      "loss": 2.3156,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9759610891342163,
+      "rewards/margins": 0.5398927330970764,
+      "rewards/rejected": -2.5158536434173584,
+      "step": 279
+    },
+    {
+      "epoch": 0.508628519527702,
+      "grad_norm": 1.7999261617660522,
+      "learning_rate": 9.645056439016827e-06,
+      "logits/chosen": 0.07349395751953125,
+      "logits/rejected": 0.07518415153026581,
+      "logps/chosen": -77.56079864501953,
+      "logps/rejected": -84.97645568847656,
+      "loss": 2.8085,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9408115148544312,
+      "rewards/margins": 0.3470652997493744,
+      "rewards/rejected": -2.287877082824707,
+      "step": 280
+    },
+    {
+      "epoch": 0.5104450499545867,
+      "grad_norm": 1.5684200525283813,
+      "learning_rate": 9.642550769767342e-06,
+      "logits/chosen": 0.16188879311084747,
+      "logits/rejected": 0.12772323191165924,
+      "logps/chosen": -89.29315185546875,
+      "logps/rejected": -94.35065460205078,
+      "loss": 2.2314,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8958841562271118,
+      "rewards/margins": 0.5102108716964722,
+      "rewards/rejected": -2.406095027923584,
+      "step": 281
+    },
+    {
+      "epoch": 0.5122615803814714,
+      "grad_norm": 1.7878178358078003,
+      "learning_rate": 9.640036653135548e-06,
+      "logits/chosen": 0.060573749244213104,
+      "logits/rejected": 0.13457715511322021,
+      "logps/chosen": -68.9404525756836,
+      "logps/rejected": -74.77693176269531,
+      "loss": 2.4359,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.889530897140503,
+      "rewards/margins": 0.45911547541618347,
+      "rewards/rejected": -2.348646402359009,
+      "step": 282
+    },
+    {
+      "epoch": 0.5140781108083561,
+      "grad_norm": 1.4741288423538208,
+      "learning_rate": 9.637514094248172e-06,
+      "logits/chosen": 0.10433132946491241,
+      "logits/rejected": 0.1439315229654312,
+      "logps/chosen": -74.89447021484375,
+      "logps/rejected": -86.58551788330078,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9090807437896729,
+      "rewards/margins": 0.6733560562133789,
+      "rewards/rejected": -2.5824368000030518,
+      "step": 283
+    },
+    {
+      "epoch": 0.5158946412352406,
+      "grad_norm": 1.2864018678665161,
+      "learning_rate": 9.634983098249146e-06,
+      "logits/chosen": 0.10891089588403702,
+      "logits/rejected": 0.11755162477493286,
+      "logps/chosen": -66.90985107421875,
+      "logps/rejected": -71.8125228881836,
+      "loss": 2.1538,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.7779250144958496,
+      "rewards/margins": 0.5625147223472595,
+      "rewards/rejected": -2.340439796447754,
+      "step": 284
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 1.465747594833374,
+      "learning_rate": 9.632443670299616e-06,
+      "logits/chosen": 0.08224496245384216,
+      "logits/rejected": 0.12130744755268097,
+      "logps/chosen": -75.4281997680664,
+      "logps/rejected": -85.0781021118164,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7217226028442383,
+      "rewards/margins": 0.4750482439994812,
+      "rewards/rejected": -2.196770668029785,
+      "step": 285
+    },
+    {
+      "epoch": 0.51952770208901,
+      "grad_norm": 1.492859959602356,
+      "learning_rate": 9.629895815577915e-06,
+      "logits/chosen": 0.06619664281606674,
+      "logits/rejected": 0.13152630627155304,
+      "logps/chosen": -96.65383911132812,
+      "logps/rejected": -111.93521881103516,
+      "loss": 2.2831,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7611618041992188,
+      "rewards/margins": 0.5886337757110596,
+      "rewards/rejected": -2.349795341491699,
+      "step": 286
+    },
+    {
+      "epoch": 0.5213442325158947,
+      "grad_norm": 1.5534065961837769,
+      "learning_rate": 9.627339539279564e-06,
+      "logits/chosen": 0.06637927144765854,
+      "logits/rejected": 0.09107412397861481,
+      "logps/chosen": -71.92534637451172,
+      "logps/rejected": -82.98391723632812,
+      "loss": 2.5101,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7867075204849243,
+      "rewards/margins": 0.39493298530578613,
+      "rewards/rejected": -2.181640625,
+      "step": 287
+    },
+    {
+      "epoch": 0.5231607629427792,
+      "grad_norm": 1.6778221130371094,
+      "learning_rate": 9.624774846617254e-06,
+      "logits/chosen": 0.14700329303741455,
+      "logits/rejected": 0.12778782844543457,
+      "logps/chosen": -65.2364730834961,
+      "logps/rejected": -74.39017486572266,
+      "loss": 2.5236,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6942007541656494,
+      "rewards/margins": 0.35334131121635437,
+      "rewards/rejected": -2.047542095184326,
+      "step": 288
+    },
+    {
+      "epoch": 0.5249772933696639,
+      "grad_norm": 1.462215542793274,
+      "learning_rate": 9.622201742820839e-06,
+      "logits/chosen": 0.11467991769313812,
+      "logits/rejected": 0.1028795838356018,
+      "logps/chosen": -67.85939025878906,
+      "logps/rejected": -74.3462142944336,
+      "loss": 2.3891,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7194863557815552,
+      "rewards/margins": 0.3846975564956665,
+      "rewards/rejected": -2.1041836738586426,
+      "step": 289
+    },
+    {
+      "epoch": 0.5267938237965486,
+      "grad_norm": 1.7591735124588013,
+      "learning_rate": 9.619620233137328e-06,
+      "logits/chosen": 0.08407986164093018,
+      "logits/rejected": 0.12201236933469772,
+      "logps/chosen": -82.57340240478516,
+      "logps/rejected": -86.94192504882812,
+      "loss": 2.4569,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8536537885665894,
+      "rewards/margins": 0.38806283473968506,
+      "rewards/rejected": -2.2417166233062744,
+      "step": 290
+    },
+    {
+      "epoch": 0.5286103542234333,
+      "grad_norm": 1.5584774017333984,
+      "learning_rate": 9.617030322830868e-06,
+      "logits/chosen": 0.1126309484243393,
+      "logits/rejected": 0.08907752484083176,
+      "logps/chosen": -73.36901092529297,
+      "logps/rejected": -76.65506744384766,
+      "loss": 2.4021,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8091729879379272,
+      "rewards/margins": 0.3603411316871643,
+      "rewards/rejected": -2.1695141792297363,
+      "step": 291
+    },
+    {
+      "epoch": 0.5304268846503178,
+      "grad_norm": 1.487938642501831,
+      "learning_rate": 9.614432017182736e-06,
+      "logits/chosen": 0.09620118141174316,
+      "logits/rejected": 0.08514149487018585,
+      "logps/chosen": -88.66265106201172,
+      "logps/rejected": -92.26091766357422,
+      "loss": 2.3636,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.5980674028396606,
+      "rewards/margins": 0.4888036549091339,
+      "rewards/rejected": -2.0868711471557617,
+      "step": 292
+    },
+    {
+      "epoch": 0.5322434150772025,
+      "grad_norm": 1.445408582687378,
+      "learning_rate": 9.611825321491331e-06,
+      "logits/chosen": 0.10014252364635468,
+      "logits/rejected": 0.12857215106487274,
+      "logps/chosen": -82.46649932861328,
+      "logps/rejected": -92.84205627441406,
+      "loss": 2.343,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4620139598846436,
+      "rewards/margins": 0.5616316199302673,
+      "rewards/rejected": -2.0236456394195557,
+      "step": 293
+    },
+    {
+      "epoch": 0.5340599455040872,
+      "grad_norm": 1.262193202972412,
+      "learning_rate": 9.609210241072158e-06,
+      "logits/chosen": 0.09868282079696655,
+      "logits/rejected": 0.16074486076831818,
+      "logps/chosen": -67.0407485961914,
+      "logps/rejected": -82.36201477050781,
+      "loss": 2.1163,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5362765789031982,
+      "rewards/margins": 0.7091963887214661,
+      "rewards/rejected": -2.2454731464385986,
+      "step": 294
+    },
+    {
+      "epoch": 0.5358764759309719,
+      "grad_norm": 1.601891279220581,
+      "learning_rate": 9.606586781257822e-06,
+      "logits/chosen": 0.1236480325460434,
+      "logits/rejected": 0.12187935411930084,
+      "logps/chosen": -78.2726058959961,
+      "logps/rejected": -84.49217987060547,
+      "loss": 2.5075,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.779215931892395,
+      "rewards/margins": 0.38987410068511963,
+      "rewards/rejected": -2.1690900325775146,
+      "step": 295
+    },
+    {
+      "epoch": 0.5376930063578564,
+      "grad_norm": 1.4211548566818237,
+      "learning_rate": 9.603954947398016e-06,
+      "logits/chosen": 0.18375667929649353,
+      "logits/rejected": 0.21945703029632568,
+      "logps/chosen": -70.29829406738281,
+      "logps/rejected": -78.42284393310547,
+      "loss": 2.1959,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.7448753118515015,
+      "rewards/margins": 0.5106840133666992,
+      "rewards/rejected": -2.2555594444274902,
+      "step": 296
+    },
+    {
+      "epoch": 0.5395095367847411,
+      "grad_norm": 1.458617925643921,
+      "learning_rate": 9.601314744859504e-06,
+      "logits/chosen": 0.07360847294330597,
+      "logits/rejected": 0.1596718430519104,
+      "logps/chosen": -72.28546142578125,
+      "logps/rejected": -87.13935089111328,
+      "loss": 2.3538,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8007893562316895,
+      "rewards/margins": 0.5433144569396973,
+      "rewards/rejected": -2.3441038131713867,
+      "step": 297
+    },
+    {
+      "epoch": 0.5413260672116258,
+      "grad_norm": 1.6526339054107666,
+      "learning_rate": 9.598666179026123e-06,
+      "logits/chosen": 0.12013350427150726,
+      "logits/rejected": 0.10237178206443787,
+      "logps/chosen": -83.06564331054688,
+      "logps/rejected": -85.60771942138672,
+      "loss": 2.4531,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.911988615989685,
+      "rewards/margins": 0.35885563492774963,
+      "rewards/rejected": -2.2708444595336914,
+      "step": 298
+    },
+    {
+      "epoch": 0.5431425976385105,
+      "grad_norm": 1.9808340072631836,
+      "learning_rate": 9.596009255298755e-06,
+      "logits/chosen": 0.062342576682567596,
+      "logits/rejected": 0.09053834527730942,
+      "logps/chosen": -96.33489990234375,
+      "logps/rejected": -93.43024444580078,
+      "loss": 2.8549,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.845227599143982,
+      "rewards/margins": 0.22291362285614014,
+      "rewards/rejected": -2.068141222000122,
+      "step": 299
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 1.6276236772537231,
+      "learning_rate": 9.593343979095334e-06,
+      "logits/chosen": 0.2073422521352768,
+      "logits/rejected": 0.13107003271579742,
+      "logps/chosen": -72.45758056640625,
+      "logps/rejected": -77.42770385742188,
+      "loss": 2.4644,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9407453536987305,
+      "rewards/margins": 0.3134301006793976,
+      "rewards/rejected": -2.2541751861572266,
+      "step": 300
+    },
+    {
+      "epoch": 0.5467756584922797,
+      "grad_norm": 1.5001753568649292,
+      "learning_rate": 9.590670355850819e-06,
+      "logits/chosen": 0.15097060799598694,
+      "logits/rejected": 0.16254279017448425,
+      "logps/chosen": -70.44524383544922,
+      "logps/rejected": -70.64158630371094,
+      "loss": 2.4159,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.6921097040176392,
+      "rewards/margins": 0.3336741626262665,
+      "rewards/rejected": -2.0257837772369385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5485921889191644,
+      "grad_norm": 1.8360233306884766,
+      "learning_rate": 9.587988391017198e-06,
+      "logits/chosen": 0.14593760669231415,
+      "logits/rejected": 0.12311654537916183,
+      "logps/chosen": -78.32576751708984,
+      "logps/rejected": -88.25840759277344,
+      "loss": 2.5826,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.6591644287109375,
+      "rewards/margins": 0.5071319341659546,
+      "rewards/rejected": -2.1662964820861816,
+      "step": 302
+    },
+    {
+      "epoch": 0.5504087193460491,
+      "grad_norm": 1.4822838306427002,
+      "learning_rate": 9.585298090063459e-06,
+      "logits/chosen": 0.20818498730659485,
+      "logits/rejected": 0.16436657309532166,
+      "logps/chosen": -67.59429931640625,
+      "logps/rejected": -77.14763641357422,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6798239946365356,
+      "rewards/margins": 0.5954271554946899,
+      "rewards/rejected": -2.2752511501312256,
+      "step": 303
+    },
+    {
+      "epoch": 0.5522252497729337,
+      "grad_norm": 1.6118305921554565,
+      "learning_rate": 9.582599458475598e-06,
+      "logits/chosen": 0.0827561467885971,
+      "logits/rejected": 0.09151773154735565,
+      "logps/chosen": -73.28964233398438,
+      "logps/rejected": -76.90730285644531,
+      "loss": 2.3792,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.8056440353393555,
+      "rewards/margins": 0.39551618695259094,
+      "rewards/rejected": -2.201160192489624,
+      "step": 304
+    },
+    {
+      "epoch": 0.5540417801998183,
+      "grad_norm": 1.4609856605529785,
+      "learning_rate": 9.579892501756593e-06,
+      "logits/chosen": 0.05332394689321518,
+      "logits/rejected": 0.1590057611465454,
+      "logps/chosen": -77.51653289794922,
+      "logps/rejected": -95.51261138916016,
+      "loss": 2.2105,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.740310549736023,
+      "rewards/margins": 0.6771562695503235,
+      "rewards/rejected": -2.4174671173095703,
+      "step": 305
+    },
+    {
+      "epoch": 0.555858310626703,
+      "grad_norm": 1.7181960344314575,
+      "learning_rate": 9.5771772254264e-06,
+      "logits/chosen": 0.05865276977419853,
+      "logits/rejected": 0.03500773757696152,
+      "logps/chosen": -83.72650146484375,
+      "logps/rejected": -93.69804382324219,
+      "loss": 2.1687,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6025928258895874,
+      "rewards/margins": 0.7339029312133789,
+      "rewards/rejected": -2.3364956378936768,
+      "step": 306
+    },
+    {
+      "epoch": 0.5576748410535877,
+      "grad_norm": 1.497698187828064,
+      "learning_rate": 9.57445363502194e-06,
+      "logits/chosen": 0.1225174218416214,
+      "logits/rejected": 0.09580346196889877,
+      "logps/chosen": -76.27323150634766,
+      "logps/rejected": -76.90716552734375,
+      "loss": 2.2116,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.715336799621582,
+      "rewards/margins": 0.4648086428642273,
+      "rewards/rejected": -2.180145263671875,
+      "step": 307
+    },
+    {
+      "epoch": 0.5594913714804723,
+      "grad_norm": 1.3957064151763916,
+      "learning_rate": 9.571721736097089e-06,
+      "logits/chosen": 0.07596514374017715,
+      "logits/rejected": 0.14832191169261932,
+      "logps/chosen": -84.65274810791016,
+      "logps/rejected": -105.76422119140625,
+      "loss": 1.9942,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.6643587350845337,
+      "rewards/margins": 0.7884883284568787,
+      "rewards/rejected": -2.4528470039367676,
+      "step": 308
+    },
+    {
+      "epoch": 0.5613079019073569,
+      "grad_norm": 1.6824349164962769,
+      "learning_rate": 9.568981534222664e-06,
+      "logits/chosen": 0.049865882843732834,
+      "logits/rejected": 0.023348212242126465,
+      "logps/chosen": -80.3697509765625,
+      "logps/rejected": -86.07074737548828,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.009049654006958,
+      "rewards/margins": 0.5470253825187683,
+      "rewards/rejected": -2.556075096130371,
+      "step": 309
+    },
+    {
+      "epoch": 0.5631244323342416,
+      "grad_norm": 1.8752086162567139,
+      "learning_rate": 9.566233034986413e-06,
+      "logits/chosen": 0.07250591367483139,
+      "logits/rejected": 0.12809628248214722,
+      "logps/chosen": -74.84794616699219,
+      "logps/rejected": -81.79740905761719,
+      "loss": 2.6286,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.75887131690979,
+      "rewards/margins": 0.33979111909866333,
+      "rewards/rejected": -2.0986623764038086,
+      "step": 310
+    },
+    {
+      "epoch": 0.5649409627611263,
+      "grad_norm": 1.7387233972549438,
+      "learning_rate": 9.563476243993008e-06,
+      "logits/chosen": 0.130618155002594,
+      "logits/rejected": 0.12952059507369995,
+      "logps/chosen": -80.75495910644531,
+      "logps/rejected": -90.4281234741211,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.836004614830017,
+      "rewards/margins": 0.4418962597846985,
+      "rewards/rejected": -2.2779006958007812,
+      "step": 311
+    },
+    {
+      "epoch": 0.5667574931880109,
+      "grad_norm": 1.5147444009780884,
+      "learning_rate": 9.56071116686402e-06,
+      "logits/chosen": 0.10193713754415512,
+      "logits/rejected": 0.22481802105903625,
+      "logps/chosen": -73.21926879882812,
+      "logps/rejected": -81.27864837646484,
+      "loss": 2.578,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6358015537261963,
+      "rewards/margins": 0.38279739022254944,
+      "rewards/rejected": -2.018598794937134,
+      "step": 312
+    },
+    {
+      "epoch": 0.5685740236148955,
+      "grad_norm": 1.6253665685653687,
+      "learning_rate": 9.557937809237927e-06,
+      "logits/chosen": 0.09468917548656464,
+      "logits/rejected": 0.09415112435817719,
+      "logps/chosen": -80.62995147705078,
+      "logps/rejected": -86.65946197509766,
+      "loss": 2.3854,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8635404109954834,
+      "rewards/margins": 0.4438764154911041,
+      "rewards/rejected": -2.3074169158935547,
+      "step": 313
+    },
+    {
+      "epoch": 0.5703905540417802,
+      "grad_norm": 1.7893344163894653,
+      "learning_rate": 9.555156176770087e-06,
+      "logits/chosen": 0.15863659977912903,
+      "logits/rejected": 0.09485571831464767,
+      "logps/chosen": -76.10442352294922,
+      "logps/rejected": -72.73162078857422,
+      "loss": 2.6795,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.935206413269043,
+      "rewards/margins": 0.22100940346717834,
+      "rewards/rejected": -2.1562156677246094,
+      "step": 314
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 1.696327805519104,
+      "learning_rate": 9.552366275132733e-06,
+      "logits/chosen": 0.07012113183736801,
+      "logits/rejected": 0.13891686499118805,
+      "logps/chosen": -78.08012390136719,
+      "logps/rejected": -83.05044555664062,
+      "loss": 2.6181,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7869051694869995,
+      "rewards/margins": 0.23424415290355682,
+      "rewards/rejected": -2.0211493968963623,
+      "step": 315
+    },
+    {
+      "epoch": 0.5740236148955495,
+      "grad_norm": 1.6044729948043823,
+      "learning_rate": 9.54956811001496e-06,
+      "logits/chosen": 0.11298641562461853,
+      "logits/rejected": 0.2595204710960388,
+      "logps/chosen": -75.5998306274414,
+      "logps/rejected": -85.31849670410156,
+      "loss": 2.5849,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.5757699012756348,
+      "rewards/margins": 0.35041162371635437,
+      "rewards/rejected": -1.9261815547943115,
+      "step": 316
+    },
+    {
+      "epoch": 0.5758401453224341,
+      "grad_norm": 1.7399548292160034,
+      "learning_rate": 9.546761687122715e-06,
+      "logits/chosen": 0.14933931827545166,
+      "logits/rejected": 0.1868455857038498,
+      "logps/chosen": -68.05965423583984,
+      "logps/rejected": -82.37442779541016,
+      "loss": 2.3208,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8474600315093994,
+      "rewards/margins": 0.5766161680221558,
+      "rewards/rejected": -2.4240763187408447,
+      "step": 317
+    },
+    {
+      "epoch": 0.5776566757493188,
+      "grad_norm": 1.597495675086975,
+      "learning_rate": 9.54394701217878e-06,
+      "logits/chosen": 0.04729313403367996,
+      "logits/rejected": 0.05815067142248154,
+      "logps/chosen": -97.36541748046875,
+      "logps/rejected": -104.02650451660156,
+      "loss": 2.2106,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8631618022918701,
+      "rewards/margins": 0.6027428507804871,
+      "rewards/rejected": -2.465904474258423,
+      "step": 318
+    },
+    {
+      "epoch": 0.5794732061762035,
+      "grad_norm": 1.8646149635314941,
+      "learning_rate": 9.541124090922771e-06,
+      "logits/chosen": 0.1399674415588379,
+      "logits/rejected": 0.11107950657606125,
+      "logps/chosen": -82.39740753173828,
+      "logps/rejected": -76.80902099609375,
+      "loss": 2.6435,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.7670300006866455,
+      "rewards/margins": 0.15345275402069092,
+      "rewards/rejected": -1.920482873916626,
+      "step": 319
+    },
+    {
+      "epoch": 0.5812897366030881,
+      "grad_norm": 1.502447247505188,
+      "learning_rate": 9.538292929111114e-06,
+      "logits/chosen": 0.10672347247600555,
+      "logits/rejected": 0.08290571719408035,
+      "logps/chosen": -74.72683715820312,
+      "logps/rejected": -86.37158203125,
+      "loss": 2.2137,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8041073083877563,
+      "rewards/margins": 0.596272885799408,
+      "rewards/rejected": -2.4003803730010986,
+      "step": 320
+    },
+    {
+      "epoch": 0.5831062670299727,
+      "grad_norm": 1.7287745475769043,
+      "learning_rate": 9.535453532517039e-06,
+      "logits/chosen": 0.13742896914482117,
+      "logits/rejected": 0.07607734203338623,
+      "logps/chosen": -86.82475280761719,
+      "logps/rejected": -77.8934555053711,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7918696403503418,
+      "rewards/margins": 0.2591831088066101,
+      "rewards/rejected": -2.0510525703430176,
+      "step": 321
+    },
+    {
+      "epoch": 0.5849227974568574,
+      "grad_norm": 1.5914596319198608,
+      "learning_rate": 9.532605906930575e-06,
+      "logits/chosen": 0.11301672458648682,
+      "logits/rejected": 0.1944851577281952,
+      "logps/chosen": -68.9898452758789,
+      "logps/rejected": -73.17451477050781,
+      "loss": 2.4565,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6999026536941528,
+      "rewards/margins": 0.30462026596069336,
+      "rewards/rejected": -2.0045228004455566,
+      "step": 322
+    },
+    {
+      "epoch": 0.5867393278837421,
+      "grad_norm": 1.6072031259536743,
+      "learning_rate": 9.529750058158522e-06,
+      "logits/chosen": 0.07092760503292084,
+      "logits/rejected": 0.08035591244697571,
+      "logps/chosen": -81.5555419921875,
+      "logps/rejected": -81.93560028076172,
+      "loss": 2.2729,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6834638118743896,
+      "rewards/margins": 0.48882579803466797,
+      "rewards/rejected": -2.1722893714904785,
+      "step": 323
+    },
+    {
+      "epoch": 0.5885558583106267,
+      "grad_norm": 1.6433221101760864,
+      "learning_rate": 9.526885992024453e-06,
+      "logits/chosen": 0.13823899626731873,
+      "logits/rejected": 0.13610433042049408,
+      "logps/chosen": -78.9625244140625,
+      "logps/rejected": -83.2666015625,
+      "loss": 2.2075,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6939194202423096,
+      "rewards/margins": 0.5755473971366882,
+      "rewards/rejected": -2.2694668769836426,
+      "step": 324
+    },
+    {
+      "epoch": 0.5903723887375113,
+      "grad_norm": 1.7674918174743652,
+      "learning_rate": 9.524013714368702e-06,
+      "logits/chosen": 0.1704932153224945,
+      "logits/rejected": 0.1073535829782486,
+      "logps/chosen": -72.47895050048828,
+      "logps/rejected": -73.0809326171875,
+      "loss": 2.5473,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8390110731124878,
+      "rewards/margins": 0.34467947483062744,
+      "rewards/rejected": -2.1836905479431152,
+      "step": 325
+    },
+    {
+      "epoch": 0.592188919164396,
+      "grad_norm": 1.8219250440597534,
+      "learning_rate": 9.521133231048338e-06,
+      "logits/chosen": 0.07941028475761414,
+      "logits/rejected": 0.13045310974121094,
+      "logps/chosen": -80.73494720458984,
+      "logps/rejected": -91.58990478515625,
+      "loss": 2.4866,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8594518899917603,
+      "rewards/margins": 0.4554288685321808,
+      "rewards/rejected": -2.314880847930908,
+      "step": 326
+    },
+    {
+      "epoch": 0.5940054495912807,
+      "grad_norm": 1.3753328323364258,
+      "learning_rate": 9.51824454793717e-06,
+      "logits/chosen": 0.08879546821117401,
+      "logits/rejected": 0.04693777486681938,
+      "logps/chosen": -82.9569091796875,
+      "logps/rejected": -91.44571685791016,
+      "loss": 2.2602,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.549119472503662,
+      "rewards/margins": 0.5145683288574219,
+      "rewards/rejected": -2.063688039779663,
+      "step": 327
+    },
+    {
+      "epoch": 0.5958219800181653,
+      "grad_norm": 1.4296562671661377,
+      "learning_rate": 9.515347670925728e-06,
+      "logits/chosen": 0.15614314377307892,
+      "logits/rejected": 0.1598319411277771,
+      "logps/chosen": -75.60345458984375,
+      "logps/rejected": -80.81770324707031,
+      "loss": 2.3946,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.556883454322815,
+      "rewards/margins": 0.4525718092918396,
+      "rewards/rejected": -2.0094552040100098,
+      "step": 328
+    },
+    {
+      "epoch": 0.59763851044505,
+      "grad_norm": 1.7932195663452148,
+      "learning_rate": 9.512442605921245e-06,
+      "logits/chosen": 0.05863601714372635,
+      "logits/rejected": 0.0788806602358818,
+      "logps/chosen": -75.95397186279297,
+      "logps/rejected": -84.50177001953125,
+      "loss": 2.5085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9159862995147705,
+      "rewards/margins": 0.4600725769996643,
+      "rewards/rejected": -2.376059055328369,
+      "step": 329
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 1.6558383703231812,
+      "learning_rate": 9.509529358847655e-06,
+      "logits/chosen": 0.08205496519804001,
+      "logits/rejected": 0.13091425597667694,
+      "logps/chosen": -82.73474884033203,
+      "logps/rejected": -94.02273559570312,
+      "loss": 2.4419,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.083285093307495,
+      "rewards/margins": 0.5300815105438232,
+      "rewards/rejected": -2.6133666038513184,
+      "step": 330
+    },
+    {
+      "epoch": 0.6012715712988193,
+      "grad_norm": 1.7888520956039429,
+      "learning_rate": 9.506607935645579e-06,
+      "logits/chosen": 0.11793217062950134,
+      "logits/rejected": 0.19568441808223724,
+      "logps/chosen": -79.86200714111328,
+      "logps/rejected": -89.25471496582031,
+      "loss": 2.4273,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9273384809494019,
+      "rewards/margins": 0.5115458965301514,
+      "rewards/rejected": -2.4388844966888428,
+      "step": 331
+    },
+    {
+      "epoch": 0.6030881017257039,
+      "grad_norm": 2.0436160564422607,
+      "learning_rate": 9.503678342272306e-06,
+      "logits/chosen": 0.043473344296216965,
+      "logits/rejected": 0.15554016828536987,
+      "logps/chosen": -75.2901611328125,
+      "logps/rejected": -83.20555877685547,
+      "loss": 2.6941,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.725769281387329,
+      "rewards/margins": 0.2789250314235687,
+      "rewards/rejected": -2.0046942234039307,
+      "step": 332
+    },
+    {
+      "epoch": 0.6049046321525886,
+      "grad_norm": 1.665578842163086,
+      "learning_rate": 9.500740584701785e-06,
+      "logits/chosen": 0.17119848728179932,
+      "logits/rejected": 0.14128939807415009,
+      "logps/chosen": -83.1322021484375,
+      "logps/rejected": -93.38603210449219,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.8975954055786133,
+      "rewards/margins": 0.7810046672821045,
+      "rewards/rejected": -2.6786000728607178,
+      "step": 333
+    },
+    {
+      "epoch": 0.6067211625794732,
+      "grad_norm": 1.9258419275283813,
+      "learning_rate": 9.497794668924617e-06,
+      "logits/chosen": 0.028591612353920937,
+      "logits/rejected": 0.08894480764865875,
+      "logps/chosen": -78.12230682373047,
+      "logps/rejected": -90.00184631347656,
+      "loss": 2.4006,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8113291263580322,
+      "rewards/margins": 0.5277000665664673,
+      "rewards/rejected": -2.339029550552368,
+      "step": 334
+    },
+    {
+      "epoch": 0.6085376930063578,
+      "grad_norm": 1.4878523349761963,
+      "learning_rate": 9.494840600948038e-06,
+      "logits/chosen": 0.03548199310898781,
+      "logits/rejected": 0.09170147776603699,
+      "logps/chosen": -73.15607452392578,
+      "logps/rejected": -83.77317810058594,
+      "loss": 2.3301,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6781296730041504,
+      "rewards/margins": 0.4735565185546875,
+      "rewards/rejected": -2.151685953140259,
+      "step": 335
+    },
+    {
+      "epoch": 0.6103542234332425,
+      "grad_norm": 1.8042774200439453,
+      "learning_rate": 9.491878386795906e-06,
+      "logits/chosen": 0.046132348477840424,
+      "logits/rejected": 0.0721711814403534,
+      "logps/chosen": -81.01045227050781,
+      "logps/rejected": -89.28679656982422,
+      "loss": 2.5593,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.012852191925049,
+      "rewards/margins": 0.44549623131752014,
+      "rewards/rejected": -2.458348512649536,
+      "step": 336
+    },
+    {
+      "epoch": 0.6121707538601272,
+      "grad_norm": 2.048952102661133,
+      "learning_rate": 9.488908032508691e-06,
+      "logits/chosen": 0.10774732381105423,
+      "logits/rejected": 0.11935572326183319,
+      "logps/chosen": -91.25210571289062,
+      "logps/rejected": -90.42224884033203,
+      "loss": 2.8282,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0127718448638916,
+      "rewards/margins": 0.16889013350009918,
+      "rewards/rejected": -2.18166184425354,
+      "step": 337
+    },
+    {
+      "epoch": 0.6139872842870118,
+      "grad_norm": 1.6447219848632812,
+      "learning_rate": 9.485929544143462e-06,
+      "logits/chosen": 0.12260966747999191,
+      "logits/rejected": 0.1517799198627472,
+      "logps/chosen": -72.49166870117188,
+      "logps/rejected": -82.3083724975586,
+      "loss": 2.4029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.624878168106079,
+      "rewards/margins": 0.5487539768218994,
+      "rewards/rejected": -2.1736321449279785,
+      "step": 338
+    },
+    {
+      "epoch": 0.6158038147138964,
+      "grad_norm": 1.599530577659607,
+      "learning_rate": 9.482942927773876e-06,
+      "logits/chosen": 0.09051798284053802,
+      "logits/rejected": 0.11202570050954819,
+      "logps/chosen": -84.42019653320312,
+      "logps/rejected": -92.61677551269531,
+      "loss": 2.2269,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6424546241760254,
+      "rewards/margins": 0.5715836882591248,
+      "rewards/rejected": -2.214038133621216,
+      "step": 339
+    },
+    {
+      "epoch": 0.6176203451407811,
+      "grad_norm": 1.5411081314086914,
+      "learning_rate": 9.479948189490164e-06,
+      "logits/chosen": 0.06563656777143478,
+      "logits/rejected": 0.17349205911159515,
+      "logps/chosen": -70.98219299316406,
+      "logps/rejected": -78.88846588134766,
+      "loss": 2.3328,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.824373722076416,
+      "rewards/margins": 0.46550822257995605,
+      "rewards/rejected": -2.289881706237793,
+      "step": 340
+    },
+    {
+      "epoch": 0.6194368755676658,
+      "grad_norm": 1.8855030536651611,
+      "learning_rate": 9.476945335399122e-06,
+      "logits/chosen": 0.11399642378091812,
+      "logits/rejected": 0.12690946459770203,
+      "logps/chosen": -85.86933898925781,
+      "logps/rejected": -89.75601196289062,
+      "loss": 2.4269,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9560539722442627,
+      "rewards/margins": 0.38427016139030457,
+      "rewards/rejected": -2.3403239250183105,
+      "step": 341
+    },
+    {
+      "epoch": 0.6212534059945504,
+      "grad_norm": 1.9973480701446533,
+      "learning_rate": 9.473934371624087e-06,
+      "logits/chosen": 0.029105912894010544,
+      "logits/rejected": 0.12763622403144836,
+      "logps/chosen": -80.68119049072266,
+      "logps/rejected": -93.90747833251953,
+      "loss": 2.4709,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.906477928161621,
+      "rewards/margins": 0.6422742009162903,
+      "rewards/rejected": -2.5487518310546875,
+      "step": 342
+    },
+    {
+      "epoch": 0.623069936421435,
+      "grad_norm": 1.7752224206924438,
+      "learning_rate": 9.47091530430494e-06,
+      "logits/chosen": 0.14096824824810028,
+      "logits/rejected": 0.1774113029241562,
+      "logps/chosen": -71.80216217041016,
+      "logps/rejected": -79.65301513671875,
+      "loss": 2.2678,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.5735952854156494,
+      "rewards/margins": 0.5823659300804138,
+      "rewards/rejected": -2.155961036682129,
+      "step": 343
+    },
+    {
+      "epoch": 0.6248864668483197,
+      "grad_norm": 1.855377197265625,
+      "learning_rate": 9.467888139598086e-06,
+      "logits/chosen": 0.10637074708938599,
+      "logits/rejected": 0.0798158049583435,
+      "logps/chosen": -79.01347351074219,
+      "logps/rejected": -80.77239990234375,
+      "loss": 2.7867,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9598251581192017,
+      "rewards/margins": 0.3184160590171814,
+      "rewards/rejected": -2.2782411575317383,
+      "step": 344
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 1.423040747642517,
+      "learning_rate": 9.464852883676441e-06,
+      "logits/chosen": 0.1304859220981598,
+      "logits/rejected": 0.1759231686592102,
+      "logps/chosen": -74.09733581542969,
+      "logps/rejected": -89.27589416503906,
+      "loss": 2.2777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7697277069091797,
+      "rewards/margins": 0.635611355304718,
+      "rewards/rejected": -2.405339002609253,
+      "step": 345
+    },
+    {
+      "epoch": 0.628519527702089,
+      "grad_norm": 1.6246038675308228,
+      "learning_rate": 9.461809542729421e-06,
+      "logits/chosen": 0.03750050812959671,
+      "logits/rejected": 0.10337980091571808,
+      "logps/chosen": -81.89110565185547,
+      "logps/rejected": -95.36811828613281,
+      "loss": 2.0438,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7299752235412598,
+      "rewards/margins": 0.7404756546020508,
+      "rewards/rejected": -2.4704508781433105,
+      "step": 346
+    },
+    {
+      "epoch": 0.6303360581289736,
+      "grad_norm": 1.6618752479553223,
+      "learning_rate": 9.458758122962926e-06,
+      "logits/chosen": 0.05359608680009842,
+      "logits/rejected": 0.10455545783042908,
+      "logps/chosen": -85.34078216552734,
+      "logps/rejected": -90.46200561523438,
+      "loss": 2.3299,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9331045150756836,
+      "rewards/margins": 0.4755082130432129,
+      "rewards/rejected": -2.4086129665374756,
+      "step": 347
+    },
+    {
+      "epoch": 0.6321525885558583,
+      "grad_norm": 1.5805696249008179,
+      "learning_rate": 9.455698630599332e-06,
+      "logits/chosen": 0.10048776119947433,
+      "logits/rejected": 0.1140337884426117,
+      "logps/chosen": -80.25875854492188,
+      "logps/rejected": -94.44998168945312,
+      "loss": 2.3439,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8421945571899414,
+      "rewards/margins": 0.5977468490600586,
+      "rewards/rejected": -2.43994140625,
+      "step": 348
+    },
+    {
+      "epoch": 0.633969118982743,
+      "grad_norm": 1.37093985080719,
+      "learning_rate": 9.452631071877478e-06,
+      "logits/chosen": 0.11764326691627502,
+      "logits/rejected": 0.10735289752483368,
+      "logps/chosen": -72.02367401123047,
+      "logps/rejected": -77.9522705078125,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.6527085304260254,
+      "rewards/margins": 0.6065118908882141,
+      "rewards/rejected": -2.259220600128174,
+      "step": 349
+    },
+    {
+      "epoch": 0.6357856494096276,
+      "grad_norm": 2.0214192867279053,
+      "learning_rate": 9.449555453052652e-06,
+      "logits/chosen": 0.13177426159381866,
+      "logits/rejected": 0.12408209592103958,
+      "logps/chosen": -76.62931823730469,
+      "logps/rejected": -81.16517639160156,
+      "loss": 2.9175,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.9575048685073853,
+      "rewards/margins": 0.21441945433616638,
+      "rewards/rejected": -2.171924352645874,
+      "step": 350
+    },
+    {
+      "epoch": 0.6376021798365122,
+      "grad_norm": 1.6041687726974487,
+      "learning_rate": 9.446471780396573e-06,
+      "logits/chosen": 0.18574532866477966,
+      "logits/rejected": 0.22683225572109222,
+      "logps/chosen": -73.95652770996094,
+      "logps/rejected": -80.29524993896484,
+      "loss": 2.3833,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8684748411178589,
+      "rewards/margins": 0.37268272042274475,
+      "rewards/rejected": -2.2411575317382812,
+      "step": 351
+    },
+    {
+      "epoch": 0.6394187102633969,
+      "grad_norm": 1.7410838603973389,
+      "learning_rate": 9.443380060197387e-06,
+      "logits/chosen": 0.07876043766736984,
+      "logits/rejected": 0.14103996753692627,
+      "logps/chosen": -74.64432525634766,
+      "logps/rejected": -83.6999740600586,
+      "loss": 2.5558,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.029916763305664,
+      "rewards/margins": 0.4088842272758484,
+      "rewards/rejected": -2.438800811767578,
+      "step": 352
+    },
+    {
+      "epoch": 0.6412352406902816,
+      "grad_norm": 1.87971031665802,
+      "learning_rate": 9.440280298759653e-06,
+      "logits/chosen": 0.13997013866901398,
+      "logits/rejected": 0.1322249174118042,
+      "logps/chosen": -88.78776550292969,
+      "logps/rejected": -88.22732543945312,
+      "loss": 2.6547,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9983258247375488,
+      "rewards/margins": 0.2458382397890091,
+      "rewards/rejected": -2.244164228439331,
+      "step": 353
+    },
+    {
+      "epoch": 0.6430517711171662,
+      "grad_norm": 1.7056363821029663,
+      "learning_rate": 9.437172502404318e-06,
+      "logits/chosen": 0.07248476892709732,
+      "logits/rejected": 0.13019773364067078,
+      "logps/chosen": -78.4591064453125,
+      "logps/rejected": -80.49126434326172,
+      "loss": 2.3809,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9740041494369507,
+      "rewards/margins": 0.4853326082229614,
+      "rewards/rejected": -2.459336757659912,
+      "step": 354
+    },
+    {
+      "epoch": 0.6448683015440508,
+      "grad_norm": 1.435718059539795,
+      "learning_rate": 9.434056677468726e-06,
+      "logits/chosen": 0.09164869040250778,
+      "logits/rejected": 0.14243285357952118,
+      "logps/chosen": -77.83367156982422,
+      "logps/rejected": -86.18836212158203,
+      "loss": 2.006,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.7355122566223145,
+      "rewards/margins": 0.6869419813156128,
+      "rewards/rejected": -2.422454357147217,
+      "step": 355
+    },
+    {
+      "epoch": 0.6466848319709355,
+      "grad_norm": 1.6553188562393188,
+      "learning_rate": 9.430932830306587e-06,
+      "logits/chosen": 0.04967673122882843,
+      "logits/rejected": 0.1456151008605957,
+      "logps/chosen": -73.50204467773438,
+      "logps/rejected": -85.31363677978516,
+      "loss": 2.4388,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.978384256362915,
+      "rewards/margins": 0.48918718099594116,
+      "rewards/rejected": -2.467571496963501,
+      "step": 356
+    },
+    {
+      "epoch": 0.6485013623978202,
+      "grad_norm": 1.8443480730056763,
+      "learning_rate": 9.427800967287963e-06,
+      "logits/chosen": 0.06455090641975403,
+      "logits/rejected": 0.12608012557029724,
+      "logps/chosen": -75.49232482910156,
+      "logps/rejected": -77.49136352539062,
+      "loss": 2.706,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.8730812072753906,
+      "rewards/margins": 0.2989741861820221,
+      "rewards/rejected": -2.17205548286438,
+      "step": 357
+    },
+    {
+      "epoch": 0.6503178928247049,
+      "grad_norm": 1.6912306547164917,
+      "learning_rate": 9.424661094799273e-06,
+      "logits/chosen": 0.11878645420074463,
+      "logits/rejected": 0.13628609478473663,
+      "logps/chosen": -73.09883880615234,
+      "logps/rejected": -84.19624328613281,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.928421139717102,
+      "rewards/margins": 0.5017052292823792,
+      "rewards/rejected": -2.430126190185547,
+      "step": 358
+    },
+    {
+      "epoch": 0.6521344232515894,
+      "grad_norm": 1.3643461465835571,
+      "learning_rate": 9.421513219243262e-06,
+      "logits/chosen": 0.07683826237916946,
+      "logits/rejected": 0.15765298902988434,
+      "logps/chosen": -76.0871353149414,
+      "logps/rejected": -97.47781372070312,
+      "loss": 1.7844,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.8300602436065674,
+      "rewards/margins": 0.911116361618042,
+      "rewards/rejected": -2.7411766052246094,
+      "step": 359
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 1.6327749490737915,
+      "learning_rate": 9.418357347038999e-06,
+      "logits/chosen": 0.1078951433300972,
+      "logits/rejected": 0.12233921140432358,
+      "logps/chosen": -75.89913177490234,
+      "logps/rejected": -78.4587173461914,
+      "loss": 2.4976,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.7992844581604004,
+      "rewards/margins": 0.311392605304718,
+      "rewards/rejected": -2.1106772422790527,
+      "step": 360
+    },
+    {
+      "epoch": 0.6557674841053588,
+      "grad_norm": 1.463025689125061,
+      "learning_rate": 9.415193484621852e-06,
+      "logits/chosen": 0.057331383228302,
+      "logits/rejected": 0.1563551127910614,
+      "logps/chosen": -80.7889633178711,
+      "logps/rejected": -93.21602630615234,
+      "loss": 2.1398,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.744189739227295,
+      "rewards/margins": 0.576555073261261,
+      "rewards/rejected": -2.320744752883911,
+      "step": 361
+    },
+    {
+      "epoch": 0.6575840145322435,
+      "grad_norm": 1.743695616722107,
+      "learning_rate": 9.412021638443491e-06,
+      "logits/chosen": 0.03781123086810112,
+      "logits/rejected": 0.154897540807724,
+      "logps/chosen": -77.72047424316406,
+      "logps/rejected": -88.80599975585938,
+      "loss": 2.2361,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9849369525909424,
+      "rewards/margins": 0.6562294363975525,
+      "rewards/rejected": -2.6411664485931396,
+      "step": 362
+    },
+    {
+      "epoch": 0.659400544959128,
+      "grad_norm": 1.5648279190063477,
+      "learning_rate": 9.408841814971862e-06,
+      "logits/chosen": 0.10401102900505066,
+      "logits/rejected": 0.12699122726917267,
+      "logps/chosen": -83.10671997070312,
+      "logps/rejected": -86.8663101196289,
+      "loss": 2.4009,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7048468589782715,
+      "rewards/margins": 0.4090174734592438,
+      "rewards/rejected": -2.1138644218444824,
+      "step": 363
+    },
+    {
+      "epoch": 0.6612170753860127,
+      "grad_norm": 1.3385239839553833,
+      "learning_rate": 9.405654020691178e-06,
+      "logits/chosen": 0.06100422143936157,
+      "logits/rejected": 0.10872650146484375,
+      "logps/chosen": -76.72837829589844,
+      "logps/rejected": -87.63470458984375,
+      "loss": 2.141,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.747226595878601,
+      "rewards/margins": 0.615075945854187,
+      "rewards/rejected": -2.362302780151367,
+      "step": 364
+    },
+    {
+      "epoch": 0.6630336058128974,
+      "grad_norm": 1.4060734510421753,
+      "learning_rate": 9.402458262101906e-06,
+      "logits/chosen": 0.10638861358165741,
+      "logits/rejected": 0.16951681673526764,
+      "logps/chosen": -77.35757446289062,
+      "logps/rejected": -92.03012084960938,
+      "loss": 2.0511,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8359463214874268,
+      "rewards/margins": 0.7444422841072083,
+      "rewards/rejected": -2.5803885459899902,
+      "step": 365
+    },
+    {
+      "epoch": 0.6648501362397821,
+      "grad_norm": 1.4371014833450317,
+      "learning_rate": 9.399254545720757e-06,
+      "logits/chosen": 0.0383220911026001,
+      "logits/rejected": 0.05856206640601158,
+      "logps/chosen": -82.40064239501953,
+      "logps/rejected": -92.1117935180664,
+      "loss": 2.1756,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8580785989761353,
+      "rewards/margins": 0.6475991010665894,
+      "rewards/rejected": -2.5056777000427246,
+      "step": 366
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.6673945188522339,
+      "learning_rate": 9.396042878080661e-06,
+      "logits/chosen": 0.14657820761203766,
+      "logits/rejected": 0.19021111726760864,
+      "logps/chosen": -73.56724548339844,
+      "logps/rejected": -78.85279846191406,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9079951047897339,
+      "rewards/margins": 0.3709837794303894,
+      "rewards/rejected": -2.2789790630340576,
+      "step": 367
+    },
+    {
+      "epoch": 0.6684831970935513,
+      "grad_norm": 1.7402448654174805,
+      "learning_rate": 9.392823265730775e-06,
+      "logits/chosen": 0.16515754163265228,
+      "logits/rejected": 0.12460774928331375,
+      "logps/chosen": -69.89266204833984,
+      "logps/rejected": -74.53837585449219,
+      "loss": 2.4262,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.0627622604370117,
+      "rewards/margins": 0.5317320823669434,
+      "rewards/rejected": -2.594494342803955,
+      "step": 368
+    },
+    {
+      "epoch": 0.670299727520436,
+      "grad_norm": 1.5290364027023315,
+      "learning_rate": 9.389595715236446e-06,
+      "logits/chosen": 0.0954434722661972,
+      "logits/rejected": 0.16226956248283386,
+      "logps/chosen": -77.17019653320312,
+      "logps/rejected": -91.07398986816406,
+      "loss": 2.2511,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.241170644760132,
+      "rewards/margins": 0.6319360136985779,
+      "rewards/rejected": -2.8731067180633545,
+      "step": 369
+    },
+    {
+      "epoch": 0.6721162579473207,
+      "grad_norm": 1.4633573293685913,
+      "learning_rate": 9.386360233179206e-06,
+      "logits/chosen": 0.06517557799816132,
+      "logits/rejected": 0.07250035554170609,
+      "logps/chosen": -79.23770141601562,
+      "logps/rejected": -88.50403594970703,
+      "loss": 2.3858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.751698613166809,
+      "rewards/margins": 0.49529415369033813,
+      "rewards/rejected": -2.246993064880371,
+      "step": 370
+    },
+    {
+      "epoch": 0.6739327883742052,
+      "grad_norm": 2.007215976715088,
+      "learning_rate": 9.383116826156775e-06,
+      "logits/chosen": 0.13584929704666138,
+      "logits/rejected": 0.10950647294521332,
+      "logps/chosen": -80.75408935546875,
+      "logps/rejected": -79.3903579711914,
+      "loss": 2.6967,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.2421724796295166,
+      "rewards/margins": 0.34328368306159973,
+      "rewards/rejected": -2.585456132888794,
+      "step": 371
+    },
+    {
+      "epoch": 0.6757493188010899,
+      "grad_norm": 1.7323477268218994,
+      "learning_rate": 9.37986550078302e-06,
+      "logits/chosen": 0.011626070365309715,
+      "logits/rejected": 0.06207559257745743,
+      "logps/chosen": -75.51494598388672,
+      "logps/rejected": -85.83218383789062,
+      "loss": 2.4608,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9569413661956787,
+      "rewards/margins": 0.5160585641860962,
+      "rewards/rejected": -2.4729998111724854,
+      "step": 372
+    },
+    {
+      "epoch": 0.6775658492279746,
+      "grad_norm": 1.6365752220153809,
+      "learning_rate": 9.376606263687959e-06,
+      "logits/chosen": 0.10213632136583328,
+      "logits/rejected": 0.08355780690908432,
+      "logps/chosen": -79.01370239257812,
+      "logps/rejected": -81.9417953491211,
+      "loss": 2.4701,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.064467668533325,
+      "rewards/margins": 0.4157037138938904,
+      "rewards/rejected": -2.4801712036132812,
+      "step": 373
+    },
+    {
+      "epoch": 0.6793823796548593,
+      "grad_norm": 1.5918922424316406,
+      "learning_rate": 9.373339121517748e-06,
+      "logits/chosen": 0.09486684203147888,
+      "logits/rejected": 0.07311725616455078,
+      "logps/chosen": -82.15351867675781,
+      "logps/rejected": -91.7690658569336,
+      "loss": 2.1297,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0153114795684814,
+      "rewards/margins": 0.669613242149353,
+      "rewards/rejected": -2.684924840927124,
+      "step": 374
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 1.6511566638946533,
+      "learning_rate": 9.370064080934654e-06,
+      "logits/chosen": 0.1406637728214264,
+      "logits/rejected": 0.20172299444675446,
+      "logps/chosen": -69.41023254394531,
+      "logps/rejected": -74.65138244628906,
+      "loss": 2.4704,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.061182737350464,
+      "rewards/margins": 0.3090137243270874,
+      "rewards/rejected": -2.3701963424682617,
+      "step": 375
+    },
+    {
+      "epoch": 0.6830154405086285,
+      "grad_norm": 1.770624041557312,
+      "learning_rate": 9.366781148617056e-06,
+      "logits/chosen": 0.10267248749732971,
+      "logits/rejected": 0.07900385558605194,
+      "logps/chosen": -77.27003479003906,
+      "logps/rejected": -88.09040832519531,
+      "loss": 2.2813,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.132927656173706,
+      "rewards/margins": 0.6629016399383545,
+      "rewards/rejected": -2.7958290576934814,
+      "step": 376
+    },
+    {
+      "epoch": 0.6848319709355132,
+      "grad_norm": 1.5172028541564941,
+      "learning_rate": 9.363490331259426e-06,
+      "logits/chosen": 0.05240853130817413,
+      "logits/rejected": 0.10445387661457062,
+      "logps/chosen": -75.5218276977539,
+      "logps/rejected": -85.49366760253906,
+      "loss": 2.1442,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9302880764007568,
+      "rewards/margins": 0.5776917338371277,
+      "rewards/rejected": -2.5079798698425293,
+      "step": 377
+    },
+    {
+      "epoch": 0.6866485013623979,
+      "grad_norm": 1.6540950536727905,
+      "learning_rate": 9.360191635572313e-06,
+      "logits/chosen": 0.14195458590984344,
+      "logits/rejected": 0.0978107899427414,
+      "logps/chosen": -85.00608825683594,
+      "logps/rejected": -85.1456298828125,
+      "loss": 2.423,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8765841722488403,
+      "rewards/margins": 0.49441370368003845,
+      "rewards/rejected": -2.370997905731201,
+      "step": 378
+    },
+    {
+      "epoch": 0.6884650317892824,
+      "grad_norm": 1.6463801860809326,
+      "learning_rate": 9.356885068282334e-06,
+      "logits/chosen": 0.13197994232177734,
+      "logits/rejected": 0.09924699366092682,
+      "logps/chosen": -86.94219970703125,
+      "logps/rejected": -86.85704803466797,
+      "loss": 2.4801,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9674016237258911,
+      "rewards/margins": 0.4530283510684967,
+      "rewards/rejected": -2.4204299449920654,
+      "step": 379
+    },
+    {
+      "epoch": 0.6902815622161671,
+      "grad_norm": 1.5428895950317383,
+      "learning_rate": 9.353570636132151e-06,
+      "logits/chosen": 0.10434838384389877,
+      "logits/rejected": 0.12619757652282715,
+      "logps/chosen": -78.56607818603516,
+      "logps/rejected": -86.33320617675781,
+      "loss": 2.2711,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9667280912399292,
+      "rewards/margins": 0.6350463628768921,
+      "rewards/rejected": -2.601774215698242,
+      "step": 380
+    },
+    {
+      "epoch": 0.6920980926430518,
+      "grad_norm": 1.6211251020431519,
+      "learning_rate": 9.350248345880471e-06,
+      "logits/chosen": 0.14081251621246338,
+      "logits/rejected": 0.1453506350517273,
+      "logps/chosen": -74.38184356689453,
+      "logps/rejected": -80.61524200439453,
+      "loss": 2.3863,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.0189671516418457,
+      "rewards/margins": 0.4528267979621887,
+      "rewards/rejected": -2.4717938899993896,
+      "step": 381
+    },
+    {
+      "epoch": 0.6939146230699365,
+      "grad_norm": 2.6346240043640137,
+      "learning_rate": 9.346918204302022e-06,
+      "logits/chosen": 0.0519830696284771,
+      "logits/rejected": 0.04284125566482544,
+      "logps/chosen": -82.38153839111328,
+      "logps/rejected": -86.35488891601562,
+      "loss": 2.5434,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9824391603469849,
+      "rewards/margins": 0.4156208634376526,
+      "rewards/rejected": -2.3980600833892822,
+      "step": 382
+    },
+    {
+      "epoch": 0.695731153496821,
+      "grad_norm": 1.4762026071548462,
+      "learning_rate": 9.343580218187544e-06,
+      "logits/chosen": 0.02196469157934189,
+      "logits/rejected": 0.04618150740861893,
+      "logps/chosen": -72.66182708740234,
+      "logps/rejected": -88.68983459472656,
+      "loss": 2.1699,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9227774143218994,
+      "rewards/margins": 0.9200201630592346,
+      "rewards/rejected": -2.8427975177764893,
+      "step": 383
+    },
+    {
+      "epoch": 0.6975476839237057,
+      "grad_norm": 1.9675711393356323,
+      "learning_rate": 9.340234394343768e-06,
+      "logits/chosen": 0.12956203520298004,
+      "logits/rejected": 0.16079677641391754,
+      "logps/chosen": -87.38492584228516,
+      "logps/rejected": -87.80662536621094,
+      "loss": 2.3138,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.144991397857666,
+      "rewards/margins": 0.4893474578857422,
+      "rewards/rejected": -2.634338855743408,
+      "step": 384
+    },
+    {
+      "epoch": 0.6993642143505904,
+      "grad_norm": 1.9116860628128052,
+      "learning_rate": 9.336880739593415e-06,
+      "logits/chosen": 0.06013559550046921,
+      "logits/rejected": 0.09889352321624756,
+      "logps/chosen": -86.55718231201172,
+      "logps/rejected": -91.01994323730469,
+      "loss": 2.4864,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9528815746307373,
+      "rewards/margins": 0.44100096821784973,
+      "rewards/rejected": -2.3938825130462646,
+      "step": 385
+    },
+    {
+      "epoch": 0.701180744777475,
+      "grad_norm": 2.0582289695739746,
+      "learning_rate": 9.33351926077517e-06,
+      "logits/chosen": 0.11538423597812653,
+      "logits/rejected": 0.17568480968475342,
+      "logps/chosen": -76.24241638183594,
+      "logps/rejected": -87.1357650756836,
+      "loss": 2.6878,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.102504253387451,
+      "rewards/margins": 0.3671773076057434,
+      "rewards/rejected": -2.469681739807129,
+      "step": 386
+    },
+    {
+      "epoch": 0.7029972752043597,
+      "grad_norm": 1.6962640285491943,
+      "learning_rate": 9.330149964743674e-06,
+      "logits/chosen": 0.12425235658884048,
+      "logits/rejected": 0.1852879822254181,
+      "logps/chosen": -82.67887115478516,
+      "logps/rejected": -98.66133117675781,
+      "loss": 2.343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.182525634765625,
+      "rewards/margins": 0.6602460145950317,
+      "rewards/rejected": -2.842771530151367,
+      "step": 387
+    },
+    {
+      "epoch": 0.7048138056312443,
+      "grad_norm": 1.6582266092300415,
+      "learning_rate": 9.326772858369506e-06,
+      "logits/chosen": 0.14438432455062866,
+      "logits/rejected": 0.0922938883304596,
+      "logps/chosen": -78.87794494628906,
+      "logps/rejected": -84.929443359375,
+      "loss": 2.3036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.988234281539917,
+      "rewards/margins": 0.4737466275691986,
+      "rewards/rejected": -2.4619810581207275,
+      "step": 388
+    },
+    {
+      "epoch": 0.706630336058129,
+      "grad_norm": 1.6689919233322144,
+      "learning_rate": 9.323387948539176e-06,
+      "logits/chosen": 0.1282673329114914,
+      "logits/rejected": 0.13633723556995392,
+      "logps/chosen": -70.22606658935547,
+      "logps/rejected": -79.26126861572266,
+      "loss": 2.241,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9162399768829346,
+      "rewards/margins": 0.5345122814178467,
+      "rewards/rejected": -2.4507524967193604,
+      "step": 389
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 1.5861207246780396,
+      "learning_rate": 9.319995242155102e-06,
+      "logits/chosen": 0.13024169206619263,
+      "logits/rejected": 0.14390775561332703,
+      "logps/chosen": -87.00923919677734,
+      "logps/rejected": -91.24594116210938,
+      "loss": 2.284,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9127002954483032,
+      "rewards/margins": 0.5265774130821228,
+      "rewards/rejected": -2.4392776489257812,
+      "step": 390
+    },
+    {
+      "epoch": 0.7102633969118983,
+      "grad_norm": 2.0118792057037354,
+      "learning_rate": 9.316594746135608e-06,
+      "logits/chosen": 0.14648675918579102,
+      "logits/rejected": 0.11786539107561111,
+      "logps/chosen": -82.68962097167969,
+      "logps/rejected": -88.26126861572266,
+      "loss": 2.7281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9463748931884766,
+      "rewards/margins": 0.20282456278800964,
+      "rewards/rejected": -2.1491994857788086,
+      "step": 391
+    },
+    {
+      "epoch": 0.7120799273387829,
+      "grad_norm": 1.8176332712173462,
+      "learning_rate": 9.313186467414892e-06,
+      "logits/chosen": 0.12084021419286728,
+      "logits/rejected": 0.1433698982000351,
+      "logps/chosen": -73.46707153320312,
+      "logps/rejected": -82.46024322509766,
+      "loss": 2.5247,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0802624225616455,
+      "rewards/margins": 0.41601985692977905,
+      "rewards/rejected": -2.4962823390960693,
+      "step": 392
+    },
+    {
+      "epoch": 0.7138964577656676,
+      "grad_norm": 1.7207351922988892,
+      "learning_rate": 9.30977041294303e-06,
+      "logits/chosen": 0.11983273923397064,
+      "logits/rejected": 0.16106371581554413,
+      "logps/chosen": -76.09751892089844,
+      "logps/rejected": -85.20323944091797,
+      "loss": 2.5284,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9018669128417969,
+      "rewards/margins": 0.4894491136074066,
+      "rewards/rejected": -2.3913159370422363,
+      "step": 393
+    },
+    {
+      "epoch": 0.7157129881925522,
+      "grad_norm": 1.6489367485046387,
+      "learning_rate": 9.306346589685956e-06,
+      "logits/chosen": 0.14845696091651917,
+      "logits/rejected": 0.10356368124485016,
+      "logps/chosen": -83.94526672363281,
+      "logps/rejected": -86.53378295898438,
+      "loss": 2.3394,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9089899063110352,
+      "rewards/margins": 0.4619474411010742,
+      "rewards/rejected": -2.3709375858306885,
+      "step": 394
+    },
+    {
+      "epoch": 0.7175295186194369,
+      "grad_norm": 1.792077898979187,
+      "learning_rate": 9.302915004625435e-06,
+      "logits/chosen": 0.06936248391866684,
+      "logits/rejected": 0.13349927961826324,
+      "logps/chosen": -91.82075500488281,
+      "logps/rejected": -98.7391357421875,
+      "loss": 2.3875,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9371806383132935,
+      "rewards/margins": 0.6781354546546936,
+      "rewards/rejected": -2.6153156757354736,
+      "step": 395
+    },
+    {
+      "epoch": 0.7193460490463215,
+      "grad_norm": 2.049694538116455,
+      "learning_rate": 9.29947566475907e-06,
+      "logits/chosen": 0.08474650233983994,
+      "logits/rejected": 0.13380834460258484,
+      "logps/chosen": -84.59528350830078,
+      "logps/rejected": -96.24267578125,
+      "loss": 2.7001,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9375882148742676,
+      "rewards/margins": 0.4607384204864502,
+      "rewards/rejected": -2.3983266353607178,
+      "step": 396
+    },
+    {
+      "epoch": 0.7211625794732062,
+      "grad_norm": 1.9235875606536865,
+      "learning_rate": 9.296028577100271e-06,
+      "logits/chosen": 0.06255945563316345,
+      "logits/rejected": 0.17074811458587646,
+      "logps/chosen": -68.99498748779297,
+      "logps/rejected": -84.45899963378906,
+      "loss": 2.347,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9344629049301147,
+      "rewards/margins": 0.688963770866394,
+      "rewards/rejected": -2.6234264373779297,
+      "step": 397
+    },
+    {
+      "epoch": 0.7229791099000908,
+      "grad_norm": 1.6617276668548584,
+      "learning_rate": 9.292573748678254e-06,
+      "logits/chosen": 0.11962493509054184,
+      "logits/rejected": 0.11647717654705048,
+      "logps/chosen": -85.3626708984375,
+      "logps/rejected": -95.70562744140625,
+      "loss": 2.2229,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.8968594074249268,
+      "rewards/margins": 0.6918852925300598,
+      "rewards/rejected": -2.588744640350342,
+      "step": 398
+    },
+    {
+      "epoch": 0.7247956403269755,
+      "grad_norm": 1.4650444984436035,
+      "learning_rate": 9.289111186538013e-06,
+      "logits/chosen": 0.06390775740146637,
+      "logits/rejected": 0.11076060682535172,
+      "logps/chosen": -69.9189453125,
+      "logps/rejected": -81.5618667602539,
+      "loss": 2.2271,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.857581377029419,
+      "rewards/margins": 0.5540103316307068,
+      "rewards/rejected": -2.4115917682647705,
+      "step": 399
+    },
+    {
+      "epoch": 0.7266121707538601,
+      "grad_norm": 1.5728726387023926,
+      "learning_rate": 9.285640897740316e-06,
+      "logits/chosen": 0.08816932141780853,
+      "logits/rejected": 0.1330798715353012,
+      "logps/chosen": -78.04735565185547,
+      "logps/rejected": -87.89193725585938,
+      "loss": 2.3595,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.823161244392395,
+      "rewards/margins": 0.4568220376968384,
+      "rewards/rejected": -2.2799830436706543,
+      "step": 400
+    },
+    {
+      "epoch": 0.7284287011807448,
+      "grad_norm": 1.6291779279708862,
+      "learning_rate": 9.282162889361686e-06,
+      "logits/chosen": 0.17718347907066345,
+      "logits/rejected": 0.20603135228157043,
+      "logps/chosen": -67.99884033203125,
+      "logps/rejected": -75.93153381347656,
+      "loss": 2.3002,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.671633243560791,
+      "rewards/margins": 0.45543500781059265,
+      "rewards/rejected": -2.127068281173706,
+      "step": 401
+    },
+    {
+      "epoch": 0.7302452316076294,
+      "grad_norm": 1.5066276788711548,
+      "learning_rate": 9.278677168494388e-06,
+      "logits/chosen": 0.16840124130249023,
+      "logits/rejected": 0.17102479934692383,
+      "logps/chosen": -78.1556625366211,
+      "logps/rejected": -84.91495513916016,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9279460906982422,
+      "rewards/margins": 0.6659258008003235,
+      "rewards/rejected": -2.593871831893921,
+      "step": 402
+    },
+    {
+      "epoch": 0.7320617620345141,
+      "grad_norm": 1.71995210647583,
+      "learning_rate": 9.275183742246412e-06,
+      "logits/chosen": -0.0005522281862795353,
+      "logits/rejected": 0.1412215530872345,
+      "logps/chosen": -73.37832641601562,
+      "logps/rejected": -92.71226501464844,
+      "loss": 2.3315,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6880759000778198,
+      "rewards/margins": 0.5844776630401611,
+      "rewards/rejected": -2.2725534439086914,
+      "step": 403
+    },
+    {
+      "epoch": 0.7338782924613987,
+      "grad_norm": 1.676321029663086,
+      "learning_rate": 9.271682617741466e-06,
+      "logits/chosen": 0.15319044888019562,
+      "logits/rejected": 0.14503052830696106,
+      "logps/chosen": -75.85136413574219,
+      "logps/rejected": -82.61698150634766,
+      "loss": 2.3528,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.819403052330017,
+      "rewards/margins": 0.3919827342033386,
+      "rewards/rejected": -2.211385726928711,
+      "step": 404
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 1.4811294078826904,
+      "learning_rate": 9.268173802118949e-06,
+      "logits/chosen": 0.07122528553009033,
+      "logits/rejected": 0.12146291136741638,
+      "logps/chosen": -76.78916931152344,
+      "logps/rejected": -85.33805084228516,
+      "loss": 2.1505,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9886473417282104,
+      "rewards/margins": 0.5911746025085449,
+      "rewards/rejected": -2.579822301864624,
+      "step": 405
+    },
+    {
+      "epoch": 0.737511353315168,
+      "grad_norm": 1.629601001739502,
+      "learning_rate": 9.264657302533947e-06,
+      "logits/chosen": 0.11274963617324829,
+      "logits/rejected": 0.0651293620467186,
+      "logps/chosen": -82.86985778808594,
+      "logps/rejected": -88.68803405761719,
+      "loss": 2.1873,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8538885116577148,
+      "rewards/margins": 0.6726438403129578,
+      "rewards/rejected": -2.5265324115753174,
+      "step": 406
+    },
+    {
+      "epoch": 0.7393278837420527,
+      "grad_norm": 1.5687483549118042,
+      "learning_rate": 9.261133126157218e-06,
+      "logits/chosen": 0.06054290384054184,
+      "logits/rejected": 0.1543959081172943,
+      "logps/chosen": -68.5821304321289,
+      "logps/rejected": -80.544921875,
+      "loss": 2.3587,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7550634145736694,
+      "rewards/margins": 0.46927812695503235,
+      "rewards/rejected": -2.224341630935669,
+      "step": 407
+    },
+    {
+      "epoch": 0.7411444141689373,
+      "grad_norm": 1.849908471107483,
+      "learning_rate": 9.257601280175167e-06,
+      "logits/chosen": 0.15177126228809357,
+      "logits/rejected": 0.1485670655965805,
+      "logps/chosen": -72.81680297851562,
+      "logps/rejected": -80.58684539794922,
+      "loss": 2.444,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9739688634872437,
+      "rewards/margins": 0.6166737079620361,
+      "rewards/rejected": -2.5906424522399902,
+      "step": 408
+    },
+    {
+      "epoch": 0.742960944595822,
+      "grad_norm": 1.779342770576477,
+      "learning_rate": 9.254061771789847e-06,
+      "logits/chosen": 0.05093669891357422,
+      "logits/rejected": 0.1048060953617096,
+      "logps/chosen": -84.48141479492188,
+      "logps/rejected": -91.49285125732422,
+      "loss": 2.5324,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9952929019927979,
+      "rewards/margins": 0.3900759220123291,
+      "rewards/rejected": -2.385368585586548,
+      "step": 409
+    },
+    {
+      "epoch": 0.7447774750227066,
+      "grad_norm": 1.71133291721344,
+      "learning_rate": 9.25051460821893e-06,
+      "logits/chosen": 0.1308833658695221,
+      "logits/rejected": 0.08994461596012115,
+      "logps/chosen": -74.09046936035156,
+      "logps/rejected": -80.80438995361328,
+      "loss": 2.1895,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8457674980163574,
+      "rewards/margins": 0.5731581449508667,
+      "rewards/rejected": -2.4189257621765137,
+      "step": 410
+    },
+    {
+      "epoch": 0.7465940054495913,
+      "grad_norm": 1.6980112791061401,
+      "learning_rate": 9.2469597966957e-06,
+      "logits/chosen": 0.05856658145785332,
+      "logits/rejected": 0.08371179550886154,
+      "logps/chosen": -78.58245849609375,
+      "logps/rejected": -88.30054473876953,
+      "loss": 2.46,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9803493022918701,
+      "rewards/margins": 0.6115507483482361,
+      "rewards/rejected": -2.591899871826172,
+      "step": 411
+    },
+    {
+      "epoch": 0.7484105358764759,
+      "grad_norm": 1.8350476026535034,
+      "learning_rate": 9.243397344469037e-06,
+      "logits/chosen": 0.05285171419382095,
+      "logits/rejected": 0.13240401446819305,
+      "logps/chosen": -73.79881286621094,
+      "logps/rejected": -86.56880187988281,
+      "loss": 2.2929,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.060075283050537,
+      "rewards/margins": 0.54813152551651,
+      "rewards/rejected": -2.6082065105438232,
+      "step": 412
+    },
+    {
+      "epoch": 0.7502270663033606,
+      "grad_norm": 1.4403142929077148,
+      "learning_rate": 9.239827258803402e-06,
+      "logits/chosen": 0.07327421009540558,
+      "logits/rejected": 0.18387822806835175,
+      "logps/chosen": -68.3246078491211,
+      "logps/rejected": -81.33067321777344,
+      "loss": 1.8794,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8658640384674072,
+      "rewards/margins": 0.7600168585777283,
+      "rewards/rejected": -2.625880718231201,
+      "step": 413
+    },
+    {
+      "epoch": 0.7520435967302452,
+      "grad_norm": 1.883155345916748,
+      "learning_rate": 9.23624954697882e-06,
+      "logits/chosen": 0.10411994159221649,
+      "logits/rejected": 0.06234448403120041,
+      "logps/chosen": -78.31743621826172,
+      "logps/rejected": -80.0100326538086,
+      "loss": 2.5646,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.934099555015564,
+      "rewards/margins": 0.28737103939056396,
+      "rewards/rejected": -2.221470832824707,
+      "step": 414
+    },
+    {
+      "epoch": 0.7538601271571299,
+      "grad_norm": 1.5926934480667114,
+      "learning_rate": 9.232664216290868e-06,
+      "logits/chosen": 0.09589770436286926,
+      "logits/rejected": 0.041689179837703705,
+      "logps/chosen": -77.89552307128906,
+      "logps/rejected": -84.02517700195312,
+      "loss": 2.2284,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.787453532218933,
+      "rewards/margins": 0.49444466829299927,
+      "rewards/rejected": -2.281898260116577,
+      "step": 415
+    },
+    {
+      "epoch": 0.7556766575840145,
+      "grad_norm": 1.511832356452942,
+      "learning_rate": 9.229071274050663e-06,
+      "logits/chosen": 0.07688678801059723,
+      "logits/rejected": 0.14106512069702148,
+      "logps/chosen": -78.53334045410156,
+      "logps/rejected": -86.92496490478516,
+      "loss": 2.2211,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0686798095703125,
+      "rewards/margins": 0.5378819704055786,
+      "rewards/rejected": -2.6065618991851807,
+      "step": 416
+    },
+    {
+      "epoch": 0.7574931880108992,
+      "grad_norm": 1.7858117818832397,
+      "learning_rate": 9.225470727584835e-06,
+      "logits/chosen": 0.14854730665683746,
+      "logits/rejected": 0.06465649604797363,
+      "logps/chosen": -87.31832122802734,
+      "logps/rejected": -79.3746566772461,
+      "loss": 2.4446,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9429931640625,
+      "rewards/margins": 0.38861486315727234,
+      "rewards/rejected": -2.3316078186035156,
+      "step": 417
+    },
+    {
+      "epoch": 0.7593097184377838,
+      "grad_norm": 1.9230687618255615,
+      "learning_rate": 9.221862584235527e-06,
+      "logits/chosen": 0.07233145087957382,
+      "logits/rejected": 0.13825634121894836,
+      "logps/chosen": -81.53103637695312,
+      "logps/rejected": -90.54696655273438,
+      "loss": 2.7171,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.065859794616699,
+      "rewards/margins": 0.30683329701423645,
+      "rewards/rejected": -2.3726933002471924,
+      "step": 418
+    },
+    {
+      "epoch": 0.7611262488646685,
+      "grad_norm": 2.063098192214966,
+      "learning_rate": 9.218246851360374e-06,
+      "logits/chosen": 0.10360075533390045,
+      "logits/rejected": 0.14677830040454865,
+      "logps/chosen": -85.08120727539062,
+      "logps/rejected": -89.35393524169922,
+      "loss": 2.4292,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.300565719604492,
+      "rewards/margins": 0.4277467727661133,
+      "rewards/rejected": -2.7283124923706055,
+      "step": 419
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 2.4728786945343018,
+      "learning_rate": 9.214623536332483e-06,
+      "logits/chosen": 0.08115474879741669,
+      "logits/rejected": 0.07649822533130646,
+      "logps/chosen": -83.79824829101562,
+      "logps/rejected": -87.47264099121094,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.0838184356689453,
+      "rewards/margins": 0.41770505905151367,
+      "rewards/rejected": -2.501523494720459,
+      "step": 420
+    },
+    {
+      "epoch": 0.7647593097184378,
+      "grad_norm": 1.7078697681427002,
+      "learning_rate": 9.210992646540425e-06,
+      "logits/chosen": 0.08155008405447006,
+      "logits/rejected": 0.08723931759595871,
+      "logps/chosen": -74.4200210571289,
+      "logps/rejected": -87.49038696289062,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9076974391937256,
+      "rewards/margins": 0.6807244420051575,
+      "rewards/rejected": -2.5884220600128174,
+      "step": 421
+    },
+    {
+      "epoch": 0.7665758401453224,
+      "grad_norm": 1.5644993782043457,
+      "learning_rate": 9.207354189388214e-06,
+      "logits/chosen": 0.16107802093029022,
+      "logits/rejected": 0.15951679646968842,
+      "logps/chosen": -74.85030364990234,
+      "logps/rejected": -81.25472259521484,
+      "loss": 2.2389,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0082590579986572,
+      "rewards/margins": 0.5728211402893066,
+      "rewards/rejected": -2.581080198287964,
+      "step": 422
+    },
+    {
+      "epoch": 0.7683923705722071,
+      "grad_norm": 1.6519557237625122,
+      "learning_rate": 9.203708172295299e-06,
+      "logits/chosen": 0.08725707978010178,
+      "logits/rejected": 0.11393023282289505,
+      "logps/chosen": -79.56071472167969,
+      "logps/rejected": -90.27279663085938,
+      "loss": 2.2853,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9312469959259033,
+      "rewards/margins": 0.6229196786880493,
+      "rewards/rejected": -2.554166793823242,
+      "step": 423
+    },
+    {
+      "epoch": 0.7702089009990918,
+      "grad_norm": 1.668648600578308,
+      "learning_rate": 9.200054602696544e-06,
+      "logits/chosen": 0.1369432657957077,
+      "logits/rejected": 0.08549812436103821,
+      "logps/chosen": -80.69041442871094,
+      "logps/rejected": -85.47969055175781,
+      "loss": 2.3022,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7405030727386475,
+      "rewards/margins": 0.4421282112598419,
+      "rewards/rejected": -2.182631492614746,
+      "step": 424
+    },
+    {
+      "epoch": 0.7720254314259763,
+      "grad_norm": 1.7834622859954834,
+      "learning_rate": 9.196393488042213e-06,
+      "logits/chosen": 0.06559979915618896,
+      "logits/rejected": 0.05396304652094841,
+      "logps/chosen": -76.51896667480469,
+      "logps/rejected": -88.19717407226562,
+      "loss": 2.5216,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.050562620162964,
+      "rewards/margins": 0.5811472535133362,
+      "rewards/rejected": -2.631709575653076,
+      "step": 425
+    },
+    {
+      "epoch": 0.773841961852861,
+      "grad_norm": 1.5034058094024658,
+      "learning_rate": 9.192724835797956e-06,
+      "logits/chosen": 0.10820607095956802,
+      "logits/rejected": 0.14427171647548676,
+      "logps/chosen": -81.0079116821289,
+      "logps/rejected": -95.6849136352539,
+      "loss": 2.0472,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.19939923286438,
+      "rewards/margins": 0.8476728796958923,
+      "rewards/rejected": -3.047071933746338,
+      "step": 426
+    },
+    {
+      "epoch": 0.7756584922797457,
+      "grad_norm": 1.7205973863601685,
+      "learning_rate": 9.189048653444796e-06,
+      "logits/chosen": 0.08522580564022064,
+      "logits/rejected": 0.09135682880878448,
+      "logps/chosen": -77.4021224975586,
+      "logps/rejected": -85.47643280029297,
+      "loss": 2.2618,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.002034902572632,
+      "rewards/margins": 0.5705874562263489,
+      "rewards/rejected": -2.572622776031494,
+      "step": 427
+    },
+    {
+      "epoch": 0.7774750227066304,
+      "grad_norm": 2.131197214126587,
+      "learning_rate": 9.185364948479109e-06,
+      "logits/chosen": 0.05514610558748245,
+      "logits/rejected": 0.0597330704331398,
+      "logps/chosen": -86.83938598632812,
+      "logps/rejected": -94.71990966796875,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.088331699371338,
+      "rewards/margins": 0.47913864254951477,
+      "rewards/rejected": -2.567470073699951,
+      "step": 428
+    },
+    {
+      "epoch": 0.779291553133515,
+      "grad_norm": 3.0228030681610107,
+      "learning_rate": 9.181673728412605e-06,
+      "logits/chosen": 0.09502460807561874,
+      "logits/rejected": 0.10438862442970276,
+      "logps/chosen": -77.20478820800781,
+      "logps/rejected": -80.07164764404297,
+      "loss": 2.5924,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.107954978942871,
+      "rewards/margins": 0.4795181155204773,
+      "rewards/rejected": -2.5874733924865723,
+      "step": 429
+    },
+    {
+      "epoch": 0.7811080835603996,
+      "grad_norm": 2.630490779876709,
+      "learning_rate": 9.17797500077233e-06,
+      "logits/chosen": 0.053769052028656006,
+      "logits/rejected": 0.1357191652059555,
+      "logps/chosen": -74.01277160644531,
+      "logps/rejected": -86.05644989013672,
+      "loss": 2.6208,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0179696083068848,
+      "rewards/margins": 0.3262583613395691,
+      "rewards/rejected": -2.3442280292510986,
+      "step": 430
+    },
+    {
+      "epoch": 0.7829246139872843,
+      "grad_norm": 1.7841664552688599,
+      "learning_rate": 9.174268773100636e-06,
+      "logits/chosen": 0.09468546509742737,
+      "logits/rejected": 0.16804733872413635,
+      "logps/chosen": -80.07379913330078,
+      "logps/rejected": -87.17518615722656,
+      "loss": 2.4676,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.27384877204895,
+      "rewards/margins": 0.4068894386291504,
+      "rewards/rejected": -2.6807379722595215,
+      "step": 431
+    },
+    {
+      "epoch": 0.784741144414169,
+      "grad_norm": 1.7030746936798096,
+      "learning_rate": 9.170555052955158e-06,
+      "logits/chosen": 0.10446357727050781,
+      "logits/rejected": 0.1235620528459549,
+      "logps/chosen": -89.81584930419922,
+      "logps/rejected": -94.43669128417969,
+      "loss": 2.3366,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.98858642578125,
+      "rewards/margins": 0.5402320623397827,
+      "rewards/rejected": -2.5288188457489014,
+      "step": 432
+    },
+    {
+      "epoch": 0.7865576748410535,
+      "grad_norm": 1.4258465766906738,
+      "learning_rate": 9.166833847908825e-06,
+      "logits/chosen": 0.10827025771141052,
+      "logits/rejected": 0.10875076055526733,
+      "logps/chosen": -69.29400634765625,
+      "logps/rejected": -80.72003936767578,
+      "loss": 2.2501,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9478650093078613,
+      "rewards/margins": 0.5940475463867188,
+      "rewards/rejected": -2.54191255569458,
+      "step": 433
+    },
+    {
+      "epoch": 0.7883742052679382,
+      "grad_norm": 1.8069814443588257,
+      "learning_rate": 9.163105165549819e-06,
+      "logits/chosen": 0.08856840431690216,
+      "logits/rejected": 0.07843751460313797,
+      "logps/chosen": -79.93275451660156,
+      "logps/rejected": -83.75475311279297,
+      "loss": 2.5214,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0315518379211426,
+      "rewards/margins": 0.45457565784454346,
+      "rewards/rejected": -2.4861276149749756,
+      "step": 434
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 1.7881503105163574,
+      "learning_rate": 9.159369013481574e-06,
+      "logits/chosen": 0.08166562020778656,
+      "logits/rejected": 0.08240097761154175,
+      "logps/chosen": -92.98298645019531,
+      "logps/rejected": -96.15763092041016,
+      "loss": 2.053,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4435133934020996,
+      "rewards/margins": 0.7140956521034241,
+      "rewards/rejected": -3.157609462738037,
+      "step": 435
+    },
+    {
+      "epoch": 0.7920072661217076,
+      "grad_norm": 1.7574496269226074,
+      "learning_rate": 9.155625399322754e-06,
+      "logits/chosen": 0.08114133030176163,
+      "logits/rejected": 0.11279730498790741,
+      "logps/chosen": -74.07096862792969,
+      "logps/rejected": -80.94386291503906,
+      "loss": 2.4289,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2497410774230957,
+      "rewards/margins": 0.4424628019332886,
+      "rewards/rejected": -2.6922037601470947,
+      "step": 436
+    },
+    {
+      "epoch": 0.7938237965485921,
+      "grad_norm": 1.4814083576202393,
+      "learning_rate": 9.15187433070724e-06,
+      "logits/chosen": 0.026073571294546127,
+      "logits/rejected": 0.1382063925266266,
+      "logps/chosen": -72.39479064941406,
+      "logps/rejected": -86.76889038085938,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9130635261535645,
+      "rewards/margins": 0.6362709999084473,
+      "rewards/rejected": -2.549334764480591,
+      "step": 437
+    },
+    {
+      "epoch": 0.7956403269754768,
+      "grad_norm": 1.6441129446029663,
+      "learning_rate": 9.148115815284113e-06,
+      "logits/chosen": 0.07432619482278824,
+      "logits/rejected": 0.10215617716312408,
+      "logps/chosen": -77.05310821533203,
+      "logps/rejected": -80.81717681884766,
+      "loss": 2.338,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9012683629989624,
+      "rewards/margins": 0.43776583671569824,
+      "rewards/rejected": -2.339034080505371,
+      "step": 438
+    },
+    {
+      "epoch": 0.7974568574023615,
+      "grad_norm": 1.8839560747146606,
+      "learning_rate": 9.144349860717643e-06,
+      "logits/chosen": 0.07576426863670349,
+      "logits/rejected": 0.08868670463562012,
+      "logps/chosen": -74.13190460205078,
+      "logps/rejected": -80.07232666015625,
+      "loss": 2.3485,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9554617404937744,
+      "rewards/margins": 0.531810998916626,
+      "rewards/rejected": -2.4872727394104004,
+      "step": 439
+    },
+    {
+      "epoch": 0.7992733878292462,
+      "grad_norm": 1.826811671257019,
+      "learning_rate": 9.140576474687263e-06,
+      "logits/chosen": 0.09120994061231613,
+      "logits/rejected": 0.08987519890069962,
+      "logps/chosen": -75.1444320678711,
+      "logps/rejected": -78.58987426757812,
+      "loss": 2.6513,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158261299133301,
+      "rewards/margins": 0.28016677498817444,
+      "rewards/rejected": -2.4384284019470215,
+      "step": 440
+    },
+    {
+      "epoch": 0.8010899182561307,
+      "grad_norm": 1.8647096157073975,
+      "learning_rate": 9.13679566488757e-06,
+      "logits/chosen": 0.12438184767961502,
+      "logits/rejected": 0.13951222598552704,
+      "logps/chosen": -83.79389953613281,
+      "logps/rejected": -90.33230590820312,
+      "loss": 2.3956,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0185930728912354,
+      "rewards/margins": 0.43685024976730347,
+      "rewards/rejected": -2.4554433822631836,
+      "step": 441
+    },
+    {
+      "epoch": 0.8029064486830154,
+      "grad_norm": 1.7610801458358765,
+      "learning_rate": 9.133007439028288e-06,
+      "logits/chosen": 0.08218151330947876,
+      "logits/rejected": 0.072816863656044,
+      "logps/chosen": -86.35633087158203,
+      "logps/rejected": -84.08189392089844,
+      "loss": 2.243,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8836033344268799,
+      "rewards/margins": 0.516740083694458,
+      "rewards/rejected": -2.400343418121338,
+      "step": 442
+    },
+    {
+      "epoch": 0.8047229791099001,
+      "grad_norm": 1.6049703359603882,
+      "learning_rate": 9.129211804834271e-06,
+      "logits/chosen": 0.059631846845149994,
+      "logits/rejected": 0.09744230657815933,
+      "logps/chosen": -76.90062713623047,
+      "logps/rejected": -82.29356384277344,
+      "loss": 2.2574,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9436204433441162,
+      "rewards/margins": 0.47444695234298706,
+      "rewards/rejected": -2.418067216873169,
+      "step": 443
+    },
+    {
+      "epoch": 0.8065395095367848,
+      "grad_norm": 1.9819791316986084,
+      "learning_rate": 9.12540877004548e-06,
+      "logits/chosen": 0.07371871173381805,
+      "logits/rejected": 0.058843065053224564,
+      "logps/chosen": -78.42516326904297,
+      "logps/rejected": -82.23822784423828,
+      "loss": 2.6959,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.273909330368042,
+      "rewards/margins": 0.3514346182346344,
+      "rewards/rejected": -2.6253440380096436,
+      "step": 444
+    },
+    {
+      "epoch": 0.8083560399636693,
+      "grad_norm": 1.9540653228759766,
+      "learning_rate": 9.12159834241696e-06,
+      "logits/chosen": 0.13155074417591095,
+      "logits/rejected": 0.11775672435760498,
+      "logps/chosen": -69.17151641845703,
+      "logps/rejected": -77.9190902709961,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.073258638381958,
+      "rewards/margins": 0.5955159664154053,
+      "rewards/rejected": -2.6687746047973633,
+      "step": 445
+    },
+    {
+      "epoch": 0.810172570390554,
+      "grad_norm": 1.9634339809417725,
+      "learning_rate": 9.117780529718843e-06,
+      "logits/chosen": 0.05265370383858681,
+      "logits/rejected": 0.09651105850934982,
+      "logps/chosen": -78.77818298339844,
+      "logps/rejected": -84.63584899902344,
+      "loss": 2.2157,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.084688186645508,
+      "rewards/margins": 0.541061282157898,
+      "rewards/rejected": -2.625749349594116,
+      "step": 446
+    },
+    {
+      "epoch": 0.8119891008174387,
+      "grad_norm": 1.9086997509002686,
+      "learning_rate": 9.113955339736309e-06,
+      "logits/chosen": 0.08473093807697296,
+      "logits/rejected": 0.047919195145368576,
+      "logps/chosen": -83.44226837158203,
+      "logps/rejected": -87.24287414550781,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9542481899261475,
+      "rewards/margins": 0.4577428698539734,
+      "rewards/rejected": -2.4119908809661865,
+      "step": 447
+    },
+    {
+      "epoch": 0.8138056312443234,
+      "grad_norm": 1.6967483758926392,
+      "learning_rate": 9.11012278026959e-06,
+      "logits/chosen": 0.062431350350379944,
+      "logits/rejected": 0.10980932414531708,
+      "logps/chosen": -90.98541259765625,
+      "logps/rejected": -102.51634979248047,
+      "loss": 2.3339,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1866049766540527,
+      "rewards/margins": 0.6549904346466064,
+      "rewards/rejected": -2.841595411300659,
+      "step": 448
+    },
+    {
+      "epoch": 0.815622161671208,
+      "grad_norm": 2.295504331588745,
+      "learning_rate": 9.106282859133936e-06,
+      "logits/chosen": 0.04481218010187149,
+      "logits/rejected": 0.002619542181491852,
+      "logps/chosen": -81.81380462646484,
+      "logps/rejected": -80.60113525390625,
+      "loss": 2.8324,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.1727848052978516,
+      "rewards/margins": 0.2777697443962097,
+      "rewards/rejected": -2.450554609298706,
+      "step": 449
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 1.7951968908309937,
+      "learning_rate": 9.102435584159623e-06,
+      "logits/chosen": 0.029065577313303947,
+      "logits/rejected": 0.11651361733675003,
+      "logps/chosen": -78.08231353759766,
+      "logps/rejected": -88.59742736816406,
+      "loss": 2.5562,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.2762022018432617,
+      "rewards/margins": 0.4410257041454315,
+      "rewards/rejected": -2.7172276973724365,
+      "step": 450
+    },
+    {
+      "epoch": 0.8192552225249773,
+      "grad_norm": 1.6301803588867188,
+      "learning_rate": 9.098580963191908e-06,
+      "logits/chosen": 0.06471782922744751,
+      "logits/rejected": 0.11030608415603638,
+      "logps/chosen": -78.3387451171875,
+      "logps/rejected": -91.82540130615234,
+      "loss": 2.1826,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.885233998298645,
+      "rewards/margins": 0.6971657872200012,
+      "rewards/rejected": -2.582399845123291,
+      "step": 451
+    },
+    {
+      "epoch": 0.821071752951862,
+      "grad_norm": 1.705962061882019,
+      "learning_rate": 9.094719004091039e-06,
+      "logits/chosen": 0.12591060996055603,
+      "logits/rejected": 0.1107010543346405,
+      "logps/chosen": -76.82783508300781,
+      "logps/rejected": -84.23365020751953,
+      "loss": 2.2843,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.04795241355896,
+      "rewards/margins": 0.5103356242179871,
+      "rewards/rejected": -2.558288335800171,
+      "step": 452
+    },
+    {
+      "epoch": 0.8228882833787466,
+      "grad_norm": 1.6535886526107788,
+      "learning_rate": 9.090849714732217e-06,
+      "logits/chosen": 0.11322569847106934,
+      "logits/rejected": 0.13092264533042908,
+      "logps/chosen": -80.16563415527344,
+      "logps/rejected": -82.96965026855469,
+      "loss": 2.3556,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9074076414108276,
+      "rewards/margins": 0.39431121945381165,
+      "rewards/rejected": -2.3017189502716064,
+      "step": 453
+    },
+    {
+      "epoch": 0.8247048138056312,
+      "grad_norm": 1.7194557189941406,
+      "learning_rate": 9.086973103005602e-06,
+      "logits/chosen": 0.1069367378950119,
+      "logits/rejected": 0.06237747147679329,
+      "logps/chosen": -72.59986877441406,
+      "logps/rejected": -77.62615203857422,
+      "loss": 2.2768,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.023085355758667,
+      "rewards/margins": 0.6052648425102234,
+      "rewards/rejected": -2.6283504962921143,
+      "step": 454
+    },
+    {
+      "epoch": 0.8265213442325159,
+      "grad_norm": 1.7921446561813354,
+      "learning_rate": 9.08308917681628e-06,
+      "logits/chosen": 0.12843580543994904,
+      "logits/rejected": 0.08002308011054993,
+      "logps/chosen": -80.6107177734375,
+      "logps/rejected": -80.92134857177734,
+      "loss": 2.3881,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9404933452606201,
+      "rewards/margins": 0.37478840351104736,
+      "rewards/rejected": -2.315281629562378,
+      "step": 455
+    },
+    {
+      "epoch": 0.8283378746594006,
+      "grad_norm": 2.3353819847106934,
+      "learning_rate": 9.07919794408425e-06,
+      "logits/chosen": 0.05452323704957962,
+      "logits/rejected": 0.12883196771144867,
+      "logps/chosen": -83.56001281738281,
+      "logps/rejected": -95.23291015625,
+      "loss": 2.8503,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.2597994804382324,
+      "rewards/margins": 0.4092620015144348,
+      "rewards/rejected": -2.6690614223480225,
+      "step": 456
+    },
+    {
+      "epoch": 0.8301544050862852,
+      "grad_norm": 1.6829804182052612,
+      "learning_rate": 9.075299412744417e-06,
+      "logits/chosen": 0.1223950982093811,
+      "logits/rejected": 0.10596577078104019,
+      "logps/chosen": -77.75115203857422,
+      "logps/rejected": -83.63490295410156,
+      "loss": 2.1639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.026258945465088,
+      "rewards/margins": 0.6107547879219055,
+      "rewards/rejected": -2.6370139122009277,
+      "step": 457
+    },
+    {
+      "epoch": 0.8319709355131698,
+      "grad_norm": 2.053755283355713,
+      "learning_rate": 9.07139359074656e-06,
+      "logits/chosen": 0.08021271228790283,
+      "logits/rejected": 0.11617676913738251,
+      "logps/chosen": -75.00261688232422,
+      "logps/rejected": -87.25222778320312,
+      "loss": 2.4048,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3207621574401855,
+      "rewards/margins": 0.5956254005432129,
+      "rewards/rejected": -2.9163873195648193,
+      "step": 458
+    },
+    {
+      "epoch": 0.8337874659400545,
+      "grad_norm": 2.1006064414978027,
+      "learning_rate": 9.067480486055333e-06,
+      "logits/chosen": 0.1115679144859314,
+      "logits/rejected": 0.08167213946580887,
+      "logps/chosen": -81.78447723388672,
+      "logps/rejected": -82.21318817138672,
+      "loss": 2.8725,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.165053367614746,
+      "rewards/margins": 0.2777022123336792,
+      "rewards/rejected": -2.442755699157715,
+      "step": 459
+    },
+    {
+      "epoch": 0.8356039963669392,
+      "grad_norm": 1.6339340209960938,
+      "learning_rate": 9.063560106650238e-06,
+      "logits/chosen": 0.0966312363743782,
+      "logits/rejected": 0.09943810105323792,
+      "logps/chosen": -71.0906982421875,
+      "logps/rejected": -76.22811126708984,
+      "loss": 2.2922,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.0007636547088623,
+      "rewards/margins": 0.47423097491264343,
+      "rewards/rejected": -2.474994659423828,
+      "step": 460
+    },
+    {
+      "epoch": 0.8374205267938238,
+      "grad_norm": 1.8504972457885742,
+      "learning_rate": 9.059632460525613e-06,
+      "logits/chosen": 0.09517084062099457,
+      "logits/rejected": 0.14127981662750244,
+      "logps/chosen": -79.06185913085938,
+      "logps/rejected": -89.11494445800781,
+      "loss": 2.3288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0016026496887207,
+      "rewards/margins": 0.5282158255577087,
+      "rewards/rejected": -2.529818534851074,
+      "step": 461
+    },
+    {
+      "epoch": 0.8392370572207084,
+      "grad_norm": 1.8151304721832275,
+      "learning_rate": 9.055697555690607e-06,
+      "logits/chosen": 0.08868349343538284,
+      "logits/rejected": 0.14090785384178162,
+      "logps/chosen": -77.26272583007812,
+      "logps/rejected": -85.14620971679688,
+      "loss": 2.3694,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.087087869644165,
+      "rewards/margins": 0.5032299757003784,
+      "rewards/rejected": -2.590317964553833,
+      "step": 462
+    },
+    {
+      "epoch": 0.8410535876475931,
+      "grad_norm": 1.8596280813217163,
+      "learning_rate": 9.051755400169182e-06,
+      "logits/chosen": 0.06272133439779282,
+      "logits/rejected": 0.12342572212219238,
+      "logps/chosen": -75.46456909179688,
+      "logps/rejected": -81.9759292602539,
+      "loss": 2.5167,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0463178157806396,
+      "rewards/margins": 0.3146142363548279,
+      "rewards/rejected": -2.3609323501586914,
+      "step": 463
+    },
+    {
+      "epoch": 0.8428701180744778,
+      "grad_norm": 1.7658417224884033,
+      "learning_rate": 9.047806002000075e-06,
+      "logits/chosen": -0.007259421981871128,
+      "logits/rejected": 0.09611339122056961,
+      "logps/chosen": -73.88439178466797,
+      "logps/rejected": -80.77063751220703,
+      "loss": 2.4938,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9998183250427246,
+      "rewards/margins": 0.4115346670150757,
+      "rewards/rejected": -2.4113528728485107,
+      "step": 464
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 1.6596916913986206,
+      "learning_rate": 9.043849369236799e-06,
+      "logits/chosen": 0.03223409131169319,
+      "logits/rejected": 0.05010119825601578,
+      "logps/chosen": -66.32562255859375,
+      "logps/rejected": -69.36509704589844,
+      "loss": 2.5881,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.909837245941162,
+      "rewards/margins": 0.2923263907432556,
+      "rewards/rejected": -2.2021636962890625,
+      "step": 465
+    },
+    {
+      "epoch": 0.846503178928247,
+      "grad_norm": 1.6832820177078247,
+      "learning_rate": 9.039885509947616e-06,
+      "logits/chosen": 0.06287454813718796,
+      "logits/rejected": 0.12746769189834595,
+      "logps/chosen": -76.6998519897461,
+      "logps/rejected": -87.36117553710938,
+      "loss": 2.3914,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.029526710510254,
+      "rewards/margins": 0.48344865441322327,
+      "rewards/rejected": -2.5129752159118652,
+      "step": 466
+    },
+    {
+      "epoch": 0.8483197093551317,
+      "grad_norm": 1.6972366571426392,
+      "learning_rate": 9.035914432215527e-06,
+      "logits/chosen": 0.1179896891117096,
+      "logits/rejected": 0.14000467956066132,
+      "logps/chosen": -75.3287353515625,
+      "logps/rejected": -78.60227966308594,
+      "loss": 2.4194,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.088515043258667,
+      "rewards/margins": 0.42544490098953247,
+      "rewards/rejected": -2.513960123062134,
+      "step": 467
+    },
+    {
+      "epoch": 0.8501362397820164,
+      "grad_norm": 2.5104973316192627,
+      "learning_rate": 9.031936144138247e-06,
+      "logits/chosen": 0.08423591405153275,
+      "logits/rejected": 0.1387752890586853,
+      "logps/chosen": -78.6176528930664,
+      "logps/rejected": -83.52703094482422,
+      "loss": 2.7262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1089184284210205,
+      "rewards/margins": 0.44402289390563965,
+      "rewards/rejected": -2.55294132232666,
+      "step": 468
+    },
+    {
+      "epoch": 0.851952770208901,
+      "grad_norm": 1.944931149482727,
+      "learning_rate": 9.027950653828202e-06,
+      "logits/chosen": 0.12252221256494522,
+      "logits/rejected": 0.14959384500980377,
+      "logps/chosen": -72.63184356689453,
+      "logps/rejected": -70.89137268066406,
+      "loss": 2.7323,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9612714052200317,
+      "rewards/margins": 0.24262896180152893,
+      "rewards/rejected": -2.2039003372192383,
+      "step": 469
+    },
+    {
+      "epoch": 0.8537693006357856,
+      "grad_norm": 1.6699979305267334,
+      "learning_rate": 9.0239579694125e-06,
+      "logits/chosen": 0.11827167868614197,
+      "logits/rejected": 0.11933022737503052,
+      "logps/chosen": -70.45188903808594,
+      "logps/rejected": -76.2336654663086,
+      "loss": 2.457,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.854758381843567,
+      "rewards/margins": 0.47020024061203003,
+      "rewards/rejected": -2.3249588012695312,
+      "step": 470
+    },
+    {
+      "epoch": 0.8555858310626703,
+      "grad_norm": 1.5790536403656006,
+      "learning_rate": 9.019958099032919e-06,
+      "logits/chosen": 0.04856352508068085,
+      "logits/rejected": 0.11418008804321289,
+      "logps/chosen": -75.7929458618164,
+      "logps/rejected": -86.55818176269531,
+      "loss": 2.2809,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.003899335861206,
+      "rewards/margins": 0.5960561037063599,
+      "rewards/rejected": -2.5999553203582764,
+      "step": 471
+    },
+    {
+      "epoch": 0.857402361489555,
+      "grad_norm": 1.3594144582748413,
+      "learning_rate": 9.015951050845891e-06,
+      "logits/chosen": 0.10341258347034454,
+      "logits/rejected": 0.12394269555807114,
+      "logps/chosen": -78.8187255859375,
+      "logps/rejected": -84.04349517822266,
+      "loss": 1.9536,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9372440576553345,
+      "rewards/margins": 0.6742948889732361,
+      "rewards/rejected": -2.611538887023926,
+      "step": 472
+    },
+    {
+      "epoch": 0.8592188919164396,
+      "grad_norm": 1.7592689990997314,
+      "learning_rate": 9.011936833022485e-06,
+      "logits/chosen": 0.10970023274421692,
+      "logits/rejected": 0.09040558338165283,
+      "logps/chosen": -80.59556579589844,
+      "logps/rejected": -84.2996826171875,
+      "loss": 2.5658,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.935511827468872,
+      "rewards/margins": 0.35643547773361206,
+      "rewards/rejected": -2.29194712638855,
+      "step": 473
+    },
+    {
+      "epoch": 0.8610354223433242,
+      "grad_norm": 1.6418087482452393,
+      "learning_rate": 9.00791545374839e-06,
+      "logits/chosen": 0.16755647957324982,
+      "logits/rejected": 0.1745520681142807,
+      "logps/chosen": -82.31010437011719,
+      "logps/rejected": -85.66128540039062,
+      "loss": 2.1854,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.9519617557525635,
+      "rewards/margins": 0.5059336423873901,
+      "rewards/rejected": -2.457895278930664,
+      "step": 474
+    },
+    {
+      "epoch": 0.8628519527702089,
+      "grad_norm": 1.4616092443466187,
+      "learning_rate": 9.0038869212239e-06,
+      "logits/chosen": 0.03978392109274864,
+      "logits/rejected": 0.1235240027308464,
+      "logps/chosen": -77.78251647949219,
+      "logps/rejected": -85.66397094726562,
+      "loss": 2.1995,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8721034526824951,
+      "rewards/margins": 0.45375847816467285,
+      "rewards/rejected": -2.325861930847168,
+      "step": 475
+    },
+    {
+      "epoch": 0.8646684831970936,
+      "grad_norm": 1.838180422782898,
+      "learning_rate": 8.99985124366389e-06,
+      "logits/chosen": 0.08277668058872223,
+      "logits/rejected": 0.035900432616472244,
+      "logps/chosen": -83.23967742919922,
+      "logps/rejected": -83.16535949707031,
+      "loss": 2.7347,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.2339279651641846,
+      "rewards/margins": 0.16711921989917755,
+      "rewards/rejected": -2.4010472297668457,
+      "step": 476
+    },
+    {
+      "epoch": 0.8664850136239782,
+      "grad_norm": 1.6746636629104614,
+      "learning_rate": 8.995808429297815e-06,
+      "logits/chosen": 0.14687396585941315,
+      "logits/rejected": 0.09570137411355972,
+      "logps/chosen": -78.79898071289062,
+      "logps/rejected": -77.9619369506836,
+      "loss": 2.4449,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8334993124008179,
+      "rewards/margins": 0.3442971706390381,
+      "rewards/rejected": -2.1777963638305664,
+      "step": 477
+    },
+    {
+      "epoch": 0.8683015440508629,
+      "grad_norm": 1.7042535543441772,
+      "learning_rate": 8.991758486369675e-06,
+      "logits/chosen": 0.047281138598918915,
+      "logits/rejected": -0.023666318506002426,
+      "logps/chosen": -83.798583984375,
+      "logps/rejected": -91.88795471191406,
+      "loss": 2.3072,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9728612899780273,
+      "rewards/margins": 0.5659723281860352,
+      "rewards/rejected": -2.5388338565826416,
+      "step": 478
+    },
+    {
+      "epoch": 0.8701180744777475,
+      "grad_norm": 1.5552817583084106,
+      "learning_rate": 8.987701423138007e-06,
+      "logits/chosen": 0.0772751122713089,
+      "logits/rejected": 0.08708472549915314,
+      "logps/chosen": -72.42391204833984,
+      "logps/rejected": -78.93070220947266,
+      "loss": 2.2815,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9803509712219238,
+      "rewards/margins": 0.5386512279510498,
+      "rewards/rejected": -2.5190021991729736,
+      "step": 479
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.6372044086456299,
+      "learning_rate": 8.983637247875872e-06,
+      "logits/chosen": -0.03352706879377365,
+      "logits/rejected": 0.01581352949142456,
+      "logps/chosen": -81.29138946533203,
+      "logps/rejected": -89.1744155883789,
+      "loss": 2.09,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9441754817962646,
+      "rewards/margins": 0.6137264370918274,
+      "rewards/rejected": -2.5579018592834473,
+      "step": 480
+    },
+    {
+      "epoch": 0.8737511353315168,
+      "grad_norm": 1.6375274658203125,
+      "learning_rate": 8.979565968870831e-06,
+      "logits/chosen": 0.07507088780403137,
+      "logits/rejected": 0.1289597451686859,
+      "logps/chosen": -74.78109741210938,
+      "logps/rejected": -84.55204772949219,
+      "loss": 2.4566,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.9041385650634766,
+      "rewards/margins": 0.42814433574676514,
+      "rewards/rejected": -2.332282781600952,
+      "step": 481
+    },
+    {
+      "epoch": 0.8755676657584015,
+      "grad_norm": 1.5601258277893066,
+      "learning_rate": 8.975487594424927e-06,
+      "logits/chosen": 0.012452262453734875,
+      "logits/rejected": 0.06930352002382278,
+      "logps/chosen": -83.01705169677734,
+      "logps/rejected": -92.38286590576172,
+      "loss": 2.2736,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.013011932373047,
+      "rewards/margins": 0.5085774064064026,
+      "rewards/rejected": -2.5215890407562256,
+      "step": 482
+    },
+    {
+      "epoch": 0.8773841961852861,
+      "grad_norm": 1.8685195446014404,
+      "learning_rate": 8.971402132854677e-06,
+      "logits/chosen": 0.09415426105260849,
+      "logits/rejected": 0.12719042599201202,
+      "logps/chosen": -79.85440063476562,
+      "logps/rejected": -82.80875396728516,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.114530086517334,
+      "rewards/margins": 0.46328115463256836,
+      "rewards/rejected": -2.5778112411499023,
+      "step": 483
+    },
+    {
+      "epoch": 0.8792007266121707,
+      "grad_norm": 1.8812834024429321,
+      "learning_rate": 8.967309592491052e-06,
+      "logits/chosen": 0.13547101616859436,
+      "logits/rejected": 0.11872326582670212,
+      "logps/chosen": -74.42122650146484,
+      "logps/rejected": -81.9700927734375,
+      "loss": 2.5016,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.1828627586364746,
+      "rewards/margins": 0.44510167837142944,
+      "rewards/rejected": -2.627964496612549,
+      "step": 484
+    },
+    {
+      "epoch": 0.8810172570390554,
+      "grad_norm": 1.510517954826355,
+      "learning_rate": 8.963209981679451e-06,
+      "logits/chosen": 0.03274242952466011,
+      "logits/rejected": 0.07188954204320908,
+      "logps/chosen": -82.55563354492188,
+      "logps/rejected": -100.33609771728516,
+      "loss": 1.9177,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9487042427062988,
+      "rewards/margins": 0.8189151287078857,
+      "rewards/rejected": -2.7676191329956055,
+      "step": 485
+    },
+    {
+      "epoch": 0.8828337874659401,
+      "grad_norm": 1.5973646640777588,
+      "learning_rate": 8.959103308779696e-06,
+      "logits/chosen": 0.0365552082657814,
+      "logits/rejected": 0.051129020750522614,
+      "logps/chosen": -72.4110107421875,
+      "logps/rejected": -83.10322570800781,
+      "loss": 2.1547,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9375314712524414,
+      "rewards/margins": 0.6991211175918579,
+      "rewards/rejected": -2.6366524696350098,
+      "step": 486
+    },
+    {
+      "epoch": 0.8846503178928247,
+      "grad_norm": 1.6707032918930054,
+      "learning_rate": 8.954989582166009e-06,
+      "logits/chosen": 0.0635693296790123,
+      "logits/rejected": 0.009854275733232498,
+      "logps/chosen": -87.12749481201172,
+      "logps/rejected": -88.9491195678711,
+      "loss": 2.2043,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9257593154907227,
+      "rewards/margins": 0.5429801344871521,
+      "rewards/rejected": -2.4687397480010986,
+      "step": 487
+    },
+    {
+      "epoch": 0.8864668483197093,
+      "grad_norm": 1.4422773122787476,
+      "learning_rate": 8.95086881022699e-06,
+      "logits/chosen": -0.024460218846797943,
+      "logits/rejected": 0.14137038588523865,
+      "logps/chosen": -74.60330200195312,
+      "logps/rejected": -95.01730346679688,
+      "loss": 2.0246,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.054253339767456,
+      "rewards/margins": 0.8260326981544495,
+      "rewards/rejected": -2.88028621673584,
+      "step": 488
+    },
+    {
+      "epoch": 0.888283378746594,
+      "grad_norm": 1.8121236562728882,
+      "learning_rate": 8.946741001365616e-06,
+      "logits/chosen": 0.05854415148496628,
+      "logits/rejected": 0.13758361339569092,
+      "logps/chosen": -74.46614074707031,
+      "logps/rejected": -83.96118927001953,
+      "loss": 2.5066,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.105980157852173,
+      "rewards/margins": 0.41478782892227173,
+      "rewards/rejected": -2.520768165588379,
+      "step": 489
+    },
+    {
+      "epoch": 0.8900999091734787,
+      "grad_norm": 2.023728370666504,
+      "learning_rate": 8.942606163999205e-06,
+      "logits/chosen": 0.04619833081960678,
+      "logits/rejected": 0.09330146014690399,
+      "logps/chosen": -75.35594940185547,
+      "logps/rejected": -89.18836975097656,
+      "loss": 2.2655,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9396051168441772,
+      "rewards/margins": 0.6726698875427246,
+      "rewards/rejected": -2.6122751235961914,
+      "step": 490
+    },
+    {
+      "epoch": 0.8919164396003633,
+      "grad_norm": 1.559735894203186,
+      "learning_rate": 8.938464306559412e-06,
+      "logits/chosen": 0.093504399061203,
+      "logits/rejected": 0.09722252935171127,
+      "logps/chosen": -80.77140808105469,
+      "logps/rejected": -87.86373901367188,
+      "loss": 2.1676,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.1073269844055176,
+      "rewards/margins": 0.5740070939064026,
+      "rewards/rejected": -2.6813340187072754,
+      "step": 491
+    },
+    {
+      "epoch": 0.8937329700272479,
+      "grad_norm": 1.6771574020385742,
+      "learning_rate": 8.934315437492203e-06,
+      "logits/chosen": 0.06368491798639297,
+      "logits/rejected": 0.04900998994708061,
+      "logps/chosen": -78.2313003540039,
+      "logps/rejected": -92.83306121826172,
+      "loss": 2.3435,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.110978364944458,
+      "rewards/margins": 0.6751725673675537,
+      "rewards/rejected": -2.786151170730591,
+      "step": 492
+    },
+    {
+      "epoch": 0.8955495004541326,
+      "grad_norm": 1.651183843612671,
+      "learning_rate": 8.930159565257846e-06,
+      "logits/chosen": 0.08393608778715134,
+      "logits/rejected": 0.0475153923034668,
+      "logps/chosen": -77.8431396484375,
+      "logps/rejected": -88.03681945800781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0274362564086914,
+      "rewards/margins": 0.8075671195983887,
+      "rewards/rejected": -2.835003137588501,
+      "step": 493
+    },
+    {
+      "epoch": 0.8973660308810173,
+      "grad_norm": 2.262662172317505,
+      "learning_rate": 8.925996698330887e-06,
+      "logits/chosen": 0.014226208440959454,
+      "logits/rejected": 0.09588178992271423,
+      "logps/chosen": -88.56185913085938,
+      "logps/rejected": -95.1131591796875,
+      "loss": 2.5853,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4549288749694824,
+      "rewards/margins": 0.5275144577026367,
+      "rewards/rejected": -2.982443332672119,
+      "step": 494
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 1.6070181131362915,
+      "learning_rate": 8.92182684520014e-06,
+      "logits/chosen": 0.11615607142448425,
+      "logits/rejected": 0.16645964980125427,
+      "logps/chosen": -78.97079467773438,
+      "logps/rejected": -90.49947357177734,
+      "loss": 2.1072,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.175706386566162,
+      "rewards/margins": 0.7435849905014038,
+      "rewards/rejected": -2.9192914962768555,
+      "step": 495
+    },
+    {
+      "epoch": 0.9009990917347865,
+      "grad_norm": 1.8343369960784912,
+      "learning_rate": 8.917650014368658e-06,
+      "logits/chosen": 0.12583515048027039,
+      "logits/rejected": 0.07308925688266754,
+      "logps/chosen": -83.38356018066406,
+      "logps/rejected": -90.73155975341797,
+      "loss": 2.3116,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3124871253967285,
+      "rewards/margins": 0.5869801640510559,
+      "rewards/rejected": -2.8994674682617188,
+      "step": 496
+    },
+    {
+      "epoch": 0.9028156221616712,
+      "grad_norm": 1.7175630331039429,
+      "learning_rate": 8.913466214353728e-06,
+      "logits/chosen": 0.14971572160720825,
+      "logits/rejected": 0.17372727394104004,
+      "logps/chosen": -86.66060638427734,
+      "logps/rejected": -94.42353820800781,
+      "loss": 2.3541,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.943577766418457,
+      "rewards/margins": 0.5955328345298767,
+      "rewards/rejected": -2.5391108989715576,
+      "step": 497
+    },
+    {
+      "epoch": 0.9046321525885559,
+      "grad_norm": 1.754708170890808,
+      "learning_rate": 8.909275453686845e-06,
+      "logits/chosen": 0.07779194414615631,
+      "logits/rejected": 0.09394712746143341,
+      "logps/chosen": -71.64151000976562,
+      "logps/rejected": -78.53131103515625,
+      "loss": 2.0436,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0102486610412598,
+      "rewards/margins": 0.6668623685836792,
+      "rewards/rejected": -2.6771109104156494,
+      "step": 498
+    },
+    {
+      "epoch": 0.9064486830154405,
+      "grad_norm": 1.9303216934204102,
+      "learning_rate": 8.905077740913701e-06,
+      "logits/chosen": 0.15853236615657806,
+      "logits/rejected": 0.11452697217464447,
+      "logps/chosen": -81.46524047851562,
+      "logps/rejected": -89.1716537475586,
+      "loss": 2.2477,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.264355182647705,
+      "rewards/margins": 0.6477693915367126,
+      "rewards/rejected": -2.9121243953704834,
+      "step": 499
+    },
+    {
+      "epoch": 0.9082652134423251,
+      "grad_norm": 2.1294379234313965,
+      "learning_rate": 8.900873084594164e-06,
+      "logits/chosen": 0.029480352997779846,
+      "logits/rejected": 0.09827219694852829,
+      "logps/chosen": -95.93231964111328,
+      "logps/rejected": -106.21436309814453,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.3455631732940674,
+      "rewards/margins": 0.5977038145065308,
+      "rewards/rejected": -2.9432668685913086,
+      "step": 500
+    },
+    {
+      "epoch": 0.9100817438692098,
+      "grad_norm": 1.7347708940505981,
+      "learning_rate": 8.896661493302258e-06,
+      "logits/chosen": 0.08531993627548218,
+      "logits/rejected": 0.14993277192115784,
+      "logps/chosen": -79.09967041015625,
+      "logps/rejected": -90.50935363769531,
+      "loss": 2.2969,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9908722639083862,
+      "rewards/margins": 0.5710910558700562,
+      "rewards/rejected": -2.5619633197784424,
+      "step": 501
+    },
+    {
+      "epoch": 0.9118982742960945,
+      "grad_norm": 2.065664529800415,
+      "learning_rate": 8.892442975626152e-06,
+      "logits/chosen": 0.18976512551307678,
+      "logits/rejected": 0.13936059176921844,
+      "logps/chosen": -82.80276489257812,
+      "logps/rejected": -84.57154083251953,
+      "loss": 2.5968,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4863569736480713,
+      "rewards/margins": 0.3651903569698334,
+      "rewards/rejected": -2.8515477180480957,
+      "step": 502
+    },
+    {
+      "epoch": 0.9137148047229791,
+      "grad_norm": 2.054955005645752,
+      "learning_rate": 8.888217540168139e-06,
+      "logits/chosen": 0.01604822278022766,
+      "logits/rejected": 0.1370609700679779,
+      "logps/chosen": -74.58015441894531,
+      "logps/rejected": -96.80826568603516,
+      "loss": 2.169,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.397087574005127,
+      "rewards/margins": 0.9368999004364014,
+      "rewards/rejected": -3.333987236022949,
+      "step": 503
+    },
+    {
+      "epoch": 0.9155313351498637,
+      "grad_norm": 2.179532527923584,
+      "learning_rate": 8.883985195544617e-06,
+      "logits/chosen": 0.03752445429563522,
+      "logits/rejected": 0.024428365752100945,
+      "logps/chosen": -83.81852722167969,
+      "logps/rejected": -88.1905288696289,
+      "loss": 2.5361,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.239609956741333,
+      "rewards/margins": 0.4188295900821686,
+      "rewards/rejected": -2.6584396362304688,
+      "step": 504
+    },
+    {
+      "epoch": 0.9173478655767484,
+      "grad_norm": 1.9171282052993774,
+      "learning_rate": 8.879745950386075e-06,
+      "logits/chosen": 0.17986616492271423,
+      "logits/rejected": 0.20335282385349274,
+      "logps/chosen": -72.679931640625,
+      "logps/rejected": -71.41309356689453,
+      "loss": 2.5452,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.261596918106079,
+      "rewards/margins": 0.3737923502922058,
+      "rewards/rejected": -2.6353893280029297,
+      "step": 505
+    },
+    {
+      "epoch": 0.9191643960036331,
+      "grad_norm": 2.3437929153442383,
+      "learning_rate": 8.87549981333707e-06,
+      "logits/chosen": 0.09010382741689682,
+      "logits/rejected": 0.040441811084747314,
+      "logps/chosen": -81.56139373779297,
+      "logps/rejected": -83.21439361572266,
+      "loss": 2.74,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2574033737182617,
+      "rewards/margins": 0.28719204664230347,
+      "rewards/rejected": -2.544595241546631,
+      "step": 506
+    },
+    {
+      "epoch": 0.9209809264305178,
+      "grad_norm": 2.4240329265594482,
+      "learning_rate": 8.871246793056215e-06,
+      "logits/chosen": 0.02888420596718788,
+      "logits/rejected": 0.11542786657810211,
+      "logps/chosen": -84.60144805908203,
+      "logps/rejected": -100.05682373046875,
+      "loss": 2.7052,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.863104820251465,
+      "rewards/margins": 0.41941094398498535,
+      "rewards/rejected": -3.28251576423645,
+      "step": 507
+    },
+    {
+      "epoch": 0.9227974568574023,
+      "grad_norm": 1.9247711896896362,
+      "learning_rate": 8.866986898216157e-06,
+      "logits/chosen": 0.06140238791704178,
+      "logits/rejected": 0.05623817816376686,
+      "logps/chosen": -77.40584564208984,
+      "logps/rejected": -86.71392822265625,
+      "loss": 2.2445,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4979231357574463,
+      "rewards/margins": 0.6161274313926697,
+      "rewards/rejected": -3.11405086517334,
+      "step": 508
+    },
+    {
+      "epoch": 0.924613987284287,
+      "grad_norm": 1.854956030845642,
+      "learning_rate": 8.862720137503568e-06,
+      "logits/chosen": 0.06755004823207855,
+      "logits/rejected": 0.03762562945485115,
+      "logps/chosen": -86.03059387207031,
+      "logps/rejected": -90.12970733642578,
+      "loss": 2.1459,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3017396926879883,
+      "rewards/margins": 0.6131142973899841,
+      "rewards/rejected": -2.914853572845459,
+      "step": 509
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.8044302463531494,
+      "learning_rate": 8.858446519619113e-06,
+      "logits/chosen": 0.08446178585290909,
+      "logits/rejected": 0.08818987011909485,
+      "logps/chosen": -77.58124542236328,
+      "logps/rejected": -85.02752685546875,
+      "loss": 2.1311,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2378087043762207,
+      "rewards/margins": 0.7067832350730896,
+      "rewards/rejected": -2.944591999053955,
+      "step": 510
+    },
+    {
+      "epoch": 0.9282470481380564,
+      "grad_norm": 1.4898866415023804,
+      "learning_rate": 8.854166053277443e-06,
+      "logits/chosen": 0.1039918065071106,
+      "logits/rejected": 0.1396111100912094,
+      "logps/chosen": -87.59841918945312,
+      "logps/rejected": -101.05659484863281,
+      "loss": 2.1092,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1167173385620117,
+      "rewards/margins": 0.7099149227142334,
+      "rewards/rejected": -2.826632499694824,
+      "step": 511
+    },
+    {
+      "epoch": 0.9300635785649409,
+      "grad_norm": 1.8009706735610962,
+      "learning_rate": 8.849878747207175e-06,
+      "logits/chosen": 0.11744043976068497,
+      "logits/rejected": 0.07621707767248154,
+      "logps/chosen": -78.37222290039062,
+      "logps/rejected": -74.32958221435547,
+      "loss": 2.6399,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.4002914428710938,
+      "rewards/margins": 0.21068021655082703,
+      "rewards/rejected": -2.610971450805664,
+      "step": 512
+    },
+    {
+      "epoch": 0.9318801089918256,
+      "grad_norm": 1.6368399858474731,
+      "learning_rate": 8.845584610150871e-06,
+      "logits/chosen": 0.09938757866621017,
+      "logits/rejected": 0.12675486505031586,
+      "logps/chosen": -82.24690246582031,
+      "logps/rejected": -92.60913848876953,
+      "loss": 1.9078,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.0825133323669434,
+      "rewards/margins": 0.8327144384384155,
+      "rewards/rejected": -2.9152278900146484,
+      "step": 513
+    },
+    {
+      "epoch": 0.9336966394187103,
+      "grad_norm": 1.8117483854293823,
+      "learning_rate": 8.841283650865027e-06,
+      "logits/chosen": 0.1355382651090622,
+      "logits/rejected": 0.14869986474514008,
+      "logps/chosen": -85.03866577148438,
+      "logps/rejected": -90.55638885498047,
+      "loss": 2.1243,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3315932750701904,
+      "rewards/margins": 0.6421911120414734,
+      "rewards/rejected": -2.9737846851348877,
+      "step": 514
+    },
+    {
+      "epoch": 0.935513169845595,
+      "grad_norm": 1.785927414894104,
+      "learning_rate": 8.836975878120046e-06,
+      "logits/chosen": 0.11557039618492126,
+      "logits/rejected": 0.11716655641794205,
+      "logps/chosen": -83.99951171875,
+      "logps/rejected": -92.36119079589844,
+      "loss": 2.4256,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.439833402633667,
+      "rewards/margins": 0.38893458247184753,
+      "rewards/rejected": -2.8287675380706787,
+      "step": 515
+    },
+    {
+      "epoch": 0.9373297002724795,
+      "grad_norm": 3.526102066040039,
+      "learning_rate": 8.832661300700228e-06,
+      "logits/chosen": 0.10778439044952393,
+      "logits/rejected": 0.18004637956619263,
+      "logps/chosen": -84.20158386230469,
+      "logps/rejected": -91.33497619628906,
+      "loss": 2.3568,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2032201290130615,
+      "rewards/margins": 0.5167797207832336,
+      "rewards/rejected": -2.7200000286102295,
+      "step": 516
+    },
+    {
+      "epoch": 0.9391462306993642,
+      "grad_norm": 1.7708752155303955,
+      "learning_rate": 8.828339927403745e-06,
+      "logits/chosen": 0.0862819030880928,
+      "logits/rejected": 0.16991934180259705,
+      "logps/chosen": -87.4437026977539,
+      "logps/rejected": -96.60844421386719,
+      "loss": 2.3041,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.230825185775757,
+      "rewards/margins": 0.5197271108627319,
+      "rewards/rejected": -2.7505524158477783,
+      "step": 517
+    },
+    {
+      "epoch": 0.9409627611262489,
+      "grad_norm": 2.1336302757263184,
+      "learning_rate": 8.824011767042631e-06,
+      "logits/chosen": 0.12070942670106888,
+      "logits/rejected": 0.19643370807170868,
+      "logps/chosen": -75.97718811035156,
+      "logps/rejected": -83.26056671142578,
+      "loss": 2.3468,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3768177032470703,
+      "rewards/margins": 0.5026894211769104,
+      "rewards/rejected": -2.879507064819336,
+      "step": 518
+    },
+    {
+      "epoch": 0.9427792915531336,
+      "grad_norm": 2.6361515522003174,
+      "learning_rate": 8.819676828442758e-06,
+      "logits/chosen": 0.04481849446892738,
+      "logits/rejected": -0.011357773095369339,
+      "logps/chosen": -84.09026336669922,
+      "logps/rejected": -87.71636199951172,
+      "loss": 2.8427,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.4249284267425537,
+      "rewards/margins": 0.41216185688972473,
+      "rewards/rejected": -2.837090015411377,
+      "step": 519
+    },
+    {
+      "epoch": 0.9445958219800181,
+      "grad_norm": 2.1068665981292725,
+      "learning_rate": 8.815335120443822e-06,
+      "logits/chosen": 0.1595858335494995,
+      "logits/rejected": 0.14017102122306824,
+      "logps/chosen": -94.00933837890625,
+      "logps/rejected": -96.91961669921875,
+      "loss": 2.6559,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.4731483459472656,
+      "rewards/margins": 0.3141450881958008,
+      "rewards/rejected": -2.7872931957244873,
+      "step": 520
+    },
+    {
+      "epoch": 0.9464123524069028,
+      "grad_norm": 1.7320570945739746,
+      "learning_rate": 8.810986651899322e-06,
+      "logits/chosen": 0.10371927917003632,
+      "logits/rejected": 0.17988254129886627,
+      "logps/chosen": -78.32308197021484,
+      "logps/rejected": -88.74859619140625,
+      "loss": 2.1866,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3973731994628906,
+      "rewards/margins": 0.7199669480323792,
+      "rewards/rejected": -3.117340087890625,
+      "step": 521
+    },
+    {
+      "epoch": 0.9482288828337875,
+      "grad_norm": 1.6995161771774292,
+      "learning_rate": 8.80663143167654e-06,
+      "logits/chosen": 0.14279219508171082,
+      "logits/rejected": 0.1058904379606247,
+      "logps/chosen": -87.79972839355469,
+      "logps/rejected": -94.35545349121094,
+      "loss": 2.1217,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.979785680770874,
+      "rewards/margins": 0.6210839152336121,
+      "rewards/rejected": -2.600869655609131,
+      "step": 522
+    },
+    {
+      "epoch": 0.9500454132606722,
+      "grad_norm": 2.1860764026641846,
+      "learning_rate": 8.80226946865653e-06,
+      "logits/chosen": 0.11601082980632782,
+      "logits/rejected": 0.15693408250808716,
+      "logps/chosen": -88.86433410644531,
+      "logps/rejected": -92.43590545654297,
+      "loss": 2.5629,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.2875983715057373,
+      "rewards/margins": 0.39402255415916443,
+      "rewards/rejected": -2.6816210746765137,
+      "step": 523
+    },
+    {
+      "epoch": 0.9518619436875567,
+      "grad_norm": 1.829032063484192,
+      "learning_rate": 8.797900771734094e-06,
+      "logits/chosen": 0.12328370660543442,
+      "logits/rejected": 0.042389824986457825,
+      "logps/chosen": -77.34864807128906,
+      "logps/rejected": -78.61405944824219,
+      "loss": 2.5327,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2514185905456543,
+      "rewards/margins": 0.3621898591518402,
+      "rewards/rejected": -2.6136088371276855,
+      "step": 524
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 2.235381841659546,
+      "learning_rate": 8.793525349817765e-06,
+      "logits/chosen": 0.12495981156826019,
+      "logits/rejected": 0.13291439414024353,
+      "logps/chosen": -76.7927474975586,
+      "logps/rejected": -86.1208267211914,
+      "loss": 2.5188,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3426284790039062,
+      "rewards/margins": 0.46077418327331543,
+      "rewards/rejected": -2.8034026622772217,
+      "step": 525
+    },
+    {
+      "epoch": 0.9554950045413261,
+      "grad_norm": 1.8071073293685913,
+      "learning_rate": 8.78914321182979e-06,
+      "logits/chosen": 0.13614074885845184,
+      "logits/rejected": 0.1179000735282898,
+      "logps/chosen": -74.44164276123047,
+      "logps/rejected": -77.69221496582031,
+      "loss": 2.6082,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0528411865234375,
+      "rewards/margins": 0.29493850469589233,
+      "rewards/rejected": -2.3477797508239746,
+      "step": 526
+    },
+    {
+      "epoch": 0.9573115349682108,
+      "grad_norm": 1.8623837232589722,
+      "learning_rate": 8.784754366706115e-06,
+      "logits/chosen": 0.02768833190202713,
+      "logits/rejected": 0.09584817290306091,
+      "logps/chosen": -77.13914489746094,
+      "logps/rejected": -86.53968048095703,
+      "loss": 2.2244,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.199755907058716,
+      "rewards/margins": 0.6271941661834717,
+      "rewards/rejected": -2.8269503116607666,
+      "step": 527
+    },
+    {
+      "epoch": 0.9591280653950953,
+      "grad_norm": 1.5299558639526367,
+      "learning_rate": 8.780358823396352e-06,
+      "logits/chosen": 0.11699292808771133,
+      "logits/rejected": 0.10016686469316483,
+      "logps/chosen": -86.1636962890625,
+      "logps/rejected": -87.16045379638672,
+      "loss": 2.0645,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.163857936859131,
+      "rewards/margins": 0.697452187538147,
+      "rewards/rejected": -2.8613100051879883,
+      "step": 528
+    },
+    {
+      "epoch": 0.96094459582198,
+      "grad_norm": 1.6236025094985962,
+      "learning_rate": 8.775956590863785e-06,
+      "logits/chosen": 0.11622033268213272,
+      "logits/rejected": 0.06629584729671478,
+      "logps/chosen": -76.42916870117188,
+      "logps/rejected": -80.24000549316406,
+      "loss": 1.8904,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.102374315261841,
+      "rewards/margins": 0.7230857014656067,
+      "rewards/rejected": -2.825460195541382,
+      "step": 529
+    },
+    {
+      "epoch": 0.9627611262488647,
+      "grad_norm": 1.6119427680969238,
+      "learning_rate": 8.771547678085332e-06,
+      "logits/chosen": 0.07009143382310867,
+      "logits/rejected": 0.16355563700199127,
+      "logps/chosen": -72.83528137207031,
+      "logps/rejected": -87.72441101074219,
+      "loss": 2.0136,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1697566509246826,
+      "rewards/margins": 0.8312336206436157,
+      "rewards/rejected": -3.000990390777588,
+      "step": 530
+    },
+    {
+      "epoch": 0.9645776566757494,
+      "grad_norm": 2.0324013233184814,
+      "learning_rate": 8.767132094051534e-06,
+      "logits/chosen": 0.14837129414081573,
+      "logits/rejected": 0.0700330063700676,
+      "logps/chosen": -84.65350341796875,
+      "logps/rejected": -91.84485626220703,
+      "loss": 2.2551,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5343966484069824,
+      "rewards/margins": 0.6457280516624451,
+      "rewards/rejected": -3.180124521255493,
+      "step": 531
+    },
+    {
+      "epoch": 0.9663941871026339,
+      "grad_norm": 1.4526044130325317,
+      "learning_rate": 8.762709847766532e-06,
+      "logits/chosen": 0.013620391488075256,
+      "logits/rejected": 0.10376289486885071,
+      "logps/chosen": -70.80522155761719,
+      "logps/rejected": -84.26114654541016,
+      "loss": 1.8607,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.249685287475586,
+      "rewards/margins": 0.8104004859924316,
+      "rewards/rejected": -3.0600855350494385,
+      "step": 532
+    },
+    {
+      "epoch": 0.9682107175295186,
+      "grad_norm": 2.0312576293945312,
+      "learning_rate": 8.758280948248059e-06,
+      "logits/chosen": 0.0965358167886734,
+      "logits/rejected": 0.10093791037797928,
+      "logps/chosen": -81.83265686035156,
+      "logps/rejected": -88.52181243896484,
+      "loss": 2.6128,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4901068210601807,
+      "rewards/margins": 0.30889907479286194,
+      "rewards/rejected": -2.7990057468414307,
+      "step": 533
+    },
+    {
+      "epoch": 0.9700272479564033,
+      "grad_norm": 2.547189712524414,
+      "learning_rate": 8.753845404527413e-06,
+      "logits/chosen": 0.020555848255753517,
+      "logits/rejected": 0.0945032387971878,
+      "logps/chosen": -72.16207122802734,
+      "logps/rejected": -86.27986907958984,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.148913860321045,
+      "rewards/margins": 0.7298768162727356,
+      "rewards/rejected": -2.8787906169891357,
+      "step": 534
+    },
+    {
+      "epoch": 0.971843778383288,
+      "grad_norm": 1.996645450592041,
+      "learning_rate": 8.74940322564944e-06,
+      "logits/chosen": 0.0345986932516098,
+      "logits/rejected": 0.08187264204025269,
+      "logps/chosen": -78.76763916015625,
+      "logps/rejected": -91.45555114746094,
+      "loss": 2.4587,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.4772286415100098,
+      "rewards/margins": 0.4609605073928833,
+      "rewards/rejected": -2.9381890296936035,
+      "step": 535
+    },
+    {
+      "epoch": 0.9736603088101726,
+      "grad_norm": 1.8145204782485962,
+      "learning_rate": 8.744954420672514e-06,
+      "logits/chosen": 0.03380711376667023,
+      "logits/rejected": 0.09039584547281265,
+      "logps/chosen": -79.83094787597656,
+      "logps/rejected": -90.87137603759766,
+      "loss": 2.3589,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.1433308124542236,
+      "rewards/margins": 0.5286959409713745,
+      "rewards/rejected": -2.6720268726348877,
+      "step": 536
+    },
+    {
+      "epoch": 0.9754768392370572,
+      "grad_norm": 1.9745460748672485,
+      "learning_rate": 8.740498998668523e-06,
+      "logits/chosen": 0.07233883440494537,
+      "logits/rejected": 0.17097605764865875,
+      "logps/chosen": -72.3458251953125,
+      "logps/rejected": -83.47319793701172,
+      "loss": 2.3469,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199948310852051,
+      "rewards/margins": 0.47777751088142395,
+      "rewards/rejected": -2.6777257919311523,
+      "step": 537
+    },
+    {
+      "epoch": 0.9772933696639419,
+      "grad_norm": 1.665255069732666,
+      "learning_rate": 8.736036968722851e-06,
+      "logits/chosen": 0.09596951305866241,
+      "logits/rejected": 0.06625551730394363,
+      "logps/chosen": -80.11689758300781,
+      "logps/rejected": -88.4573974609375,
+      "loss": 2.1889,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1831865310668945,
+      "rewards/margins": 0.6106584668159485,
+      "rewards/rejected": -2.7938451766967773,
+      "step": 538
+    },
+    {
+      "epoch": 0.9791099000908265,
+      "grad_norm": 1.5587108135223389,
+      "learning_rate": 8.73156833993435e-06,
+      "logits/chosen": 0.07950340211391449,
+      "logits/rejected": 0.0903201699256897,
+      "logps/chosen": -78.9182357788086,
+      "logps/rejected": -92.290283203125,
+      "loss": 2.048,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.12032413482666,
+      "rewards/margins": 0.7009568214416504,
+      "rewards/rejected": -2.8212809562683105,
+      "step": 539
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 1.7177670001983643,
+      "learning_rate": 8.727093121415338e-06,
+      "logits/chosen": 0.06052964925765991,
+      "logits/rejected": 0.14043518900871277,
+      "logps/chosen": -79.04943084716797,
+      "logps/rejected": -88.881591796875,
+      "loss": 2.2089,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2680561542510986,
+      "rewards/margins": 0.5380589962005615,
+      "rewards/rejected": -2.806114912033081,
+      "step": 540
+    },
+    {
+      "epoch": 0.9827429609445958,
+      "grad_norm": 1.7630436420440674,
+      "learning_rate": 8.722611322291558e-06,
+      "logits/chosen": 0.02032918483018875,
+      "logits/rejected": 0.11728623509407043,
+      "logps/chosen": -78.1409912109375,
+      "logps/rejected": -89.3685073852539,
+      "loss": 2.3303,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.356957197189331,
+      "rewards/margins": 0.5139226317405701,
+      "rewards/rejected": -2.870880126953125,
+      "step": 541
+    },
+    {
+      "epoch": 0.9845594913714805,
+      "grad_norm": 1.852418065071106,
+      "learning_rate": 8.718122951702183e-06,
+      "logits/chosen": 0.0439969077706337,
+      "logits/rejected": 0.16671347618103027,
+      "logps/chosen": -69.96346282958984,
+      "logps/rejected": -85.24410247802734,
+      "loss": 2.1393,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3295838832855225,
+      "rewards/margins": 0.6544574499130249,
+      "rewards/rejected": -2.984041213989258,
+      "step": 542
+    },
+    {
+      "epoch": 0.9863760217983651,
+      "grad_norm": 1.9028044939041138,
+      "learning_rate": 8.713628018799782e-06,
+      "logits/chosen": 0.04875154793262482,
+      "logits/rejected": -0.009283583611249924,
+      "logps/chosen": -75.80145263671875,
+      "logps/rejected": -78.19249725341797,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.241586208343506,
+      "rewards/margins": 0.4672737419605255,
+      "rewards/rejected": -2.708860158920288,
+      "step": 543
+    },
+    {
+      "epoch": 0.9881925522252498,
+      "grad_norm": 1.5648458003997803,
+      "learning_rate": 8.709126532750304e-06,
+      "logits/chosen": 0.02982347458600998,
+      "logits/rejected": 0.07835812866687775,
+      "logps/chosen": -75.35140991210938,
+      "logps/rejected": -91.11531829833984,
+      "loss": 2.0945,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.062386989593506,
+      "rewards/margins": 0.6832989454269409,
+      "rewards/rejected": -2.7456860542297363,
+      "step": 544
+    },
+    {
+      "epoch": 0.9900090826521344,
+      "grad_norm": 1.6214131116867065,
+      "learning_rate": 8.704618502733064e-06,
+      "logits/chosen": 0.09551770985126495,
+      "logits/rejected": 0.11310344189405441,
+      "logps/chosen": -80.15961456298828,
+      "logps/rejected": -93.38126373291016,
+      "loss": 2.1642,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.255985736846924,
+      "rewards/margins": 0.7712126970291138,
+      "rewards/rejected": -3.027198553085327,
+      "step": 545
+    },
+    {
+      "epoch": 0.9918256130790191,
+      "grad_norm": 1.762953519821167,
+      "learning_rate": 8.70010393794072e-06,
+      "logits/chosen": 0.0971008688211441,
+      "logits/rejected": 0.16590046882629395,
+      "logps/chosen": -86.24911499023438,
+      "logps/rejected": -102.59831237792969,
+      "loss": 2.0951,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3542394638061523,
+      "rewards/margins": 0.7857018709182739,
+      "rewards/rejected": -3.139941453933716,
+      "step": 546
+    },
+    {
+      "epoch": 0.9936421435059037,
+      "grad_norm": 1.6033780574798584,
+      "learning_rate": 8.695582847579254e-06,
+      "logits/chosen": 0.15147003531455994,
+      "logits/rejected": 0.13297875225543976,
+      "logps/chosen": -78.5172348022461,
+      "logps/rejected": -86.49285888671875,
+      "loss": 2.1101,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2025821208953857,
+      "rewards/margins": 0.7471145391464233,
+      "rewards/rejected": -2.9496963024139404,
+      "step": 547
+    },
+    {
+      "epoch": 0.9954586739327884,
+      "grad_norm": 1.4905331134796143,
+      "learning_rate": 8.69105524086796e-06,
+      "logits/chosen": 0.07559307664632797,
+      "logits/rejected": 0.09286531805992126,
+      "logps/chosen": -74.92215728759766,
+      "logps/rejected": -78.23614501953125,
+      "loss": 2.2745,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.24694561958313,
+      "rewards/margins": 0.5010396838188171,
+      "rewards/rejected": -2.747985363006592,
+      "step": 548
+    },
+    {
+      "epoch": 0.997275204359673,
+      "grad_norm": 1.823449730873108,
+      "learning_rate": 8.686521127039414e-06,
+      "logits/chosen": 0.03763032704591751,
+      "logits/rejected": 0.06642220914363861,
+      "logps/chosen": -79.37518310546875,
+      "logps/rejected": -87.17401885986328,
+      "loss": 2.5954,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158803701400757,
+      "rewards/margins": 0.4884320795536041,
+      "rewards/rejected": -2.647235870361328,
+      "step": 549
+    },
+    {
+      "epoch": 0.9990917347865577,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.681980515339464e-06,
+      "logits/chosen": 0.08186414837837219,
+      "logits/rejected": 0.08158639073371887,
+      "logps/chosen": -84.89871215820312,
+      "logps/rejected": -85.87324523925781,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.1884236335754395,
+      "rewards/margins": 0.2923518121242523,
+      "rewards/rejected": -2.4807753562927246,
+      "step": 550
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.677433415027209e-06,
+      "logits/chosen": 0.18737269937992096,
+      "logits/rejected": 0.11028257012367249,
+      "logps/chosen": -76.3554916381836,
+      "logps/rejected": -86.39812469482422,
+      "loss": 1.1517,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1043050289154053,
+      "rewards/margins": 0.7949569225311279,
+      "rewards/rejected": -2.8992621898651123,
+      "step": 551
+    },
+    {
+      "epoch": 1.0018165304268847,
+      "grad_norm": 1.6554900407791138,
+      "learning_rate": 8.672879835374976e-06,
+      "logits/chosen": 0.11259499937295914,
+      "logits/rejected": 0.07095417380332947,
+      "logps/chosen": -84.55628967285156,
+      "logps/rejected": -88.55287170410156,
+      "loss": 2.1275,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.2375316619873047,
+      "rewards/margins": 0.5453992486000061,
+      "rewards/rejected": -2.782930850982666,
+      "step": 552
+    },
+    {
+      "epoch": 1.0036330608537694,
+      "grad_norm": 1.6807721853256226,
+      "learning_rate": 8.668319785668307e-06,
+      "logits/chosen": 0.14847862720489502,
+      "logits/rejected": 0.1863957643508911,
+      "logps/chosen": -87.27869415283203,
+      "logps/rejected": -89.04036712646484,
+      "loss": 2.3209,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.087535858154297,
+      "rewards/margins": 0.5243443846702576,
+      "rewards/rejected": -2.61188006401062,
+      "step": 553
+    },
+    {
+      "epoch": 1.005449591280654,
+      "grad_norm": 1.8480761051177979,
+      "learning_rate": 8.66375327520594e-06,
+      "logits/chosen": 0.07236558198928833,
+      "logits/rejected": 0.13397395610809326,
+      "logps/chosen": -85.43444061279297,
+      "logps/rejected": -101.47248840332031,
+      "loss": 1.7106,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.157266616821289,
+      "rewards/margins": 0.9501473903656006,
+      "rewards/rejected": -3.1074140071868896,
+      "step": 554
+    },
+    {
+      "epoch": 1.0072661217075387,
+      "grad_norm": 1.586917757987976,
+      "learning_rate": 8.659180313299783e-06,
+      "logits/chosen": 0.05113527923822403,
+      "logits/rejected": 0.14182163774967194,
+      "logps/chosen": -78.31108856201172,
+      "logps/rejected": -90.46333312988281,
+      "loss": 1.8965,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.945315957069397,
+      "rewards/margins": 0.8104385137557983,
+      "rewards/rejected": -2.755754232406616,
+      "step": 555
+    },
+    {
+      "epoch": 1.0090826521344232,
+      "grad_norm": 1.406339406967163,
+      "learning_rate": 8.6546009092749e-06,
+      "logits/chosen": -0.01422208547592163,
+      "logits/rejected": 0.10859756916761398,
+      "logps/chosen": -75.02066040039062,
+      "logps/rejected": -87.69715118408203,
+      "loss": 2.027,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9996552467346191,
+      "rewards/margins": 0.7365851402282715,
+      "rewards/rejected": -2.7362406253814697,
+      "step": 556
+    },
+    {
+      "epoch": 1.0108991825613078,
+      "grad_norm": 1.5511010885238647,
+      "learning_rate": 8.650015072469496e-06,
+      "logits/chosen": 0.07154582440853119,
+      "logits/rejected": 0.11479673534631729,
+      "logps/chosen": -88.33872985839844,
+      "logps/rejected": -96.75527954101562,
+      "loss": 1.966,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2190096378326416,
+      "rewards/margins": 0.861441969871521,
+      "rewards/rejected": -3.080451250076294,
+      "step": 557
+    },
+    {
+      "epoch": 1.0127157129881925,
+      "grad_norm": 1.5951570272445679,
+      "learning_rate": 8.645422812234888e-06,
+      "logits/chosen": 0.17342886328697205,
+      "logits/rejected": 0.11909954994916916,
+      "logps/chosen": -79.63985443115234,
+      "logps/rejected": -83.41281127929688,
+      "loss": 2.1115,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3469128608703613,
+      "rewards/margins": 0.7317838668823242,
+      "rewards/rejected": -3.0786967277526855,
+      "step": 558
+    },
+    {
+      "epoch": 1.0145322434150772,
+      "grad_norm": 1.891517162322998,
+      "learning_rate": 8.640824137935498e-06,
+      "logits/chosen": 0.14266839623451233,
+      "logits/rejected": 0.14209085702896118,
+      "logps/chosen": -84.77491760253906,
+      "logps/rejected": -87.46759796142578,
+      "loss": 2.2691,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.022730827331543,
+      "rewards/margins": 0.42655885219573975,
+      "rewards/rejected": -2.449289560317993,
+      "step": 559
+    },
+    {
+      "epoch": 1.0163487738419619,
+      "grad_norm": 1.7919303178787231,
+      "learning_rate": 8.636219058948823e-06,
+      "logits/chosen": 0.01973957009613514,
+      "logits/rejected": 0.12187746167182922,
+      "logps/chosen": -78.26383209228516,
+      "logps/rejected": -92.714599609375,
+      "loss": 1.974,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2774088382720947,
+      "rewards/margins": 0.8670985698699951,
+      "rewards/rejected": -3.1445071697235107,
+      "step": 560
+    },
+    {
+      "epoch": 1.0181653042688465,
+      "grad_norm": 1.5417025089263916,
+      "learning_rate": 8.631607584665415e-06,
+      "logits/chosen": 0.07740732282400131,
+      "logits/rejected": 0.17052005231380463,
+      "logps/chosen": -83.54193115234375,
+      "logps/rejected": -95.1084976196289,
+      "loss": 1.9963,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1598939895629883,
+      "rewards/margins": 0.802270233631134,
+      "rewards/rejected": -2.9621646404266357,
+      "step": 561
+    },
+    {
+      "epoch": 1.0199818346957312,
+      "grad_norm": 1.9338072538375854,
+      "learning_rate": 8.62698972448888e-06,
+      "logits/chosen": 0.09191437065601349,
+      "logits/rejected": 0.1776769608259201,
+      "logps/chosen": -82.68082427978516,
+      "logps/rejected": -96.89566802978516,
+      "loss": 2.0399,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3505523204803467,
+      "rewards/margins": 0.7893310785293579,
+      "rewards/rejected": -3.1398837566375732,
+      "step": 562
+    },
+    {
+      "epoch": 1.021798365122616,
+      "grad_norm": 1.609601378440857,
+      "learning_rate": 8.622365487835836e-06,
+      "logits/chosen": 0.13105043768882751,
+      "logits/rejected": 0.138211190700531,
+      "logps/chosen": -77.19727325439453,
+      "logps/rejected": -84.1854476928711,
+      "loss": 2.3184,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.13655948638916,
+      "rewards/margins": 0.5744473934173584,
+      "rewards/rejected": -2.7110068798065186,
+      "step": 563
+    },
+    {
+      "epoch": 1.0236148955495004,
+      "grad_norm": 2.1222517490386963,
+      "learning_rate": 8.617734884135906e-06,
+      "logits/chosen": 0.14744582772254944,
+      "logits/rejected": 0.053739145398139954,
+      "logps/chosen": -81.11271667480469,
+      "logps/rejected": -81.40655517578125,
+      "loss": 2.2,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.419705390930176,
+      "rewards/margins": 0.7198007702827454,
+      "rewards/rejected": -3.1395058631896973,
+      "step": 564
+    },
+    {
+      "epoch": 1.025431425976385,
+      "grad_norm": 1.8075231313705444,
+      "learning_rate": 8.613097922831699e-06,
+      "logits/chosen": 0.07210052758455276,
+      "logits/rejected": 0.07022904604673386,
+      "logps/chosen": -88.72612762451172,
+      "logps/rejected": -98.2815933227539,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1057255268096924,
+      "rewards/margins": 0.8747395873069763,
+      "rewards/rejected": -2.9804654121398926,
+      "step": 565
+    },
+    {
+      "epoch": 1.0272479564032697,
+      "grad_norm": 1.3381479978561401,
+      "learning_rate": 8.608454613378783e-06,
+      "logits/chosen": 0.08187520503997803,
+      "logits/rejected": 0.1098201796412468,
+      "logps/chosen": -77.1299819946289,
+      "logps/rejected": -89.75099182128906,
+      "loss": 1.8687,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3585519790649414,
+      "rewards/margins": 0.8574986457824707,
+      "rewards/rejected": -3.216050386428833,
+      "step": 566
+    },
+    {
+      "epoch": 1.0290644868301544,
+      "grad_norm": 1.6325740814208984,
+      "learning_rate": 8.603804965245678e-06,
+      "logits/chosen": 0.12077341973781586,
+      "logits/rejected": 0.09103713929653168,
+      "logps/chosen": -74.03280639648438,
+      "logps/rejected": -79.61197662353516,
+      "loss": 2.0884,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.430152654647827,
+      "rewards/margins": 0.6819649934768677,
+      "rewards/rejected": -3.1121177673339844,
+      "step": 567
+    },
+    {
+      "epoch": 1.030881017257039,
+      "grad_norm": 1.782224178314209,
+      "learning_rate": 8.599148987913829e-06,
+      "logits/chosen": -0.008054995909333229,
+      "logits/rejected": 0.08848709613084793,
+      "logps/chosen": -88.39682006835938,
+      "logps/rejected": -98.20311737060547,
+      "loss": 2.3203,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.21012282371521,
+      "rewards/margins": 0.5451631546020508,
+      "rewards/rejected": -2.75528621673584,
+      "step": 568
+    },
+    {
+      "epoch": 1.0326975476839237,
+      "grad_norm": 1.7691792249679565,
+      "learning_rate": 8.594486690877577e-06,
+      "logits/chosen": 0.05038870871067047,
+      "logits/rejected": -0.0020756814628839493,
+      "logps/chosen": -88.35281372070312,
+      "logps/rejected": -87.42422485351562,
+      "loss": 2.1789,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6035892963409424,
+      "rewards/margins": 0.559741735458374,
+      "rewards/rejected": -3.1633315086364746,
+      "step": 569
+    },
+    {
+      "epoch": 1.0345140781108084,
+      "grad_norm": 1.7002201080322266,
+      "learning_rate": 8.589818083644161e-06,
+      "logits/chosen": 0.030099213123321533,
+      "logits/rejected": 0.10112117975950241,
+      "logps/chosen": -74.28582763671875,
+      "logps/rejected": -85.93130493164062,
+      "loss": 2.2068,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.324796676635742,
+      "rewards/margins": 0.5959895849227905,
+      "rewards/rejected": -2.9207863807678223,
+      "step": 570
+    },
+    {
+      "epoch": 1.036330608537693,
+      "grad_norm": 1.5611207485198975,
+      "learning_rate": 8.585143175733686e-06,
+      "logits/chosen": 0.09465142339468002,
+      "logits/rejected": 0.13904502987861633,
+      "logps/chosen": -86.5047378540039,
+      "logps/rejected": -95.09537506103516,
+      "loss": 1.9628,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5564661026000977,
+      "rewards/margins": 0.8365517258644104,
+      "rewards/rejected": -3.3930182456970215,
+      "step": 571
+    },
+    {
+      "epoch": 1.0381471389645776,
+      "grad_norm": 1.7271509170532227,
+      "learning_rate": 8.5804619766791e-06,
+      "logits/chosen": 0.08665300905704498,
+      "logits/rejected": 0.10898350924253464,
+      "logps/chosen": -85.17437744140625,
+      "logps/rejected": -94.50398254394531,
+      "loss": 2.2717,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3690195083618164,
+      "rewards/margins": 0.6290315389633179,
+      "rewards/rejected": -2.9980509281158447,
+      "step": 572
+    },
+    {
+      "epoch": 1.0399636693914622,
+      "grad_norm": 1.7832714319229126,
+      "learning_rate": 8.575774496026184e-06,
+      "logits/chosen": 0.033456090837717056,
+      "logits/rejected": 0.09503711760044098,
+      "logps/chosen": -87.27224731445312,
+      "logps/rejected": -94.98229217529297,
+      "loss": 2.1781,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.509322166442871,
+      "rewards/margins": 0.5943110585212708,
+      "rewards/rejected": -3.103632926940918,
+      "step": 573
+    },
+    {
+      "epoch": 1.041780199818347,
+      "grad_norm": 1.8239610195159912,
+      "learning_rate": 8.571080743333528e-06,
+      "logits/chosen": 0.11436322331428528,
+      "logits/rejected": 0.06905262917280197,
+      "logps/chosen": -81.12911987304688,
+      "logps/rejected": -88.73529052734375,
+      "loss": 2.1358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2652292251586914,
+      "rewards/margins": 0.6765917539596558,
+      "rewards/rejected": -2.9418206214904785,
+      "step": 574
+    },
+    {
+      "epoch": 1.0435967302452316,
+      "grad_norm": 1.9338902235031128,
+      "learning_rate": 8.566380728172512e-06,
+      "logits/chosen": 0.11198662221431732,
+      "logits/rejected": 0.11845988035202026,
+      "logps/chosen": -82.55985260009766,
+      "logps/rejected": -94.05873107910156,
+      "loss": 1.9609,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3074758052825928,
+      "rewards/margins": 0.8309231400489807,
+      "rewards/rejected": -3.1383988857269287,
+      "step": 575
+    },
+    {
+      "epoch": 1.0454132606721163,
+      "grad_norm": 1.6932177543640137,
+      "learning_rate": 8.561674460127281e-06,
+      "logits/chosen": 0.0954650342464447,
+      "logits/rejected": 0.139107346534729,
+      "logps/chosen": -83.95409393310547,
+      "logps/rejected": -98.31000518798828,
+      "loss": 2.311,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.476616859436035,
+      "rewards/margins": 0.7193690538406372,
+      "rewards/rejected": -3.195985794067383,
+      "step": 576
+    },
+    {
+      "epoch": 1.047229791099001,
+      "grad_norm": 1.6585469245910645,
+      "learning_rate": 8.556961948794738e-06,
+      "logits/chosen": 0.050676412880420685,
+      "logits/rejected": 0.13941077888011932,
+      "logps/chosen": -91.46603393554688,
+      "logps/rejected": -111.38130187988281,
+      "loss": 1.9574,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.630396604537964,
+      "rewards/margins": 0.8525002002716064,
+      "rewards/rejected": -3.4828968048095703,
+      "step": 577
+    },
+    {
+      "epoch": 1.0490463215258856,
+      "grad_norm": 1.8747247457504272,
+      "learning_rate": 8.552243203784514e-06,
+      "logits/chosen": 0.06109774485230446,
+      "logits/rejected": 0.03656052052974701,
+      "logps/chosen": -84.25686645507812,
+      "logps/rejected": -99.20081329345703,
+      "loss": 2.0244,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4541397094726562,
+      "rewards/margins": 1.0175232887268066,
+      "rewards/rejected": -3.471662759780884,
+      "step": 578
+    },
+    {
+      "epoch": 1.0508628519527703,
+      "grad_norm": 1.714175820350647,
+      "learning_rate": 8.547518234718954e-06,
+      "logits/chosen": 0.07959345728158951,
+      "logits/rejected": 0.17609870433807373,
+      "logps/chosen": -79.58773040771484,
+      "logps/rejected": -98.85990142822266,
+      "loss": 1.7738,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.431654930114746,
+      "rewards/margins": 1.0072871446609497,
+      "rewards/rejected": -3.4389421939849854,
+      "step": 579
+    },
+    {
+      "epoch": 1.052679382379655,
+      "grad_norm": 1.6469632387161255,
+      "learning_rate": 8.542787051233088e-06,
+      "logits/chosen": 0.06128916144371033,
+      "logits/rejected": 0.11901555210351944,
+      "logps/chosen": -90.99638366699219,
+      "logps/rejected": -104.79571533203125,
+      "loss": 1.8789,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3995859622955322,
+      "rewards/margins": 0.9405049085617065,
+      "rewards/rejected": -3.340090751647949,
+      "step": 580
+    },
+    {
+      "epoch": 1.0544959128065394,
+      "grad_norm": 1.8093998432159424,
+      "learning_rate": 8.538049662974625e-06,
+      "logits/chosen": 0.136922687292099,
+      "logits/rejected": 0.06489068269729614,
+      "logps/chosen": -84.12126159667969,
+      "logps/rejected": -85.51757049560547,
+      "loss": 2.2372,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7452549934387207,
+      "rewards/margins": 0.6496641635894775,
+      "rewards/rejected": -3.394918918609619,
+      "step": 581
+    },
+    {
+      "epoch": 1.056312443233424,
+      "grad_norm": 1.761516809463501,
+      "learning_rate": 8.533306079603928e-06,
+      "logits/chosen": 0.02220803126692772,
+      "logits/rejected": 0.03564752638339996,
+      "logps/chosen": -91.6943359375,
+      "logps/rejected": -105.33677673339844,
+      "loss": 1.831,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.1495776176452637,
+      "rewards/margins": 0.8695197701454163,
+      "rewards/rejected": -3.019097328186035,
+      "step": 582
+    },
+    {
+      "epoch": 1.0581289736603088,
+      "grad_norm": 1.8054577112197876,
+      "learning_rate": 8.52855631079398e-06,
+      "logits/chosen": 0.044406406581401825,
+      "logits/rejected": 0.13649103045463562,
+      "logps/chosen": -83.07669830322266,
+      "logps/rejected": -99.42033386230469,
+      "loss": 2.1295,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.674705743789673,
+      "rewards/margins": 0.8526960015296936,
+      "rewards/rejected": -3.52740216255188,
+      "step": 583
+    },
+    {
+      "epoch": 1.0599455040871935,
+      "grad_norm": 2.092067241668701,
+      "learning_rate": 8.523800366230397e-06,
+      "logits/chosen": 0.11083705723285675,
+      "logits/rejected": 0.13336583971977234,
+      "logps/chosen": -82.55166625976562,
+      "logps/rejected": -88.78050994873047,
+      "loss": 2.783,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.7093071937561035,
+      "rewards/margins": 0.34429922699928284,
+      "rewards/rejected": -3.0536062717437744,
+      "step": 584
+    },
+    {
+      "epoch": 1.0617620345140781,
+      "grad_norm": 2.0747568607330322,
+      "learning_rate": 8.519038255611372e-06,
+      "logits/chosen": 0.06060079485177994,
+      "logits/rejected": 0.06341060996055603,
+      "logps/chosen": -82.93423461914062,
+      "logps/rejected": -85.8175048828125,
+      "loss": 2.1046,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6574935913085938,
+      "rewards/margins": 0.7763374447822571,
+      "rewards/rejected": -3.433830738067627,
+      "step": 585
+    },
+    {
+      "epoch": 1.0635785649409628,
+      "grad_norm": 2.1435482501983643,
+      "learning_rate": 8.51426998864768e-06,
+      "logits/chosen": 0.044051673263311386,
+      "logits/rejected": 0.08015372604131699,
+      "logps/chosen": -86.49735260009766,
+      "logps/rejected": -97.29399108886719,
+      "loss": 2.1365,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.620311737060547,
+      "rewards/margins": 0.6845528483390808,
+      "rewards/rejected": -3.3048648834228516,
+      "step": 586
+    },
+    {
+      "epoch": 1.0653950953678475,
+      "grad_norm": 1.65945303440094,
+      "learning_rate": 8.509495575062647e-06,
+      "logits/chosen": 0.15405744314193726,
+      "logits/rejected": 0.12232419848442078,
+      "logps/chosen": -73.09071350097656,
+      "logps/rejected": -74.24645233154297,
+      "loss": 2.1912,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3465700149536133,
+      "rewards/margins": 0.6234737634658813,
+      "rewards/rejected": -2.970043420791626,
+      "step": 587
+    },
+    {
+      "epoch": 1.067211625794732,
+      "grad_norm": 2.166276216506958,
+      "learning_rate": 8.504715024592132e-06,
+      "logits/chosen": -0.05482687056064606,
+      "logits/rejected": -0.003489813767373562,
+      "logps/chosen": -84.82199096679688,
+      "logps/rejected": -97.53748321533203,
+      "loss": 2.1753,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5944690704345703,
+      "rewards/margins": 0.6753374934196472,
+      "rewards/rejected": -3.269806385040283,
+      "step": 588
+    },
+    {
+      "epoch": 1.0690281562216166,
+      "grad_norm": 1.8869906663894653,
+      "learning_rate": 8.499928346984512e-06,
+      "logits/chosen": 0.13489654660224915,
+      "logits/rejected": 0.052815720438957214,
+      "logps/chosen": -90.98455047607422,
+      "logps/rejected": -91.64305877685547,
+      "loss": 2.3005,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.8705921173095703,
+      "rewards/margins": 0.555056095123291,
+      "rewards/rejected": -3.4256482124328613,
+      "step": 589
+    },
+    {
+      "epoch": 1.0708446866485013,
+      "grad_norm": 2.286853790283203,
+      "learning_rate": 8.495135552000657e-06,
+      "logits/chosen": -0.0365738607943058,
+      "logits/rejected": 0.09564587473869324,
+      "logps/chosen": -93.99796295166016,
+      "logps/rejected": -111.09819793701172,
+      "loss": 2.1574,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.484596014022827,
+      "rewards/margins": 0.840925395488739,
+      "rewards/rejected": -3.325521469116211,
+      "step": 590
+    },
+    {
+      "epoch": 1.072661217075386,
+      "grad_norm": 1.7640435695648193,
+      "learning_rate": 8.490336649413909e-06,
+      "logits/chosen": 0.06625102460384369,
+      "logits/rejected": 0.14394444227218628,
+      "logps/chosen": -74.65644836425781,
+      "logps/rejected": -92.41899871826172,
+      "loss": 2.0387,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.453733205795288,
+      "rewards/margins": 0.7862347364425659,
+      "rewards/rejected": -3.2399678230285645,
+      "step": 591
+    },
+    {
+      "epoch": 1.0744777475022707,
+      "grad_norm": 1.5737273693084717,
+      "learning_rate": 8.485531649010063e-06,
+      "logits/chosen": -0.0012233639135956764,
+      "logits/rejected": 0.021886199712753296,
+      "logps/chosen": -79.36224365234375,
+      "logps/rejected": -94.19145965576172,
+      "loss": 2.3886,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.6413750648498535,
+      "rewards/margins": 0.7800988554954529,
+      "rewards/rejected": -3.421473979949951,
+      "step": 592
+    },
+    {
+      "epoch": 1.0762942779291553,
+      "grad_norm": 2.4172728061676025,
+      "learning_rate": 8.480720560587356e-06,
+      "logits/chosen": 0.04091513529419899,
+      "logits/rejected": 0.16047289967536926,
+      "logps/chosen": -76.38783264160156,
+      "logps/rejected": -94.16845703125,
+      "loss": 2.2219,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.719104290008545,
+      "rewards/margins": 0.9459499716758728,
+      "rewards/rejected": -3.6650545597076416,
+      "step": 593
+    },
+    {
+      "epoch": 1.07811080835604,
+      "grad_norm": 1.6836284399032593,
+      "learning_rate": 8.475903393956434e-06,
+      "logits/chosen": 0.07094185054302216,
+      "logits/rejected": 0.09420361369848251,
+      "logps/chosen": -82.47905731201172,
+      "logps/rejected": -90.06558990478516,
+      "loss": 2.1268,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3819637298583984,
+      "rewards/margins": 0.5947781801223755,
+      "rewards/rejected": -2.9767420291900635,
+      "step": 594
+    },
+    {
+      "epoch": 1.0799273387829247,
+      "grad_norm": 1.808720588684082,
+      "learning_rate": 8.471080158940336e-06,
+      "logits/chosen": 0.010593242943286896,
+      "logits/rejected": 0.03635484725236893,
+      "logps/chosen": -82.94917297363281,
+      "logps/rejected": -94.52174377441406,
+      "loss": 1.9352,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.120471239089966,
+      "rewards/margins": 0.8991338014602661,
+      "rewards/rejected": -3.0196051597595215,
+      "step": 595
+    },
+    {
+      "epoch": 1.0817438692098094,
+      "grad_norm": 1.809515118598938,
+      "learning_rate": 8.46625086537448e-06,
+      "logits/chosen": 0.0202183797955513,
+      "logits/rejected": 0.1138802096247673,
+      "logps/chosen": -84.1740951538086,
+      "logps/rejected": -99.66446685791016,
+      "loss": 2.1384,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7415525913238525,
+      "rewards/margins": 0.8209500908851624,
+      "rewards/rejected": -3.562502861022949,
+      "step": 596
+    },
+    {
+      "epoch": 1.0835603996366938,
+      "grad_norm": 2.023786783218384,
+      "learning_rate": 8.46141552310664e-06,
+      "logits/chosen": 0.11318185180425644,
+      "logits/rejected": 0.1275079846382141,
+      "logps/chosen": -78.02458953857422,
+      "logps/rejected": -89.36206817626953,
+      "loss": 1.9695,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5616073608398438,
+      "rewards/margins": 0.7586690783500671,
+      "rewards/rejected": -3.3202764987945557,
+      "step": 597
+    },
+    {
+      "epoch": 1.0853769300635785,
+      "grad_norm": 1.7670155763626099,
+      "learning_rate": 8.456574141996913e-06,
+      "logits/chosen": 0.09054378420114517,
+      "logits/rejected": 0.17597083747386932,
+      "logps/chosen": -79.7802505493164,
+      "logps/rejected": -95.21873474121094,
+      "loss": 1.9499,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.336801767349243,
+      "rewards/margins": 0.7675365209579468,
+      "rewards/rejected": -3.1043381690979004,
+      "step": 598
+    },
+    {
+      "epoch": 1.0871934604904632,
+      "grad_norm": 2.033724308013916,
+      "learning_rate": 8.451726731917722e-06,
+      "logits/chosen": 0.08810828626155853,
+      "logits/rejected": 0.1388731151819229,
+      "logps/chosen": -75.62480163574219,
+      "logps/rejected": -90.9974365234375,
+      "loss": 1.993,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2493839263916016,
+      "rewards/margins": 0.9519067406654358,
+      "rewards/rejected": -3.2012906074523926,
+      "step": 599
+    },
+    {
+      "epoch": 1.0890099909173478,
+      "grad_norm": 2.0463502407073975,
+      "learning_rate": 8.446873302753783e-06,
+      "logits/chosen": -0.017319753766059875,
+      "logits/rejected": 0.026935823261737823,
+      "logps/chosen": -91.20410919189453,
+      "logps/rejected": -97.56501770019531,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5798678398132324,
+      "rewards/margins": 0.46915650367736816,
+      "rewards/rejected": -3.0490241050720215,
+      "step": 600
+    },
+    {
+      "epoch": 1.0908265213442325,
+      "grad_norm": 1.944875717163086,
+      "learning_rate": 8.44201386440208e-06,
+      "logits/chosen": 0.1267091929912567,
+      "logits/rejected": 0.12764671444892883,
+      "logps/chosen": -82.59441375732422,
+      "logps/rejected": -91.89214324951172,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.495068311691284,
+      "rewards/margins": 0.6413819789886475,
+      "rewards/rejected": -3.1364500522613525,
+      "step": 601
+    },
+    {
+      "epoch": 1.0926430517711172,
+      "grad_norm": 1.8729236125946045,
+      "learning_rate": 8.437148426771852e-06,
+      "logits/chosen": 0.15159347653388977,
+      "logits/rejected": 0.16930025815963745,
+      "logps/chosen": -80.54000854492188,
+      "logps/rejected": -93.16973114013672,
+      "loss": 1.8605,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4500646591186523,
+      "rewards/margins": 0.9991447925567627,
+      "rewards/rejected": -3.449209690093994,
+      "step": 602
+    },
+    {
+      "epoch": 1.0944595821980019,
+      "grad_norm": 1.9284056425094604,
+      "learning_rate": 8.432276999784578e-06,
+      "logits/chosen": 0.16441625356674194,
+      "logits/rejected": 0.17125487327575684,
+      "logps/chosen": -83.31804656982422,
+      "logps/rejected": -91.39764404296875,
+      "loss": 1.9416,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.577606201171875,
+      "rewards/margins": 0.8178077936172485,
+      "rewards/rejected": -3.395413875579834,
+      "step": 603
+    },
+    {
+      "epoch": 1.0962761126248866,
+      "grad_norm": 2.0095536708831787,
+      "learning_rate": 8.427399593373941e-06,
+      "logits/chosen": 0.03275620937347412,
+      "logits/rejected": 0.033026132732629776,
+      "logps/chosen": -83.01813507080078,
+      "logps/rejected": -93.54053497314453,
+      "loss": 2.0358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.488004684448242,
+      "rewards/margins": 0.6699528694152832,
+      "rewards/rejected": -3.1579573154449463,
+      "step": 604
+    },
+    {
+      "epoch": 1.0980926430517712,
+      "grad_norm": 1.8138859272003174,
+      "learning_rate": 8.422516217485826e-06,
+      "logits/chosen": 0.1295854151248932,
+      "logits/rejected": 0.15053214132785797,
+      "logps/chosen": -87.74354553222656,
+      "logps/rejected": -94.38423919677734,
+      "loss": 2.6379,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.538604259490967,
+      "rewards/margins": 0.44896796345710754,
+      "rewards/rejected": -2.987572193145752,
+      "step": 605
+    },
+    {
+      "epoch": 1.0999091734786557,
+      "grad_norm": 2.262779951095581,
+      "learning_rate": 8.417626882078287e-06,
+      "logits/chosen": 0.07311846315860748,
+      "logits/rejected": 0.09329289197921753,
+      "logps/chosen": -88.92317199707031,
+      "logps/rejected": -105.15555572509766,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.468625783920288,
+      "rewards/margins": 1.0499695539474487,
+      "rewards/rejected": -3.5185956954956055,
+      "step": 606
+    },
+    {
+      "epoch": 1.1017257039055404,
+      "grad_norm": 1.7983455657958984,
+      "learning_rate": 8.412731597121527e-06,
+      "logits/chosen": 0.06970179080963135,
+      "logits/rejected": 0.170780211687088,
+      "logps/chosen": -84.99759674072266,
+      "logps/rejected": -91.03089904785156,
+      "loss": 2.3611,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.31594181060791,
+      "rewards/margins": 0.46637701988220215,
+      "rewards/rejected": -2.7823188304901123,
+      "step": 607
+    },
+    {
+      "epoch": 1.103542234332425,
+      "grad_norm": 1.8828606605529785,
+      "learning_rate": 8.407830372597884e-06,
+      "logits/chosen": 0.10343387722969055,
+      "logits/rejected": 0.15009143948554993,
+      "logps/chosen": -74.23157501220703,
+      "logps/rejected": -85.4384994506836,
+      "loss": 2.0535,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3184473514556885,
+      "rewards/margins": 0.7085882425308228,
+      "rewards/rejected": -3.027035713195801,
+      "step": 608
+    },
+    {
+      "epoch": 1.1053587647593097,
+      "grad_norm": 1.8570231199264526,
+      "learning_rate": 8.402923218501813e-06,
+      "logits/chosen": 0.0035794638097286224,
+      "logits/rejected": 0.044682763516902924,
+      "logps/chosen": -82.44414520263672,
+      "logps/rejected": -93.22355651855469,
+      "loss": 2.1751,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.584066867828369,
+      "rewards/margins": 0.7579058408737183,
+      "rewards/rejected": -3.3419723510742188,
+      "step": 609
+    },
+    {
+      "epoch": 1.1071752951861944,
+      "grad_norm": 1.9866671562194824,
+      "learning_rate": 8.39801014483985e-06,
+      "logits/chosen": -0.015922199934720993,
+      "logits/rejected": 0.10281078517436981,
+      "logps/chosen": -99.91399383544922,
+      "logps/rejected": -113.79450988769531,
+      "loss": 2.2522,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.595137119293213,
+      "rewards/margins": 0.7955703139305115,
+      "rewards/rejected": -3.390707492828369,
+      "step": 610
+    },
+    {
+      "epoch": 1.108991825613079,
+      "grad_norm": 2.148136854171753,
+      "learning_rate": 8.393091161630612e-06,
+      "logits/chosen": 0.04068015143275261,
+      "logits/rejected": 0.09121442586183548,
+      "logps/chosen": -85.49508666992188,
+      "logps/rejected": -97.27752685546875,
+      "loss": 1.9467,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.673704147338867,
+      "rewards/margins": 0.8281559348106384,
+      "rewards/rejected": -3.5018603801727295,
+      "step": 611
+    },
+    {
+      "epoch": 1.1108083560399638,
+      "grad_norm": 2.0157570838928223,
+      "learning_rate": 8.388166278904759e-06,
+      "logits/chosen": 0.03451048582792282,
+      "logits/rejected": 0.10373912751674652,
+      "logps/chosen": -79.24334716796875,
+      "logps/rejected": -91.64152526855469,
+      "loss": 2.041,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4238996505737305,
+      "rewards/margins": 0.7874577045440674,
+      "rewards/rejected": -3.211357593536377,
+      "step": 612
+    },
+    {
+      "epoch": 1.1126248864668482,
+      "grad_norm": 1.7337363958358765,
+      "learning_rate": 8.383235506704986e-06,
+      "logits/chosen": 0.06933741271495819,
+      "logits/rejected": 0.10624522715806961,
+      "logps/chosen": -77.62388610839844,
+      "logps/rejected": -91.96408081054688,
+      "loss": 2.0593,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.304389476776123,
+      "rewards/margins": 0.6717057228088379,
+      "rewards/rejected": -2.97609543800354,
+      "step": 613
+    },
+    {
+      "epoch": 1.1144414168937329,
+      "grad_norm": 3.6357574462890625,
+      "learning_rate": 8.378298855085996e-06,
+      "logits/chosen": 0.08138547092676163,
+      "logits/rejected": 0.0961306244134903,
+      "logps/chosen": -87.44918823242188,
+      "logps/rejected": -94.0655746459961,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5593717098236084,
+      "rewards/margins": 0.6174351572990417,
+      "rewards/rejected": -3.176806688308716,
+      "step": 614
+    },
+    {
+      "epoch": 1.1162579473206176,
+      "grad_norm": 2.273991107940674,
+      "learning_rate": 8.373356334114484e-06,
+      "logits/chosen": 0.029560726135969162,
+      "logits/rejected": 0.11155828088521957,
+      "logps/chosen": -91.85681915283203,
+      "logps/rejected": -103.11241912841797,
+      "loss": 2.2401,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5434679985046387,
+      "rewards/margins": 0.7232552170753479,
+      "rewards/rejected": -3.266723394393921,
+      "step": 615
+    },
+    {
+      "epoch": 1.1180744777475022,
+      "grad_norm": 2.1391029357910156,
+      "learning_rate": 8.368407953869105e-06,
+      "logits/chosen": 0.05583259090781212,
+      "logits/rejected": 0.04287164285778999,
+      "logps/chosen": -92.72290802001953,
+      "logps/rejected": -99.79463195800781,
+      "loss": 1.8293,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.395390272140503,
+      "rewards/margins": 0.9053142666816711,
+      "rewards/rejected": -3.3007044792175293,
+      "step": 616
+    },
+    {
+      "epoch": 1.119891008174387,
+      "grad_norm": 1.6499502658843994,
+      "learning_rate": 8.363453724440471e-06,
+      "logits/chosen": 0.06658023595809937,
+      "logits/rejected": 0.04786865413188934,
+      "logps/chosen": -88.62269592285156,
+      "logps/rejected": -95.76766967773438,
+      "loss": 1.9608,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.522400379180908,
+      "rewards/margins": 0.8233134746551514,
+      "rewards/rejected": -3.3457136154174805,
+      "step": 617
+    },
+    {
+      "epoch": 1.1217075386012716,
+      "grad_norm": 2.018575668334961,
+      "learning_rate": 8.358493655931119e-06,
+      "logits/chosen": 0.06770970672369003,
+      "logits/rejected": 0.08037856221199036,
+      "logps/chosen": -82.23643493652344,
+      "logps/rejected": -97.71478271484375,
+      "loss": 1.8448,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.201833724975586,
+      "rewards/margins": 1.0529935359954834,
+      "rewards/rejected": -3.2548277378082275,
+      "step": 618
+    },
+    {
+      "epoch": 1.1235240690281563,
+      "grad_norm": 2.0033349990844727,
+      "learning_rate": 8.353527758455491e-06,
+      "logits/chosen": 0.11088447272777557,
+      "logits/rejected": 0.09635978937149048,
+      "logps/chosen": -81.90861511230469,
+      "logps/rejected": -92.48001098632812,
+      "loss": 2.2057,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.494230270385742,
+      "rewards/margins": 0.7774206399917603,
+      "rewards/rejected": -3.271650791168213,
+      "step": 619
+    },
+    {
+      "epoch": 1.125340599455041,
+      "grad_norm": 1.700210452079773,
+      "learning_rate": 8.348556042139918e-06,
+      "logits/chosen": 0.08965969830751419,
+      "logits/rejected": 0.06873422861099243,
+      "logps/chosen": -80.28114318847656,
+      "logps/rejected": -84.04791259765625,
+      "loss": 2.4366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.773008108139038,
+      "rewards/margins": 0.5039690732955933,
+      "rewards/rejected": -3.276977300643921,
+      "step": 620
+    },
+    {
+      "epoch": 1.1271571298819256,
+      "grad_norm": 2.623145580291748,
+      "learning_rate": 8.343578517122594e-06,
+      "logits/chosen": 0.08407651633024216,
+      "logits/rejected": 0.09678006917238235,
+      "logps/chosen": -81.90699005126953,
+      "logps/rejected": -94.30622863769531,
+      "loss": 2.126,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4066429138183594,
+      "rewards/margins": 0.8733029365539551,
+      "rewards/rejected": -3.2799456119537354,
+      "step": 621
+    },
+    {
+      "epoch": 1.12897366030881,
+      "grad_norm": 1.8616830110549927,
+      "learning_rate": 8.338595193553559e-06,
+      "logits/chosen": 0.002603452652692795,
+      "logits/rejected": 0.13452833890914917,
+      "logps/chosen": -73.35641479492188,
+      "logps/rejected": -84.70191192626953,
+      "loss": 2.084,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3735039234161377,
+      "rewards/margins": 0.7769300937652588,
+      "rewards/rejected": -3.1504342555999756,
+      "step": 622
+    },
+    {
+      "epoch": 1.1307901907356948,
+      "grad_norm": 1.8286879062652588,
+      "learning_rate": 8.333606081594678e-06,
+      "logits/chosen": 0.06104264035820961,
+      "logits/rejected": 0.10199623554944992,
+      "logps/chosen": -76.39376831054688,
+      "logps/rejected": -88.08850860595703,
+      "loss": 1.7885,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.130338191986084,
+      "rewards/margins": 0.848843514919281,
+      "rewards/rejected": -2.9791817665100098,
+      "step": 623
+    },
+    {
+      "epoch": 1.1326067211625794,
+      "grad_norm": 1.4803669452667236,
+      "learning_rate": 8.32861119141962e-06,
+      "logits/chosen": 0.07387635856866837,
+      "logits/rejected": 0.11718127131462097,
+      "logps/chosen": -82.92483520507812,
+      "logps/rejected": -91.46949005126953,
+      "loss": 1.7624,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.291823148727417,
+      "rewards/margins": 0.9700915217399597,
+      "rewards/rejected": -3.2619142532348633,
+      "step": 624
+    },
+    {
+      "epoch": 1.134423251589464,
+      "grad_norm": 1.6750752925872803,
+      "learning_rate": 8.323610533213835e-06,
+      "logits/chosen": 0.10784655809402466,
+      "logits/rejected": 0.17897658050060272,
+      "logps/chosen": -75.01900482177734,
+      "logps/rejected": -88.1203384399414,
+      "loss": 2.1956,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4190964698791504,
+      "rewards/margins": 0.688554584980011,
+      "rewards/rejected": -3.1076512336730957,
+      "step": 625
+    },
+    {
+      "epoch": 1.1362397820163488,
+      "grad_norm": 1.8641157150268555,
+      "learning_rate": 8.318604117174533e-06,
+      "logits/chosen": 0.05499793961644173,
+      "logits/rejected": 0.0563310906291008,
+      "logps/chosen": -83.42897033691406,
+      "logps/rejected": -94.2099380493164,
+      "loss": 1.675,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2912588119506836,
+      "rewards/margins": 0.9781790375709534,
+      "rewards/rejected": -3.2694382667541504,
+      "step": 626
+    },
+    {
+      "epoch": 1.1380563124432335,
+      "grad_norm": 1.7625656127929688,
+      "learning_rate": 8.313591953510674e-06,
+      "logits/chosen": 0.0831976979970932,
+      "logits/rejected": 0.12448858469724655,
+      "logps/chosen": -80.08613586425781,
+      "logps/rejected": -85.6810302734375,
+      "loss": 2.3784,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.453763008117676,
+      "rewards/margins": 0.46461760997772217,
+      "rewards/rejected": -2.9183804988861084,
+      "step": 627
+    },
+    {
+      "epoch": 1.1398728428701181,
+      "grad_norm": 2.10945463180542,
+      "learning_rate": 8.308574052442928e-06,
+      "logits/chosen": 0.09317971020936966,
+      "logits/rejected": 0.12335637211799622,
+      "logps/chosen": -82.36027526855469,
+      "logps/rejected": -94.93681335449219,
+      "loss": 2.1623,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5503153800964355,
+      "rewards/margins": 0.6077253222465515,
+      "rewards/rejected": -3.158040761947632,
+      "step": 628
+    },
+    {
+      "epoch": 1.1416893732970028,
+      "grad_norm": 1.8368538618087769,
+      "learning_rate": 8.30355042420367e-06,
+      "logits/chosen": 0.08310627192258835,
+      "logits/rejected": 0.11056395620107651,
+      "logps/chosen": -81.20982360839844,
+      "logps/rejected": -92.69551086425781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2482149600982666,
+      "rewards/margins": 0.9621560573577881,
+      "rewards/rejected": -3.2103710174560547,
+      "step": 629
+    },
+    {
+      "epoch": 1.1435059037238875,
+      "grad_norm": 1.7961188554763794,
+      "learning_rate": 8.29852107903695e-06,
+      "logits/chosen": 0.13976231217384338,
+      "logits/rejected": 0.11740799248218536,
+      "logps/chosen": -70.01700592041016,
+      "logps/rejected": -76.1086654663086,
+      "loss": 2.3401,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.239321231842041,
+      "rewards/margins": 0.5643521547317505,
+      "rewards/rejected": -2.80367374420166,
+      "step": 630
+    },
+    {
+      "epoch": 1.145322434150772,
+      "grad_norm": 2.041840076446533,
+      "learning_rate": 8.293486027198483e-06,
+      "logits/chosen": 0.06489598006010056,
+      "logits/rejected": 0.11150355637073517,
+      "logps/chosen": -95.31941223144531,
+      "logps/rejected": -96.98028564453125,
+      "loss": 2.2151,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5194008350372314,
+      "rewards/margins": 0.5945936441421509,
+      "rewards/rejected": -3.113994836807251,
+      "step": 631
+    },
+    {
+      "epoch": 1.1471389645776566,
+      "grad_norm": 2.7394556999206543,
+      "learning_rate": 8.288445278955615e-06,
+      "logits/chosen": 0.10089049488306046,
+      "logits/rejected": 0.08729755133390427,
+      "logps/chosen": -95.64276123046875,
+      "logps/rejected": -99.41658020019531,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.5987305641174316,
+      "rewards/margins": 0.4612637162208557,
+      "rewards/rejected": -3.0599937438964844,
+      "step": 632
+    },
+    {
+      "epoch": 1.1489554950045413,
+      "grad_norm": 2.024982213973999,
+      "learning_rate": 8.283398844587308e-06,
+      "logits/chosen": 0.023156292736530304,
+      "logits/rejected": 0.06789979338645935,
+      "logps/chosen": -89.90803527832031,
+      "logps/rejected": -96.06836700439453,
+      "loss": 1.9289,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.1415646076202393,
+      "rewards/margins": 0.7636609673500061,
+      "rewards/rejected": -2.9052255153656006,
+      "step": 633
+    },
+    {
+      "epoch": 1.150772025431426,
+      "grad_norm": 1.8750615119934082,
+      "learning_rate": 8.278346734384122e-06,
+      "logits/chosen": 0.02243964746594429,
+      "logits/rejected": 0.12660253047943115,
+      "logps/chosen": -71.34333038330078,
+      "logps/rejected": -86.67922973632812,
+      "loss": 1.8554,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.2015068531036377,
+      "rewards/margins": 0.9946616291999817,
+      "rewards/rejected": -3.1961684226989746,
+      "step": 634
+    },
+    {
+      "epoch": 1.1525885558583107,
+      "grad_norm": 1.9947502613067627,
+      "learning_rate": 8.27328895864819e-06,
+      "logits/chosen": 0.05387189984321594,
+      "logits/rejected": 0.08354412019252777,
+      "logps/chosen": -85.38788604736328,
+      "logps/rejected": -97.08065795898438,
+      "loss": 2.3917,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.595083236694336,
+      "rewards/margins": 0.5549375414848328,
+      "rewards/rejected": -3.1500213146209717,
+      "step": 635
+    },
+    {
+      "epoch": 1.1544050862851953,
+      "grad_norm": 2.145918607711792,
+      "learning_rate": 8.268225527693193e-06,
+      "logits/chosen": 0.10176774114370346,
+      "logits/rejected": 0.14395561814308167,
+      "logps/chosen": -76.02487182617188,
+      "logps/rejected": -86.05180358886719,
+      "loss": 1.8909,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.0001683235168457,
+      "rewards/margins": 0.8724240064620972,
+      "rewards/rejected": -2.8725922107696533,
+      "step": 636
+    },
+    {
+      "epoch": 1.15622161671208,
+      "grad_norm": 1.6830824613571167,
+      "learning_rate": 8.263156451844353e-06,
+      "logits/chosen": 0.10138621181249619,
+      "logits/rejected": 0.11966821551322937,
+      "logps/chosen": -76.56681060791016,
+      "logps/rejected": -88.1928939819336,
+      "loss": 1.9508,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.247589588165283,
+      "rewards/margins": 0.8286643028259277,
+      "rewards/rejected": -3.076253890991211,
+      "step": 637
+    },
+    {
+      "epoch": 1.1580381471389645,
+      "grad_norm": 2.0992536544799805,
+      "learning_rate": 8.258081741438396e-06,
+      "logits/chosen": 0.07541397213935852,
+      "logits/rejected": 0.09907414764165878,
+      "logps/chosen": -90.69495391845703,
+      "logps/rejected": -101.71548461914062,
+      "loss": 2.1933,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4336788654327393,
+      "rewards/margins": 0.7834519147872925,
+      "rewards/rejected": -3.217130422592163,
+      "step": 638
+    },
+    {
+      "epoch": 1.1598546775658491,
+      "grad_norm": 1.8066256046295166,
+      "learning_rate": 8.25300140682354e-06,
+      "logits/chosen": 0.14558832347393036,
+      "logits/rejected": 0.14358943700790405,
+      "logps/chosen": -82.73606872558594,
+      "logps/rejected": -86.27526092529297,
+      "loss": 2.0025,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.393693208694458,
+      "rewards/margins": 0.7131556272506714,
+      "rewards/rejected": -3.106848955154419,
+      "step": 639
+    },
+    {
+      "epoch": 1.1616712079927338,
+      "grad_norm": 2.3521695137023926,
+      "learning_rate": 8.247915458359473e-06,
+      "logits/chosen": 0.05236164107918739,
+      "logits/rejected": 0.1095740869641304,
+      "logps/chosen": -83.70420837402344,
+      "logps/rejected": -100.09117889404297,
+      "loss": 2.3348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5044684410095215,
+      "rewards/margins": 0.7525762319564819,
+      "rewards/rejected": -3.257044553756714,
+      "step": 640
+    },
+    {
+      "epoch": 1.1634877384196185,
+      "grad_norm": 2.343412160873413,
+      "learning_rate": 8.242823906417329e-06,
+      "logits/chosen": -0.00503757456317544,
+      "logits/rejected": 0.06355556100606918,
+      "logps/chosen": -80.41883087158203,
+      "logps/rejected": -95.32416534423828,
+      "loss": 2.2004,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3209891319274902,
+      "rewards/margins": 0.745349109172821,
+      "rewards/rejected": -3.066338539123535,
+      "step": 641
+    },
+    {
+      "epoch": 1.1653042688465032,
+      "grad_norm": 1.9915282726287842,
+      "learning_rate": 8.237726761379668e-06,
+      "logits/chosen": 0.02087419107556343,
+      "logits/rejected": 0.08556030690670013,
+      "logps/chosen": -84.41456604003906,
+      "logps/rejected": -97.61527252197266,
+      "loss": 2.0052,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.385272741317749,
+      "rewards/margins": 0.7285323739051819,
+      "rewards/rejected": -3.113805055618286,
+      "step": 642
+    },
+    {
+      "epoch": 1.1671207992733879,
+      "grad_norm": 2.4018874168395996,
+      "learning_rate": 8.232624033640458e-06,
+      "logits/chosen": 0.0840807855129242,
+      "logits/rejected": 0.03736239671707153,
+      "logps/chosen": -83.6908950805664,
+      "logps/rejected": -88.65196228027344,
+      "loss": 2.5072,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.494255542755127,
+      "rewards/margins": 0.5595706701278687,
+      "rewards/rejected": -3.053826332092285,
+      "step": 643
+    },
+    {
+      "epoch": 1.1689373297002725,
+      "grad_norm": 2.3055484294891357,
+      "learning_rate": 8.227515733605049e-06,
+      "logits/chosen": 0.05399080738425255,
+      "logits/rejected": 0.13454601168632507,
+      "logps/chosen": -77.62064361572266,
+      "logps/rejected": -86.54924011230469,
+      "loss": 2.0759,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3616065979003906,
+      "rewards/margins": 0.7132005095481873,
+      "rewards/rejected": -3.0748066902160645,
+      "step": 644
+    },
+    {
+      "epoch": 1.1707538601271572,
+      "grad_norm": 1.989559292793274,
+      "learning_rate": 8.222401871690153e-06,
+      "logits/chosen": 0.03339620307087898,
+      "logits/rejected": 0.06584692001342773,
+      "logps/chosen": -74.14878845214844,
+      "logps/rejected": -89.0107421875,
+      "loss": 1.9992,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3560426235198975,
+      "rewards/margins": 0.7924355864524841,
+      "rewards/rejected": -3.1484780311584473,
+      "step": 645
+    },
+    {
+      "epoch": 1.1725703905540419,
+      "grad_norm": 1.7572104930877686,
+      "learning_rate": 8.217282458323825e-06,
+      "logits/chosen": 0.09702566266059875,
+      "logits/rejected": 0.06184859201312065,
+      "logps/chosen": -82.05592346191406,
+      "logps/rejected": -87.59920501708984,
+      "loss": 2.0353,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.121622085571289,
+      "rewards/margins": 0.707017183303833,
+      "rewards/rejected": -2.828639268875122,
+      "step": 646
+    },
+    {
+      "epoch": 1.1743869209809263,
+      "grad_norm": 1.8265388011932373,
+      "learning_rate": 8.21215750394544e-06,
+      "logits/chosen": 0.12195339798927307,
+      "logits/rejected": 0.09588825702667236,
+      "logps/chosen": -76.14807891845703,
+      "logps/rejected": -85.36613464355469,
+      "loss": 2.0684,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2251272201538086,
+      "rewards/margins": 0.8967663049697876,
+      "rewards/rejected": -3.1218934059143066,
+      "step": 647
+    },
+    {
+      "epoch": 1.176203451407811,
+      "grad_norm": 1.9061174392700195,
+      "learning_rate": 8.207027019005675e-06,
+      "logits/chosen": 0.09651574492454529,
+      "logits/rejected": 0.10208628326654434,
+      "logps/chosen": -86.28733825683594,
+      "logps/rejected": -97.5046157836914,
+      "loss": 2.1319,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3506147861480713,
+      "rewards/margins": 0.7415947318077087,
+      "rewards/rejected": -3.0922091007232666,
+      "step": 648
+    },
+    {
+      "epoch": 1.1780199818346957,
+      "grad_norm": 1.999624252319336,
+      "learning_rate": 8.20189101396648e-06,
+      "logits/chosen": 0.08931821584701538,
+      "logits/rejected": 0.09364073723554611,
+      "logps/chosen": -84.32461547851562,
+      "logps/rejected": -96.70581817626953,
+      "loss": 1.7919,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4468250274658203,
+      "rewards/margins": 0.8684096932411194,
+      "rewards/rejected": -3.315234661102295,
+      "step": 649
+    },
+    {
+      "epoch": 1.1798365122615804,
+      "grad_norm": 1.5666590929031372,
+      "learning_rate": 8.196749499301062e-06,
+      "logits/chosen": 0.09945614635944366,
+      "logits/rejected": 0.12043865770101547,
+      "logps/chosen": -81.99935150146484,
+      "logps/rejected": -96.36174011230469,
+      "loss": 1.8932,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4319562911987305,
+      "rewards/margins": 0.8513790369033813,
+      "rewards/rejected": -3.2833354473114014,
+      "step": 650
+    },
+    {
+      "epoch": 1.181653042688465,
+      "grad_norm": 1.9403899908065796,
+      "learning_rate": 8.191602485493868e-06,
+      "logits/chosen": 0.12361271679401398,
+      "logits/rejected": 0.14710690081119537,
+      "logps/chosen": -82.11424255371094,
+      "logps/rejected": -89.37992858886719,
+      "loss": 2.0615,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.161297082901001,
+      "rewards/margins": 0.7220393419265747,
+      "rewards/rejected": -2.8833365440368652,
+      "step": 651
+    },
+    {
+      "epoch": 1.1834695731153497,
+      "grad_norm": 1.8558425903320312,
+      "learning_rate": 8.186449983040552e-06,
+      "logits/chosen": 0.07215605676174164,
+      "logits/rejected": 0.06390087306499481,
+      "logps/chosen": -81.64322662353516,
+      "logps/rejected": -85.39015197753906,
+      "loss": 2.0064,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.320831775665283,
+      "rewards/margins": 0.7350802421569824,
+      "rewards/rejected": -3.0559120178222656,
+      "step": 652
+    },
+    {
+      "epoch": 1.1852861035422344,
+      "grad_norm": 1.7956428527832031,
+      "learning_rate": 8.181292002447966e-06,
+      "logits/chosen": 0.11276388168334961,
+      "logits/rejected": 0.06517648696899414,
+      "logps/chosen": -76.14613342285156,
+      "logps/rejected": -82.24524688720703,
+      "loss": 2.0473,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2219183444976807,
+      "rewards/margins": 0.6231520175933838,
+      "rewards/rejected": -2.8450703620910645,
+      "step": 653
+    },
+    {
+      "epoch": 1.1871026339691189,
+      "grad_norm": 2.1648740768432617,
+      "learning_rate": 8.17612855423413e-06,
+      "logits/chosen": 0.0334198959171772,
+      "logits/rejected": 0.013921715319156647,
+      "logps/chosen": -83.02323150634766,
+      "logps/rejected": -92.95353698730469,
+      "loss": 1.9884,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.0769619941711426,
+      "rewards/margins": 0.851192057132721,
+      "rewards/rejected": -2.9281537532806396,
+      "step": 654
+    },
+    {
+      "epoch": 1.1889191643960035,
+      "grad_norm": 1.921157717704773,
+      "learning_rate": 8.170959648928214e-06,
+      "logits/chosen": 0.073361836373806,
+      "logits/rejected": 0.13112977147102356,
+      "logps/chosen": -83.41218566894531,
+      "logps/rejected": -92.03260803222656,
+      "loss": 2.0203,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7512502670288086,
+      "rewards/margins": 0.8279457092285156,
+      "rewards/rejected": -3.579195737838745,
+      "step": 655
+    },
+    {
+      "epoch": 1.1907356948228882,
+      "grad_norm": 2.4273247718811035,
+      "learning_rate": 8.165785297070516e-06,
+      "logits/chosen": 0.027622584253549576,
+      "logits/rejected": 0.08521658927202225,
+      "logps/chosen": -84.34210205078125,
+      "logps/rejected": -101.76324462890625,
+      "loss": 1.9023,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6731176376342773,
+      "rewards/margins": 1.004132866859436,
+      "rewards/rejected": -3.677250385284424,
+      "step": 656
+    },
+    {
+      "epoch": 1.192552225249773,
+      "grad_norm": 1.8262163400650024,
+      "learning_rate": 8.16060550921244e-06,
+      "logits/chosen": 0.03431132063269615,
+      "logits/rejected": 0.06454990059137344,
+      "logps/chosen": -76.59049987792969,
+      "logps/rejected": -97.41972351074219,
+      "loss": 2.0105,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3002851009368896,
+      "rewards/margins": 0.7388467788696289,
+      "rewards/rejected": -3.0391316413879395,
+      "step": 657
+    },
+    {
+      "epoch": 1.1943687556766576,
+      "grad_norm": 1.9906315803527832,
+      "learning_rate": 8.155420295916474e-06,
+      "logits/chosen": 0.08427983522415161,
+      "logits/rejected": 0.025464089587330818,
+      "logps/chosen": -84.34020233154297,
+      "logps/rejected": -92.31369018554688,
+      "loss": 1.9895,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3555400371551514,
+      "rewards/margins": 0.8297086954116821,
+      "rewards/rejected": -3.185249090194702,
+      "step": 658
+    },
+    {
+      "epoch": 1.1961852861035422,
+      "grad_norm": 1.9980812072753906,
+      "learning_rate": 8.150229667756172e-06,
+      "logits/chosen": 0.07556813955307007,
+      "logits/rejected": 0.07593435049057007,
+      "logps/chosen": -85.46858215332031,
+      "logps/rejected": -90.94758605957031,
+      "loss": 2.2687,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.679569959640503,
+      "rewards/margins": 0.5908665657043457,
+      "rewards/rejected": -3.2704365253448486,
+      "step": 659
+    },
+    {
+      "epoch": 1.198001816530427,
+      "grad_norm": 2.0758330821990967,
+      "learning_rate": 8.14503363531613e-06,
+      "logits/chosen": 0.11312856525182724,
+      "logits/rejected": 0.08777043223381042,
+      "logps/chosen": -80.32341003417969,
+      "logps/rejected": -84.89898681640625,
+      "loss": 2.1371,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5170207023620605,
+      "rewards/margins": 0.632743239402771,
+      "rewards/rejected": -3.149764060974121,
+      "step": 660
+    },
+    {
+      "epoch": 1.1998183469573116,
+      "grad_norm": 2.765526533126831,
+      "learning_rate": 8.139832209191961e-06,
+      "logits/chosen": 0.17475244402885437,
+      "logits/rejected": 0.13103297352790833,
+      "logps/chosen": -92.03308868408203,
+      "logps/rejected": -94.15422821044922,
+      "loss": 2.4151,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.3920867443084717,
+      "rewards/margins": 0.518372118473053,
+      "rewards/rejected": -2.910458564758301,
+      "step": 661
+    },
+    {
+      "epoch": 1.2016348773841963,
+      "grad_norm": 2.0190646648406982,
+      "learning_rate": 8.13462539999028e-06,
+      "logits/chosen": 0.040579523891210556,
+      "logits/rejected": 0.060216568410396576,
+      "logps/chosen": -96.32977294921875,
+      "logps/rejected": -105.37261962890625,
+      "loss": 2.1216,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5936977863311768,
+      "rewards/margins": 1.040100336074829,
+      "rewards/rejected": -3.6337978839874268,
+      "step": 662
+    },
+    {
+      "epoch": 1.2034514078110807,
+      "grad_norm": 2.53027606010437,
+      "learning_rate": 8.129413218328674e-06,
+      "logits/chosen": 0.07498883455991745,
+      "logits/rejected": 0.042916588485240936,
+      "logps/chosen": -83.17558288574219,
+      "logps/rejected": -89.55461883544922,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6350390911102295,
+      "rewards/margins": 0.7082449793815613,
+      "rewards/rejected": -3.3432838916778564,
+      "step": 663
+    },
+    {
+      "epoch": 1.2052679382379654,
+      "grad_norm": 1.6480835676193237,
+      "learning_rate": 8.124195674835695e-06,
+      "logits/chosen": 0.1062261089682579,
+      "logits/rejected": 0.09598012268543243,
+      "logps/chosen": -84.5193099975586,
+      "logps/rejected": -95.05423736572266,
+      "loss": 1.7544,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.1563069820404053,
+      "rewards/margins": 0.8859087228775024,
+      "rewards/rejected": -3.042215585708618,
+      "step": 664
+    },
+    {
+      "epoch": 1.20708446866485,
+      "grad_norm": 2.126664161682129,
+      "learning_rate": 8.118972780150817e-06,
+      "logits/chosen": -0.01870904304087162,
+      "logits/rejected": 0.1347496211528778,
+      "logps/chosen": -79.35530853271484,
+      "logps/rejected": -105.41546630859375,
+      "loss": 2.0524,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6006765365600586,
+      "rewards/margins": 1.0105364322662354,
+      "rewards/rejected": -3.611213207244873,
+      "step": 665
+    },
+    {
+      "epoch": 1.2089009990917348,
+      "grad_norm": 2.117645263671875,
+      "learning_rate": 8.113744544924434e-06,
+      "logits/chosen": 0.01287349034100771,
+      "logits/rejected": 0.019422955811023712,
+      "logps/chosen": -105.86250305175781,
+      "logps/rejected": -117.42082977294922,
+      "loss": 1.8973,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.636453151702881,
+      "rewards/margins": 0.8021495342254639,
+      "rewards/rejected": -3.4386026859283447,
+      "step": 666
+    },
+    {
+      "epoch": 1.2107175295186194,
+      "grad_norm": 2.956285238265991,
+      "learning_rate": 8.108510979817828e-06,
+      "logits/chosen": 0.047061942517757416,
+      "logits/rejected": 0.11715231090784073,
+      "logps/chosen": -89.56261444091797,
+      "logps/rejected": -103.57279968261719,
+      "loss": 2.0627,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5317227840423584,
+      "rewards/margins": 0.8301056623458862,
+      "rewards/rejected": -3.361828327178955,
+      "step": 667
+    },
+    {
+      "epoch": 1.2125340599455041,
+      "grad_norm": 1.897623062133789,
+      "learning_rate": 8.10327209550315e-06,
+      "logits/chosen": 0.017857536673545837,
+      "logits/rejected": 0.12088865786790848,
+      "logps/chosen": -95.75634002685547,
+      "logps/rejected": -113.0296630859375,
+      "loss": 1.9091,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6507534980773926,
+      "rewards/margins": 0.963871955871582,
+      "rewards/rejected": -3.6146254539489746,
+      "step": 668
+    },
+    {
+      "epoch": 1.2143505903723888,
+      "grad_norm": 2.176003932952881,
+      "learning_rate": 8.098027902663396e-06,
+      "logits/chosen": -0.0014993082731962204,
+      "logits/rejected": 0.11425516754388809,
+      "logps/chosen": -79.22991943359375,
+      "logps/rejected": -96.39210510253906,
+      "loss": 1.8977,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.582876443862915,
+      "rewards/margins": 0.9855522513389587,
+      "rewards/rejected": -3.5684289932250977,
+      "step": 669
+    },
+    {
+      "epoch": 1.2161671207992735,
+      "grad_norm": 2.140666961669922,
+      "learning_rate": 8.092778411992388e-06,
+      "logits/chosen": 0.04144468903541565,
+      "logits/rejected": 0.0556509830057621,
+      "logps/chosen": -80.35415649414062,
+      "logps/rejected": -97.16475677490234,
+      "loss": 1.9182,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.66176176071167,
+      "rewards/margins": 0.9624839425086975,
+      "rewards/rejected": -3.6242458820343018,
+      "step": 670
+    },
+    {
+      "epoch": 1.2179836512261581,
+      "grad_norm": 1.937705636024475,
+      "learning_rate": 8.087523634194755e-06,
+      "logits/chosen": 0.03470912575721741,
+      "logits/rejected": 0.12844812870025635,
+      "logps/chosen": -81.73896026611328,
+      "logps/rejected": -98.39558410644531,
+      "loss": 2.032,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.406296730041504,
+      "rewards/margins": 0.9326379895210266,
+      "rewards/rejected": -3.3389346599578857,
+      "step": 671
+    },
+    {
+      "epoch": 1.2198001816530426,
+      "grad_norm": 1.8903594017028809,
+      "learning_rate": 8.082263579985898e-06,
+      "logits/chosen": 0.03509046137332916,
+      "logits/rejected": 0.0663415864109993,
+      "logps/chosen": -81.79936981201172,
+      "logps/rejected": -96.15756225585938,
+      "loss": 1.7671,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2738354206085205,
+      "rewards/margins": 1.1403576135635376,
+      "rewards/rejected": -3.4141931533813477,
+      "step": 672
+    },
+    {
+      "epoch": 1.2216167120799273,
+      "grad_norm": 2.206265687942505,
+      "learning_rate": 8.076998260091989e-06,
+      "logits/chosen": 0.016251683235168457,
+      "logits/rejected": 0.07814528048038483,
+      "logps/chosen": -80.23258209228516,
+      "logps/rejected": -85.3895034790039,
+      "loss": 2.2589,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6450579166412354,
+      "rewards/margins": 0.5741185545921326,
+      "rewards/rejected": -3.2191762924194336,
+      "step": 673
+    },
+    {
+      "epoch": 1.223433242506812,
+      "grad_norm": 2.3217952251434326,
+      "learning_rate": 8.071727685249929e-06,
+      "logits/chosen": 0.15181973576545715,
+      "logits/rejected": 0.11617043614387512,
+      "logps/chosen": -85.15460205078125,
+      "logps/rejected": -83.82669830322266,
+      "loss": 2.2534,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.530156135559082,
+      "rewards/margins": 0.6635551452636719,
+      "rewards/rejected": -3.193711519241333,
+      "step": 674
+    },
+    {
+      "epoch": 1.2252497729336966,
+      "grad_norm": 2.011244297027588,
+      "learning_rate": 8.06645186620734e-06,
+      "logits/chosen": 0.02763158269226551,
+      "logits/rejected": 0.04441550746560097,
+      "logps/chosen": -81.14067840576172,
+      "logps/rejected": -95.907958984375,
+      "loss": 1.8979,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.693267583847046,
+      "rewards/margins": 0.9938480854034424,
+      "rewards/rejected": -3.6871156692504883,
+      "step": 675
+    },
+    {
+      "epoch": 1.2270663033605813,
+      "grad_norm": 2.3554768562316895,
+      "learning_rate": 8.061170813722533e-06,
+      "logits/chosen": 0.06379646062850952,
+      "logits/rejected": 0.052463918924331665,
+      "logps/chosen": -92.42412567138672,
+      "logps/rejected": -92.37240600585938,
+      "loss": 2.4093,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.7538669109344482,
+      "rewards/margins": 0.5145044326782227,
+      "rewards/rejected": -3.268371105194092,
+      "step": 676
+    },
+    {
+      "epoch": 1.228882833787466,
+      "grad_norm": 2.2725136280059814,
+      "learning_rate": 8.055884538564493e-06,
+      "logits/chosen": 0.028828933835029602,
+      "logits/rejected": 0.08082776516675949,
+      "logps/chosen": -76.20146179199219,
+      "logps/rejected": -94.82791900634766,
+      "loss": 1.7849,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.348304510116577,
+      "rewards/margins": 1.1067698001861572,
+      "rewards/rejected": -3.4550740718841553,
+      "step": 677
+    },
+    {
+      "epoch": 1.2306993642143507,
+      "grad_norm": 2.3565657138824463,
+      "learning_rate": 8.050593051512859e-06,
+      "logits/chosen": 0.09367989003658295,
+      "logits/rejected": 0.06132106855511665,
+      "logps/chosen": -84.85393524169922,
+      "logps/rejected": -91.87137603759766,
+      "loss": 2.162,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5945935249328613,
+      "rewards/margins": 0.7115421295166016,
+      "rewards/rejected": -3.306135416030884,
+      "step": 678
+    },
+    {
+      "epoch": 1.2325158946412351,
+      "grad_norm": 2.2214298248291016,
+      "learning_rate": 8.045296363357891e-06,
+      "logits/chosen": 0.015750454738736153,
+      "logits/rejected": -0.001254781149327755,
+      "logps/chosen": -94.99285125732422,
+      "logps/rejected": -107.42745971679688,
+      "loss": 2.0311,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.769968032836914,
+      "rewards/margins": 0.9211719632148743,
+      "rewards/rejected": -3.6911399364471436,
+      "step": 679
+    },
+    {
+      "epoch": 1.2343324250681198,
+      "grad_norm": 1.9377881288528442,
+      "learning_rate": 8.039994484900463e-06,
+      "logits/chosen": -0.011005287989974022,
+      "logits/rejected": 0.05877537280321121,
+      "logps/chosen": -84.62346649169922,
+      "logps/rejected": -99.59872436523438,
+      "loss": 1.8343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.554427146911621,
+      "rewards/margins": 1.086029291152954,
+      "rewards/rejected": -3.640456199645996,
+      "step": 680
+    },
+    {
+      "epoch": 1.2361489554950045,
+      "grad_norm": 2.665781259536743,
+      "learning_rate": 8.034687426952023e-06,
+      "logits/chosen": 0.030874190852046013,
+      "logits/rejected": 0.09339022636413574,
+      "logps/chosen": -82.25083923339844,
+      "logps/rejected": -99.68885803222656,
+      "loss": 2.0966,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.847683906555176,
+      "rewards/margins": 0.8753347396850586,
+      "rewards/rejected": -3.7230184078216553,
+      "step": 681
+    },
+    {
+      "epoch": 1.2379654859218892,
+      "grad_norm": 1.8999613523483276,
+      "learning_rate": 8.029375200334588e-06,
+      "logits/chosen": 0.07361680269241333,
+      "logits/rejected": 0.08493717014789581,
+      "logps/chosen": -88.70954895019531,
+      "logps/rejected": -102.10923767089844,
+      "loss": 1.7236,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.518230676651001,
+      "rewards/margins": 1.0492362976074219,
+      "rewards/rejected": -3.567467212677002,
+      "step": 682
+    },
+    {
+      "epoch": 1.2397820163487738,
+      "grad_norm": 2.179049491882324,
+      "learning_rate": 8.024057815880716e-06,
+      "logits/chosen": 0.12260061502456665,
+      "logits/rejected": 0.09896722435951233,
+      "logps/chosen": -82.61366271972656,
+      "logps/rejected": -89.37825775146484,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3436989784240723,
+      "rewards/margins": 0.7696498036384583,
+      "rewards/rejected": -3.1133487224578857,
+      "step": 683
+    },
+    {
+      "epoch": 1.2415985467756585,
+      "grad_norm": 1.8527759313583374,
+      "learning_rate": 8.018735284433475e-06,
+      "logits/chosen": 0.007166730239987373,
+      "logits/rejected": 0.04468049108982086,
+      "logps/chosen": -88.59947204589844,
+      "logps/rejected": -97.66645050048828,
+      "loss": 1.9353,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.669034957885742,
+      "rewards/margins": 0.8667396306991577,
+      "rewards/rejected": -3.5357747077941895,
+      "step": 684
+    },
+    {
+      "epoch": 1.2434150772025432,
+      "grad_norm": 2.013792037963867,
+      "learning_rate": 8.013407616846436e-06,
+      "logits/chosen": 0.05774620547890663,
+      "logits/rejected": 0.06977065652608871,
+      "logps/chosen": -83.88308715820312,
+      "logps/rejected": -93.36985778808594,
+      "loss": 1.9716,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5362911224365234,
+      "rewards/margins": 0.9699288010597229,
+      "rewards/rejected": -3.5062198638916016,
+      "step": 685
+    },
+    {
+      "epoch": 1.2452316076294279,
+      "grad_norm": 1.9147852659225464,
+      "learning_rate": 8.008074823983642e-06,
+      "logits/chosen": 0.10071337223052979,
+      "logits/rejected": 0.09696929156780243,
+      "logps/chosen": -87.06709289550781,
+      "logps/rejected": -98.84783172607422,
+      "loss": 1.7989,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7123823165893555,
+      "rewards/margins": 0.9586774110794067,
+      "rewards/rejected": -3.6710598468780518,
+      "step": 686
+    },
+    {
+      "epoch": 1.2470481380563125,
+      "grad_norm": 3.8477859497070312,
+      "learning_rate": 8.002736916719582e-06,
+      "logits/chosen": 0.09164638817310333,
+      "logits/rejected": 0.14311164617538452,
+      "logps/chosen": -94.75381469726562,
+      "logps/rejected": -101.14093017578125,
+      "loss": 2.3677,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7352569103240967,
+      "rewards/margins": 0.5952449440956116,
+      "rewards/rejected": -3.3305020332336426,
+      "step": 687
+    },
+    {
+      "epoch": 1.248864668483197,
+      "grad_norm": 2.1601197719573975,
+      "learning_rate": 7.997393905939183e-06,
+      "logits/chosen": 0.14653781056404114,
+      "logits/rejected": 0.217861607670784,
+      "logps/chosen": -78.05258178710938,
+      "logps/rejected": -87.37406921386719,
+      "loss": 2.2395,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4850776195526123,
+      "rewards/margins": 0.5597268342971802,
+      "rewards/rejected": -3.044804573059082,
+      "step": 688
+    },
+    {
+      "epoch": 1.2506811989100817,
+      "grad_norm": 2.1657280921936035,
+      "learning_rate": 7.992045802537772e-06,
+      "logits/chosen": 0.008183173835277557,
+      "logits/rejected": 0.07762319594621658,
+      "logps/chosen": -76.81200408935547,
+      "logps/rejected": -89.77823638916016,
+      "loss": 2.0968,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.395460844039917,
+      "rewards/margins": 0.7663928270339966,
+      "rewards/rejected": -3.161853790283203,
+      "step": 689
+    },
+    {
+      "epoch": 1.2524977293369663,
+      "grad_norm": 2.0229945182800293,
+      "learning_rate": 7.986692617421063e-06,
+      "logits/chosen": 0.049828238785266876,
+      "logits/rejected": 0.12920916080474854,
+      "logps/chosen": -89.49681854248047,
+      "logps/rejected": -94.05538177490234,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5402889251708984,
+      "rewards/margins": 0.6041077375411987,
+      "rewards/rejected": -3.1443967819213867,
+      "step": 690
+    },
+    {
+      "epoch": 1.254314259763851,
+      "grad_norm": 2.1632978916168213,
+      "learning_rate": 7.981334361505131e-06,
+      "logits/chosen": 0.01283620297908783,
+      "logits/rejected": 0.06363464891910553,
+      "logps/chosen": -90.30138397216797,
+      "logps/rejected": -109.25923156738281,
+      "loss": 1.9464,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5346744060516357,
+      "rewards/margins": 1.0385441780090332,
+      "rewards/rejected": -3.573218584060669,
+      "step": 691
+    },
+    {
+      "epoch": 1.2561307901907357,
+      "grad_norm": 2.1562981605529785,
+      "learning_rate": 7.975971045716398e-06,
+      "logits/chosen": 0.12652680277824402,
+      "logits/rejected": 0.15553821623325348,
+      "logps/chosen": -74.45803833007812,
+      "logps/rejected": -90.87694549560547,
+      "loss": 1.8599,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.415904998779297,
+      "rewards/margins": 0.9220768809318542,
+      "rewards/rejected": -3.337982177734375,
+      "step": 692
+    },
+    {
+      "epoch": 1.2579473206176204,
+      "grad_norm": 2.007781982421875,
+      "learning_rate": 7.970602680991594e-06,
+      "logits/chosen": 0.05228351801633835,
+      "logits/rejected": 0.12716761231422424,
+      "logps/chosen": -96.6414794921875,
+      "logps/rejected": -115.6935043334961,
+      "loss": 2.0777,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5941665172576904,
+      "rewards/margins": 1.015256643295288,
+      "rewards/rejected": -3.6094233989715576,
+      "step": 693
+    },
+    {
+      "epoch": 1.259763851044505,
+      "grad_norm": 2.2393176555633545,
+      "learning_rate": 7.965229278277749e-06,
+      "logits/chosen": 0.11106812953948975,
+      "logits/rejected": 0.10329913347959518,
+      "logps/chosen": -91.51345825195312,
+      "logps/rejected": -95.78245544433594,
+      "loss": 2.1322,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6126160621643066,
+      "rewards/margins": 0.8137804865837097,
+      "rewards/rejected": -3.426396608352661,
+      "step": 694
+    },
+    {
+      "epoch": 1.2615803814713895,
+      "grad_norm": 1.9000614881515503,
+      "learning_rate": 7.959850848532172e-06,
+      "logits/chosen": 0.053980953991413116,
+      "logits/rejected": 0.1285300850868225,
+      "logps/chosen": -80.65451049804688,
+      "logps/rejected": -96.31999206542969,
+      "loss": 1.8157,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.449226140975952,
+      "rewards/margins": 0.9224156141281128,
+      "rewards/rejected": -3.3716418743133545,
+      "step": 695
+    },
+    {
+      "epoch": 1.2633969118982744,
+      "grad_norm": 2.1616997718811035,
+      "learning_rate": 7.95446740272241e-06,
+      "logits/chosen": 0.0943533331155777,
+      "logits/rejected": 0.09738673269748688,
+      "logps/chosen": -86.31460571289062,
+      "logps/rejected": -95.93961334228516,
+      "loss": 2.3068,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4675066471099854,
+      "rewards/margins": 0.5155179500579834,
+      "rewards/rejected": -2.9830245971679688,
+      "step": 696
+    },
+    {
+      "epoch": 1.2652134423251589,
+      "grad_norm": 2.040947675704956,
+      "learning_rate": 7.949078951826254e-06,
+      "logits/chosen": 0.05092187598347664,
+      "logits/rejected": 0.10424528270959854,
+      "logps/chosen": -78.41385650634766,
+      "logps/rejected": -96.27998352050781,
+      "loss": 1.9729,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.392333507537842,
+      "rewards/margins": 0.8541865348815918,
+      "rewards/rejected": -3.246520519256592,
+      "step": 697
+    },
+    {
+      "epoch": 1.2670299727520435,
+      "grad_norm": 2.123605251312256,
+      "learning_rate": 7.943685506831689e-06,
+      "logits/chosen": 0.1472083032131195,
+      "logits/rejected": 0.07182697206735611,
+      "logps/chosen": -87.48722839355469,
+      "logps/rejected": -89.31311798095703,
+      "loss": 2.1536,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.34977388381958,
+      "rewards/margins": 0.6212242245674133,
+      "rewards/rejected": -2.9709980487823486,
+      "step": 698
+    },
+    {
+      "epoch": 1.2688465031789282,
+      "grad_norm": 2.11053729057312,
+      "learning_rate": 7.938287078736889e-06,
+      "logits/chosen": 0.08341722190380096,
+      "logits/rejected": 0.12655527889728546,
+      "logps/chosen": -79.16002655029297,
+      "logps/rejected": -91.57137298583984,
+      "loss": 2.0396,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4659664630889893,
+      "rewards/margins": 0.7776908874511719,
+      "rewards/rejected": -3.2436575889587402,
+      "step": 699
+    },
+    {
+      "epoch": 1.270663033605813,
+      "grad_norm": 1.9555423259735107,
+      "learning_rate": 7.932883678550191e-06,
+      "logits/chosen": -0.06148137152194977,
+      "logits/rejected": 0.055927492678165436,
+      "logps/chosen": -81.51953125,
+      "logps/rejected": -95.884521484375,
+      "loss": 1.9223,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5550904273986816,
+      "rewards/margins": 0.8093082904815674,
+      "rewards/rejected": -3.364398956298828,
+      "step": 700
+    },
+    {
+      "epoch": 1.2724795640326976,
+      "grad_norm": 2.031662702560425,
+      "learning_rate": 7.927475317290065e-06,
+      "logits/chosen": 0.04235261678695679,
+      "logits/rejected": 0.1255102902650833,
+      "logps/chosen": -74.28742218017578,
+      "logps/rejected": -86.79241180419922,
+      "loss": 1.9669,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.138885498046875,
+      "rewards/margins": 0.8455573916435242,
+      "rewards/rejected": -2.984442949295044,
+      "step": 701
+    },
+    {
+      "epoch": 1.2742960944595823,
+      "grad_norm": 2.131220817565918,
+      "learning_rate": 7.922062005985105e-06,
+      "logits/chosen": 0.15667293965816498,
+      "logits/rejected": 0.12293773889541626,
+      "logps/chosen": -76.34903717041016,
+      "logps/rejected": -80.57623291015625,
+      "loss": 2.2422,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.408144474029541,
+      "rewards/margins": 0.5897243618965149,
+      "rewards/rejected": -2.997868776321411,
+      "step": 702
+    },
+    {
+      "epoch": 1.276112624886467,
+      "grad_norm": 1.9162285327911377,
+      "learning_rate": 7.916643755673994e-06,
+      "logits/chosen": -0.00397343747317791,
+      "logits/rejected": -0.01795889064669609,
+      "logps/chosen": -72.12644958496094,
+      "logps/rejected": -80.48601531982422,
+      "loss": 2.286,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1793017387390137,
+      "rewards/margins": 0.6844039559364319,
+      "rewards/rejected": -2.863705635070801,
+      "step": 703
+    },
+    {
+      "epoch": 1.2779291553133514,
+      "grad_norm": 2.877826690673828,
+      "learning_rate": 7.911220577405485e-06,
+      "logits/chosen": 0.031911637634038925,
+      "logits/rejected": -0.02112415060400963,
+      "logps/chosen": -92.17039489746094,
+      "logps/rejected": -100.05145263671875,
+      "loss": 2.2664,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5374324321746826,
+      "rewards/margins": 0.6277021765708923,
+      "rewards/rejected": -3.1651346683502197,
+      "step": 704
+    },
+    {
+      "epoch": 1.2797456857402363,
+      "grad_norm": 2.437596082687378,
+      "learning_rate": 7.905792482238387e-06,
+      "logits/chosen": 0.05161967873573303,
+      "logits/rejected": 0.07132750749588013,
+      "logps/chosen": -83.6578140258789,
+      "logps/rejected": -91.72476959228516,
+      "loss": 2.2835,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.62272310256958,
+      "rewards/margins": 0.7025635242462158,
+      "rewards/rejected": -3.325286865234375,
+      "step": 705
+    },
+    {
+      "epoch": 1.2815622161671207,
+      "grad_norm": 1.9355360269546509,
+      "learning_rate": 7.900359481241531e-06,
+      "logits/chosen": 0.11531795561313629,
+      "logits/rejected": 0.1957155019044876,
+      "logps/chosen": -70.70565795898438,
+      "logps/rejected": -83.15331268310547,
+      "loss": 1.9752,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.568023681640625,
+      "rewards/margins": 0.748910129070282,
+      "rewards/rejected": -3.3169333934783936,
+      "step": 706
+    },
+    {
+      "epoch": 1.2833787465940054,
+      "grad_norm": 1.9978153705596924,
+      "learning_rate": 7.89492158549375e-06,
+      "logits/chosen": 0.04371386021375656,
+      "logits/rejected": 0.07748796790838242,
+      "logps/chosen": -80.5776138305664,
+      "logps/rejected": -90.11798095703125,
+      "loss": 2.0422,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1579325199127197,
+      "rewards/margins": 0.726677656173706,
+      "rewards/rejected": -2.8846099376678467,
+      "step": 707
+    },
+    {
+      "epoch": 1.28519527702089,
+      "grad_norm": 1.982696771621704,
+      "learning_rate": 7.889478806083859e-06,
+      "logits/chosen": 0.11542633175849915,
+      "logits/rejected": 0.09008399397134781,
+      "logps/chosen": -80.96235656738281,
+      "logps/rejected": -84.60002899169922,
+      "loss": 2.1817,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.0073299407958984,
+      "rewards/margins": 0.626833438873291,
+      "rewards/rejected": -2.6341633796691895,
+      "step": 708
+    },
+    {
+      "epoch": 1.2870118074477748,
+      "grad_norm": 2.2399492263793945,
+      "learning_rate": 7.884031154110636e-06,
+      "logits/chosen": 0.13016283512115479,
+      "logits/rejected": 0.10277393460273743,
+      "logps/chosen": -76.04215240478516,
+      "logps/rejected": -82.029296875,
+      "loss": 2.1021,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.4932732582092285,
+      "rewards/margins": 0.7273231744766235,
+      "rewards/rejected": -3.2205960750579834,
+      "step": 709
+    },
+    {
+      "epoch": 1.2888283378746594,
+      "grad_norm": 1.7780169248580933,
+      "learning_rate": 7.87857864068279e-06,
+      "logits/chosen": 0.015022790059447289,
+      "logits/rejected": 0.08707302063703537,
+      "logps/chosen": -85.66944122314453,
+      "logps/rejected": -101.31831359863281,
+      "loss": 1.8779,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4976251125335693,
+      "rewards/margins": 0.9009003639221191,
+      "rewards/rejected": -3.3985257148742676,
+      "step": 710
+    },
+    {
+      "epoch": 1.2906448683015441,
+      "grad_norm": 2.2210092544555664,
+      "learning_rate": 7.873121276918947e-06,
+      "logits/chosen": 0.02679639309644699,
+      "logits/rejected": 0.00411232328042388,
+      "logps/chosen": -87.88114929199219,
+      "logps/rejected": -98.84779357910156,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.56654953956604,
+      "rewards/margins": 0.7424182295799255,
+      "rewards/rejected": -3.3089675903320312,
+      "step": 711
+    },
+    {
+      "epoch": 1.2924613987284288,
+      "grad_norm": 2.3803563117980957,
+      "learning_rate": 7.867659073947618e-06,
+      "logits/chosen": 0.12869741022586823,
+      "logits/rejected": 0.17313969135284424,
+      "logps/chosen": -78.2127685546875,
+      "logps/rejected": -89.54530334472656,
+      "loss": 2.2313,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2719531059265137,
+      "rewards/margins": 0.6112346649169922,
+      "rewards/rejected": -2.883187770843506,
+      "step": 712
+    },
+    {
+      "epoch": 1.2942779291553133,
+      "grad_norm": 1.8979357481002808,
+      "learning_rate": 7.862192042907191e-06,
+      "logits/chosen": -0.002289058640599251,
+      "logits/rejected": 0.01941063068807125,
+      "logps/chosen": -79.63894653320312,
+      "logps/rejected": -99.8336410522461,
+      "loss": 1.8824,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.396660566329956,
+      "rewards/margins": 1.1458237171173096,
+      "rewards/rejected": -3.5424838066101074,
+      "step": 713
+    },
+    {
+      "epoch": 1.296094459582198,
+      "grad_norm": 2.069608449935913,
+      "learning_rate": 7.856720194945892e-06,
+      "logits/chosen": 0.061066634953022,
+      "logits/rejected": 0.0689641609787941,
+      "logps/chosen": -78.00001525878906,
+      "logps/rejected": -87.9764404296875,
+      "loss": 2.0765,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3631246089935303,
+      "rewards/margins": 0.8202179670333862,
+      "rewards/rejected": -3.183342695236206,
+      "step": 714
+    },
+    {
+      "epoch": 1.2979109900090826,
+      "grad_norm": 2.080536365509033,
+      "learning_rate": 7.85124354122177e-06,
+      "logits/chosen": 0.055848486721515656,
+      "logits/rejected": 0.09090499579906464,
+      "logps/chosen": -86.07003784179688,
+      "logps/rejected": -95.0962905883789,
+      "loss": 2.0101,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5897908210754395,
+      "rewards/margins": 0.6455370187759399,
+      "rewards/rejected": -3.23532772064209,
+      "step": 715
+    },
+    {
+      "epoch": 1.2997275204359673,
+      "grad_norm": 1.9409488439559937,
+      "learning_rate": 7.845762092902678e-06,
+      "logits/chosen": 0.17071956396102905,
+      "logits/rejected": 0.11957748234272003,
+      "logps/chosen": -85.08267211914062,
+      "logps/rejected": -100.11673736572266,
+      "loss": 1.9544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2637064456939697,
+      "rewards/margins": 0.9579347968101501,
+      "rewards/rejected": -3.2216408252716064,
+      "step": 716
+    },
+    {
+      "epoch": 1.301544050862852,
+      "grad_norm": 2.3988044261932373,
+      "learning_rate": 7.840275861166243e-06,
+      "logits/chosen": 0.07587475329637527,
+      "logits/rejected": 0.018805254250764847,
+      "logps/chosen": -92.8540267944336,
+      "logps/rejected": -97.47180938720703,
+      "loss": 2.3578,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.429079294204712,
+      "rewards/margins": 0.5412003993988037,
+      "rewards/rejected": -2.9702794551849365,
+      "step": 717
+    },
+    {
+      "epoch": 1.3033605812897366,
+      "grad_norm": 2.509941577911377,
+      "learning_rate": 7.834784857199846e-06,
+      "logits/chosen": 0.0847601592540741,
+      "logits/rejected": 0.039042674005031586,
+      "logps/chosen": -80.2916030883789,
+      "logps/rejected": -81.85697174072266,
+      "loss": 2.0909,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4719648361206055,
+      "rewards/margins": 0.6977176666259766,
+      "rewards/rejected": -3.169682264328003,
+      "step": 718
+    },
+    {
+      "epoch": 1.3051771117166213,
+      "grad_norm": 2.192031145095825,
+      "learning_rate": 7.8292890922006e-06,
+      "logits/chosen": 0.048360832035541534,
+      "logits/rejected": 0.10979942232370377,
+      "logps/chosen": -72.04978942871094,
+      "logps/rejected": -85.74111938476562,
+      "loss": 2.1544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.367837905883789,
+      "rewards/margins": 0.6129733920097351,
+      "rewards/rejected": -2.98081111907959,
+      "step": 719
+    },
+    {
+      "epoch": 1.3069936421435058,
+      "grad_norm": 2.33526873588562,
+      "learning_rate": 7.82378857737533e-06,
+      "logits/chosen": 0.09924662113189697,
+      "logits/rejected": 0.10467047989368439,
+      "logps/chosen": -85.21525573730469,
+      "logps/rejected": -94.86468505859375,
+      "loss": 2.0783,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.466061592102051,
+      "rewards/margins": 0.7927428483963013,
+      "rewards/rejected": -3.2588045597076416,
+      "step": 720
+    },
+    {
+      "epoch": 1.3088101725703907,
+      "grad_norm": 2.381256341934204,
+      "learning_rate": 7.818283323940539e-06,
+      "logits/chosen": 0.1228807270526886,
+      "logits/rejected": 0.1338506042957306,
+      "logps/chosen": -90.72122192382812,
+      "logps/rejected": -96.53125,
+      "loss": 2.0252,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4308130741119385,
+      "rewards/margins": 0.6035470366477966,
+      "rewards/rejected": -3.03436017036438,
+      "step": 721
+    },
+    {
+      "epoch": 1.3106267029972751,
+      "grad_norm": 1.942866325378418,
+      "learning_rate": 7.8127733431224e-06,
+      "logits/chosen": 0.05967065691947937,
+      "logits/rejected": 0.08946052193641663,
+      "logps/chosen": -78.91156768798828,
+      "logps/rejected": -90.6188735961914,
+      "loss": 1.9435,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.333373546600342,
+      "rewards/margins": 0.7353364825248718,
+      "rewards/rejected": -3.0687103271484375,
+      "step": 722
+    },
+    {
+      "epoch": 1.3124432334241598,
+      "grad_norm": 2.068608283996582,
+      "learning_rate": 7.80725864615672e-06,
+      "logits/chosen": 0.01783425733447075,
+      "logits/rejected": 0.06930528581142426,
+      "logps/chosen": -81.48194885253906,
+      "logps/rejected": -96.17157745361328,
+      "loss": 2.0554,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.423696756362915,
+      "rewards/margins": 0.8400081396102905,
+      "rewards/rejected": -3.263705253601074,
+      "step": 723
+    },
+    {
+      "epoch": 1.3142597638510445,
+      "grad_norm": 2.349059820175171,
+      "learning_rate": 7.801739244288929e-06,
+      "logits/chosen": 0.08266536891460419,
+      "logits/rejected": 0.09768007695674896,
+      "logps/chosen": -82.81443786621094,
+      "logps/rejected": -92.47273254394531,
+      "loss": 1.8288,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.347062587738037,
+      "rewards/margins": 0.839061975479126,
+      "rewards/rejected": -3.186124563217163,
+      "step": 724
+    },
+    {
+      "epoch": 1.3160762942779292,
+      "grad_norm": 2.1312410831451416,
+      "learning_rate": 7.79621514877405e-06,
+      "logits/chosen": 0.05873649939894676,
+      "logits/rejected": 0.10713300108909607,
+      "logps/chosen": -99.603759765625,
+      "logps/rejected": -110.6041259765625,
+      "loss": 1.8875,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4636247158050537,
+      "rewards/margins": 0.8619670271873474,
+      "rewards/rejected": -3.325591564178467,
+      "step": 725
+    },
+    {
+      "epoch": 1.3178928247048138,
+      "grad_norm": 2.137871742248535,
+      "learning_rate": 7.79068637087667e-06,
+      "logits/chosen": 0.07839182764291763,
+      "logits/rejected": 0.1225314810872078,
+      "logps/chosen": -82.73197937011719,
+      "logps/rejected": -101.63593292236328,
+      "loss": 1.7664,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.491330862045288,
+      "rewards/margins": 1.1549298763275146,
+      "rewards/rejected": -3.646261215209961,
+      "step": 726
+    },
+    {
+      "epoch": 1.3197093551316985,
+      "grad_norm": 2.030689001083374,
+      "learning_rate": 7.785152921870939e-06,
+      "logits/chosen": 0.034951481968164444,
+      "logits/rejected": 0.06388817727565765,
+      "logps/chosen": -85.28425598144531,
+      "logps/rejected": -101.43257904052734,
+      "loss": 2.3856,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6969833374023438,
+      "rewards/margins": 0.727889895439148,
+      "rewards/rejected": -3.4248733520507812,
+      "step": 727
+    },
+    {
+      "epoch": 1.3215258855585832,
+      "grad_norm": 2.65952205657959,
+      "learning_rate": 7.779614813040513e-06,
+      "logits/chosen": 0.1679934561252594,
+      "logits/rejected": 0.18826396763324738,
+      "logps/chosen": -80.64529418945312,
+      "logps/rejected": -87.35059356689453,
+      "loss": 1.9723,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.259064197540283,
+      "rewards/margins": 0.7135547995567322,
+      "rewards/rejected": -2.97261905670166,
+      "step": 728
+    },
+    {
+      "epoch": 1.3233424159854676,
+      "grad_norm": 2.1841068267822266,
+      "learning_rate": 7.774072055678565e-06,
+      "logits/chosen": 0.11853618919849396,
+      "logits/rejected": 0.1412927508354187,
+      "logps/chosen": -80.26649475097656,
+      "logps/rejected": -88.19939422607422,
+      "loss": 2.1344,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5820248126983643,
+      "rewards/margins": 0.8093788623809814,
+      "rewards/rejected": -3.3914034366607666,
+      "step": 729
+    },
+    {
+      "epoch": 1.3251589464123525,
+      "grad_norm": 2.463083505630493,
+      "learning_rate": 7.768524661087744e-06,
+      "logits/chosen": 0.05979528650641441,
+      "logits/rejected": 0.10566453635692596,
+      "logps/chosen": -88.6331787109375,
+      "logps/rejected": -97.57271575927734,
+      "loss": 2.1545,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4209208488464355,
+      "rewards/margins": 0.7680116295814514,
+      "rewards/rejected": -3.188931941986084,
+      "step": 730
+    },
+    {
+      "epoch": 1.326975476839237,
+      "grad_norm": 2.3693671226501465,
+      "learning_rate": 7.76297264058015e-06,
+      "logits/chosen": 0.05617160722613335,
+      "logits/rejected": 0.017447492107748985,
+      "logps/chosen": -91.15104675292969,
+      "logps/rejected": -91.32392883300781,
+      "loss": 2.0652,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.469024181365967,
+      "rewards/margins": 0.7454814910888672,
+      "rewards/rejected": -3.214505672454834,
+      "step": 731
+    },
+    {
+      "epoch": 1.3287920072661217,
+      "grad_norm": 2.0553715229034424,
+      "learning_rate": 7.757416005477323e-06,
+      "logits/chosen": 0.10786914825439453,
+      "logits/rejected": 0.1281609684228897,
+      "logps/chosen": -83.35604095458984,
+      "logps/rejected": -91.0313949584961,
+      "loss": 2.0932,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.75126314163208,
+      "rewards/margins": 0.7679738402366638,
+      "rewards/rejected": -3.5192370414733887,
+      "step": 732
+    },
+    {
+      "epoch": 1.3306085376930064,
+      "grad_norm": 2.6259214878082275,
+      "learning_rate": 7.751854767110205e-06,
+      "logits/chosen": 0.025147411972284317,
+      "logits/rejected": 0.11643064022064209,
+      "logps/chosen": -74.00305938720703,
+      "logps/rejected": -99.70087432861328,
+      "loss": 1.9757,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.617929220199585,
+      "rewards/margins": 1.131290078163147,
+      "rewards/rejected": -3.7492191791534424,
+      "step": 733
+    },
+    {
+      "epoch": 1.332425068119891,
+      "grad_norm": 2.21140193939209,
+      "learning_rate": 7.746288936819134e-06,
+      "logits/chosen": 0.036435484886169434,
+      "logits/rejected": 0.09736932814121246,
+      "logps/chosen": -79.6103515625,
+      "logps/rejected": -97.5123519897461,
+      "loss": 1.999,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.716804027557373,
+      "rewards/margins": 0.9280580282211304,
+      "rewards/rejected": -3.644861936569214,
+      "step": 734
+    },
+    {
+      "epoch": 1.3342415985467757,
+      "grad_norm": 1.8063753843307495,
+      "learning_rate": 7.740718525953804e-06,
+      "logits/chosen": 0.012957969680428505,
+      "logits/rejected": 0.06407644599676132,
+      "logps/chosen": -86.7850112915039,
+      "logps/rejected": -104.12468719482422,
+      "loss": 1.8674,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5402910709381104,
+      "rewards/margins": 0.9172405004501343,
+      "rewards/rejected": -3.457531690597534,
+      "step": 735
+    },
+    {
+      "epoch": 1.3360581289736604,
+      "grad_norm": 2.384291410446167,
+      "learning_rate": 7.735143545873252e-06,
+      "logits/chosen": 0.15138015151023865,
+      "logits/rejected": 0.1674022376537323,
+      "logps/chosen": -78.28397369384766,
+      "logps/rejected": -88.92375183105469,
+      "loss": 1.8622,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6077139377593994,
+      "rewards/margins": 0.9488463401794434,
+      "rewards/rejected": -3.5565602779388428,
+      "step": 736
+    },
+    {
+      "epoch": 1.337874659400545,
+      "grad_norm": 2.1717827320098877,
+      "learning_rate": 7.729564007945834e-06,
+      "logits/chosen": 0.07187292724847794,
+      "logits/rejected": 0.12309285253286362,
+      "logps/chosen": -87.5180892944336,
+      "logps/rejected": -93.48638153076172,
+      "loss": 2.2686,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4702401161193848,
+      "rewards/margins": 0.6287782192230225,
+      "rewards/rejected": -3.099018096923828,
+      "step": 737
+    },
+    {
+      "epoch": 1.3396911898274295,
+      "grad_norm": 2.6149754524230957,
+      "learning_rate": 7.723979923549201e-06,
+      "logits/chosen": 0.10500997304916382,
+      "logits/rejected": 0.13640286028385162,
+      "logps/chosen": -76.4052963256836,
+      "logps/rejected": -84.28958892822266,
+      "loss": 2.4316,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.9436187744140625,
+      "rewards/margins": 0.5192191004753113,
+      "rewards/rejected": -3.4628381729125977,
+      "step": 738
+    },
+    {
+      "epoch": 1.3415077202543142,
+      "grad_norm": 2.2604503631591797,
+      "learning_rate": 7.718391304070274e-06,
+      "logits/chosen": 0.019504152238368988,
+      "logits/rejected": 0.0677918866276741,
+      "logps/chosen": -96.57980346679688,
+      "logps/rejected": -111.43043518066406,
+      "loss": 2.1929,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.63236141204834,
+      "rewards/margins": 0.8397969007492065,
+      "rewards/rejected": -3.4721579551696777,
+      "step": 739
+    },
+    {
+      "epoch": 1.3433242506811989,
+      "grad_norm": 2.827580451965332,
+      "learning_rate": 7.712798160905218e-06,
+      "logits/chosen": 0.028927450999617577,
+      "logits/rejected": 0.060569021850824356,
+      "logps/chosen": -85.33325958251953,
+      "logps/rejected": -87.87812042236328,
+      "loss": 2.2098,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.669008255004883,
+      "rewards/margins": 0.6498023271560669,
+      "rewards/rejected": -3.31881046295166,
+      "step": 740
+    },
+    {
+      "epoch": 1.3451407811080835,
+      "grad_norm": 2.375645160675049,
+      "learning_rate": 7.707200505459427e-06,
+      "logits/chosen": 0.04526009410619736,
+      "logits/rejected": 0.0799727588891983,
+      "logps/chosen": -89.62974548339844,
+      "logps/rejected": -99.64259338378906,
+      "loss": 2.2352,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.709292411804199,
+      "rewards/margins": 0.6063066720962524,
+      "rewards/rejected": -3.3155994415283203,
+      "step": 741
+    },
+    {
+      "epoch": 1.3469573115349682,
+      "grad_norm": 2.4972751140594482,
+      "learning_rate": 7.701598349147496e-06,
+      "logits/chosen": -0.020723650231957436,
+      "logits/rejected": 0.04764425382018089,
+      "logps/chosen": -85.17987060546875,
+      "logps/rejected": -98.80146789550781,
+      "loss": 2.0034,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.656331777572632,
+      "rewards/margins": 0.9565580487251282,
+      "rewards/rejected": -3.6128900051116943,
+      "step": 742
+    },
+    {
+      "epoch": 1.348773841961853,
+      "grad_norm": 2.23154616355896,
+      "learning_rate": 7.695991703393196e-06,
+      "logits/chosen": 0.062117062509059906,
+      "logits/rejected": 0.018935926258563995,
+      "logps/chosen": -99.47203063964844,
+      "logps/rejected": -104.25543212890625,
+      "loss": 2.3257,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.538417339324951,
+      "rewards/margins": 0.5829829573631287,
+      "rewards/rejected": -3.1214003562927246,
+      "step": 743
+    },
+    {
+      "epoch": 1.3505903723887376,
+      "grad_norm": 2.1097710132598877,
+      "learning_rate": 7.690380579629454e-06,
+      "logits/chosen": 0.07178185135126114,
+      "logits/rejected": 0.13060711324214935,
+      "logps/chosen": -82.14419555664062,
+      "logps/rejected": -94.78469848632812,
+      "loss": 1.9158,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.5844056606292725,
+      "rewards/margins": 0.95278000831604,
+      "rewards/rejected": -3.5371859073638916,
+      "step": 744
+    },
+    {
+      "epoch": 1.352406902815622,
+      "grad_norm": 2.6826353073120117,
+      "learning_rate": 7.684764989298329e-06,
+      "logits/chosen": 0.10977941751480103,
+      "logits/rejected": 0.11871352046728134,
+      "logps/chosen": -87.78453826904297,
+      "logps/rejected": -88.22404479980469,
+      "loss": 2.405,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.647420644760132,
+      "rewards/margins": 0.5518589019775391,
+      "rewards/rejected": -3.199279546737671,
+      "step": 745
+    },
+    {
+      "epoch": 1.354223433242507,
+      "grad_norm": 2.235769510269165,
+      "learning_rate": 7.679144943850986e-06,
+      "logits/chosen": 0.05031013488769531,
+      "logits/rejected": 0.01593434251844883,
+      "logps/chosen": -94.69499206542969,
+      "logps/rejected": -99.25582885742188,
+      "loss": 2.4038,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.653748035430908,
+      "rewards/margins": 0.7265360355377197,
+      "rewards/rejected": -3.380283832550049,
+      "step": 746
+    },
+    {
+      "epoch": 1.3560399636693914,
+      "grad_norm": 2.4417803287506104,
+      "learning_rate": 7.673520454747681e-06,
+      "logits/chosen": 0.04040508344769478,
+      "logits/rejected": 0.07281962782144547,
+      "logps/chosen": -77.71315002441406,
+      "logps/rejected": -94.82330322265625,
+      "loss": 1.6934,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4112367630004883,
+      "rewards/margins": 1.1407581567764282,
+      "rewards/rejected": -3.551994800567627,
+      "step": 747
+    },
+    {
+      "epoch": 1.357856494096276,
+      "grad_norm": 1.7434958219528198,
+      "learning_rate": 7.667891533457718e-06,
+      "logits/chosen": 0.1453830599784851,
+      "logits/rejected": 0.10198508203029633,
+      "logps/chosen": -79.32072448730469,
+      "logps/rejected": -88.78571319580078,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6976542472839355,
+      "rewards/margins": 0.8518580198287964,
+      "rewards/rejected": -3.5495121479034424,
+      "step": 748
+    },
+    {
+      "epoch": 1.3596730245231607,
+      "grad_norm": 2.364915609359741,
+      "learning_rate": 7.662258191459456e-06,
+      "logits/chosen": 0.07202528417110443,
+      "logits/rejected": 0.07270742952823639,
+      "logps/chosen": -82.5189208984375,
+      "logps/rejected": -87.38330841064453,
+      "loss": 2.2639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.583577871322632,
+      "rewards/margins": 0.5575169324874878,
+      "rewards/rejected": -3.141094923019409,
+      "step": 749
+    },
+    {
+      "epoch": 1.3614895549500454,
+      "grad_norm": 1.9622827768325806,
+      "learning_rate": 7.656620440240257e-06,
+      "logits/chosen": 0.0625351071357727,
+      "logits/rejected": 0.1307908296585083,
+      "logps/chosen": -84.89385223388672,
+      "logps/rejected": -92.00534057617188,
+      "loss": 2.0835,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.679396629333496,
+      "rewards/margins": 0.7283416986465454,
+      "rewards/rejected": -3.407738208770752,
+      "step": 750
+    },
+    {
+      "epoch": 1.36330608537693,
+      "grad_norm": 3.2049853801727295,
+      "learning_rate": 7.650978291296475e-06,
+      "logits/chosen": 0.06182331219315529,
+      "logits/rejected": 0.08293959498405457,
+      "logps/chosen": -77.00846099853516,
+      "logps/rejected": -85.75076293945312,
+      "loss": 2.1026,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4312846660614014,
+      "rewards/margins": 0.7030216455459595,
+      "rewards/rejected": -3.134305953979492,
+      "step": 751
+    },
+    {
+      "epoch": 1.3651226158038148,
+      "grad_norm": 2.065100908279419,
+      "learning_rate": 7.645331756133438e-06,
+      "logits/chosen": 0.12972712516784668,
+      "logits/rejected": 0.12206097692251205,
+      "logps/chosen": -88.62772369384766,
+      "logps/rejected": -98.35625457763672,
+      "loss": 1.8117,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.59391450881958,
+      "rewards/margins": 0.8645298480987549,
+      "rewards/rejected": -3.458444833755493,
+      "step": 752
+    },
+    {
+      "epoch": 1.3669391462306995,
+      "grad_norm": 2.078524112701416,
+      "learning_rate": 7.639680846265411e-06,
+      "logits/chosen": 0.09852910041809082,
+      "logits/rejected": 0.12210299074649811,
+      "logps/chosen": -81.41858673095703,
+      "logps/rejected": -92.6720199584961,
+      "loss": 1.9148,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4295332431793213,
+      "rewards/margins": 0.8047889471054077,
+      "rewards/rejected": -3.2343220710754395,
+      "step": 753
+    },
+    {
+      "epoch": 1.368755676657584,
+      "grad_norm": 1.991837739944458,
+      "learning_rate": 7.634025573215585e-06,
+      "logits/chosen": -0.02604215405881405,
+      "logits/rejected": 0.048631928861141205,
+      "logps/chosen": -95.89967346191406,
+      "logps/rejected": -109.60630798339844,
+      "loss": 1.9067,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.491032361984253,
+      "rewards/margins": 1.0193102359771729,
+      "rewards/rejected": -3.5103423595428467,
+      "step": 754
+    },
+    {
+      "epoch": 1.3705722070844686,
+      "grad_norm": 2.560757875442505,
+      "learning_rate": 7.628365948516047e-06,
+      "logits/chosen": 0.12844571471214294,
+      "logits/rejected": 0.11145608872175217,
+      "logps/chosen": -81.85939025878906,
+      "logps/rejected": -92.21177673339844,
+      "loss": 2.195,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6527202129364014,
+      "rewards/margins": 0.7286384105682373,
+      "rewards/rejected": -3.3813586235046387,
+      "step": 755
+    },
+    {
+      "epoch": 1.3723887375113533,
+      "grad_norm": 2.5189437866210938,
+      "learning_rate": 7.622701983707756e-06,
+      "logits/chosen": 0.049444377422332764,
+      "logits/rejected": 0.07628104835748672,
+      "logps/chosen": -80.98538970947266,
+      "logps/rejected": -89.72940063476562,
+      "loss": 2.1411,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7696704864501953,
+      "rewards/margins": 0.7333941459655762,
+      "rewards/rejected": -3.5030641555786133,
+      "step": 756
+    },
+    {
+      "epoch": 1.374205267938238,
+      "grad_norm": 2.5741689205169678,
+      "learning_rate": 7.617033690340523e-06,
+      "logits/chosen": 0.02790415659546852,
+      "logits/rejected": 0.0068018268793821335,
+      "logps/chosen": -83.76559448242188,
+      "logps/rejected": -91.61640167236328,
+      "loss": 2.203,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.734015941619873,
+      "rewards/margins": 0.6714599132537842,
+      "rewards/rejected": -3.405475616455078,
+      "step": 757
+    },
+    {
+      "epoch": 1.3760217983651226,
+      "grad_norm": 1.7192715406417847,
+      "learning_rate": 7.611361079972984e-06,
+      "logits/chosen": 0.1048116460442543,
+      "logits/rejected": 0.13254055380821228,
+      "logps/chosen": -67.2003402709961,
+      "logps/rejected": -91.35997009277344,
+      "loss": 1.5709,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.3325514793395996,
+      "rewards/margins": 1.2846351861953735,
+      "rewards/rejected": -3.617187023162842,
+      "step": 758
+    },
+    {
+      "epoch": 1.3778383287920073,
+      "grad_norm": 2.5107131004333496,
+      "learning_rate": 7.605684164172581e-06,
+      "logits/chosen": 0.04970414191484451,
+      "logits/rejected": 0.06727412343025208,
+      "logps/chosen": -78.22857666015625,
+      "logps/rejected": -91.56071472167969,
+      "loss": 2.2132,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.5088820457458496,
+      "rewards/margins": 0.8137573003768921,
+      "rewards/rejected": -3.322639226913452,
+      "step": 759
+    },
+    {
+      "epoch": 1.379654859218892,
+      "grad_norm": 2.327336311340332,
+      "learning_rate": 7.600002954515532e-06,
+      "logits/chosen": 0.09795573353767395,
+      "logits/rejected": 0.04695986211299896,
+      "logps/chosen": -85.50299835205078,
+      "logps/rejected": -92.48783111572266,
+      "loss": 1.9836,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.55703067779541,
+      "rewards/margins": 0.7251678109169006,
+      "rewards/rejected": -3.282198429107666,
+      "step": 760
+    },
+    {
+      "epoch": 1.3814713896457766,
+      "grad_norm": 2.2488017082214355,
+      "learning_rate": 7.5943174625868146e-06,
+      "logits/chosen": 0.00528271310031414,
+      "logits/rejected": 0.05780588090419769,
+      "logps/chosen": -90.95349884033203,
+      "logps/rejected": -98.3586654663086,
+      "loss": 2.4998,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.74594783782959,
+      "rewards/margins": 0.4672420024871826,
+      "rewards/rejected": -3.2131900787353516,
+      "step": 761
+    },
+    {
+      "epoch": 1.3832879200726613,
+      "grad_norm": 2.4573490619659424,
+      "learning_rate": 7.588627699980134e-06,
+      "logits/chosen": 0.023196734488010406,
+      "logits/rejected": 0.07495579123497009,
+      "logps/chosen": -77.51380157470703,
+      "logps/rejected": -95.95264434814453,
+      "loss": 1.8657,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.519115447998047,
+      "rewards/margins": 1.0751680135726929,
+      "rewards/rejected": -3.59428334236145,
+      "step": 762
+    },
+    {
+      "epoch": 1.3851044504995458,
+      "grad_norm": 2.718414783477783,
+      "learning_rate": 7.582933678297911e-06,
+      "logits/chosen": -0.006191038526594639,
+      "logits/rejected": 0.08378598839044571,
+      "logps/chosen": -85.86323547363281,
+      "logps/rejected": -98.59344482421875,
+      "loss": 2.3292,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.69970703125,
+      "rewards/margins": 0.7494994401931763,
+      "rewards/rejected": -3.4492063522338867,
+      "step": 763
+    },
+    {
+      "epoch": 1.3869209809264305,
+      "grad_norm": 3.296757936477661,
+      "learning_rate": 7.577235409151242e-06,
+      "logits/chosen": 0.03207925334572792,
+      "logits/rejected": 0.11028114706277847,
+      "logps/chosen": -77.22196960449219,
+      "logps/rejected": -87.56900024414062,
+      "loss": 2.5069,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.9215140342712402,
+      "rewards/margins": 0.6305817365646362,
+      "rewards/rejected": -3.552095890045166,
+      "step": 764
+    },
+    {
+      "epoch": 1.3887375113533151,
+      "grad_norm": 2.6371073722839355,
+      "learning_rate": 7.5715329041598955e-06,
+      "logits/chosen": 0.06983280181884766,
+      "logits/rejected": 0.0031841248273849487,
+      "logps/chosen": -93.26805877685547,
+      "logps/rejected": -100.43709564208984,
+      "loss": 2.1158,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.602299928665161,
+      "rewards/margins": 0.8267786502838135,
+      "rewards/rejected": -3.4290785789489746,
+      "step": 765
+    },
+    {
+      "epoch": 1.3905540417801998,
+      "grad_norm": 2.557405471801758,
+      "learning_rate": 7.565826174952267e-06,
+      "logits/chosen": 0.0866529792547226,
+      "logits/rejected": 0.11220981180667877,
+      "logps/chosen": -81.10444641113281,
+      "logps/rejected": -89.5517807006836,
+      "loss": 2.0158,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.469362258911133,
+      "rewards/margins": 0.6984151601791382,
+      "rewards/rejected": -3.1677775382995605,
+      "step": 766
+    },
+    {
+      "epoch": 1.3923705722070845,
+      "grad_norm": 2.013252019882202,
+      "learning_rate": 7.560115233165375e-06,
+      "logits/chosen": 0.10826343297958374,
+      "logits/rejected": 0.09555835276842117,
+      "logps/chosen": -90.81183624267578,
+      "logps/rejected": -99.46356201171875,
+      "loss": 1.9386,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6625754833221436,
+      "rewards/margins": 0.9187412858009338,
+      "rewards/rejected": -3.5813164710998535,
+      "step": 767
+    },
+    {
+      "epoch": 1.3941871026339692,
+      "grad_norm": 2.1441266536712646,
+      "learning_rate": 7.554400090444822e-06,
+      "logits/chosen": 0.04958781227469444,
+      "logits/rejected": 0.10418076813220978,
+      "logps/chosen": -73.46155548095703,
+      "logps/rejected": -83.24677276611328,
+      "loss": 1.9933,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4550204277038574,
+      "rewards/margins": 0.7676887512207031,
+      "rewards/rejected": -3.2227089405059814,
+      "step": 768
+    },
+    {
+      "epoch": 1.3960036330608538,
+      "grad_norm": 2.401198387145996,
+      "learning_rate": 7.54868075844478e-06,
+      "logits/chosen": 0.10376621782779694,
+      "logits/rejected": 0.1655123233795166,
+      "logps/chosen": -80.3701171875,
+      "logps/rejected": -95.8973159790039,
+      "loss": 1.784,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2211532592773438,
+      "rewards/margins": 1.070261836051941,
+      "rewards/rejected": -3.291415214538574,
+      "step": 769
+    },
+    {
+      "epoch": 1.3978201634877383,
+      "grad_norm": 1.93374502658844,
+      "learning_rate": 7.5429572488279615e-06,
+      "logits/chosen": 0.081766776740551,
+      "logits/rejected": 0.1179564967751503,
+      "logps/chosen": -73.5924301147461,
+      "logps/rejected": -85.84019470214844,
+      "loss": 2.2637,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3363659381866455,
+      "rewards/margins": 0.6945525407791138,
+      "rewards/rejected": -3.030918598175049,
+      "step": 770
+    },
+    {
+      "epoch": 1.3996366939146232,
+      "grad_norm": 2.4809932708740234,
+      "learning_rate": 7.5372295732656e-06,
+      "logits/chosen": 0.11731807887554169,
+      "logits/rejected": 0.05732544884085655,
+      "logps/chosen": -89.25398254394531,
+      "logps/rejected": -90.94881439208984,
+      "loss": 2.412,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.435483694076538,
+      "rewards/margins": 0.5316743850708008,
+      "rewards/rejected": -2.967158317565918,
+      "step": 771
+    },
+    {
+      "epoch": 1.4014532243415077,
+      "grad_norm": 2.622253894805908,
+      "learning_rate": 7.531497743437424e-06,
+      "logits/chosen": -0.02307087928056717,
+      "logits/rejected": 0.034644536674022675,
+      "logps/chosen": -79.53402709960938,
+      "logps/rejected": -95.38764953613281,
+      "loss": 2.1841,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5010921955108643,
+      "rewards/margins": 0.7504414916038513,
+      "rewards/rejected": -3.2515335083007812,
+      "step": 772
+    },
+    {
+      "epoch": 1.4032697547683923,
+      "grad_norm": 2.341421365737915,
+      "learning_rate": 7.525761771031632e-06,
+      "logits/chosen": 0.09784074872732162,
+      "logits/rejected": 0.1302230954170227,
+      "logps/chosen": -78.1878433227539,
+      "logps/rejected": -88.39629364013672,
+      "loss": 2.064,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.63604736328125,
+      "rewards/margins": 0.6544671654701233,
+      "rewards/rejected": -3.2905144691467285,
+      "step": 773
+    },
+    {
+      "epoch": 1.405086285195277,
+      "grad_norm": 2.4410409927368164,
+      "learning_rate": 7.520021667744869e-06,
+      "logits/chosen": 0.03553074598312378,
+      "logits/rejected": 0.024142932146787643,
+      "logps/chosen": -92.7217025756836,
+      "logps/rejected": -99.62353515625,
+      "loss": 2.077,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5202903747558594,
+      "rewards/margins": 0.9078298211097717,
+      "rewards/rejected": -3.4281198978424072,
+      "step": 774
+    },
+    {
+      "epoch": 1.4069028156221617,
+      "grad_norm": 2.1428725719451904,
+      "learning_rate": 7.514277445282206e-06,
+      "logits/chosen": 0.12562477588653564,
+      "logits/rejected": 0.11905661970376968,
+      "logps/chosen": -85.4616470336914,
+      "logps/rejected": -101.02755737304688,
+      "loss": 1.8298,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.236478328704834,
+      "rewards/margins": 0.9414302110671997,
+      "rewards/rejected": -3.177908420562744,
+      "step": 775
+    },
+    {
+      "epoch": 1.4087193460490464,
+      "grad_norm": 2.2605643272399902,
+      "learning_rate": 7.508529115357115e-06,
+      "logits/chosen": 0.08080364018678665,
+      "logits/rejected": 0.14329738914966583,
+      "logps/chosen": -86.23857116699219,
+      "logps/rejected": -94.77734375,
+      "loss": 2.1025,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3885529041290283,
+      "rewards/margins": 0.8040337562561035,
+      "rewards/rejected": -3.1925864219665527,
+      "step": 776
+    },
+    {
+      "epoch": 1.410535876475931,
+      "grad_norm": 2.1767449378967285,
+      "learning_rate": 7.502776689691439e-06,
+      "logits/chosen": 0.061047837138175964,
+      "logits/rejected": 0.049486950039863586,
+      "logps/chosen": -79.15450286865234,
+      "logps/rejected": -89.98368072509766,
+      "loss": 2.032,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.129742383956909,
+      "rewards/margins": 0.7515033483505249,
+      "rewards/rejected": -2.8812458515167236,
+      "step": 777
+    },
+    {
+      "epoch": 1.4123524069028157,
+      "grad_norm": 2.2512662410736084,
+      "learning_rate": 7.4970201800153765e-06,
+      "logits/chosen": 0.08744536340236664,
+      "logits/rejected": 0.10895106941461563,
+      "logps/chosen": -68.41326141357422,
+      "logps/rejected": -82.72618103027344,
+      "loss": 2.0954,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.360013008117676,
+      "rewards/margins": 0.856154203414917,
+      "rewards/rejected": -3.2161672115325928,
+      "step": 778
+    },
+    {
+      "epoch": 1.4141689373297002,
+      "grad_norm": 2.1528515815734863,
+      "learning_rate": 7.491259598067452e-06,
+      "logits/chosen": 0.10324890911579132,
+      "logits/rejected": 0.017273597419261932,
+      "logps/chosen": -101.93807983398438,
+      "logps/rejected": -110.01669311523438,
+      "loss": 1.8012,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2577784061431885,
+      "rewards/margins": 0.8829382658004761,
+      "rewards/rejected": -3.140716552734375,
+      "step": 779
+    },
+    {
+      "epoch": 1.4159854677565848,
+      "grad_norm": 2.24306321144104,
+      "learning_rate": 7.485494955594495e-06,
+      "logits/chosen": 0.0638672262430191,
+      "logits/rejected": 0.08847285807132721,
+      "logps/chosen": -83.1998291015625,
+      "logps/rejected": -87.86129760742188,
+      "loss": 2.215,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.975947380065918,
+      "rewards/margins": 0.5703001022338867,
+      "rewards/rejected": -2.546247720718384,
+      "step": 780
+    },
+    {
+      "epoch": 1.4178019981834695,
+      "grad_norm": 2.103116035461426,
+      "learning_rate": 7.479726264351619e-06,
+      "logits/chosen": 0.09146749973297119,
+      "logits/rejected": 0.12321461737155914,
+      "logps/chosen": -81.99695587158203,
+      "logps/rejected": -89.16120910644531,
+      "loss": 1.9022,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.159066677093506,
+      "rewards/margins": 0.8253484964370728,
+      "rewards/rejected": -2.984415054321289,
+      "step": 781
+    },
+    {
+      "epoch": 1.4196185286103542,
+      "grad_norm": 2.1531925201416016,
+      "learning_rate": 7.473953536102185e-06,
+      "logits/chosen": 0.1091499999165535,
+      "logits/rejected": 0.13049989938735962,
+      "logps/chosen": -81.74616241455078,
+      "logps/rejected": -94.27207946777344,
+      "loss": 1.9222,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.445281982421875,
+      "rewards/margins": 0.8645297288894653,
+      "rewards/rejected": -3.30981183052063,
+      "step": 782
+    },
+    {
+      "epoch": 1.4214350590372389,
+      "grad_norm": 2.489908218383789,
+      "learning_rate": 7.468176782617793e-06,
+      "logits/chosen": -0.04253482073545456,
+      "logits/rejected": 0.10589614510536194,
+      "logps/chosen": -83.19894409179688,
+      "logps/rejected": -99.28282928466797,
+      "loss": 1.7606,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2363641262054443,
+      "rewards/margins": 1.0474493503570557,
+      "rewards/rejected": -3.2838134765625,
+      "step": 783
+    },
+    {
+      "epoch": 1.4232515894641236,
+      "grad_norm": 2.4771764278411865,
+      "learning_rate": 7.462396015678249e-06,
+      "logits/chosen": 0.02688070759177208,
+      "logits/rejected": 0.05978942662477493,
+      "logps/chosen": -75.48037719726562,
+      "logps/rejected": -87.0803451538086,
+      "loss": 1.892,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3419995307922363,
+      "rewards/margins": 0.8640164732933044,
+      "rewards/rejected": -3.2060160636901855,
+      "step": 784
+    },
+    {
+      "epoch": 1.4250681198910082,
+      "grad_norm": 2.3507192134857178,
+      "learning_rate": 7.456611247071544e-06,
+      "logits/chosen": 0.13474830985069275,
+      "logits/rejected": 0.08549933135509491,
+      "logps/chosen": -78.67109680175781,
+      "logps/rejected": -80.96075439453125,
+      "loss": 2.1411,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.477113723754883,
+      "rewards/margins": 0.652525007724762,
+      "rewards/rejected": -3.129638671875,
+      "step": 785
+    },
+    {
+      "epoch": 1.4268846503178927,
+      "grad_norm": 2.009307861328125,
+      "learning_rate": 7.450822488593829e-06,
+      "logits/chosen": 0.04329455643892288,
+      "logits/rejected": 0.05247782543301582,
+      "logps/chosen": -89.56121063232422,
+      "logps/rejected": -99.5736083984375,
+      "loss": 1.8014,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2546510696411133,
+      "rewards/margins": 0.9092133045196533,
+      "rewards/rejected": -3.1638646125793457,
+      "step": 786
+    },
+    {
+      "epoch": 1.4287011807447776,
+      "grad_norm": 2.3687586784362793,
+      "learning_rate": 7.445029752049387e-06,
+      "logits/chosen": 0.04104599729180336,
+      "logits/rejected": 0.08001314103603363,
+      "logps/chosen": -77.9354248046875,
+      "logps/rejected": -92.67184448242188,
+      "loss": 1.7508,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2342448234558105,
+      "rewards/margins": 0.9252514839172363,
+      "rewards/rejected": -3.159496307373047,
+      "step": 787
+    },
+    {
+      "epoch": 1.430517711171662,
+      "grad_norm": 1.8127901554107666,
+      "learning_rate": 7.439233049250622e-06,
+      "logits/chosen": 0.07227849960327148,
+      "logits/rejected": 0.0907898098230362,
+      "logps/chosen": -71.49815368652344,
+      "logps/rejected": -82.32070922851562,
+      "loss": 1.8207,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.33886456489563,
+      "rewards/margins": 0.906674325466156,
+      "rewards/rejected": -3.2455389499664307,
+      "step": 788
+    },
+    {
+      "epoch": 1.4323342415985467,
+      "grad_norm": 2.0310990810394287,
+      "learning_rate": 7.433432392018013e-06,
+      "logits/chosen": 0.07254546135663986,
+      "logits/rejected": 0.08750150352716446,
+      "logps/chosen": -87.65977478027344,
+      "logps/rejected": -105.3295669555664,
+      "loss": 1.781,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.4844467639923096,
+      "rewards/margins": 0.8295308351516724,
+      "rewards/rejected": -3.3139777183532715,
+      "step": 789
+    },
+    {
+      "epoch": 1.4341507720254314,
+      "grad_norm": 2.093871831893921,
+      "learning_rate": 7.427627792180116e-06,
+      "logits/chosen": 0.09880155324935913,
+      "logits/rejected": 0.15118218958377838,
+      "logps/chosen": -75.81531524658203,
+      "logps/rejected": -88.76158142089844,
+      "loss": 2.2499,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6412038803100586,
+      "rewards/margins": 0.7693830132484436,
+      "rewards/rejected": -3.4105868339538574,
+      "step": 790
+    },
+    {
+      "epoch": 1.435967302452316,
+      "grad_norm": 2.393105983734131,
+      "learning_rate": 7.4218192615735185e-06,
+      "logits/chosen": 0.08802710473537445,
+      "logits/rejected": 0.04610452428460121,
+      "logps/chosen": -99.03189849853516,
+      "logps/rejected": -103.37187194824219,
+      "loss": 1.8155,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3743648529052734,
+      "rewards/margins": 0.8601830005645752,
+      "rewards/rejected": -3.2345480918884277,
+      "step": 791
+    },
+    {
+      "epoch": 1.4377838328792008,
+      "grad_norm": 2.148087978363037,
+      "learning_rate": 7.416006812042827e-06,
+      "logits/chosen": 0.05752583593130112,
+      "logits/rejected": 0.034685466438531876,
+      "logps/chosen": -89.9745101928711,
+      "logps/rejected": -97.91671752929688,
+      "loss": 1.9892,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.483428716659546,
+      "rewards/margins": 0.7583959102630615,
+      "rewards/rejected": -3.2418243885040283,
+      "step": 792
+    },
+    {
+      "epoch": 1.4396003633060854,
+      "grad_norm": 2.605093479156494,
+      "learning_rate": 7.410190455440638e-06,
+      "logits/chosen": 0.07939667254686356,
+      "logits/rejected": 0.05258120596408844,
+      "logps/chosen": -81.38958740234375,
+      "logps/rejected": -92.0197982788086,
+      "loss": 1.8609,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4127330780029297,
+      "rewards/margins": 0.8712365627288818,
+      "rewards/rejected": -3.2839698791503906,
+      "step": 793
+    },
+    {
+      "epoch": 1.44141689373297,
+      "grad_norm": 2.3762154579162598,
+      "learning_rate": 7.4043702036275136e-06,
+      "logits/chosen": 0.04353059083223343,
+      "logits/rejected": 0.08619395643472672,
+      "logps/chosen": -83.58119201660156,
+      "logps/rejected": -94.14823150634766,
+      "loss": 2.4321,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.5087859630584717,
+      "rewards/margins": 0.4951530396938324,
+      "rewards/rejected": -3.003939151763916,
+      "step": 794
+    },
+    {
+      "epoch": 1.4432334241598546,
+      "grad_norm": 2.350267171859741,
+      "learning_rate": 7.398546068471961e-06,
+      "logits/chosen": 0.18240118026733398,
+      "logits/rejected": 0.14964665472507477,
+      "logps/chosen": -79.20641326904297,
+      "logps/rejected": -86.13677978515625,
+      "loss": 2.1999,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4007363319396973,
+      "rewards/margins": 0.6775516867637634,
+      "rewards/rejected": -3.0782880783081055,
+      "step": 795
+    },
+    {
+      "epoch": 1.4450499545867395,
+      "grad_norm": 2.183232545852661,
+      "learning_rate": 7.392718061850407e-06,
+      "logits/chosen": 0.06714559346437454,
+      "logits/rejected": 0.046004436910152435,
+      "logps/chosen": -95.4114990234375,
+      "logps/rejected": -101.2634048461914,
+      "loss": 1.9802,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4638829231262207,
+      "rewards/margins": 0.7256942987442017,
+      "rewards/rejected": -3.189577579498291,
+      "step": 796
+    },
+    {
+      "epoch": 1.446866485013624,
+      "grad_norm": 2.2636830806732178,
+      "learning_rate": 7.3868861956471725e-06,
+      "logits/chosen": 0.025870440527796745,
+      "logits/rejected": 0.05159129202365875,
+      "logps/chosen": -80.84700775146484,
+      "logps/rejected": -88.84320831298828,
+      "loss": 1.7361,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.451986312866211,
+      "rewards/margins": 0.9137101173400879,
+      "rewards/rejected": -3.3656961917877197,
+      "step": 797
+    },
+    {
+      "epoch": 1.4486830154405086,
+      "grad_norm": 2.3395354747772217,
+      "learning_rate": 7.381050481754447e-06,
+      "logits/chosen": 0.06643021106719971,
+      "logits/rejected": 0.1107374057173729,
+      "logps/chosen": -88.24702453613281,
+      "logps/rejected": -97.00286865234375,
+      "loss": 2.3738,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.6666600704193115,
+      "rewards/margins": 0.5626463294029236,
+      "rewards/rejected": -3.22930645942688,
+      "step": 798
+    },
+    {
+      "epoch": 1.4504995458673933,
+      "grad_norm": 2.936711549758911,
+      "learning_rate": 7.375210932072265e-06,
+      "logits/chosen": 0.04520021751523018,
+      "logits/rejected": 0.06952523440122604,
+      "logps/chosen": -82.35380554199219,
+      "logps/rejected": -89.27215576171875,
+      "loss": 2.4222,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5209689140319824,
+      "rewards/margins": 0.5571924448013306,
+      "rewards/rejected": -3.0781612396240234,
+      "step": 799
+    },
+    {
+      "epoch": 1.452316076294278,
+      "grad_norm": 2.342952013015747,
+      "learning_rate": 7.36936755850849e-06,
+      "logits/chosen": 0.016199974343180656,
+      "logits/rejected": 0.07721687853336334,
+      "logps/chosen": -83.52679443359375,
+      "logps/rejected": -95.04873657226562,
+      "loss": 1.8844,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.487191677093506,
+      "rewards/margins": 0.9364355206489563,
+      "rewards/rejected": -3.4236273765563965,
+      "step": 800
+    },
+    {
+      "epoch": 1.4541326067211626,
+      "grad_norm": 2.611804246902466,
+      "learning_rate": 7.363520372978774e-06,
+      "logits/chosen": 0.0884896069765091,
+      "logits/rejected": 0.12684395909309387,
+      "logps/chosen": -73.71459197998047,
+      "logps/rejected": -78.7441635131836,
+      "loss": 2.5897,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.3413586616516113,
+      "rewards/margins": 0.3767651617527008,
+      "rewards/rejected": -2.7181236743927,
+      "step": 801
+    },
+    {
+      "epoch": 1.4559491371480473,
+      "grad_norm": 2.2446863651275635,
+      "learning_rate": 7.357669387406548e-06,
+      "logits/chosen": 0.04801754653453827,
+      "logits/rejected": 0.14114932715892792,
+      "logps/chosen": -72.67373657226562,
+      "logps/rejected": -88.53012084960938,
+      "loss": 1.8134,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4397833347320557,
+      "rewards/margins": 0.9794785976409912,
+      "rewards/rejected": -3.419261932373047,
+      "step": 802
+    },
+    {
+      "epoch": 1.457765667574932,
+      "grad_norm": 2.2070155143737793,
+      "learning_rate": 7.351814613722991e-06,
+      "logits/chosen": 0.09963800758123398,
+      "logits/rejected": 0.09044703096151352,
+      "logps/chosen": -93.37407684326172,
+      "logps/rejected": -101.1263656616211,
+      "loss": 2.0345,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4976069927215576,
+      "rewards/margins": 0.7321794033050537,
+      "rewards/rejected": -3.2297863960266113,
+      "step": 803
+    },
+    {
+      "epoch": 1.4595821980018164,
+      "grad_norm": 2.339240789413452,
+      "learning_rate": 7.3459560638670035e-06,
+      "logits/chosen": 0.06331755220890045,
+      "logits/rejected": 0.09320802986621857,
+      "logps/chosen": -84.83074951171875,
+      "logps/rejected": -90.84583282470703,
+      "loss": 1.7825,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.2360713481903076,
+      "rewards/margins": 0.8683581352233887,
+      "rewards/rejected": -3.104429244995117,
+      "step": 804
+    },
+    {
+      "epoch": 1.461398728428701,
+      "grad_norm": 2.2252860069274902,
+      "learning_rate": 7.3400937497851906e-06,
+      "logits/chosen": 0.06259419769048691,
+      "logits/rejected": 0.05668449029326439,
+      "logps/chosen": -76.97444915771484,
+      "logps/rejected": -95.10333251953125,
+      "loss": 1.9808,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.475806713104248,
+      "rewards/margins": 0.9982489943504333,
+      "rewards/rejected": -3.4740560054779053,
+      "step": 805
+    },
+    {
+      "epoch": 1.4632152588555858,
+      "grad_norm": 2.056358814239502,
+      "learning_rate": 7.334227683431832e-06,
+      "logits/chosen": 0.0795949399471283,
+      "logits/rejected": 0.11871220171451569,
+      "logps/chosen": -82.24653625488281,
+      "logps/rejected": -104.25836944580078,
+      "loss": 1.8753,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.397820234298706,
+      "rewards/margins": 1.0821467638015747,
+      "rewards/rejected": -3.479966878890991,
+      "step": 806
+    },
+    {
+      "epoch": 1.4650317892824705,
+      "grad_norm": 2.7462871074676514,
+      "learning_rate": 7.328357876768858e-06,
+      "logits/chosen": 0.08270560950040817,
+      "logits/rejected": 0.11420280486345291,
+      "logps/chosen": -76.36333465576172,
+      "logps/rejected": -86.0474853515625,
+      "loss": 2.3324,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.573794364929199,
+      "rewards/margins": 0.7034318447113037,
+      "rewards/rejected": -3.277225971221924,
+      "step": 807
+    },
+    {
+      "epoch": 1.4668483197093551,
+      "grad_norm": 2.5223159790039062,
+      "learning_rate": 7.322484341765824e-06,
+      "logits/chosen": 0.17336739599704742,
+      "logits/rejected": 0.15756890177726746,
+      "logps/chosen": -82.3572998046875,
+      "logps/rejected": -94.081787109375,
+      "loss": 2.0898,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3327016830444336,
+      "rewards/margins": 0.8286820650100708,
+      "rewards/rejected": -3.161383867263794,
+      "step": 808
+    },
+    {
+      "epoch": 1.4686648501362398,
+      "grad_norm": 2.4014031887054443,
+      "learning_rate": 7.316607090399894e-06,
+      "logits/chosen": 0.0774412751197815,
+      "logits/rejected": 0.016062507405877113,
+      "logps/chosen": -79.8453598022461,
+      "logps/rejected": -80.36003875732422,
+      "loss": 2.0743,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6673779487609863,
+      "rewards/margins": 0.6888259649276733,
+      "rewards/rejected": -3.356204032897949,
+      "step": 809
+    },
+    {
+      "epoch": 1.4704813805631245,
+      "grad_norm": 2.3294990062713623,
+      "learning_rate": 7.310726134655807e-06,
+      "logits/chosen": 0.03737390413880348,
+      "logits/rejected": 0.1001349613070488,
+      "logps/chosen": -81.04705810546875,
+      "logps/rejected": -96.8980712890625,
+      "loss": 1.7474,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.27017879486084,
+      "rewards/margins": 1.0432392358779907,
+      "rewards/rejected": -3.313417673110962,
+      "step": 810
+    },
+    {
+      "epoch": 1.472297910990009,
+      "grad_norm": 2.287733793258667,
+      "learning_rate": 7.30484148652585e-06,
+      "logits/chosen": 0.0469367690384388,
+      "logits/rejected": 0.06735274195671082,
+      "logps/chosen": -80.7060546875,
+      "logps/rejected": -93.6329345703125,
+      "loss": 2.0516,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.463186740875244,
+      "rewards/margins": 0.7543611526489258,
+      "rewards/rejected": -3.21754789352417,
+      "step": 811
+    },
+    {
+      "epoch": 1.4741144414168939,
+      "grad_norm": 3.1462066173553467,
+      "learning_rate": 7.298953158009852e-06,
+      "logits/chosen": 0.09930360317230225,
+      "logits/rejected": 0.13193194568157196,
+      "logps/chosen": -82.47021484375,
+      "logps/rejected": -94.59288024902344,
+      "loss": 1.9513,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.323047161102295,
+      "rewards/margins": 0.8602778911590576,
+      "rewards/rejected": -3.1833250522613525,
+      "step": 812
+    },
+    {
+      "epoch": 1.4759309718437783,
+      "grad_norm": 2.2134501934051514,
+      "learning_rate": 7.293061161115138e-06,
+      "logits/chosen": 0.04882989823818207,
+      "logits/rejected": 0.07393565028905869,
+      "logps/chosen": -69.4496841430664,
+      "logps/rejected": -79.9421157836914,
+      "loss": 1.9357,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.233206272125244,
+      "rewards/margins": 0.7889159321784973,
+      "rewards/rejected": -3.0221221446990967,
+      "step": 813
+    },
+    {
+      "epoch": 1.477747502270663,
+      "grad_norm": 2.3149659633636475,
+      "learning_rate": 7.287165507856513e-06,
+      "logits/chosen": 0.05155353993177414,
+      "logits/rejected": 0.036280907690525055,
+      "logps/chosen": -87.41694641113281,
+      "logps/rejected": -97.0438461303711,
+      "loss": 1.9108,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.27425217628479,
+      "rewards/margins": 0.8441011309623718,
+      "rewards/rejected": -3.1183536052703857,
+      "step": 814
+    },
+    {
+      "epoch": 1.4795640326975477,
+      "grad_norm": 2.4460105895996094,
+      "learning_rate": 7.281266210256244e-06,
+      "logits/chosen": 0.06415215134620667,
+      "logits/rejected": 0.08172871172428131,
+      "logps/chosen": -94.13727569580078,
+      "logps/rejected": -108.59651184082031,
+      "loss": 1.9764,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.476318120956421,
+      "rewards/margins": 0.8187392354011536,
+      "rewards/rejected": -3.2950570583343506,
+      "step": 815
+    },
+    {
+      "epoch": 1.4813805631244323,
+      "grad_norm": 2.5813546180725098,
+      "learning_rate": 7.275363280344023e-06,
+      "logits/chosen": 0.10900135338306427,
+      "logits/rejected": 0.06012497842311859,
+      "logps/chosen": -86.26004028320312,
+      "logps/rejected": -92.2495346069336,
+      "loss": 2.1791,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4611434936523438,
+      "rewards/margins": 0.6075427532196045,
+      "rewards/rejected": -3.0686862468719482,
+      "step": 816
+    },
+    {
+      "epoch": 1.483197093551317,
+      "grad_norm": 2.3455967903137207,
+      "learning_rate": 7.269456730156954e-06,
+      "logits/chosen": 0.08659522235393524,
+      "logits/rejected": 0.1016351729631424,
+      "logps/chosen": -75.75826263427734,
+      "logps/rejected": -85.91856384277344,
+      "loss": 2.0077,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4886691570281982,
+      "rewards/margins": 0.9004117846488953,
+      "rewards/rejected": -3.3890810012817383,
+      "step": 817
+    },
+    {
+      "epoch": 1.4850136239782017,
+      "grad_norm": 2.612858533859253,
+      "learning_rate": 7.26354657173952e-06,
+      "logits/chosen": 0.08592739701271057,
+      "logits/rejected": 0.15154355764389038,
+      "logps/chosen": -86.36177062988281,
+      "logps/rejected": -95.97001647949219,
+      "loss": 2.1468,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6705503463745117,
+      "rewards/margins": 0.7056388854980469,
+      "rewards/rejected": -3.376189708709717,
+      "step": 818
+    },
+    {
+      "epoch": 1.4868301544050864,
+      "grad_norm": 2.39699387550354,
+      "learning_rate": 7.257632817143566e-06,
+      "logits/chosen": 0.03663618862628937,
+      "logits/rejected": 0.06003079563379288,
+      "logps/chosen": -73.82808685302734,
+      "logps/rejected": -86.11414337158203,
+      "loss": 2.0372,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.251981258392334,
+      "rewards/margins": 0.8644815683364868,
+      "rewards/rejected": -3.1164627075195312,
+      "step": 819
+    },
+    {
+      "epoch": 1.4886466848319708,
+      "grad_norm": 2.1887619495391846,
+      "learning_rate": 7.251715478428263e-06,
+      "logits/chosen": -0.026794860139489174,
+      "logits/rejected": 0.11405564099550247,
+      "logps/chosen": -96.14276123046875,
+      "logps/rejected": -116.06658172607422,
+      "loss": 1.786,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.527817964553833,
+      "rewards/margins": 1.0386499166488647,
+      "rewards/rejected": -3.5664680004119873,
+      "step": 820
+    },
+    {
+      "epoch": 1.4904632152588557,
+      "grad_norm": 4.235897064208984,
+      "learning_rate": 7.2457945676600975e-06,
+      "logits/chosen": 0.10090241581201553,
+      "logits/rejected": 0.10542615503072739,
+      "logps/chosen": -78.00686645507812,
+      "logps/rejected": -86.51849365234375,
+      "loss": 2.1653,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6344261169433594,
+      "rewards/margins": 0.6512963771820068,
+      "rewards/rejected": -3.285722494125366,
+      "step": 821
+    },
+    {
+      "epoch": 1.4922797456857402,
+      "grad_norm": 2.189833402633667,
+      "learning_rate": 7.239870096912835e-06,
+      "logits/chosen": 0.0326794758439064,
+      "logits/rejected": 0.0896943062543869,
+      "logps/chosen": -89.783447265625,
+      "logps/rejected": -102.53228759765625,
+      "loss": 2.0422,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5607566833496094,
+      "rewards/margins": 1.009847640991211,
+      "rewards/rejected": -3.5706043243408203,
+      "step": 822
+    },
+    {
+      "epoch": 1.4940962761126249,
+      "grad_norm": 2.7656984329223633,
+      "learning_rate": 7.233942078267507e-06,
+      "logits/chosen": 0.03892774134874344,
+      "logits/rejected": 0.10934557020664215,
+      "logps/chosen": -87.70173645019531,
+      "logps/rejected": -98.54985046386719,
+      "loss": 1.9344,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.796427011489868,
+      "rewards/margins": 0.7776911854743958,
+      "rewards/rejected": -3.574118137359619,
+      "step": 823
+    },
+    {
+      "epoch": 1.4959128065395095,
+      "grad_norm": 2.1388044357299805,
+      "learning_rate": 7.228010523812372e-06,
+      "logits/chosen": 0.04929535835981369,
+      "logits/rejected": 0.05947191268205643,
+      "logps/chosen": -100.03392791748047,
+      "logps/rejected": -106.81358337402344,
+      "loss": 2.2769,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.775858163833618,
+      "rewards/margins": 0.8662734031677246,
+      "rewards/rejected": -3.6421315670013428,
+      "step": 824
+    },
+    {
+      "epoch": 1.4977293369663942,
+      "grad_norm": 2.8936967849731445,
+      "learning_rate": 7.222075445642904e-06,
+      "logits/chosen": 0.05339030176401138,
+      "logits/rejected": 0.09673048555850983,
+      "logps/chosen": -91.19949340820312,
+      "logps/rejected": -98.978515625,
+      "loss": 2.0686,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.4693431854248047,
+      "rewards/margins": 0.6974996328353882,
+      "rewards/rejected": -3.1668429374694824,
+      "step": 825
+    },
+    {
+      "epoch": 1.4995458673932789,
+      "grad_norm": 2.069953203201294,
+      "learning_rate": 7.216136855861763e-06,
+      "logits/chosen": 0.07213248312473297,
+      "logits/rejected": 0.14326205849647522,
+      "logps/chosen": -82.97685241699219,
+      "logps/rejected": -95.03224182128906,
+      "loss": 1.4687,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": -2.4275951385498047,
+      "rewards/margins": 1.1413557529449463,
+      "rewards/rejected": -3.568950653076172,
+      "step": 826
+    },
+    {
+      "epoch": 1.5013623978201633,
+      "grad_norm": 2.1628522872924805,
+      "learning_rate": 7.2101947665787605e-06,
+      "logits/chosen": 0.03247791528701782,
+      "logits/rejected": 0.13113634288311005,
+      "logps/chosen": -83.87806701660156,
+      "logps/rejected": -96.8788833618164,
+      "loss": 1.9276,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.425245761871338,
+      "rewards/margins": 0.938535213470459,
+      "rewards/rejected": -3.363780975341797,
+      "step": 827
+    },
+    {
+      "epoch": 1.5031789282470482,
+      "grad_norm": 2.1415798664093018,
+      "learning_rate": 7.20424918991086e-06,
+      "logits/chosen": 0.03731315955519676,
+      "logits/rejected": 0.038553059101104736,
+      "logps/chosen": -83.6593017578125,
+      "logps/rejected": -93.7667007446289,
+      "loss": 1.8234,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.623645305633545,
+      "rewards/margins": 1.0124547481536865,
+      "rewards/rejected": -3.6361002922058105,
+      "step": 828
+    },
+    {
+      "epoch": 1.5049954586739327,
+      "grad_norm": 2.155017375946045,
+      "learning_rate": 7.198300137982121e-06,
+      "logits/chosen": 0.05089723318815231,
+      "logits/rejected": 0.0881565734744072,
+      "logps/chosen": -87.8398666381836,
+      "logps/rejected": -99.52027130126953,
+      "loss": 1.8965,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.6992077827453613,
+      "rewards/margins": 0.9189518690109253,
+      "rewards/rejected": -3.618159770965576,
+      "step": 829
+    },
+    {
+      "epoch": 1.5068119891008176,
+      "grad_norm": 2.2586467266082764,
+      "learning_rate": 7.1923476229237e-06,
+      "logits/chosen": 0.032588109374046326,
+      "logits/rejected": 0.041385456919670105,
+      "logps/chosen": -88.36325073242188,
+      "logps/rejected": -97.6312026977539,
+      "loss": 1.8165,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.743462562561035,
+      "rewards/margins": 0.9551791548728943,
+      "rewards/rejected": -3.698641300201416,
+      "step": 830
+    },
+    {
+      "epoch": 1.508628519527702,
+      "grad_norm": 2.1857669353485107,
+      "learning_rate": 7.186391656873809e-06,
+      "logits/chosen": 0.09781689196825027,
+      "logits/rejected": 0.15005499124526978,
+      "logps/chosen": -85.13274383544922,
+      "logps/rejected": -98.7428207397461,
+      "loss": 1.8323,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.4565935134887695,
+      "rewards/margins": 0.9222261309623718,
+      "rewards/rejected": -3.378819704055786,
+      "step": 831
+    },
+    {
+      "epoch": 1.5104450499545867,
+      "grad_norm": 2.0807337760925293,
+      "learning_rate": 7.180432251977702e-06,
+      "logits/chosen": 0.0358676053583622,
+      "logits/rejected": 0.07221710681915283,
+      "logps/chosen": -82.19658660888672,
+      "logps/rejected": -93.883056640625,
+      "loss": 2.0109,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4384050369262695,
+      "rewards/margins": 0.8013818860054016,
+      "rewards/rejected": -3.2397866249084473,
+      "step": 832
+    },
+    {
+      "epoch": 1.5122615803814714,
+      "grad_norm": 2.5140273571014404,
+      "learning_rate": 7.1744694203876405e-06,
+      "logits/chosen": 0.057662852108478546,
+      "logits/rejected": 0.07171612232923508,
+      "logps/chosen": -83.77603912353516,
+      "logps/rejected": -94.08065795898438,
+      "loss": 1.8424,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.402648687362671,
+      "rewards/margins": 0.952576756477356,
+      "rewards/rejected": -3.3552253246307373,
+      "step": 833
+    },
+    {
+      "epoch": 1.514078110808356,
+      "grad_norm": 2.1008050441741943,
+      "learning_rate": 7.168503174262881e-06,
+      "logits/chosen": 0.0810992568731308,
+      "logits/rejected": 0.13944530487060547,
+      "logps/chosen": -77.3492660522461,
+      "logps/rejected": -90.80574035644531,
+      "loss": 1.8164,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4869396686553955,
+      "rewards/margins": 1.0017035007476807,
+      "rewards/rejected": -3.488643169403076,
+      "step": 834
+    },
+    {
+      "epoch": 1.5158946412352408,
+      "grad_norm": 2.228290319442749,
+      "learning_rate": 7.162533525769633e-06,
+      "logits/chosen": -0.05040167272090912,
+      "logits/rejected": 0.09420110285282135,
+      "logps/chosen": -84.03011322021484,
+      "logps/rejected": -93.4085464477539,
+      "loss": 2.0479,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6713035106658936,
+      "rewards/margins": 0.8382018804550171,
+      "rewards/rejected": -3.509505033493042,
+      "step": 835
+    },
+    {
+      "epoch": 1.5177111716621252,
+      "grad_norm": 2.6196677684783936,
+      "learning_rate": 7.156560487081052e-06,
+      "logits/chosen": 0.06445930898189545,
+      "logits/rejected": 0.09075027704238892,
+      "logps/chosen": -81.3261489868164,
+      "logps/rejected": -97.19158935546875,
+      "loss": 1.83,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7067627906799316,
+      "rewards/margins": 1.0383878946304321,
+      "rewards/rejected": -3.745150566101074,
+      "step": 836
+    },
+    {
+      "epoch": 1.5195277020890101,
+      "grad_norm": 2.3788716793060303,
+      "learning_rate": 7.150584070377202e-06,
+      "logits/chosen": 0.023997552692890167,
+      "logits/rejected": 0.084604412317276,
+      "logps/chosen": -86.31644439697266,
+      "logps/rejected": -103.74675750732422,
+      "loss": 1.9901,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.7836720943450928,
+      "rewards/margins": 0.9425742626190186,
+      "rewards/rejected": -3.7262461185455322,
+      "step": 837
+    },
+    {
+      "epoch": 1.5213442325158946,
+      "grad_norm": 2.6834628582000732,
+      "learning_rate": 7.144604287845041e-06,
+      "logits/chosen": 0.015507448464632034,
+      "logits/rejected": 0.07179054617881775,
+      "logps/chosen": -79.04088592529297,
+      "logps/rejected": -95.68939208984375,
+      "loss": 1.7565,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.706390142440796,
+      "rewards/margins": 1.0823572874069214,
+      "rewards/rejected": -3.7887473106384277,
+      "step": 838
+    },
+    {
+      "epoch": 1.5231607629427792,
+      "grad_norm": 2.56843638420105,
+      "learning_rate": 7.1386211516783835e-06,
+      "logits/chosen": -0.017521001398563385,
+      "logits/rejected": -0.061632271856069565,
+      "logps/chosen": -80.90260314941406,
+      "logps/rejected": -93.43807220458984,
+      "loss": 1.8708,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.495241165161133,
+      "rewards/margins": 1.119477391242981,
+      "rewards/rejected": -3.614718437194824,
+      "step": 839
+    },
+    {
+      "epoch": 1.524977293369664,
+      "grad_norm": 2.1130528450012207,
+      "learning_rate": 7.132634674077884e-06,
+      "logits/chosen": 0.0070226918905973434,
+      "logits/rejected": 0.12307396531105042,
+      "logps/chosen": -80.71980285644531,
+      "logps/rejected": -101.35388946533203,
+      "loss": 1.8225,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7102115154266357,
+      "rewards/margins": 1.0462895631790161,
+      "rewards/rejected": -3.7565011978149414,
+      "step": 840
+    },
+    {
+      "epoch": 1.5267938237965486,
+      "grad_norm": 2.270901918411255,
+      "learning_rate": 7.12664486725102e-06,
+      "logits/chosen": 0.09991536289453506,
+      "logits/rejected": 0.11764054000377655,
+      "logps/chosen": -84.9771499633789,
+      "logps/rejected": -100.33434295654297,
+      "loss": 2.2036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7242026329040527,
+      "rewards/margins": 0.846297562122345,
+      "rewards/rejected": -3.570500373840332,
+      "step": 841
+    },
+    {
+      "epoch": 1.5286103542234333,
+      "grad_norm": 2.743037462234497,
+      "learning_rate": 7.120651743412043e-06,
+      "logits/chosen": 0.025931421667337418,
+      "logits/rejected": -0.005241448059678078,
+      "logps/chosen": -85.05397033691406,
+      "logps/rejected": -92.75206756591797,
+      "loss": 2.0745,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6730573177337646,
+      "rewards/margins": 1.065488338470459,
+      "rewards/rejected": -3.7385458946228027,
+      "step": 842
+    },
+    {
+      "epoch": 1.5304268846503177,
+      "grad_norm": 2.7490670680999756,
+      "learning_rate": 7.114655314781979e-06,
+      "logits/chosen": 0.07127739489078522,
+      "logits/rejected": 0.04277154803276062,
+      "logps/chosen": -83.83795166015625,
+      "logps/rejected": -84.39205169677734,
+      "loss": 2.9054,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7217230796813965,
+      "rewards/margins": 0.3551695644855499,
+      "rewards/rejected": -3.076892852783203,
+      "step": 843
+    },
+    {
+      "epoch": 1.5322434150772026,
+      "grad_norm": 2.204655170440674,
+      "learning_rate": 7.10865559358859e-06,
+      "logits/chosen": 0.034612249583005905,
+      "logits/rejected": 0.07071523368358612,
+      "logps/chosen": -77.57713317871094,
+      "logps/rejected": -88.81067657470703,
+      "loss": 2.0337,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.794020175933838,
+      "rewards/margins": 0.7705952525138855,
+      "rewards/rejected": -3.564615488052368,
+      "step": 844
+    },
+    {
+      "epoch": 1.534059945504087,
+      "grad_norm": 2.373863935470581,
+      "learning_rate": 7.102652592066354e-06,
+      "logits/chosen": 0.1263967752456665,
+      "logits/rejected": 0.10894232988357544,
+      "logps/chosen": -78.27210998535156,
+      "logps/rejected": -84.76968383789062,
+      "loss": 2.0159,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7315664291381836,
+      "rewards/margins": 0.8217055201530457,
+      "rewards/rejected": -3.553272247314453,
+      "step": 845
+    },
+    {
+      "epoch": 1.535876475930972,
+      "grad_norm": 2.7803144454956055,
+      "learning_rate": 7.096646322456433e-06,
+      "logits/chosen": 0.12285368889570236,
+      "logits/rejected": 0.13744983077049255,
+      "logps/chosen": -82.08031463623047,
+      "logps/rejected": -88.31858825683594,
+      "loss": 2.2342,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5975961685180664,
+      "rewards/margins": 0.7215418219566345,
+      "rewards/rejected": -3.3191380500793457,
+      "step": 846
+    },
+    {
+      "epoch": 1.5376930063578564,
+      "grad_norm": 2.3556156158447266,
+      "learning_rate": 7.090636797006658e-06,
+      "logits/chosen": 0.10716073215007782,
+      "logits/rejected": 0.146171435713768,
+      "logps/chosen": -90.718994140625,
+      "logps/rejected": -105.76301574707031,
+      "loss": 1.9024,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.929590940475464,
+      "rewards/margins": 0.8752865195274353,
+      "rewards/rejected": -3.804877281188965,
+      "step": 847
+    },
+    {
+      "epoch": 1.5395095367847411,
+      "grad_norm": 2.5544545650482178,
+      "learning_rate": 7.0846240279714984e-06,
+      "logits/chosen": 0.0810776948928833,
+      "logits/rejected": 0.0847952738404274,
+      "logps/chosen": -85.46900939941406,
+      "logps/rejected": -91.54446411132812,
+      "loss": 1.8777,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6484267711639404,
+      "rewards/margins": 0.8561803102493286,
+      "rewards/rejected": -3.5046072006225586,
+      "step": 848
+    },
+    {
+      "epoch": 1.5413260672116258,
+      "grad_norm": 2.667482852935791,
+      "learning_rate": 7.078608027612035e-06,
+      "logits/chosen": 0.05214187875390053,
+      "logits/rejected": 0.009745856747031212,
+      "logps/chosen": -83.0794677734375,
+      "logps/rejected": -96.57014465332031,
+      "loss": 2.0238,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.473628282546997,
+      "rewards/margins": 1.0539703369140625,
+      "rewards/rejected": -3.5275986194610596,
+      "step": 849
+    },
+    {
+      "epoch": 1.5431425976385105,
+      "grad_norm": 2.492583751678467,
+      "learning_rate": 7.072588808195944e-06,
+      "logits/chosen": 0.10039053112268448,
+      "logits/rejected": 0.05160602927207947,
+      "logps/chosen": -83.8985366821289,
+      "logps/rejected": -87.81637573242188,
+      "loss": 2.2203,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.873837471008301,
+      "rewards/margins": 0.702816367149353,
+      "rewards/rejected": -3.5766537189483643,
+      "step": 850
+    },
+    {
+      "epoch": 1.5449591280653951,
+      "grad_norm": 2.280651807785034,
+      "learning_rate": 7.066566381997457e-06,
+      "logits/chosen": 0.08965125679969788,
+      "logits/rejected": 0.08822537213563919,
+      "logps/chosen": -90.04351043701172,
+      "logps/rejected": -101.63990783691406,
+      "loss": 1.7834,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.407850742340088,
+      "rewards/margins": 1.1601994037628174,
+      "rewards/rejected": -3.5680501461029053,
+      "step": 851
+    },
+    {
+      "epoch": 1.5467756584922796,
+      "grad_norm": 1.9988317489624023,
+      "learning_rate": 7.060540761297353e-06,
+      "logits/chosen": 0.0349108949303627,
+      "logits/rejected": 0.12820343673229218,
+      "logps/chosen": -80.97975158691406,
+      "logps/rejected": -99.21882629394531,
+      "loss": 1.7877,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5994269847869873,
+      "rewards/margins": 0.9676470756530762,
+      "rewards/rejected": -3.5670742988586426,
+      "step": 852
+    },
+    {
+      "epoch": 1.5485921889191645,
+      "grad_norm": 2.590162992477417,
+      "learning_rate": 7.05451195838292e-06,
+      "logits/chosen": 0.10045134276151657,
+      "logits/rejected": 0.14448747038841248,
+      "logps/chosen": -77.86140441894531,
+      "logps/rejected": -89.10052490234375,
+      "loss": 2.1282,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7747833728790283,
+      "rewards/margins": 0.7405270934104919,
+      "rewards/rejected": -3.515310764312744,
+      "step": 853
+    },
+    {
+      "epoch": 1.550408719346049,
+      "grad_norm": 2.6667673587799072,
+      "learning_rate": 7.048479985547941e-06,
+      "logits/chosen": 0.02231951802968979,
+      "logits/rejected": 0.12828901410102844,
+      "logps/chosen": -78.2810287475586,
+      "logps/rejected": -97.22885131835938,
+      "loss": 1.8773,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4572958946228027,
+      "rewards/margins": 0.9423343539237976,
+      "rewards/rejected": -3.399630069732666,
+      "step": 854
+    },
+    {
+      "epoch": 1.5522252497729339,
+      "grad_norm": 1.7987407445907593,
+      "learning_rate": 7.042444855092653e-06,
+      "logits/chosen": 0.061210960149765015,
+      "logits/rejected": 0.08236300945281982,
+      "logps/chosen": -79.21247100830078,
+      "logps/rejected": -94.42364501953125,
+      "loss": 2.144,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.486496686935425,
+      "rewards/margins": 0.8338861465454102,
+      "rewards/rejected": -3.320382833480835,
+      "step": 855
+    },
+    {
+      "epoch": 1.5540417801998183,
+      "grad_norm": 2.5392510890960693,
+      "learning_rate": 7.0364065793237406e-06,
+      "logits/chosen": 0.034715939313173294,
+      "logits/rejected": -0.004529901314526796,
+      "logps/chosen": -87.7265853881836,
+      "logps/rejected": -99.10587310791016,
+      "loss": 1.8283,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.661682605743408,
+      "rewards/margins": 0.9480312466621399,
+      "rewards/rejected": -3.609713554382324,
+      "step": 856
+    },
+    {
+      "epoch": 1.555858310626703,
+      "grad_norm": 2.442492961883545,
+      "learning_rate": 7.0303651705542985e-06,
+      "logits/chosen": 0.0268821120262146,
+      "logits/rejected": 0.09976281225681305,
+      "logps/chosen": -91.73252868652344,
+      "logps/rejected": -100.85424041748047,
+      "loss": 2.1762,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6233739852905273,
+      "rewards/margins": 0.893970787525177,
+      "rewards/rejected": -3.5173447132110596,
+      "step": 857
+    },
+    {
+      "epoch": 1.5576748410535877,
+      "grad_norm": 2.857959270477295,
+      "learning_rate": 7.024320641103812e-06,
+      "logits/chosen": 0.15326820313930511,
+      "logits/rejected": 0.1491074562072754,
+      "logps/chosen": -82.46217346191406,
+      "logps/rejected": -87.15203857421875,
+      "loss": 2.4176,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4776577949523926,
+      "rewards/margins": 0.4634871482849121,
+      "rewards/rejected": -2.9411449432373047,
+      "step": 858
+    },
+    {
+      "epoch": 1.5594913714804723,
+      "grad_norm": 2.5892186164855957,
+      "learning_rate": 7.018273003298129e-06,
+      "logits/chosen": 0.04699310287833214,
+      "logits/rejected": 0.033118702471256256,
+      "logps/chosen": -79.92018127441406,
+      "logps/rejected": -89.115966796875,
+      "loss": 2.0407,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5493407249450684,
+      "rewards/margins": 0.8943923115730286,
+      "rewards/rejected": -3.443732738494873,
+      "step": 859
+    },
+    {
+      "epoch": 1.561307901907357,
+      "grad_norm": 2.2549710273742676,
+      "learning_rate": 7.0122222694694356e-06,
+      "logits/chosen": 0.12253312766551971,
+      "logits/rejected": 0.1437494158744812,
+      "logps/chosen": -75.97108459472656,
+      "logps/rejected": -90.68475341796875,
+      "loss": 1.7125,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.593079090118408,
+      "rewards/margins": 1.1156821250915527,
+      "rewards/rejected": -3.70876145362854,
+      "step": 860
+    },
+    {
+      "epoch": 1.5631244323342415,
+      "grad_norm": 2.22299861907959,
+      "learning_rate": 7.006168451956233e-06,
+      "logits/chosen": 0.11537674069404602,
+      "logits/rejected": 0.045418016612529755,
+      "logps/chosen": -86.62315368652344,
+      "logps/rejected": -87.4183120727539,
+      "loss": 2.0229,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.457634687423706,
+      "rewards/margins": 0.7262483835220337,
+      "rewards/rejected": -3.183882713317871,
+      "step": 861
+    },
+    {
+      "epoch": 1.5649409627611264,
+      "grad_norm": 2.734768867492676,
+      "learning_rate": 7.0001115631033065e-06,
+      "logits/chosen": 0.04321238026022911,
+      "logits/rejected": 0.04483964294195175,
+      "logps/chosen": -82.62739562988281,
+      "logps/rejected": -94.7513656616211,
+      "loss": 1.9925,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.520638942718506,
+      "rewards/margins": 0.8393934965133667,
+      "rewards/rejected": -3.360032320022583,
+      "step": 862
+    },
+    {
+      "epoch": 1.5667574931880108,
+      "grad_norm": 2.2604191303253174,
+      "learning_rate": 6.994051615261711e-06,
+      "logits/chosen": 0.0408606193959713,
+      "logits/rejected": 0.047267988324165344,
+      "logps/chosen": -80.66863250732422,
+      "logps/rejected": -95.75201416015625,
+      "loss": 1.8486,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.533386707305908,
+      "rewards/margins": 1.029523491859436,
+      "rewards/rejected": -3.5629100799560547,
+      "step": 863
+    },
+    {
+      "epoch": 1.5685740236148955,
+      "grad_norm": 2.274548292160034,
+      "learning_rate": 6.987988620788731e-06,
+      "logits/chosen": 0.010744954459369183,
+      "logits/rejected": 0.052073340862989426,
+      "logps/chosen": -82.55726623535156,
+      "logps/rejected": -93.57171630859375,
+      "loss": 1.9414,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.571608543395996,
+      "rewards/margins": 0.828125,
+      "rewards/rejected": -3.399733781814575,
+      "step": 864
+    },
+    {
+      "epoch": 1.5703905540417802,
+      "grad_norm": 2.2155966758728027,
+      "learning_rate": 6.98192259204787e-06,
+      "logits/chosen": 0.082489974796772,
+      "logits/rejected": 0.11105959117412567,
+      "logps/chosen": -72.24876403808594,
+      "logps/rejected": -88.81566619873047,
+      "loss": 1.7377,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.400843381881714,
+      "rewards/margins": 1.03752601146698,
+      "rewards/rejected": -3.4383695125579834,
+      "step": 865
+    },
+    {
+      "epoch": 1.5722070844686649,
+      "grad_norm": 2.4514036178588867,
+      "learning_rate": 6.9758535414088166e-06,
+      "logits/chosen": 0.029958020895719528,
+      "logits/rejected": 0.0316297747194767,
+      "logps/chosen": -88.95033264160156,
+      "logps/rejected": -92.07289123535156,
+      "loss": 2.3553,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.583791732788086,
+      "rewards/margins": 0.5526307225227356,
+      "rewards/rejected": -3.1364223957061768,
+      "step": 866
+    },
+    {
+      "epoch": 1.5740236148955495,
+      "grad_norm": 2.701751470565796,
+      "learning_rate": 6.969781481247423e-06,
+      "logits/chosen": 0.08873719722032547,
+      "logits/rejected": 0.05975634232163429,
+      "logps/chosen": -83.68199157714844,
+      "logps/rejected": -91.66279602050781,
+      "loss": 2.0972,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5379390716552734,
+      "rewards/margins": 0.720723569393158,
+      "rewards/rejected": -3.258662462234497,
+      "step": 867
+    },
+    {
+      "epoch": 1.575840145322434,
+      "grad_norm": 2.887376308441162,
+      "learning_rate": 6.963706423945674e-06,
+      "logits/chosen": 0.05533679947257042,
+      "logits/rejected": 0.02152944728732109,
+      "logps/chosen": -81.51123046875,
+      "logps/rejected": -95.27003479003906,
+      "loss": 2.0153,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.5463404655456543,
+      "rewards/margins": 0.968324601650238,
+      "rewards/rejected": -3.514665126800537,
+      "step": 868
+    },
+    {
+      "epoch": 1.577656675749319,
+      "grad_norm": 2.442810535430908,
+      "learning_rate": 6.957628381891674e-06,
+      "logits/chosen": -0.04294189438223839,
+      "logits/rejected": 0.03631982207298279,
+      "logps/chosen": -77.20703125,
+      "logps/rejected": -93.48060607910156,
+      "loss": 2.1137,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.508554220199585,
+      "rewards/margins": 0.8056586384773254,
+      "rewards/rejected": -3.3142127990722656,
+      "step": 869
+    },
+    {
+      "epoch": 1.5794732061762033,
+      "grad_norm": 2.456723928451538,
+      "learning_rate": 6.951547367479604e-06,
+      "logits/chosen": 0.025873543694615364,
+      "logits/rejected": -0.04093042016029358,
+      "logps/chosen": -86.80729675292969,
+      "logps/rejected": -96.05107116699219,
+      "loss": 2.0419,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3767457008361816,
+      "rewards/margins": 0.8237012028694153,
+      "rewards/rejected": -3.200446844100952,
+      "step": 870
+    },
+    {
+      "epoch": 1.5812897366030882,
+      "grad_norm": 2.332174062728882,
+      "learning_rate": 6.945463393109712e-06,
+      "logits/chosen": 0.006868576630949974,
+      "logits/rejected": 0.09155347943305969,
+      "logps/chosen": -85.2925796508789,
+      "logps/rejected": -104.44573211669922,
+      "loss": 2.0281,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5827579498291016,
+      "rewards/margins": 0.9421724081039429,
+      "rewards/rejected": -3.524930477142334,
+      "step": 871
+    },
+    {
+      "epoch": 1.5831062670299727,
+      "grad_norm": 3.058819055557251,
+      "learning_rate": 6.939376471188285e-06,
+      "logits/chosen": 0.07147755473852158,
+      "logits/rejected": 0.03247075155377388,
+      "logps/chosen": -85.51266479492188,
+      "logps/rejected": -96.28234100341797,
+      "loss": 2.2254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.8757948875427246,
+      "rewards/margins": 0.7729579210281372,
+      "rewards/rejected": -3.6487529277801514,
+      "step": 872
+    },
+    {
+      "epoch": 1.5849227974568574,
+      "grad_norm": 2.271951198577881,
+      "learning_rate": 6.933286614127613e-06,
+      "logits/chosen": 0.04744710773229599,
+      "logits/rejected": 0.06620746105909348,
+      "logps/chosen": -75.55758666992188,
+      "logps/rejected": -77.83596801757812,
+      "loss": 2.0876,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.806255340576172,
+      "rewards/margins": 0.6999154686927795,
+      "rewards/rejected": -3.5061709880828857,
+      "step": 873
+    },
+    {
+      "epoch": 1.586739327883742,
+      "grad_norm": 2.9459800720214844,
+      "learning_rate": 6.927193834345976e-06,
+      "logits/chosen": 0.025282222777605057,
+      "logits/rejected": 0.08691692352294922,
+      "logps/chosen": -88.87342834472656,
+      "logps/rejected": -101.53143310546875,
+      "loss": 2.0501,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.600255012512207,
+      "rewards/margins": 0.8346633911132812,
+      "rewards/rejected": -3.4349184036254883,
+      "step": 874
+    },
+    {
+      "epoch": 1.5885558583106267,
+      "grad_norm": 2.279984474182129,
+      "learning_rate": 6.921098144267613e-06,
+      "logits/chosen": 0.027605077251791954,
+      "logits/rejected": 0.09924539178609848,
+      "logps/chosen": -85.24947357177734,
+      "logps/rejected": -97.9180679321289,
+      "loss": 1.9226,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.842658281326294,
+      "rewards/margins": 0.8581206202507019,
+      "rewards/rejected": -3.7007789611816406,
+      "step": 875
+    },
+    {
+      "epoch": 1.5903723887375114,
+      "grad_norm": 3.2044014930725098,
+      "learning_rate": 6.914999556322697e-06,
+      "logits/chosen": 0.06477774679660797,
+      "logits/rejected": 0.03072739765048027,
+      "logps/chosen": -96.84561157226562,
+      "logps/rejected": -98.88067626953125,
+      "loss": 2.3308,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.960343360900879,
+      "rewards/margins": 0.545425295829773,
+      "rewards/rejected": -3.5057687759399414,
+      "step": 876
+    },
+    {
+      "epoch": 1.5921889191643959,
+      "grad_norm": 2.679797410964966,
+      "learning_rate": 6.90889808294731e-06,
+      "logits/chosen": 0.04393957555294037,
+      "logits/rejected": 0.02098749577999115,
+      "logps/chosen": -88.12010192871094,
+      "logps/rejected": -96.62691497802734,
+      "loss": 2.297,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7419230937957764,
+      "rewards/margins": 0.6746903657913208,
+      "rewards/rejected": -3.4166133403778076,
+      "step": 877
+    },
+    {
+      "epoch": 1.5940054495912808,
+      "grad_norm": 2.7355191707611084,
+      "learning_rate": 6.9027937365834206e-06,
+      "logits/chosen": 0.07665219157934189,
+      "logits/rejected": 0.10063200443983078,
+      "logps/chosen": -86.96102905273438,
+      "logps/rejected": -101.6490478515625,
+      "loss": 1.9587,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6776695251464844,
+      "rewards/margins": 0.838853120803833,
+      "rewards/rejected": -3.5165226459503174,
+      "step": 878
+    },
+    {
+      "epoch": 1.5958219800181652,
+      "grad_norm": 3.1400301456451416,
+      "learning_rate": 6.896686529678853e-06,
+      "logits/chosen": 0.10836174339056015,
+      "logits/rejected": 0.14626868069171906,
+      "logps/chosen": -85.75300598144531,
+      "logps/rejected": -101.42611694335938,
+      "loss": 2.0544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7651474475860596,
+      "rewards/margins": 0.8746659159660339,
+      "rewards/rejected": -3.63981294631958,
+      "step": 879
+    },
+    {
+      "epoch": 1.5976385104450501,
+      "grad_norm": 2.9733288288116455,
+      "learning_rate": 6.890576474687264e-06,
+      "logits/chosen": 0.05398109182715416,
+      "logits/rejected": 0.011274321004748344,
+      "logps/chosen": -91.39241790771484,
+      "logps/rejected": -98.18811798095703,
+      "loss": 2.1884,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.658442497253418,
+      "rewards/margins": 0.6335260272026062,
+      "rewards/rejected": -3.291968584060669,
+      "step": 880
+    },
+    {
+      "epoch": 1.5994550408719346,
+      "grad_norm": 2.4757211208343506,
+      "learning_rate": 6.884463584068121e-06,
+      "logits/chosen": 0.10966772586107254,
+      "logits/rejected": 0.0948777049779892,
+      "logps/chosen": -84.34974670410156,
+      "logps/rejected": -92.52738952636719,
+      "loss": 1.8273,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.648416519165039,
+      "rewards/margins": 0.8553692102432251,
+      "rewards/rejected": -3.5037853717803955,
+      "step": 881
+    },
+    {
+      "epoch": 1.6012715712988193,
+      "grad_norm": 2.846766710281372,
+      "learning_rate": 6.878347870286673e-06,
+      "logits/chosen": 0.03845957666635513,
+      "logits/rejected": 0.027499085292220116,
+      "logps/chosen": -77.8682861328125,
+      "logps/rejected": -86.57948303222656,
+      "loss": 1.9555,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.8318474292755127,
+      "rewards/margins": 0.8005434274673462,
+      "rewards/rejected": -3.6323909759521484,
+      "step": 882
+    },
+    {
+      "epoch": 1.603088101725704,
+      "grad_norm": 2.80184268951416,
+      "learning_rate": 6.872229345813924e-06,
+      "logits/chosen": 0.08280351012945175,
+      "logits/rejected": 0.11747594177722931,
+      "logps/chosen": -80.64393615722656,
+      "logps/rejected": -94.31782531738281,
+      "loss": 1.7596,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.782545804977417,
+      "rewards/margins": 0.9048705101013184,
+      "rewards/rejected": -3.6874163150787354,
+      "step": 883
+    },
+    {
+      "epoch": 1.6049046321525886,
+      "grad_norm": 2.607837200164795,
+      "learning_rate": 6.8661080231266104e-06,
+      "logits/chosen": 0.11254586279392242,
+      "logits/rejected": 0.06956940144300461,
+      "logps/chosen": -84.59523010253906,
+      "logps/rejected": -91.33222198486328,
+      "loss": 2.176,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.859100818634033,
+      "rewards/margins": 0.6652751564979553,
+      "rewards/rejected": -3.524376153945923,
+      "step": 884
+    },
+    {
+      "epoch": 1.6067211625794733,
+      "grad_norm": 3.3895885944366455,
+      "learning_rate": 6.859983914707177e-06,
+      "logits/chosen": 0.11012329161167145,
+      "logits/rejected": 0.1130778044462204,
+      "logps/chosen": -92.46974182128906,
+      "logps/rejected": -103.50257110595703,
+      "loss": 2.0558,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -3.044379234313965,
+      "rewards/margins": 0.958195686340332,
+      "rewards/rejected": -4.002574920654297,
+      "step": 885
+    },
+    {
+      "epoch": 1.6085376930063577,
+      "grad_norm": 2.8371028900146484,
+      "learning_rate": 6.853857033043745e-06,
+      "logits/chosen": 0.05022600293159485,
+      "logits/rejected": 0.0587327815592289,
+      "logps/chosen": -87.489013671875,
+      "logps/rejected": -105.74163055419922,
+      "loss": 2.1291,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5108165740966797,
+      "rewards/margins": 0.822872519493103,
+      "rewards/rejected": -3.333688735961914,
+      "step": 886
+    },
+    {
+      "epoch": 1.6103542234332426,
+      "grad_norm": 2.347508668899536,
+      "learning_rate": 6.847727390630093e-06,
+      "logits/chosen": 0.1272752583026886,
+      "logits/rejected": 0.12413996458053589,
+      "logps/chosen": -87.31372833251953,
+      "logps/rejected": -103.10004425048828,
+      "loss": 1.9655,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.733205795288086,
+      "rewards/margins": 1.1384268999099731,
+      "rewards/rejected": -3.8716328144073486,
+      "step": 887
+    },
+    {
+      "epoch": 1.612170753860127,
+      "grad_norm": 3.185014247894287,
+      "learning_rate": 6.84159499996563e-06,
+      "logits/chosen": 0.11554019898176193,
+      "logits/rejected": 0.1505330353975296,
+      "logps/chosen": -83.4138412475586,
+      "logps/rejected": -97.09869384765625,
+      "loss": 1.9864,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.906578540802002,
+      "rewards/margins": 1.0419787168502808,
+      "rewards/rejected": -3.948557138442993,
+      "step": 888
+    },
+    {
+      "epoch": 1.6139872842870118,
+      "grad_norm": 2.601205348968506,
+      "learning_rate": 6.83545987355537e-06,
+      "logits/chosen": 0.10080607980489731,
+      "logits/rejected": 0.06827899813652039,
+      "logps/chosen": -81.38069915771484,
+      "logps/rejected": -92.65945434570312,
+      "loss": 1.7634,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6402342319488525,
+      "rewards/margins": 1.1383386850357056,
+      "rewards/rejected": -3.7785730361938477,
+      "step": 889
+    },
+    {
+      "epoch": 1.6158038147138964,
+      "grad_norm": 2.91386079788208,
+      "learning_rate": 6.829322023909901e-06,
+      "logits/chosen": 0.06657513976097107,
+      "logits/rejected": 0.05152777582406998,
+      "logps/chosen": -90.99147033691406,
+      "logps/rejected": -93.03759765625,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.659510374069214,
+      "rewards/margins": 0.5215647220611572,
+      "rewards/rejected": -3.181075096130371,
+      "step": 890
+    },
+    {
+      "epoch": 1.6176203451407811,
+      "grad_norm": 2.6477596759796143,
+      "learning_rate": 6.823181463545368e-06,
+      "logits/chosen": 0.01778372749686241,
+      "logits/rejected": 0.06530646234750748,
+      "logps/chosen": -85.58574676513672,
+      "logps/rejected": -101.24583435058594,
+      "loss": 1.9453,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.543752670288086,
+      "rewards/margins": 0.894145667552948,
+      "rewards/rejected": -3.4378981590270996,
+      "step": 891
+    },
+    {
+      "epoch": 1.6194368755676658,
+      "grad_norm": 2.449430465698242,
+      "learning_rate": 6.817038204983444e-06,
+      "logits/chosen": 0.10009394586086273,
+      "logits/rejected": 0.11922462284564972,
+      "logps/chosen": -75.22452545166016,
+      "logps/rejected": -89.5363998413086,
+      "loss": 1.9703,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5327696800231934,
+      "rewards/margins": 1.006116271018982,
+      "rewards/rejected": -3.538886070251465,
+      "step": 892
+    },
+    {
+      "epoch": 1.6212534059945503,
+      "grad_norm": 2.757439374923706,
+      "learning_rate": 6.810892260751302e-06,
+      "logits/chosen": 0.054223865270614624,
+      "logits/rejected": 0.11705522984266281,
+      "logps/chosen": -84.47832489013672,
+      "logps/rejected": -107.8006591796875,
+      "loss": 1.9296,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.757310152053833,
+      "rewards/margins": 1.0707166194915771,
+      "rewards/rejected": -3.828026294708252,
+      "step": 893
+    },
+    {
+      "epoch": 1.6230699364214352,
+      "grad_norm": 2.5267086029052734,
+      "learning_rate": 6.804743643381592e-06,
+      "logits/chosen": 0.07622219622135162,
+      "logits/rejected": 0.05764157325029373,
+      "logps/chosen": -81.3788070678711,
+      "logps/rejected": -87.3583984375,
+      "loss": 1.857,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6934287548065186,
+      "rewards/margins": 0.9804441928863525,
+      "rewards/rejected": -3.673872947692871,
+      "step": 894
+    },
+    {
+      "epoch": 1.6248864668483196,
+      "grad_norm": 2.3378968238830566,
+      "learning_rate": 6.798592365412415e-06,
+      "logits/chosen": 0.08304903656244278,
+      "logits/rejected": 0.10661397874355316,
+      "logps/chosen": -84.71305847167969,
+      "logps/rejected": -100.10174560546875,
+      "loss": 1.8795,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6393752098083496,
+      "rewards/margins": 0.9716106057167053,
+      "rewards/rejected": -3.610985517501831,
+      "step": 895
+    },
+    {
+      "epoch": 1.6267029972752045,
+      "grad_norm": 3.1945290565490723,
+      "learning_rate": 6.792438439387299e-06,
+      "logits/chosen": 0.09845674782991409,
+      "logits/rejected": 0.037844423204660416,
+      "logps/chosen": -87.18885803222656,
+      "logps/rejected": -91.7021484375,
+      "loss": 2.4462,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.7558887004852295,
+      "rewards/margins": 0.6585051417350769,
+      "rewards/rejected": -3.414393663406372,
+      "step": 896
+    },
+    {
+      "epoch": 1.628519527702089,
+      "grad_norm": 2.678049087524414,
+      "learning_rate": 6.7862818778551695e-06,
+      "logits/chosen": -0.059538260102272034,
+      "logits/rejected": 0.028212856501340866,
+      "logps/chosen": -91.52423858642578,
+      "logps/rejected": -110.80790710449219,
+      "loss": 2.0741,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6971120834350586,
+      "rewards/margins": 0.9168665409088135,
+      "rewards/rejected": -3.6139791011810303,
+      "step": 897
+    },
+    {
+      "epoch": 1.6303360581289736,
+      "grad_norm": 2.703550338745117,
+      "learning_rate": 6.780122693370329e-06,
+      "logits/chosen": -0.007212355732917786,
+      "logits/rejected": 0.08302198350429535,
+      "logps/chosen": -78.15343475341797,
+      "logps/rejected": -88.86624908447266,
+      "loss": 2.1919,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.428589344024658,
+      "rewards/margins": 0.6562613844871521,
+      "rewards/rejected": -3.084850549697876,
+      "step": 898
+    },
+    {
+      "epoch": 1.6321525885558583,
+      "grad_norm": 2.5438737869262695,
+      "learning_rate": 6.773960898492428e-06,
+      "logits/chosen": 0.10631975531578064,
+      "logits/rejected": 0.1630961000919342,
+      "logps/chosen": -79.30477142333984,
+      "logps/rejected": -94.29690551757812,
+      "loss": 1.8887,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7491891384124756,
+      "rewards/margins": 0.8967292308807373,
+      "rewards/rejected": -3.645918369293213,
+      "step": 899
+    },
+    {
+      "epoch": 1.633969118982743,
+      "grad_norm": 2.6740238666534424,
+      "learning_rate": 6.767796505786435e-06,
+      "logits/chosen": 0.04583762586116791,
+      "logits/rejected": 0.05214417725801468,
+      "logps/chosen": -90.60303497314453,
+      "logps/rejected": -99.45308685302734,
+      "loss": 2.2367,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.576535224914551,
+      "rewards/margins": 0.6341556310653687,
+      "rewards/rejected": -3.210691213607788,
+      "step": 900
+    },
+    {
+      "epoch": 1.6357856494096277,
+      "grad_norm": 2.8195323944091797,
+      "learning_rate": 6.761629527822622e-06,
+      "logits/chosen": 0.11805088073015213,
+      "logits/rejected": 0.03617120161652565,
+      "logps/chosen": -79.40000915527344,
+      "logps/rejected": -79.39904022216797,
+      "loss": 2.0995,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.64365816116333,
+      "rewards/margins": 0.7394881248474121,
+      "rewards/rejected": -3.383146047592163,
+      "step": 901
+    },
+    {
+      "epoch": 1.6376021798365121,
+      "grad_norm": 2.666821241378784,
+      "learning_rate": 6.755459977176533e-06,
+      "logits/chosen": 0.017997367307543755,
+      "logits/rejected": 0.09890137612819672,
+      "logps/chosen": -89.78903198242188,
+      "logps/rejected": -111.10968017578125,
+      "loss": 1.8919,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.503241539001465,
+      "rewards/margins": 1.0098021030426025,
+      "rewards/rejected": -3.5130434036254883,
+      "step": 902
+    },
+    {
+      "epoch": 1.639418710263397,
+      "grad_norm": 2.3226935863494873,
+      "learning_rate": 6.749287866428953e-06,
+      "logits/chosen": 0.09748942404985428,
+      "logits/rejected": 0.09682448208332062,
+      "logps/chosen": -82.52100372314453,
+      "logps/rejected": -91.55313873291016,
+      "loss": 1.7682,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5952441692352295,
+      "rewards/margins": 1.0089478492736816,
+      "rewards/rejected": -3.604191780090332,
+      "step": 903
+    },
+    {
+      "epoch": 1.6412352406902815,
+      "grad_norm": 2.479614019393921,
+      "learning_rate": 6.743113208165893e-06,
+      "logits/chosen": 0.029578503221273422,
+      "logits/rejected": 0.07657214999198914,
+      "logps/chosen": -100.56556701660156,
+      "logps/rejected": -102.63167572021484,
+      "loss": 2.5709,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.717003345489502,
+      "rewards/margins": 0.48937955498695374,
+      "rewards/rejected": -3.206382989883423,
+      "step": 904
+    },
+    {
+      "epoch": 1.6430517711171662,
+      "grad_norm": 3.3280487060546875,
+      "learning_rate": 6.736936014978552e-06,
+      "logits/chosen": 0.04310007020831108,
+      "logits/rejected": 0.05533391237258911,
+      "logps/chosen": -102.93861389160156,
+      "logps/rejected": -110.60729217529297,
+      "loss": 1.9292,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3910446166992188,
+      "rewards/margins": 0.9039056897163391,
+      "rewards/rejected": -3.294950485229492,
+      "step": 905
+    },
+    {
+      "epoch": 1.6448683015440508,
+      "grad_norm": 2.6267921924591064,
+      "learning_rate": 6.730756299463308e-06,
+      "logits/chosen": 0.06594766676425934,
+      "logits/rejected": 0.11423090845346451,
+      "logps/chosen": -79.65184020996094,
+      "logps/rejected": -90.49372863769531,
+      "loss": 2.2204,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.69381046295166,
+      "rewards/margins": 0.8501594066619873,
+      "rewards/rejected": -3.5439698696136475,
+      "step": 906
+    },
+    {
+      "epoch": 1.6466848319709355,
+      "grad_norm": 2.4378793239593506,
+      "learning_rate": 6.724574074221673e-06,
+      "logits/chosen": 0.043015994131565094,
+      "logits/rejected": 0.14672963321208954,
+      "logps/chosen": -78.83383178710938,
+      "logps/rejected": -93.22097778320312,
+      "loss": 1.9456,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.480109691619873,
+      "rewards/margins": 0.8949618339538574,
+      "rewards/rejected": -3.3750715255737305,
+      "step": 907
+    },
+    {
+      "epoch": 1.6485013623978202,
+      "grad_norm": 2.400392770767212,
+      "learning_rate": 6.718389351860285e-06,
+      "logits/chosen": 0.08273115754127502,
+      "logits/rejected": 0.06503652036190033,
+      "logps/chosen": -87.24859619140625,
+      "logps/rejected": -95.38743591308594,
+      "loss": 1.7985,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.356613874435425,
+      "rewards/margins": 0.8771872520446777,
+      "rewards/rejected": -3.2338013648986816,
+      "step": 908
+    },
+    {
+      "epoch": 1.6503178928247049,
+      "grad_norm": 2.42429780960083,
+      "learning_rate": 6.712202144990863e-06,
+      "logits/chosen": 0.10674826800823212,
+      "logits/rejected": 0.08516664803028107,
+      "logps/chosen": -89.37960815429688,
+      "logps/rejected": -99.21661376953125,
+      "loss": 2.1353,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6561362743377686,
+      "rewards/margins": 0.7742940187454224,
+      "rewards/rejected": -3.4304304122924805,
+      "step": 909
+    },
+    {
+      "epoch": 1.6521344232515895,
+      "grad_norm": 2.6496243476867676,
+      "learning_rate": 6.7060124662302066e-06,
+      "logits/chosen": 0.14033779501914978,
+      "logits/rejected": 0.07252150774002075,
+      "logps/chosen": -80.96002197265625,
+      "logps/rejected": -82.9183120727539,
+      "loss": 2.0659,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.427546739578247,
+      "rewards/margins": 0.8414397835731506,
+      "rewards/rejected": -3.268986463546753,
+      "step": 910
+    },
+    {
+      "epoch": 1.653950953678474,
+      "grad_norm": 3.0638275146484375,
+      "learning_rate": 6.699820328200143e-06,
+      "logits/chosen": 0.058655206114053726,
+      "logits/rejected": 0.08714447915554047,
+      "logps/chosen": -88.06983947753906,
+      "logps/rejected": -95.36041259765625,
+      "loss": 2.0403,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.2383832931518555,
+      "rewards/margins": 0.9142959117889404,
+      "rewards/rejected": -3.152679443359375,
+      "step": 911
+    },
+    {
+      "epoch": 1.655767484105359,
+      "grad_norm": 2.2842650413513184,
+      "learning_rate": 6.69362574352752e-06,
+      "logits/chosen": -0.011715320870280266,
+      "logits/rejected": 0.03974226117134094,
+      "logps/chosen": -84.77818298339844,
+      "logps/rejected": -94.44766235351562,
+      "loss": 2.091,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4050097465515137,
+      "rewards/margins": 0.7936041951179504,
+      "rewards/rejected": -3.1986136436462402,
+      "step": 912
+    },
+    {
+      "epoch": 1.6575840145322434,
+      "grad_norm": 3.1531357765197754,
+      "learning_rate": 6.6874287248441794e-06,
+      "logits/chosen": 0.08675999939441681,
+      "logits/rejected": 0.06470594555139542,
+      "logps/chosen": -80.276123046875,
+      "logps/rejected": -90.92185974121094,
+      "loss": 2.0593,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.28584361076355,
+      "rewards/margins": 0.7852586507797241,
+      "rewards/rejected": -3.0711023807525635,
+      "step": 913
+    },
+    {
+      "epoch": 1.659400544959128,
+      "grad_norm": 2.7664525508880615,
+      "learning_rate": 6.681229284786915e-06,
+      "logits/chosen": 0.05306999385356903,
+      "logits/rejected": 0.051360100507736206,
+      "logps/chosen": -83.74992370605469,
+      "logps/rejected": -90.31805419921875,
+      "loss": 2.2115,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.623281955718994,
+      "rewards/margins": 0.6698784232139587,
+      "rewards/rejected": -3.2931606769561768,
+      "step": 914
+    },
+    {
+      "epoch": 1.6612170753860127,
+      "grad_norm": 2.890864610671997,
+      "learning_rate": 6.675027435997469e-06,
+      "logits/chosen": 0.04681715369224548,
+      "logits/rejected": 0.1523803025484085,
+      "logps/chosen": -74.13650512695312,
+      "logps/rejected": -93.63848876953125,
+      "loss": 1.988,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.540015697479248,
+      "rewards/margins": 0.9194749593734741,
+      "rewards/rejected": -3.4594907760620117,
+      "step": 915
+    },
+    {
+      "epoch": 1.6630336058128974,
+      "grad_norm": 2.2408931255340576,
+      "learning_rate": 6.668823191122487e-06,
+      "logits/chosen": 0.08750308305025101,
+      "logits/rejected": 0.06304562091827393,
+      "logps/chosen": -87.24713897705078,
+      "logps/rejected": -96.94601440429688,
+      "loss": 1.9141,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.300774097442627,
+      "rewards/margins": 0.8336561322212219,
+      "rewards/rejected": -3.134430170059204,
+      "step": 916
+    },
+    {
+      "epoch": 1.664850136239782,
+      "grad_norm": 2.584038019180298,
+      "learning_rate": 6.662616562813512e-06,
+      "logits/chosen": 0.0785432755947113,
+      "logits/rejected": 0.05843789130449295,
+      "logps/chosen": -88.5465087890625,
+      "logps/rejected": -96.19537353515625,
+      "loss": 1.7811,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.419790267944336,
+      "rewards/margins": 0.844237744808197,
+      "rewards/rejected": -3.2640278339385986,
+      "step": 917
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 2.4870214462280273,
+      "learning_rate": 6.6564075637269295e-06,
+      "logits/chosen": 0.06065249443054199,
+      "logits/rejected": 0.0761883556842804,
+      "logps/chosen": -76.00924682617188,
+      "logps/rejected": -85.09017944335938,
+      "loss": 1.9802,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3578128814697266,
+      "rewards/margins": 0.8296225070953369,
+      "rewards/rejected": -3.1874351501464844,
+      "step": 918
+    },
+    {
+      "epoch": 1.6684831970935514,
+      "grad_norm": 2.0682199001312256,
+      "learning_rate": 6.650196206523977e-06,
+      "logits/chosen": 0.0461968369781971,
+      "logits/rejected": 0.08261042088270187,
+      "logps/chosen": -72.26417541503906,
+      "logps/rejected": -77.92495727539062,
+      "loss": 2.0028,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1353201866149902,
+      "rewards/margins": 0.6552493572235107,
+      "rewards/rejected": -2.790569305419922,
+      "step": 919
+    },
+    {
+      "epoch": 1.6702997275204359,
+      "grad_norm": 2.482517719268799,
+      "learning_rate": 6.643982503870693e-06,
+      "logits/chosen": 0.12246346473693848,
+      "logits/rejected": 0.03192145377397537,
+      "logps/chosen": -89.27822875976562,
+      "logps/rejected": -89.58540344238281,
+      "loss": 2.0059,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4135119915008545,
+      "rewards/margins": 0.6876509785652161,
+      "rewards/rejected": -3.101162910461426,
+      "step": 920
+    },
+    {
+      "epoch": 1.6721162579473208,
+      "grad_norm": 2.84334659576416,
+      "learning_rate": 6.637766468437894e-06,
+      "logits/chosen": 0.027422966435551643,
+      "logits/rejected": 0.019883258268237114,
+      "logps/chosen": -77.09085845947266,
+      "logps/rejected": -88.27078247070312,
+      "loss": 1.9826,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.446176052093506,
+      "rewards/margins": 0.8014341592788696,
+      "rewards/rejected": -3.247610092163086,
+      "step": 921
+    },
+    {
+      "epoch": 1.6739327883742052,
+      "grad_norm": 2.386929512023926,
+      "learning_rate": 6.631548112901164e-06,
+      "logits/chosen": 0.025946754962205887,
+      "logits/rejected": 0.06377127766609192,
+      "logps/chosen": -82.12971496582031,
+      "logps/rejected": -97.94608306884766,
+      "loss": 1.9747,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.573225736618042,
+      "rewards/margins": 0.7901925444602966,
+      "rewards/rejected": -3.3634183406829834,
+      "step": 922
+    },
+    {
+      "epoch": 1.67574931880109,
+      "grad_norm": 2.3462514877319336,
+      "learning_rate": 6.6253274499408095e-06,
+      "logits/chosen": -0.009377829730510712,
+      "logits/rejected": 0.0917067900300026,
+      "logps/chosen": -83.40922546386719,
+      "logps/rejected": -103.09521484375,
+      "loss": 1.7486,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.368823766708374,
+      "rewards/margins": 0.9489405751228333,
+      "rewards/rejected": -3.3177640438079834,
+      "step": 923
+    },
+    {
+      "epoch": 1.6775658492279746,
+      "grad_norm": 2.3609023094177246,
+      "learning_rate": 6.619104492241848e-06,
+      "logits/chosen": 0.013379395008087158,
+      "logits/rejected": 0.1470850110054016,
+      "logps/chosen": -75.76339721679688,
+      "logps/rejected": -95.58364868164062,
+      "loss": 1.7618,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3512778282165527,
+      "rewards/margins": 1.085105061531067,
+      "rewards/rejected": -3.4363832473754883,
+      "step": 924
+    },
+    {
+      "epoch": 1.6793823796548593,
+      "grad_norm": 2.4780282974243164,
+      "learning_rate": 6.61287925249397e-06,
+      "logits/chosen": 0.03466716408729553,
+      "logits/rejected": 0.06602051854133606,
+      "logps/chosen": -85.4227294921875,
+      "logps/rejected": -94.31620025634766,
+      "loss": 2.1665,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6465556621551514,
+      "rewards/margins": 0.6559208035469055,
+      "rewards/rejected": -3.302476644515991,
+      "step": 925
+    },
+    {
+      "epoch": 1.681198910081744,
+      "grad_norm": 4.518651485443115,
+      "learning_rate": 6.606651743391528e-06,
+      "logits/chosen": 0.00949503481388092,
+      "logits/rejected": 0.06442085653543472,
+      "logps/chosen": -88.59455108642578,
+      "logps/rejected": -105.29449462890625,
+      "loss": 2.0174,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2852275371551514,
+      "rewards/margins": 0.9270148873329163,
+      "rewards/rejected": -3.2122421264648438,
+      "step": 926
+    },
+    {
+      "epoch": 1.6830154405086284,
+      "grad_norm": 2.9527838230133057,
+      "learning_rate": 6.6004219776334925e-06,
+      "logits/chosen": 0.06499281525611877,
+      "logits/rejected": 0.011767326854169369,
+      "logps/chosen": -87.16426849365234,
+      "logps/rejected": -91.7620620727539,
+      "loss": 2.0636,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3328614234924316,
+      "rewards/margins": 0.7728868126869202,
+      "rewards/rejected": -3.105748176574707,
+      "step": 927
+    },
+    {
+      "epoch": 1.6848319709355133,
+      "grad_norm": 2.7613911628723145,
+      "learning_rate": 6.594189967923445e-06,
+      "logits/chosen": 0.10220185667276382,
+      "logits/rejected": 0.11646111309528351,
+      "logps/chosen": -88.42827606201172,
+      "logps/rejected": -100.41016387939453,
+      "loss": 1.8203,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.3946800231933594,
+      "rewards/margins": 0.9652897119522095,
+      "rewards/rejected": -3.3599696159362793,
+      "step": 928
+    },
+    {
+      "epoch": 1.6866485013623977,
+      "grad_norm": 2.674398899078369,
+      "learning_rate": 6.587955726969533e-06,
+      "logits/chosen": 0.07171659916639328,
+      "logits/rejected": 0.04297471046447754,
+      "logps/chosen": -90.65692138671875,
+      "logps/rejected": -93.96014404296875,
+      "loss": 2.5874,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.538684606552124,
+      "rewards/margins": 0.4615382254123688,
+      "rewards/rejected": -3.00022292137146,
+      "step": 929
+    },
+    {
+      "epoch": 1.6884650317892824,
+      "grad_norm": 2.5353660583496094,
+      "learning_rate": 6.581719267484466e-06,
+      "logits/chosen": 0.13961178064346313,
+      "logits/rejected": 0.13023585081100464,
+      "logps/chosen": -77.078369140625,
+      "logps/rejected": -86.81641387939453,
+      "loss": 1.7334,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3547840118408203,
+      "rewards/margins": 0.9528897404670715,
+      "rewards/rejected": -3.307673692703247,
+      "step": 930
+    },
+    {
+      "epoch": 1.690281562216167,
+      "grad_norm": 2.8513107299804688,
+      "learning_rate": 6.575480602185468e-06,
+      "logits/chosen": 0.12496423721313477,
+      "logits/rejected": 0.14027045667171478,
+      "logps/chosen": -84.40544891357422,
+      "logps/rejected": -98.02970123291016,
+      "loss": 2.2827,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.501647472381592,
+      "rewards/margins": 0.8144313097000122,
+      "rewards/rejected": -3.3160786628723145,
+      "step": 931
+    },
+    {
+      "epoch": 1.6920980926430518,
+      "grad_norm": 2.7415926456451416,
+      "learning_rate": 6.56923974379426e-06,
+      "logits/chosen": 0.032566964626312256,
+      "logits/rejected": 0.09472635388374329,
+      "logps/chosen": -91.36210632324219,
+      "logps/rejected": -107.93646240234375,
+      "loss": 2.1175,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199713945388794,
+      "rewards/margins": 0.8857415914535522,
+      "rewards/rejected": -3.0854556560516357,
+      "step": 932
+    },
+    {
+      "epoch": 1.6939146230699365,
+      "grad_norm": 2.986032009124756,
+      "learning_rate": 6.562996705037046e-06,
+      "logits/chosen": 0.028138715773820877,
+      "logits/rejected": 0.08858595788478851,
+      "logps/chosen": -78.46945190429688,
+      "logps/rejected": -92.57295989990234,
+      "loss": 2.384,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4276278018951416,
+      "rewards/margins": 0.7274990677833557,
+      "rewards/rejected": -3.1551268100738525,
+      "step": 933
+    },
+    {
+      "epoch": 1.695731153496821,
+      "grad_norm": 2.294461488723755,
+      "learning_rate": 6.556751498644461e-06,
+      "logits/chosen": 0.1189623773097992,
+      "logits/rejected": 0.11807019263505936,
+      "logps/chosen": -72.875244140625,
+      "logps/rejected": -77.55158996582031,
+      "loss": 2.0037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3735506534576416,
+      "rewards/margins": 0.6975762844085693,
+      "rewards/rejected": -3.07112717628479,
+      "step": 934
+    },
+    {
+      "epoch": 1.6975476839237058,
+      "grad_norm": 2.706350803375244,
+      "learning_rate": 6.550504137351576e-06,
+      "logits/chosen": 0.03257827088236809,
+      "logits/rejected": 0.06685741245746613,
+      "logps/chosen": -93.078369140625,
+      "logps/rejected": -100.50369262695312,
+      "loss": 2.3693,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5124664306640625,
+      "rewards/margins": 0.7416088581085205,
+      "rewards/rejected": -3.254075050354004,
+      "step": 935
+    },
+    {
+      "epoch": 1.6993642143505903,
+      "grad_norm": 2.6172664165496826,
+      "learning_rate": 6.544254633897843e-06,
+      "logits/chosen": 0.10361647605895996,
+      "logits/rejected": 0.1024157926440239,
+      "logps/chosen": -83.01170349121094,
+      "logps/rejected": -94.26844787597656,
+      "loss": 2.02,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4206433296203613,
+      "rewards/margins": 0.8451172113418579,
+      "rewards/rejected": -3.265760660171509,
+      "step": 936
+    },
+    {
+      "epoch": 1.7011807447774752,
+      "grad_norm": 2.4271178245544434,
+      "learning_rate": 6.538003001027089e-06,
+      "logits/chosen": 0.0540219321846962,
+      "logits/rejected": 0.1337059736251831,
+      "logps/chosen": -82.78097534179688,
+      "logps/rejected": -102.0522689819336,
+      "loss": 1.8375,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.637333631515503,
+      "rewards/margins": 1.0766263008117676,
+      "rewards/rejected": -3.7139596939086914,
+      "step": 937
+    },
+    {
+      "epoch": 1.7029972752043596,
+      "grad_norm": 3.2782678604125977,
+      "learning_rate": 6.531749251487482e-06,
+      "logits/chosen": 0.09090890735387802,
+      "logits/rejected": 0.08031464368104935,
+      "logps/chosen": -77.13628387451172,
+      "logps/rejected": -79.11683654785156,
+      "loss": 2.6753,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3901565074920654,
+      "rewards/margins": 0.39642852544784546,
+      "rewards/rejected": -2.7865848541259766,
+      "step": 938
+    },
+    {
+      "epoch": 1.7048138056312443,
+      "grad_norm": 3.1739110946655273,
+      "learning_rate": 6.525493398031508e-06,
+      "logits/chosen": 0.08761344850063324,
+      "logits/rejected": 0.1552104502916336,
+      "logps/chosen": -83.78360748291016,
+      "logps/rejected": -97.63893127441406,
+      "loss": 2.2164,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.603111743927002,
+      "rewards/margins": 0.6446429491043091,
+      "rewards/rejected": -3.2477540969848633,
+      "step": 939
+    },
+    {
+      "epoch": 1.706630336058129,
+      "grad_norm": 3.4325766563415527,
+      "learning_rate": 6.5192354534159406e-06,
+      "logits/chosen": 0.018010685220360756,
+      "logits/rejected": 0.07210525125265121,
+      "logps/chosen": -96.70736694335938,
+      "logps/rejected": -112.33457946777344,
+      "loss": 1.7358,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5734646320343018,
+      "rewards/margins": 1.2066644430160522,
+      "rewards/rejected": -3.7801291942596436,
+      "step": 940
+    },
+    {
+      "epoch": 1.7084468664850136,
+      "grad_norm": 2.4167909622192383,
+      "learning_rate": 6.5129754304018165e-06,
+      "logits/chosen": 0.04140740633010864,
+      "logits/rejected": 0.08365748077630997,
+      "logps/chosen": -76.82650756835938,
+      "logps/rejected": -84.57887268066406,
+      "loss": 2.3819,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.239987373352051,
+      "rewards/margins": 0.5318999290466309,
+      "rewards/rejected": -2.7718873023986816,
+      "step": 941
+    },
+    {
+      "epoch": 1.7102633969118983,
+      "grad_norm": 2.5350871086120605,
+      "learning_rate": 6.506713341754415e-06,
+      "logits/chosen": 0.021460117772221565,
+      "logits/rejected": 0.10231097042560577,
+      "logps/chosen": -79.95002746582031,
+      "logps/rejected": -96.06900024414062,
+      "loss": 1.7958,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.4041337966918945,
+      "rewards/margins": 1.1107407808303833,
+      "rewards/rejected": -3.5148746967315674,
+      "step": 942
+    },
+    {
+      "epoch": 1.7120799273387828,
+      "grad_norm": 2.602478504180908,
+      "learning_rate": 6.500449200243224e-06,
+      "logits/chosen": 0.08819068968296051,
+      "logits/rejected": 0.10114292055368423,
+      "logps/chosen": -72.37012481689453,
+      "logps/rejected": -82.5709228515625,
+      "loss": 2.1943,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.402527332305908,
+      "rewards/margins": 0.6120179295539856,
+      "rewards/rejected": -3.014545202255249,
+      "step": 943
+    },
+    {
+      "epoch": 1.7138964577656677,
+      "grad_norm": 2.405487537384033,
+      "learning_rate": 6.494183018641919e-06,
+      "logits/chosen": 0.020005354657769203,
+      "logits/rejected": 0.07407501339912415,
+      "logps/chosen": -73.03540802001953,
+      "logps/rejected": -84.96491241455078,
+      "loss": 1.9642,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6638689041137695,
+      "rewards/margins": 0.8980711102485657,
+      "rewards/rejected": -3.5619401931762695,
+      "step": 944
+    },
+    {
+      "epoch": 1.7157129881925521,
+      "grad_norm": 2.8340280055999756,
+      "learning_rate": 6.487914809728335e-06,
+      "logits/chosen": 0.09565074741840363,
+      "logits/rejected": 0.13740883767604828,
+      "logps/chosen": -90.7655258178711,
+      "logps/rejected": -105.01213073730469,
+      "loss": 2.102,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.418067216873169,
+      "rewards/margins": 0.7691794633865356,
+      "rewards/rejected": -3.187246799468994,
+      "step": 945
+    },
+    {
+      "epoch": 1.717529518619437,
+      "grad_norm": 2.5804646015167236,
+      "learning_rate": 6.481644586284443e-06,
+      "logits/chosen": 0.06889508664608002,
+      "logits/rejected": 0.0715349018573761,
+      "logps/chosen": -90.50798034667969,
+      "logps/rejected": -96.70076751708984,
+      "loss": 1.8887,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6765780448913574,
+      "rewards/margins": 0.9487190246582031,
+      "rewards/rejected": -3.6252973079681396,
+      "step": 946
+    },
+    {
+      "epoch": 1.7193460490463215,
+      "grad_norm": 2.947023868560791,
+      "learning_rate": 6.475372361096316e-06,
+      "logits/chosen": 0.057566914707422256,
+      "logits/rejected": 0.12709926068782806,
+      "logps/chosen": -77.96834564208984,
+      "logps/rejected": -92.29097747802734,
+      "loss": 2.1631,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5991628170013428,
+      "rewards/margins": 0.8114217519760132,
+      "rewards/rejected": -3.4105844497680664,
+      "step": 947
+    },
+    {
+      "epoch": 1.7211625794732062,
+      "grad_norm": 2.337834596633911,
+      "learning_rate": 6.469098146954119e-06,
+      "logits/chosen": 0.08289645612239838,
+      "logits/rejected": 0.11055514216423035,
+      "logps/chosen": -85.81585693359375,
+      "logps/rejected": -98.47114562988281,
+      "loss": 1.6936,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.407276153564453,
+      "rewards/margins": 0.9422338008880615,
+      "rewards/rejected": -3.3495099544525146,
+      "step": 948
+    },
+    {
+      "epoch": 1.7229791099000908,
+      "grad_norm": 2.2546920776367188,
+      "learning_rate": 6.462821956652064e-06,
+      "logits/chosen": 0.035123877227306366,
+      "logits/rejected": 0.07462943345308304,
+      "logps/chosen": -76.06634521484375,
+      "logps/rejected": -84.29540252685547,
+      "loss": 2.1523,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3520922660827637,
+      "rewards/margins": 0.698824405670166,
+      "rewards/rejected": -3.050916910171509,
+      "step": 949
+    },
+    {
+      "epoch": 1.7247956403269755,
+      "grad_norm": 2.8522515296936035,
+      "learning_rate": 6.456543802988395e-06,
+      "logits/chosen": 0.07937800884246826,
+      "logits/rejected": 0.09652063250541687,
+      "logps/chosen": -83.96992492675781,
+      "logps/rejected": -97.54905700683594,
+      "loss": 1.9428,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6706998348236084,
+      "rewards/margins": 0.8279770612716675,
+      "rewards/rejected": -3.4986767768859863,
+      "step": 950
+    },
+    {
+      "epoch": 1.7266121707538602,
+      "grad_norm": 2.6932287216186523,
+      "learning_rate": 6.450263698765363e-06,
+      "logits/chosen": 0.07110100239515305,
+      "logits/rejected": 0.09598114341497421,
+      "logps/chosen": -84.68408966064453,
+      "logps/rejected": -96.74703979492188,
+      "loss": 1.7506,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.7609381675720215,
+      "rewards/margins": 0.9440479278564453,
+      "rewards/rejected": -3.7049858570098877,
+      "step": 951
+    },
+    {
+      "epoch": 1.7284287011807447,
+      "grad_norm": 2.2383601665496826,
+      "learning_rate": 6.4439816567891925e-06,
+      "logits/chosen": 0.03344205766916275,
+      "logits/rejected": 0.060385581105947495,
+      "logps/chosen": -90.16828155517578,
+      "logps/rejected": -102.0325927734375,
+      "loss": 1.6958,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.6442179679870605,
+      "rewards/margins": 1.103801965713501,
+      "rewards/rejected": -3.7480199337005615,
+      "step": 952
+    },
+    {
+      "epoch": 1.7302452316076296,
+      "grad_norm": 2.574528455734253,
+      "learning_rate": 6.43769768987006e-06,
+      "logits/chosen": 0.10802514851093292,
+      "logits/rejected": 0.06120527908205986,
+      "logps/chosen": -90.94403076171875,
+      "logps/rejected": -96.3043441772461,
+      "loss": 1.9554,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.774078845977783,
+      "rewards/margins": 0.967216968536377,
+      "rewards/rejected": -3.74129581451416,
+      "step": 953
+    },
+    {
+      "epoch": 1.732061762034514,
+      "grad_norm": 2.6094908714294434,
+      "learning_rate": 6.431411810822069e-06,
+      "logits/chosen": 0.013344529084861279,
+      "logits/rejected": 0.06106187403202057,
+      "logps/chosen": -82.14124298095703,
+      "logps/rejected": -97.54022979736328,
+      "loss": 2.1501,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.814952850341797,
+      "rewards/margins": 0.8146507143974304,
+      "rewards/rejected": -3.629603624343872,
+      "step": 954
+    },
+    {
+      "epoch": 1.7338782924613987,
+      "grad_norm": 2.762138605117798,
+      "learning_rate": 6.425124032463223e-06,
+      "logits/chosen": 0.03482529893517494,
+      "logits/rejected": 0.03158363327383995,
+      "logps/chosen": -84.45268249511719,
+      "logps/rejected": -91.9566650390625,
+      "loss": 2.1222,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.973424196243286,
+      "rewards/margins": 0.6969504952430725,
+      "rewards/rejected": -3.670374631881714,
+      "step": 955
+    },
+    {
+      "epoch": 1.7356948228882834,
+      "grad_norm": 2.802934169769287,
+      "learning_rate": 6.418834367615395e-06,
+      "logits/chosen": 0.09849456697702408,
+      "logits/rejected": 0.07233191281557083,
+      "logps/chosen": -85.41230010986328,
+      "logps/rejected": -92.97038269042969,
+      "loss": 2.1891,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4898338317871094,
+      "rewards/margins": 0.8130300045013428,
+      "rewards/rejected": -3.3028640747070312,
+      "step": 956
+    },
+    {
+      "epoch": 1.737511353315168,
+      "grad_norm": 2.2555596828460693,
+      "learning_rate": 6.4125428291043064e-06,
+      "logits/chosen": -0.0023345965892076492,
+      "logits/rejected": 0.08920536190271378,
+      "logps/chosen": -79.65885925292969,
+      "logps/rejected": -86.8980712890625,
+      "loss": 2.0766,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5075061321258545,
+      "rewards/margins": 0.6272554993629456,
+      "rewards/rejected": -3.1347618103027344,
+      "step": 957
+    },
+    {
+      "epoch": 1.7393278837420527,
+      "grad_norm": 3.097147226333618,
+      "learning_rate": 6.4062494297595025e-06,
+      "logits/chosen": 0.024910222738981247,
+      "logits/rejected": 0.07345539331436157,
+      "logps/chosen": -97.62705993652344,
+      "logps/rejected": -112.78961944580078,
+      "loss": 2.0315,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.446077823638916,
+      "rewards/margins": 0.7760088443756104,
+      "rewards/rejected": -3.2220866680145264,
+      "step": 958
+    },
+    {
+      "epoch": 1.7411444141689372,
+      "grad_norm": 2.4640140533447266,
+      "learning_rate": 6.399954182414317e-06,
+      "logits/chosen": 0.002200646325945854,
+      "logits/rejected": 0.054388273507356644,
+      "logps/chosen": -84.92884063720703,
+      "logps/rejected": -102.98699188232422,
+      "loss": 1.9766,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.6910696029663086,
+      "rewards/margins": 0.8245669603347778,
+      "rewards/rejected": -3.515636444091797,
+      "step": 959
+    },
+    {
+      "epoch": 1.742960944595822,
+      "grad_norm": 3.0036492347717285,
+      "learning_rate": 6.393657099905854e-06,
+      "logits/chosen": 0.09085743874311447,
+      "logits/rejected": 0.06018194556236267,
+      "logps/chosen": -95.91146850585938,
+      "logps/rejected": -102.4392318725586,
+      "loss": 2.6866,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -3.0499067306518555,
+      "rewards/margins": 0.4895576536655426,
+      "rewards/rejected": -3.5394644737243652,
+      "step": 960
+    },
+    {
+      "epoch": 1.7447774750227065,
+      "grad_norm": 3.2119052410125732,
+      "learning_rate": 6.387358195074968e-06,
+      "logits/chosen": 0.06521936506032944,
+      "logits/rejected": 0.10467957705259323,
+      "logps/chosen": -77.4794692993164,
+      "logps/rejected": -89.99710845947266,
+      "loss": 2.0836,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.603015899658203,
+      "rewards/margins": 0.7744656801223755,
+      "rewards/rejected": -3.377481460571289,
+      "step": 961
+    },
+    {
+      "epoch": 1.7465940054495914,
+      "grad_norm": 3.0814146995544434,
+      "learning_rate": 6.381057480766217e-06,
+      "logits/chosen": 0.08652829378843307,
+      "logits/rejected": 0.07287663966417313,
+      "logps/chosen": -83.66262817382812,
+      "logps/rejected": -100.8420639038086,
+      "loss": 1.9472,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7384371757507324,
+      "rewards/margins": 0.9687957763671875,
+      "rewards/rejected": -3.707233428955078,
+      "step": 962
+    },
+    {
+      "epoch": 1.7484105358764759,
+      "grad_norm": 2.4629218578338623,
+      "learning_rate": 6.374754969827855e-06,
+      "logits/chosen": -0.0019003197085112333,
+      "logits/rejected": 0.08570417761802673,
+      "logps/chosen": -83.80225372314453,
+      "logps/rejected": -101.95899200439453,
+      "loss": 2.0027,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.913536548614502,
+      "rewards/margins": 0.874174952507019,
+      "rewards/rejected": -3.7877113819122314,
+      "step": 963
+    },
+    {
+      "epoch": 1.7502270663033606,
+      "grad_norm": 2.850562572479248,
+      "learning_rate": 6.3684506751118016e-06,
+      "logits/chosen": 0.03824794292449951,
+      "logits/rejected": 0.10350260138511658,
+      "logps/chosen": -78.23057556152344,
+      "logps/rejected": -101.54271697998047,
+      "loss": 2.0086,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7757749557495117,
+      "rewards/margins": 1.0304248332977295,
+      "rewards/rejected": -3.806199550628662,
+      "step": 964
+    },
+    {
+      "epoch": 1.7520435967302452,
+      "grad_norm": 2.3626081943511963,
+      "learning_rate": 6.362144609473609e-06,
+      "logits/chosen": 0.04834410548210144,
+      "logits/rejected": 0.0772976502776146,
+      "logps/chosen": -90.42755126953125,
+      "logps/rejected": -96.35401916503906,
+      "loss": 2.3424,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.76436185836792,
+      "rewards/margins": 0.6245766878128052,
+      "rewards/rejected": -3.3889384269714355,
+      "step": 965
+    },
+    {
+      "epoch": 1.75386012715713,
+      "grad_norm": 2.790731191635132,
+      "learning_rate": 6.355836785772444e-06,
+      "logits/chosen": 0.05425344407558441,
+      "logits/rejected": 0.0733315572142601,
+      "logps/chosen": -87.27237701416016,
+      "logps/rejected": -100.02421569824219,
+      "loss": 1.925,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.5662331581115723,
+      "rewards/margins": 0.8153632879257202,
+      "rewards/rejected": -3.381596803665161,
+      "step": 966
+    },
+    {
+      "epoch": 1.7556766575840146,
+      "grad_norm": 2.5702576637268066,
+      "learning_rate": 6.349527216871059e-06,
+      "logits/chosen": 0.03957521542906761,
+      "logits/rejected": 0.08907577395439148,
+      "logps/chosen": -87.7530746459961,
+      "logps/rejected": -96.26228332519531,
+      "loss": 1.8814,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.9404945373535156,
+      "rewards/margins": 0.8145545125007629,
+      "rewards/rejected": -3.755048990249634,
+      "step": 967
+    },
+    {
+      "epoch": 1.757493188010899,
+      "grad_norm": 3.443202495574951,
+      "learning_rate": 6.343215915635762e-06,
+      "logits/chosen": -0.01773964986205101,
+      "logits/rejected": 0.05454108119010925,
+      "logps/chosen": -76.4457778930664,
+      "logps/rejected": -94.93628692626953,
+      "loss": 1.7882,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6354503631591797,
+      "rewards/margins": 1.137264609336853,
+      "rewards/rejected": -3.772714853286743,
+      "step": 968
+    },
+    {
+      "epoch": 1.759309718437784,
+      "grad_norm": 3.239837169647217,
+      "learning_rate": 6.336902894936394e-06,
+      "logits/chosen": 0.019338876008987427,
+      "logits/rejected": 0.024039428681135178,
+      "logps/chosen": -85.57742309570312,
+      "logps/rejected": -93.43807983398438,
+      "loss": 1.9777,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5836596488952637,
+      "rewards/margins": 0.6887004375457764,
+      "rewards/rejected": -3.272360324859619,
+      "step": 969
+    },
+    {
+      "epoch": 1.7611262488646684,
+      "grad_norm": 2.760834217071533,
+      "learning_rate": 6.330588167646305e-06,
+      "logits/chosen": 0.025167983025312424,
+      "logits/rejected": 0.045912109315395355,
+      "logps/chosen": -90.79326629638672,
+      "logps/rejected": -96.16666412353516,
+      "loss": 2.353,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.9890851974487305,
+      "rewards/margins": 0.8027018308639526,
+      "rewards/rejected": -3.7917871475219727,
+      "step": 970
+    },
+    {
+      "epoch": 1.7629427792915533,
+      "grad_norm": 2.8496274948120117,
+      "learning_rate": 6.324271746642323e-06,
+      "logits/chosen": 0.05632323399186134,
+      "logits/rejected": 0.06209895759820938,
+      "logps/chosen": -84.43073272705078,
+      "logps/rejected": -96.11868286132812,
+      "loss": 1.7381,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.424320697784424,
+      "rewards/margins": 1.1575902700424194,
+      "rewards/rejected": -3.581911087036133,
+      "step": 971
+    },
+    {
+      "epoch": 1.7647593097184378,
+      "grad_norm": 2.613065719604492,
+      "learning_rate": 6.317953644804731e-06,
+      "logits/chosen": 0.08047536015510559,
+      "logits/rejected": 0.05248764902353287,
+      "logps/chosen": -95.60069274902344,
+      "logps/rejected": -101.73131561279297,
+      "loss": 2.2215,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.8458030223846436,
+      "rewards/margins": 0.6640199422836304,
+      "rewards/rejected": -3.5098226070404053,
+      "step": 972
+    },
+    {
+      "epoch": 1.7665758401453224,
+      "grad_norm": 2.8011529445648193,
+      "learning_rate": 6.311633875017237e-06,
+      "logits/chosen": 0.10373665392398834,
+      "logits/rejected": 0.1293720006942749,
+      "logps/chosen": -80.55433654785156,
+      "logps/rejected": -91.2505874633789,
+      "loss": 2.1095,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.8390448093414307,
+      "rewards/margins": 0.7663621306419373,
+      "rewards/rejected": -3.6054069995880127,
+      "step": 973
+    },
+    {
+      "epoch": 1.768392370572207,
+      "grad_norm": 2.5334253311157227,
+      "learning_rate": 6.305312450166952e-06,
+      "logits/chosen": 0.05905047059059143,
+      "logits/rejected": 0.11380237340927124,
+      "logps/chosen": -78.28790283203125,
+      "logps/rejected": -85.13541412353516,
+      "loss": 1.9114,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6557254791259766,
+      "rewards/margins": 0.9857444763183594,
+      "rewards/rejected": -3.641469955444336,
+      "step": 974
+    },
+    {
+      "epoch": 1.7702089009990918,
+      "grad_norm": 2.603902578353882,
+      "learning_rate": 6.298989383144366e-06,
+      "logits/chosen": -0.02739763632416725,
+      "logits/rejected": 0.04656928405165672,
+      "logps/chosen": -93.05496978759766,
+      "logps/rejected": -102.54607391357422,
+      "loss": 2.3069,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.9203593730926514,
+      "rewards/margins": 0.6168675422668457,
+      "rewards/rejected": -3.537227153778076,
+      "step": 975
+    },
+    {
+      "epoch": 1.7720254314259765,
+      "grad_norm": 3.098738670349121,
+      "learning_rate": 6.292664686843308e-06,
+      "logits/chosen": 0.123380646109581,
+      "logits/rejected": 0.07776113599538803,
+      "logps/chosen": -101.1915512084961,
+      "logps/rejected": -105.4427719116211,
+      "loss": 2.0616,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.950930118560791,
+      "rewards/margins": 0.7281030416488647,
+      "rewards/rejected": -3.679033041000366,
+      "step": 976
+    },
+    {
+      "epoch": 1.773841961852861,
+      "grad_norm": 3.1999757289886475,
+      "learning_rate": 6.286338374160939e-06,
+      "logits/chosen": 0.06259569525718689,
+      "logits/rejected": 0.07669581472873688,
+      "logps/chosen": -81.40658569335938,
+      "logps/rejected": -90.0005874633789,
+      "loss": 2.0672,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.647714614868164,
+      "rewards/margins": 0.9036591053009033,
+      "rewards/rejected": -3.5513734817504883,
+      "step": 977
+    },
+    {
+      "epoch": 1.7756584922797458,
+      "grad_norm": 2.5514464378356934,
+      "learning_rate": 6.280010457997711e-06,
+      "logits/chosen": -0.02144845947623253,
+      "logits/rejected": 0.02661653235554695,
+      "logps/chosen": -90.29862213134766,
+      "logps/rejected": -102.64820861816406,
+      "loss": 1.7177,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.7331790924072266,
+      "rewards/margins": 1.1093989610671997,
+      "rewards/rejected": -3.842578411102295,
+      "step": 978
+    },
+    {
+      "epoch": 1.7774750227066303,
+      "grad_norm": 2.597369432449341,
+      "learning_rate": 6.273680951257343e-06,
+      "logits/chosen": 0.05057895556092262,
+      "logits/rejected": 0.07442101091146469,
+      "logps/chosen": -89.04737091064453,
+      "logps/rejected": -94.92131042480469,
+      "loss": 1.9793,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.585592746734619,
+      "rewards/margins": 0.7941515445709229,
+      "rewards/rejected": -3.379744291305542,
+      "step": 979
+    },
+    {
+      "epoch": 1.779291553133515,
+      "grad_norm": 2.6108598709106445,
+      "learning_rate": 6.267349866846808e-06,
+      "logits/chosen": 0.0492534302175045,
+      "logits/rejected": 0.04458750784397125,
+      "logps/chosen": -81.35431671142578,
+      "logps/rejected": -93.98658752441406,
+      "loss": 1.9557,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.794140338897705,
+      "rewards/margins": 0.9564652442932129,
+      "rewards/rejected": -3.7506051063537598,
+      "step": 980
+    },
+    {
+      "epoch": 1.7811080835603996,
+      "grad_norm": 2.4546945095062256,
+      "learning_rate": 6.2610172176762855e-06,
+      "logits/chosen": 0.010974138975143433,
+      "logits/rejected": 0.05425180122256279,
+      "logps/chosen": -81.56734466552734,
+      "logps/rejected": -99.60003662109375,
+      "loss": 1.6879,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.6332409381866455,
+      "rewards/margins": 1.0027284622192383,
+      "rewards/rejected": -3.635969400405884,
+      "step": 981
+    },
+    {
+      "epoch": 1.7829246139872843,
+      "grad_norm": 2.4584851264953613,
+      "learning_rate": 6.25468301665915e-06,
+      "logits/chosen": 0.06493166089057922,
+      "logits/rejected": 0.08574498444795609,
+      "logps/chosen": -77.84220886230469,
+      "logps/rejected": -84.97069549560547,
+      "loss": 1.9226,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.454103708267212,
+      "rewards/margins": 0.7075288891792297,
+      "rewards/rejected": -3.161632776260376,
+      "step": 982
+    },
+    {
+      "epoch": 1.784741144414169,
+      "grad_norm": 2.611698627471924,
+      "learning_rate": 6.248347276711941e-06,
+      "logits/chosen": 0.05734262615442276,
+      "logits/rejected": 0.0742194727063179,
+      "logps/chosen": -97.02610778808594,
+      "logps/rejected": -102.16181945800781,
+      "loss": 2.334,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.948233127593994,
+      "rewards/margins": 0.6351866722106934,
+      "rewards/rejected": -3.5834197998046875,
+      "step": 983
+    },
+    {
+      "epoch": 1.7865576748410534,
+      "grad_norm": 2.727557420730591,
+      "learning_rate": 6.242010010754336e-06,
+      "logits/chosen": 0.05103582888841629,
+      "logits/rejected": 0.059754204005002975,
+      "logps/chosen": -73.64825439453125,
+      "logps/rejected": -85.44071960449219,
+      "loss": 1.8675,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4483113288879395,
+      "rewards/margins": 0.932466983795166,
+      "rewards/rejected": -3.3807785511016846,
+      "step": 984
+    },
+    {
+      "epoch": 1.7883742052679383,
+      "grad_norm": 2.2711496353149414,
+      "learning_rate": 6.2356712317091205e-06,
+      "logits/chosen": 0.0846191942691803,
+      "logits/rejected": 0.08121812343597412,
+      "logps/chosen": -81.7682113647461,
+      "logps/rejected": -89.93470001220703,
+      "loss": 1.8697,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5603229999542236,
+      "rewards/margins": 0.7764644622802734,
+      "rewards/rejected": -3.336787462234497,
+      "step": 985
+    },
+    {
+      "epoch": 1.7901907356948228,
+      "grad_norm": 2.29776930809021,
+      "learning_rate": 6.229330952502174e-06,
+      "logits/chosen": 0.020622923970222473,
+      "logits/rejected": 0.10273537039756775,
+      "logps/chosen": -80.6253662109375,
+      "logps/rejected": -95.64512634277344,
+      "loss": 1.7414,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.5752944946289062,
+      "rewards/margins": 0.9720845818519592,
+      "rewards/rejected": -3.547379493713379,
+      "step": 986
+    },
+    {
+      "epoch": 1.7920072661217077,
+      "grad_norm": 2.5851526260375977,
+      "learning_rate": 6.222989186062426e-06,
+      "logits/chosen": 0.07696882635354996,
+      "logits/rejected": 0.07504019141197205,
+      "logps/chosen": -94.02040100097656,
+      "logps/rejected": -95.83682250976562,
+      "loss": 2.1574,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.6728358268737793,
+      "rewards/margins": 0.7833377718925476,
+      "rewards/rejected": -3.4561736583709717,
+      "step": 987
+    },
+    {
+      "epoch": 1.7938237965485921,
+      "grad_norm": 3.3578975200653076,
+      "learning_rate": 6.216645945321841e-06,
+      "logits/chosen": 0.09853662550449371,
+      "logits/rejected": 0.07587447762489319,
+      "logps/chosen": -90.90428924560547,
+      "logps/rejected": -97.71269226074219,
+      "loss": 2.388,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.915060043334961,
+      "rewards/margins": 0.736826479434967,
+      "rewards/rejected": -3.651886463165283,
+      "step": 988
+    },
+    {
+      "epoch": 1.7956403269754768,
+      "grad_norm": 3.560649871826172,
+      "learning_rate": 6.210301243215396e-06,
+      "logits/chosen": 0.05813619866967201,
+      "logits/rejected": 0.08984968811273575,
+      "logps/chosen": -89.59809875488281,
+      "logps/rejected": -98.89776611328125,
+      "loss": 2.0893,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.80810546875,
+      "rewards/margins": 0.8890154957771301,
+      "rewards/rejected": -3.6971209049224854,
+      "step": 989
+    },
+    {
+      "epoch": 1.7974568574023615,
+      "grad_norm": 3.1380116939544678,
+      "learning_rate": 6.20395509268104e-06,
+      "logits/chosen": 0.033310696482658386,
+      "logits/rejected": 0.07882840186357498,
+      "logps/chosen": -88.1374282836914,
+      "logps/rejected": -97.17999267578125,
+      "loss": 2.0696,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.663801670074463,
+      "rewards/margins": 0.6945213675498962,
+      "rewards/rejected": -3.358323097229004,
+      "step": 990
+    },
+    {
+      "epoch": 1.7992733878292462,
+      "grad_norm": 2.356675863265991,
+      "learning_rate": 6.197607506659678e-06,
+      "logits/chosen": -0.005927765741944313,
+      "logits/rejected": 0.055597010999917984,
+      "logps/chosen": -91.83357238769531,
+      "logps/rejected": -98.24932861328125,
+      "loss": 2.2475,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7551968097686768,
+      "rewards/margins": 0.7314962148666382,
+      "rewards/rejected": -3.4866933822631836,
+      "step": 991
+    },
+    {
+      "epoch": 1.8010899182561309,
+      "grad_norm": 2.7283682823181152,
+      "learning_rate": 6.191258498095148e-06,
+      "logits/chosen": -0.02681618742644787,
+      "logits/rejected": 0.09765288978815079,
+      "logps/chosen": -88.14956665039062,
+      "logps/rejected": -108.59061431884766,
+      "loss": 2.046,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.5922963619232178,
+      "rewards/margins": 0.9590500593185425,
+      "rewards/rejected": -3.5513463020324707,
+      "step": 992
+    },
+    {
+      "epoch": 1.8029064486830153,
+      "grad_norm": 2.527743339538574,
+      "learning_rate": 6.184908079934181e-06,
+      "logits/chosen": 0.10376276075839996,
+      "logits/rejected": 0.13840174674987793,
+      "logps/chosen": -81.53548431396484,
+      "logps/rejected": -91.71275329589844,
+      "loss": 1.7434,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.7831573486328125,
+      "rewards/margins": 1.0856074094772339,
+      "rewards/rejected": -3.868764877319336,
+      "step": 993
+    },
+    {
+      "epoch": 1.8047229791099002,
+      "grad_norm": 2.573827028274536,
+      "learning_rate": 6.178556265126383e-06,
+      "logits/chosen": 0.019435830414295197,
+      "logits/rejected": 0.0659492164850235,
+      "logps/chosen": -78.87043762207031,
+      "logps/rejected": -86.28160858154297,
+      "loss": 2.1035,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.601402759552002,
+      "rewards/margins": 0.636344850063324,
+      "rewards/rejected": -3.23774790763855,
+      "step": 994
+    },
+    {
+      "epoch": 1.8065395095367847,
+      "grad_norm": 2.3812830448150635,
+      "learning_rate": 6.172203066624213e-06,
+      "logits/chosen": 0.04023706167936325,
+      "logits/rejected": 0.04221217706799507,
+      "logps/chosen": -89.31175231933594,
+      "logps/rejected": -101.20529174804688,
+      "loss": 1.8209,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4635345935821533,
+      "rewards/margins": 0.9060386419296265,
+      "rewards/rejected": -3.3695733547210693,
+      "step": 995
+    },
+    {
+      "epoch": 1.8083560399636693,
+      "grad_norm": 2.4562618732452393,
+      "learning_rate": 6.16584849738295e-06,
+      "logits/chosen": 0.12471188604831696,
+      "logits/rejected": 0.16778936982154846,
+      "logps/chosen": -87.34111022949219,
+      "logps/rejected": -99.54594421386719,
+      "loss": 1.8418,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6531598567962646,
+      "rewards/margins": 1.1297169923782349,
+      "rewards/rejected": -3.78287672996521,
+      "step": 996
+    },
+    {
+      "epoch": 1.810172570390554,
+      "grad_norm": 2.9506518840789795,
+      "learning_rate": 6.159492570360664e-06,
+      "logits/chosen": 0.06164424866437912,
+      "logits/rejected": 0.043557893484830856,
+      "logps/chosen": -80.98966217041016,
+      "logps/rejected": -84.5133056640625,
+      "loss": 2.3627,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.7590372562408447,
+      "rewards/margins": 0.5638130903244019,
+      "rewards/rejected": -3.322849988937378,
+      "step": 997
+    },
+    {
+      "epoch": 1.8119891008174387,
+      "grad_norm": 2.8041412830352783,
+      "learning_rate": 6.1531352985182e-06,
+      "logits/chosen": 0.07123232632875443,
+      "logits/rejected": 0.1083369255065918,
+      "logps/chosen": -88.13341522216797,
+      "logps/rejected": -101.4854736328125,
+      "loss": 1.9746,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.748055934906006,
+      "rewards/margins": 0.8071377277374268,
+      "rewards/rejected": -3.5551934242248535,
+      "step": 998
+    },
+    {
+      "epoch": 1.8138056312443234,
+      "grad_norm": 2.383258104324341,
+      "learning_rate": 6.146776694819139e-06,
+      "logits/chosen": -0.025281429290771484,
+      "logits/rejected": 0.016018148511648178,
+      "logps/chosen": -83.41783905029297,
+      "logps/rejected": -96.40564727783203,
+      "loss": 1.881,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.274135112762451,
+      "rewards/margins": 0.8547357320785522,
+      "rewards/rejected": -3.128870725631714,
+      "step": 999
+    },
+    {
+      "epoch": 1.815622161671208,
+      "grad_norm": 2.6372334957122803,
+      "learning_rate": 6.140416772229785e-06,
+      "logits/chosen": 0.03502847999334335,
+      "logits/rejected": 0.1193343997001648,
+      "logps/chosen": -82.01066589355469,
+      "logps/rejected": -99.53868865966797,
+      "loss": 1.8357,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.390124797821045,
+      "rewards/margins": 0.9666082262992859,
+      "rewards/rejected": -3.3567328453063965,
+      "step": 1000
+    },
+    {
+      "epoch": 1.8174386920980927,
+      "grad_norm": 2.2228522300720215,
+      "learning_rate": 6.134055543719122e-06,
+      "logits/chosen": 0.08193308860063553,
+      "logits/rejected": 0.0918150246143341,
+      "logps/chosen": -89.3790283203125,
+      "logps/rejected": -104.6737060546875,
+      "loss": 1.9502,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5274341106414795,
+      "rewards/margins": 0.9238023161888123,
+      "rewards/rejected": -3.4512364864349365,
+      "step": 1001
+    },
+    {
+      "epoch": 1.8192552225249772,
+      "grad_norm": 3.1102793216705322,
+      "learning_rate": 6.1276930222588085e-06,
+      "logits/chosen": 0.04992733523249626,
+      "logits/rejected": 0.03953206539154053,
+      "logps/chosen": -88.82096862792969,
+      "logps/rejected": -90.49384307861328,
+      "loss": 2.1536,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5509090423583984,
+      "rewards/margins": 0.6242492198944092,
+      "rewards/rejected": -3.1751582622528076,
+      "step": 1002
+    },
+    {
+      "epoch": 1.821071752951862,
+      "grad_norm": 2.4615204334259033,
+      "learning_rate": 6.12132922082313e-06,
+      "logits/chosen": 0.034680627286434174,
+      "logits/rejected": 0.10913330316543579,
+      "logps/chosen": -75.16706848144531,
+      "logps/rejected": -88.61273956298828,
+      "loss": 1.7472,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3874828815460205,
+      "rewards/margins": 0.8909300565719604,
+      "rewards/rejected": -3.2784130573272705,
+      "step": 1003
+    },
+    {
+      "epoch": 1.8228882833787465,
+      "grad_norm": 2.435248374938965,
+      "learning_rate": 6.114964152388986e-06,
+      "logits/chosen": 0.04803591966629028,
+      "logits/rejected": 0.06991353631019592,
+      "logps/chosen": -84.33198547363281,
+      "logps/rejected": -95.14921569824219,
+      "loss": 1.9234,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6023752689361572,
+      "rewards/margins": 0.8003835082054138,
+      "rewards/rejected": -3.402759075164795,
+      "step": 1004
+    },
+    {
+      "epoch": 1.8247048138056312,
+      "grad_norm": 2.49566912651062,
+      "learning_rate": 6.10859782993586e-06,
+      "logits/chosen": 0.05515093728899956,
+      "logits/rejected": 0.08386579155921936,
+      "logps/chosen": -91.16459655761719,
+      "logps/rejected": -113.24137878417969,
+      "loss": 1.9861,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7292985916137695,
+      "rewards/margins": 0.9162713289260864,
+      "rewards/rejected": -3.6455698013305664,
+      "step": 1005
+    },
+    {
+      "epoch": 1.8265213442325159,
+      "grad_norm": 3.0968334674835205,
+      "learning_rate": 6.1022302664457935e-06,
+      "logits/chosen": 0.0909426286816597,
+      "logits/rejected": 0.13981547951698303,
+      "logps/chosen": -76.67719268798828,
+      "logps/rejected": -94.50039672851562,
+      "loss": 1.828,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.559506416320801,
+      "rewards/margins": 1.0854456424713135,
+      "rewards/rejected": -3.6449520587921143,
+      "step": 1006
+    },
+    {
+      "epoch": 1.8283378746594006,
+      "grad_norm": 2.719757080078125,
+      "learning_rate": 6.095861474903354e-06,
+      "logits/chosen": 0.010468171909451485,
+      "logits/rejected": 0.02864597924053669,
+      "logps/chosen": -85.98566436767578,
+      "logps/rejected": -98.51669311523438,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.431942939758301,
+      "rewards/margins": 0.5168296694755554,
+      "rewards/rejected": -2.948772430419922,
+      "step": 1007
+    },
+    {
+      "epoch": 1.8301544050862852,
+      "grad_norm": 3.089297294616699,
+      "learning_rate": 6.089491468295619e-06,
+      "logits/chosen": 0.06479363143444061,
+      "logits/rejected": 0.1313035637140274,
+      "logps/chosen": -93.31026458740234,
+      "logps/rejected": -103.13523864746094,
+      "loss": 2.119,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.7715659141540527,
+      "rewards/margins": 0.7990520000457764,
+      "rewards/rejected": -3.570618152618408,
+      "step": 1008
+    },
+    {
+      "epoch": 1.8319709355131697,
+      "grad_norm": 2.6829261779785156,
+      "learning_rate": 6.083120259612139e-06,
+      "logits/chosen": 0.06222057342529297,
+      "logits/rejected": 0.06457602977752686,
+      "logps/chosen": -83.79937744140625,
+      "logps/rejected": -93.75650024414062,
+      "loss": 1.8928,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6541078090667725,
+      "rewards/margins": 0.8687030076980591,
+      "rewards/rejected": -3.522810697555542,
+      "step": 1009
+    },
+    {
+      "epoch": 1.8337874659400546,
+      "grad_norm": 2.447404623031616,
+      "learning_rate": 6.076747861844919e-06,
+      "logits/chosen": 0.02118651196360588,
+      "logits/rejected": 0.00795823149383068,
+      "logps/chosen": -88.48612976074219,
+      "logps/rejected": -103.70702362060547,
+      "loss": 1.7299,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.2803008556365967,
+      "rewards/margins": 1.0432041883468628,
+      "rewards/rejected": -3.32350492477417,
+      "step": 1010
+    },
+    {
+      "epoch": 1.835603996366939,
+      "grad_norm": 2.094322681427002,
+      "learning_rate": 6.070374287988387e-06,
+      "logits/chosen": 0.05191745609045029,
+      "logits/rejected": 0.07431665807962418,
+      "logps/chosen": -80.48505401611328,
+      "logps/rejected": -93.57875061035156,
+      "loss": 1.937,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5567147731781006,
+      "rewards/margins": 0.762231171131134,
+      "rewards/rejected": -3.31894588470459,
+      "step": 1011
+    },
+    {
+      "epoch": 1.837420526793824,
+      "grad_norm": 2.8458070755004883,
+      "learning_rate": 6.06399955103937e-06,
+      "logits/chosen": 0.02621028572320938,
+      "logits/rejected": 0.024392439052462578,
+      "logps/chosen": -90.91835021972656,
+      "logps/rejected": -98.42837524414062,
+      "loss": 2.2507,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.444821357727051,
+      "rewards/margins": 0.6921339631080627,
+      "rewards/rejected": -3.1369552612304688,
+      "step": 1012
+    },
+    {
+      "epoch": 1.8392370572207084,
+      "grad_norm": 2.8486974239349365,
+      "learning_rate": 6.057623663997067e-06,
+      "logits/chosen": 0.0883512869477272,
+      "logits/rejected": 0.0882159024477005,
+      "logps/chosen": -81.09939575195312,
+      "logps/rejected": -91.97223663330078,
+      "loss": 1.8517,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5850532054901123,
+      "rewards/margins": 0.9884753227233887,
+      "rewards/rejected": -3.573528528213501,
+      "step": 1013
+    },
+    {
+      "epoch": 1.841053587647593,
+      "grad_norm": 2.216489553451538,
+      "learning_rate": 6.051246639863021e-06,
+      "logits/chosen": 0.07911565154790878,
+      "logits/rejected": 0.0659380629658699,
+      "logps/chosen": -82.15691375732422,
+      "logps/rejected": -89.2088851928711,
+      "loss": 2.0057,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5066802501678467,
+      "rewards/margins": 0.7392367720603943,
+      "rewards/rejected": -3.2459168434143066,
+      "step": 1014
+    },
+    {
+      "epoch": 1.8428701180744778,
+      "grad_norm": 2.6589348316192627,
+      "learning_rate": 6.044868491641097e-06,
+      "logits/chosen": 0.0051671527326107025,
+      "logits/rejected": 0.06533940136432648,
+      "logps/chosen": -94.15599822998047,
+      "logps/rejected": -106.89962768554688,
+      "loss": 1.9055,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5753836631774902,
+      "rewards/margins": 0.968835711479187,
+      "rewards/rejected": -3.544219493865967,
+      "step": 1015
+    },
+    {
+      "epoch": 1.8446866485013624,
+      "grad_norm": 2.1812055110931396,
+      "learning_rate": 6.038489232337447e-06,
+      "logits/chosen": 0.08165213465690613,
+      "logits/rejected": 0.07264502346515656,
+      "logps/chosen": -84.1948471069336,
+      "logps/rejected": -91.88217163085938,
+      "loss": 1.9628,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.6751532554626465,
+      "rewards/margins": 0.9011788368225098,
+      "rewards/rejected": -3.5763320922851562,
+      "step": 1016
+    },
+    {
+      "epoch": 1.8465031789282471,
+      "grad_norm": 2.577268362045288,
+      "learning_rate": 6.032108874960491e-06,
+      "logits/chosen": 0.059723783284425735,
+      "logits/rejected": 0.08904848992824554,
+      "logps/chosen": -75.84976196289062,
+      "logps/rejected": -94.64164733886719,
+      "loss": 1.856,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.532991647720337,
+      "rewards/margins": 1.01395845413208,
+      "rewards/rejected": -3.546950101852417,
+      "step": 1017
+    },
+    {
+      "epoch": 1.8483197093551316,
+      "grad_norm": 4.6174211502075195,
+      "learning_rate": 6.025727432520892e-06,
+      "logits/chosen": 0.03918137028813362,
+      "logits/rejected": 0.021953536197543144,
+      "logps/chosen": -86.72551727294922,
+      "logps/rejected": -92.44689178466797,
+      "loss": 1.8242,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.633363723754883,
+      "rewards/margins": 0.8154869675636292,
+      "rewards/rejected": -3.448850631713867,
+      "step": 1018
+    },
+    {
+      "epoch": 1.8501362397820165,
+      "grad_norm": 2.8885879516601562,
+      "learning_rate": 6.01934491803152e-06,
+      "logits/chosen": 0.02478492073714733,
+      "logits/rejected": 0.0335882194340229,
+      "logps/chosen": -97.10049438476562,
+      "logps/rejected": -117.7772216796875,
+      "loss": 1.9857,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -3.0783286094665527,
+      "rewards/margins": 0.9448035359382629,
+      "rewards/rejected": -4.023131847381592,
+      "step": 1019
+    },
+    {
+      "epoch": 1.851952770208901,
+      "grad_norm": 2.8736963272094727,
+      "learning_rate": 6.012961344507433e-06,
+      "logits/chosen": 0.05578463524580002,
+      "logits/rejected": 0.05935496836900711,
+      "logps/chosen": -76.78963470458984,
+      "logps/rejected": -85.67704772949219,
+      "loss": 2.1156,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.713351011276245,
+      "rewards/margins": 0.6802167296409607,
+      "rewards/rejected": -3.3935675621032715,
+      "step": 1020
+    },
+    {
+      "epoch": 1.8537693006357856,
+      "grad_norm": 2.5551319122314453,
+      "learning_rate": 6.0065767249658506e-06,
+      "logits/chosen": -0.018782105296850204,
+      "logits/rejected": 0.04194478690624237,
+      "logps/chosen": -95.86324310302734,
+      "logps/rejected": -108.91368103027344,
+      "loss": 1.759,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3475992679595947,
+      "rewards/margins": 0.9991620182991028,
+      "rewards/rejected": -3.3467609882354736,
+      "step": 1021
+    },
+    {
+      "epoch": 1.8555858310626703,
+      "grad_norm": 2.703446388244629,
+      "learning_rate": 6.0001910724261215e-06,
+      "logits/chosen": 0.10875789821147919,
+      "logits/rejected": 0.09087061136960983,
+      "logps/chosen": -89.86807250976562,
+      "logps/rejected": -91.68760681152344,
+      "loss": 2.3511,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.873063325881958,
+      "rewards/margins": 0.5627723336219788,
+      "rewards/rejected": -3.435835599899292,
+      "step": 1022
+    },
+    {
+      "epoch": 1.857402361489555,
+      "grad_norm": 2.9073855876922607,
+      "learning_rate": 5.993804399909704e-06,
+      "logits/chosen": 0.0655340775847435,
+      "logits/rejected": 0.14421050250530243,
+      "logps/chosen": -89.02464294433594,
+      "logps/rejected": -107.83901977539062,
+      "loss": 2.058,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5416080951690674,
+      "rewards/margins": 0.8793646097183228,
+      "rewards/rejected": -3.4209725856781006,
+      "step": 1023
+    },
+    {
+      "epoch": 1.8592188919164396,
+      "grad_norm": 2.6842780113220215,
+      "learning_rate": 5.987416720440137e-06,
+      "logits/chosen": 0.03317519277334213,
+      "logits/rejected": 0.10288646817207336,
+      "logps/chosen": -82.84213256835938,
+      "logps/rejected": -97.43663024902344,
+      "loss": 2.0877,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3678500652313232,
+      "rewards/margins": 0.7004600763320923,
+      "rewards/rejected": -3.068309783935547,
+      "step": 1024
+    },
+    {
+      "epoch": 1.861035422343324,
+      "grad_norm": 3.016253709793091,
+      "learning_rate": 5.9810280470430096e-06,
+      "logits/chosen": 0.10067103803157806,
+      "logits/rejected": 0.04382970929145813,
+      "logps/chosen": -94.9444351196289,
+      "logps/rejected": -97.62810516357422,
+      "loss": 1.9073,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.510204792022705,
+      "rewards/margins": 0.7732842564582825,
+      "rewards/rejected": -3.283489227294922,
+      "step": 1025
+    },
+    {
+      "epoch": 1.862851952770209,
+      "grad_norm": 2.667470693588257,
+      "learning_rate": 5.974638392745937e-06,
+      "logits/chosen": 0.11636751890182495,
+      "logits/rejected": 0.03302691876888275,
+      "logps/chosen": -90.78875732421875,
+      "logps/rejected": -92.41315460205078,
+      "loss": 2.5596,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.622999668121338,
+      "rewards/margins": 0.5555787682533264,
+      "rewards/rejected": -3.1785783767700195,
+      "step": 1026
+    },
+    {
+      "epoch": 1.8646684831970934,
+      "grad_norm": 3.063692808151245,
+      "learning_rate": 5.9682477705785415e-06,
+      "logits/chosen": 0.020983893424272537,
+      "logits/rejected": 0.08335445076227188,
+      "logps/chosen": -84.9983139038086,
+      "logps/rejected": -96.67384338378906,
+      "loss": 1.9632,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6382555961608887,
+      "rewards/margins": 0.8702573180198669,
+      "rewards/rejected": -3.5085127353668213,
+      "step": 1027
+    },
+    {
+      "epoch": 1.8664850136239783,
+      "grad_norm": 2.388528823852539,
+      "learning_rate": 5.961856193572411e-06,
+      "logits/chosen": 0.09512500464916229,
+      "logits/rejected": 0.08783603459596634,
+      "logps/chosen": -78.07685852050781,
+      "logps/rejected": -84.70189666748047,
+      "loss": 2.0445,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.421865701675415,
+      "rewards/margins": 0.7848958969116211,
+      "rewards/rejected": -3.206761598587036,
+      "step": 1028
+    },
+    {
+      "epoch": 1.8683015440508628,
+      "grad_norm": 2.6571311950683594,
+      "learning_rate": 5.955463674761085e-06,
+      "logits/chosen": 0.009374767541885376,
+      "logits/rejected": 0.09341388195753098,
+      "logps/chosen": -91.43545532226562,
+      "logps/rejected": -108.69731140136719,
+      "loss": 1.8366,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6931421756744385,
+      "rewards/margins": 0.9958528280258179,
+      "rewards/rejected": -3.688995122909546,
+      "step": 1029
+    },
+    {
+      "epoch": 1.8701180744777475,
+      "grad_norm": 2.6928844451904297,
+      "learning_rate": 5.9490702271800225e-06,
+      "logits/chosen": 0.015922928228974342,
+      "logits/rejected": 0.07699844241142273,
+      "logps/chosen": -83.87287902832031,
+      "logps/rejected": -98.65171813964844,
+      "loss": 1.7379,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.819220542907715,
+      "rewards/margins": 0.9954001903533936,
+      "rewards/rejected": -3.8146207332611084,
+      "step": 1030
+    },
+    {
+      "epoch": 1.8719346049046321,
+      "grad_norm": 2.33632755279541,
+      "learning_rate": 5.9426758638665775e-06,
+      "logits/chosen": -0.017036177217960358,
+      "logits/rejected": 0.05616893619298935,
+      "logps/chosen": -84.12918853759766,
+      "logps/rejected": -99.93852233886719,
+      "loss": 2.1195,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.344330072402954,
+      "rewards/margins": 0.8381181359291077,
+      "rewards/rejected": -3.182448387145996,
+      "step": 1031
+    },
+    {
+      "epoch": 1.8737511353315168,
+      "grad_norm": 2.552150011062622,
+      "learning_rate": 5.936280597859968e-06,
+      "logits/chosen": -0.020246881991624832,
+      "logits/rejected": 0.12626580893993378,
+      "logps/chosen": -84.22633361816406,
+      "logps/rejected": -107.020263671875,
+      "loss": 1.7604,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.7721216678619385,
+      "rewards/margins": 1.0607465505599976,
+      "rewards/rejected": -3.8328680992126465,
+      "step": 1032
+    },
+    {
+      "epoch": 1.8755676657584015,
+      "grad_norm": 2.8530185222625732,
+      "learning_rate": 5.929884442201255e-06,
+      "logits/chosen": 0.0990794450044632,
+      "logits/rejected": 0.08221088349819183,
+      "logps/chosen": -85.16941833496094,
+      "logps/rejected": -97.42240905761719,
+      "loss": 1.7271,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.6215596199035645,
+      "rewards/margins": 0.9904119372367859,
+      "rewards/rejected": -3.611971616744995,
+      "step": 1033
+    },
+    {
+      "epoch": 1.877384196185286,
+      "grad_norm": 2.0179224014282227,
+      "learning_rate": 5.9234874099333165e-06,
+      "logits/chosen": 0.042671047151088715,
+      "logits/rejected": 0.06781131029129028,
+      "logps/chosen": -94.68910217285156,
+      "logps/rejected": -104.71215057373047,
+      "loss": 1.9179,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3657076358795166,
+      "rewards/margins": 0.916804313659668,
+      "rewards/rejected": -3.2825117111206055,
+      "step": 1034
+    },
+    {
+      "epoch": 1.8792007266121709,
+      "grad_norm": 3.3149285316467285,
+      "learning_rate": 5.91708951410081e-06,
+      "logits/chosen": 0.016219373792409897,
+      "logits/rejected": 0.12756215035915375,
+      "logps/chosen": -92.19625091552734,
+      "logps/rejected": -107.68182373046875,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7595980167388916,
+      "rewards/margins": 0.8341989517211914,
+      "rewards/rejected": -3.593796968460083,
+      "step": 1035
+    },
+    {
+      "epoch": 1.8810172570390553,
+      "grad_norm": 2.448784589767456,
+      "learning_rate": 5.910690767750164e-06,
+      "logits/chosen": 0.0679556131362915,
+      "logits/rejected": 0.12997615337371826,
+      "logps/chosen": -86.93294525146484,
+      "logps/rejected": -96.73664093017578,
+      "loss": 2.1059,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.574408769607544,
+      "rewards/margins": 0.6823816299438477,
+      "rewards/rejected": -3.2567903995513916,
+      "step": 1036
+    },
+    {
+      "epoch": 1.8828337874659402,
+      "grad_norm": 2.479353666305542,
+      "learning_rate": 5.904291183929533e-06,
+      "logits/chosen": 0.039881881326436996,
+      "logits/rejected": 0.12234417349100113,
+      "logps/chosen": -90.86981201171875,
+      "logps/rejected": -103.53450012207031,
+      "loss": 1.7585,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.512882709503174,
+      "rewards/margins": 0.9952716827392578,
+      "rewards/rejected": -3.50815486907959,
+      "step": 1037
+    },
+    {
+      "epoch": 1.8846503178928247,
+      "grad_norm": 2.7851786613464355,
+      "learning_rate": 5.897890775688787e-06,
+      "logits/chosen": 0.1078023687005043,
+      "logits/rejected": 0.10604678839445114,
+      "logps/chosen": -89.10372924804688,
+      "logps/rejected": -102.15894317626953,
+      "loss": 2.1306,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.9036662578582764,
+      "rewards/margins": 0.8164169192314148,
+      "rewards/rejected": -3.720082998275757,
+      "step": 1038
+    },
+    {
+      "epoch": 1.8864668483197093,
+      "grad_norm": 2.5644519329071045,
+      "learning_rate": 5.891489556079467e-06,
+      "logits/chosen": 0.08762264996767044,
+      "logits/rejected": 0.08112500607967377,
+      "logps/chosen": -87.15840148925781,
+      "logps/rejected": -93.27011108398438,
+      "loss": 1.6835,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.5466814041137695,
+      "rewards/margins": 0.9623738527297974,
+      "rewards/rejected": -3.5090548992156982,
+      "step": 1039
+    },
+    {
+      "epoch": 1.888283378746594,
+      "grad_norm": 2.2946412563323975,
+      "learning_rate": 5.88508753815478e-06,
+      "logits/chosen": -0.015484964475035667,
+      "logits/rejected": 0.06506240367889404,
+      "logps/chosen": -79.59893798828125,
+      "logps/rejected": -95.95955657958984,
+      "loss": 1.7787,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.913203477859497,
+      "rewards/margins": 1.0231151580810547,
+      "rewards/rejected": -3.936318874359131,
+      "step": 1040
+    },
+    {
+      "epoch": 1.8900999091734787,
+      "grad_norm": 2.629913330078125,
+      "learning_rate": 5.878684734969551e-06,
+      "logits/chosen": 0.1164872944355011,
+      "logits/rejected": 0.13854700326919556,
+      "logps/chosen": -76.74923706054688,
+      "logps/rejected": -85.82882690429688,
+      "loss": 2.0775,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.8172569274902344,
+      "rewards/margins": 0.7402248382568359,
+      "rewards/rejected": -3.5574820041656494,
+      "step": 1041
+    },
+    {
+      "epoch": 1.8919164396003634,
+      "grad_norm": 3.180204153060913,
+      "learning_rate": 5.872281159580212e-06,
+      "logits/chosen": 0.054435305297374725,
+      "logits/rejected": 0.05697637051343918,
+      "logps/chosen": -89.08125305175781,
+      "logps/rejected": -96.02574920654297,
+      "loss": 2.4721,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.8339388370513916,
+      "rewards/margins": 0.6004619002342224,
+      "rewards/rejected": -3.434400796890259,
+      "step": 1042
+    },
+    {
+      "epoch": 1.8937329700272478,
+      "grad_norm": 2.789027214050293,
+      "learning_rate": 5.865876825044768e-06,
+      "logits/chosen": 0.05693540349602699,
+      "logits/rejected": 0.0930028110742569,
+      "logps/chosen": -90.22679138183594,
+      "logps/rejected": -103.24329376220703,
+      "loss": 1.7064,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.798175573348999,
+      "rewards/margins": 1.1202514171600342,
+      "rewards/rejected": -3.918426990509033,
+      "step": 1043
+    },
+    {
+      "epoch": 1.8955495004541327,
+      "grad_norm": 2.492196798324585,
+      "learning_rate": 5.859471744422775e-06,
+      "logits/chosen": 0.16648317873477936,
+      "logits/rejected": 0.2084580510854721,
+      "logps/chosen": -77.3297119140625,
+      "logps/rejected": -84.67134094238281,
+      "loss": 2.1084,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.466355562210083,
+      "rewards/margins": 0.6548389196395874,
+      "rewards/rejected": -3.121194362640381,
+      "step": 1044
+    },
+    {
+      "epoch": 1.8973660308810172,
+      "grad_norm": 3.025188684463501,
+      "learning_rate": 5.8530659307753034e-06,
+      "logits/chosen": 0.005150540731847286,
+      "logits/rejected": 0.07447411119937897,
+      "logps/chosen": -81.24078369140625,
+      "logps/rejected": -95.47711944580078,
+      "loss": 1.9188,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.512604236602783,
+      "rewards/margins": 0.9106044769287109,
+      "rewards/rejected": -3.423208713531494,
+      "step": 1045
+    },
+    {
+      "epoch": 1.8991825613079019,
+      "grad_norm": 2.3530304431915283,
+      "learning_rate": 5.8466593971649235e-06,
+      "logits/chosen": 0.04525354132056236,
+      "logits/rejected": 0.1054767295718193,
+      "logps/chosen": -87.58064270019531,
+      "logps/rejected": -106.40087890625,
+      "loss": 1.6973,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5296742916107178,
+      "rewards/margins": 1.0762970447540283,
+      "rewards/rejected": -3.605971097946167,
+      "step": 1046
+    },
+    {
+      "epoch": 1.9009990917347865,
+      "grad_norm": 2.4091858863830566,
+      "learning_rate": 5.840252156655676e-06,
+      "logits/chosen": 0.005571405403316021,
+      "logits/rejected": 0.06598618626594543,
+      "logps/chosen": -83.3829574584961,
+      "logps/rejected": -103.74708557128906,
+      "loss": 1.7217,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.880439281463623,
+      "rewards/margins": 1.0567747354507446,
+      "rewards/rejected": -3.937213897705078,
+      "step": 1047
+    },
+    {
+      "epoch": 1.9028156221616712,
+      "grad_norm": 3.0969154834747314,
+      "learning_rate": 5.833844222313035e-06,
+      "logits/chosen": 0.04075758904218674,
+      "logits/rejected": 0.062300361692905426,
+      "logps/chosen": -82.41780853271484,
+      "logps/rejected": -95.66773986816406,
+      "loss": 1.9223,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.86830472946167,
+      "rewards/margins": 0.9922465085983276,
+      "rewards/rejected": -3.860550880432129,
+      "step": 1048
+    },
+    {
+      "epoch": 1.904632152588556,
+      "grad_norm": 2.517890214920044,
+      "learning_rate": 5.827435607203898e-06,
+      "logits/chosen": -0.0032346658408641815,
+      "logits/rejected": 0.00909996498376131,
+      "logps/chosen": -83.48892211914062,
+      "logps/rejected": -95.3106689453125,
+      "loss": 1.8614,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.942373752593994,
+      "rewards/margins": 0.9874045848846436,
+      "rewards/rejected": -3.929778575897217,
+      "step": 1049
+    },
+    {
+      "epoch": 1.9064486830154403,
+      "grad_norm": 2.7391834259033203,
+      "learning_rate": 5.821026324396546e-06,
+      "logits/chosen": 0.06283523887395859,
+      "logits/rejected": 0.0978909358382225,
+      "logps/chosen": -78.13603210449219,
+      "logps/rejected": -87.04612731933594,
+      "loss": 2.0506,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.8059606552124023,
+      "rewards/margins": 0.8077430129051208,
+      "rewards/rejected": -3.613703489303589,
+      "step": 1050
+    },
+    {
+      "epoch": 1.9082652134423252,
+      "grad_norm": 3.022676944732666,
+      "learning_rate": 5.814616386960623e-06,
+      "logits/chosen": 0.03695928677916527,
+      "logits/rejected": 0.07273729145526886,
+      "logps/chosen": -87.82364654541016,
+      "logps/rejected": -93.47010803222656,
+      "loss": 2.2419,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.8199379444122314,
+      "rewards/margins": 0.6366495490074158,
+      "rewards/rejected": -3.456587553024292,
+      "step": 1051
+    },
+    {
+      "epoch": 1.9100817438692097,
+      "grad_norm": 2.835696220397949,
+      "learning_rate": 5.808205807967106e-06,
+      "logits/chosen": 0.040822338312864304,
+      "logits/rejected": 0.04650012031197548,
+      "logps/chosen": -88.42845916748047,
+      "logps/rejected": -98.560791015625,
+      "loss": 2.0351,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5347342491149902,
+      "rewards/margins": 0.8958336710929871,
+      "rewards/rejected": -3.430567502975464,
+      "step": 1052
+    },
+    {
+      "epoch": 1.9118982742960946,
+      "grad_norm": 2.339982748031616,
+      "learning_rate": 5.801794600488284e-06,
+      "logits/chosen": 0.07362057268619537,
+      "logits/rejected": 0.09707952290773392,
+      "logps/chosen": -89.50122833251953,
+      "logps/rejected": -103.22574615478516,
+      "loss": 1.8518,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.825887680053711,
+      "rewards/margins": 0.8704123497009277,
+      "rewards/rejected": -3.6962995529174805,
+      "step": 1053
+    },
+    {
+      "epoch": 1.913714804722979,
+      "grad_norm": 2.6183598041534424,
+      "learning_rate": 5.795382777597724e-06,
+      "logits/chosen": -0.023022204637527466,
+      "logits/rejected": 0.0642365887761116,
+      "logps/chosen": -87.64765167236328,
+      "logps/rejected": -105.07542419433594,
+      "loss": 1.8466,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.9167988300323486,
+      "rewards/margins": 1.0478229522705078,
+      "rewards/rejected": -3.9646215438842773,
+      "step": 1054
+    },
+    {
+      "epoch": 1.9155313351498637,
+      "grad_norm": 2.5414812564849854,
+      "learning_rate": 5.788970352370248e-06,
+      "logits/chosen": 0.03230145573616028,
+      "logits/rejected": 0.06198396906256676,
+      "logps/chosen": -90.33793640136719,
+      "logps/rejected": -104.18983459472656,
+      "loss": 1.583,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": -2.441845655441284,
+      "rewards/margins": 1.1811773777008057,
+      "rewards/rejected": -3.623023271560669,
+      "step": 1055
+    },
+    {
+      "epoch": 1.9173478655767484,
+      "grad_norm": 2.3225555419921875,
+      "learning_rate": 5.782557337881911e-06,
+      "logits/chosen": 0.035128459334373474,
+      "logits/rejected": 0.015921056270599365,
+      "logps/chosen": -94.31529235839844,
+      "logps/rejected": -98.8448257446289,
+      "loss": 2.325,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.809417247772217,
+      "rewards/margins": 0.6194378733634949,
+      "rewards/rejected": -3.4288549423217773,
+      "step": 1056
+    },
+    {
+      "epoch": 1.919164396003633,
+      "grad_norm": 4.9872002601623535,
+      "learning_rate": 5.776143747209963e-06,
+      "logits/chosen": 0.1352149248123169,
+      "logits/rejected": 0.11303969472646713,
+      "logps/chosen": -79.81632995605469,
+      "logps/rejected": -87.6213150024414,
+      "loss": 1.8354,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.758040189743042,
+      "rewards/margins": 0.8502681255340576,
+      "rewards/rejected": -3.6083083152770996,
+      "step": 1057
+    },
+    {
+      "epoch": 1.9209809264305178,
+      "grad_norm": 2.7759110927581787,
+      "learning_rate": 5.769729593432835e-06,
+      "logits/chosen": -0.007872266694903374,
+      "logits/rejected": 0.03391397371888161,
+      "logps/chosen": -83.1175537109375,
+      "logps/rejected": -95.22735595703125,
+      "loss": 1.908,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.699291706085205,
+      "rewards/margins": 0.8297994136810303,
+      "rewards/rejected": -3.529090642929077,
+      "step": 1058
+    },
+    {
+      "epoch": 1.9227974568574022,
+      "grad_norm": 2.7459685802459717,
+      "learning_rate": 5.763314889630098e-06,
+      "logits/chosen": 0.08921480178833008,
+      "logits/rejected": 0.09074505418539047,
+      "logps/chosen": -86.36589813232422,
+      "logps/rejected": -90.33338165283203,
+      "loss": 2.3566,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.7738795280456543,
+      "rewards/margins": 0.6300579309463501,
+      "rewards/rejected": -3.403937578201294,
+      "step": 1059
+    },
+    {
+      "epoch": 1.9246139872842871,
+      "grad_norm": 2.5738747119903564,
+      "learning_rate": 5.756899648882457e-06,
+      "logits/chosen": 0.07045866549015045,
+      "logits/rejected": 0.05316246673464775,
+      "logps/chosen": -81.53553771972656,
+      "logps/rejected": -96.67655944824219,
+      "loss": 1.9618,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.904839038848877,
+      "rewards/margins": 0.9506840109825134,
+      "rewards/rejected": -3.855523109436035,
+      "step": 1060
+    },
+    {
+      "epoch": 1.9264305177111716,
+      "grad_norm": 3.5665910243988037,
+      "learning_rate": 5.750483884271699e-06,
+      "logits/chosen": 0.003421567380428314,
+      "logits/rejected": 0.02376568131148815,
+      "logps/chosen": -81.46550750732422,
+      "logps/rejected": -94.78770446777344,
+      "loss": 2.2817,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.870290517807007,
+      "rewards/margins": 0.734123170375824,
+      "rewards/rejected": -3.6044135093688965,
+      "step": 1061
+    },
+    {
+      "epoch": 1.9282470481380565,
+      "grad_norm": 2.6111512184143066,
+      "learning_rate": 5.744067608880689e-06,
+      "logits/chosen": 0.06835394352674484,
+      "logits/rejected": 0.07780571281909943,
+      "logps/chosen": -87.38811492919922,
+      "logps/rejected": -96.62849426269531,
+      "loss": 2.0882,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.758291721343994,
+      "rewards/margins": 0.8456116318702698,
+      "rewards/rejected": -3.6039035320281982,
+      "step": 1062
+    },
+    {
+      "epoch": 1.930063578564941,
+      "grad_norm": 2.7628681659698486,
+      "learning_rate": 5.737650835793329e-06,
+      "logits/chosen": 0.049255527555942535,
+      "logits/rejected": 0.11250608414411545,
+      "logps/chosen": -77.61981201171875,
+      "logps/rejected": -95.13066864013672,
+      "loss": 1.8218,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.9048948287963867,
+      "rewards/margins": 1.1056883335113525,
+      "rewards/rejected": -4.01058292388916,
+      "step": 1063
+    },
+    {
+      "epoch": 1.9318801089918256,
+      "grad_norm": 2.6407134532928467,
+      "learning_rate": 5.731233578094534e-06,
+      "logits/chosen": 0.13734818994998932,
+      "logits/rejected": 0.13501118123531342,
+      "logps/chosen": -92.89653015136719,
+      "logps/rejected": -102.6830062866211,
+      "loss": 1.6036,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.748579263687134,
+      "rewards/margins": 0.995154857635498,
+      "rewards/rejected": -3.7437338829040527,
+      "step": 1064
+    },
+    {
+      "epoch": 1.9336966394187103,
+      "grad_norm": 2.519453763961792,
+      "learning_rate": 5.724815848870213e-06,
+      "logits/chosen": 0.1508520096540451,
+      "logits/rejected": 0.10084985196590424,
+      "logps/chosen": -79.31387329101562,
+      "logps/rejected": -87.86670684814453,
+      "loss": 1.7351,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.8119821548461914,
+      "rewards/margins": 1.0087649822235107,
+      "rewards/rejected": -3.820747137069702,
+      "step": 1065
+    },
+    {
+      "epoch": 1.935513169845595,
+      "grad_norm": 2.204310178756714,
+      "learning_rate": 5.718397661207233e-06,
+      "logits/chosen": 0.07638321816921234,
+      "logits/rejected": 0.08917201310396194,
+      "logps/chosen": -94.94502258300781,
+      "logps/rejected": -103.1028823852539,
+      "loss": 2.0475,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5278711318969727,
+      "rewards/margins": 0.8113152384757996,
+      "rewards/rejected": -3.339186429977417,
+      "step": 1066
+    },
+    {
+      "epoch": 1.9373297002724796,
+      "grad_norm": 3.3903145790100098,
+      "learning_rate": 5.7119790281933914e-06,
+      "logits/chosen": 0.09656227380037308,
+      "logits/rejected": 0.12109103053808212,
+      "logps/chosen": -87.74819946289062,
+      "logps/rejected": -100.70112609863281,
+      "loss": 1.9754,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6874074935913086,
+      "rewards/margins": 1.12949800491333,
+      "rewards/rejected": -3.8169054985046387,
+      "step": 1067
+    },
+    {
+      "epoch": 1.939146230699364,
+      "grad_norm": 2.5827860832214355,
+      "learning_rate": 5.705559962917406e-06,
+      "logits/chosen": 0.10529054701328278,
+      "logits/rejected": 0.12139459699392319,
+      "logps/chosen": -84.56909942626953,
+      "logps/rejected": -95.2447280883789,
+      "loss": 1.9098,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.517082452774048,
+      "rewards/margins": 0.7772176861763,
+      "rewards/rejected": -3.2943003177642822,
+      "step": 1068
+    },
+    {
+      "epoch": 1.940962761126249,
+      "grad_norm": 2.6466243267059326,
+      "learning_rate": 5.699140478468864e-06,
+      "logits/chosen": -0.014325696043670177,
+      "logits/rejected": 0.02049952559173107,
+      "logps/chosen": -95.5872802734375,
+      "logps/rejected": -110.2877197265625,
+      "loss": 1.7809,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.483633279800415,
+      "rewards/margins": 0.9650425314903259,
+      "rewards/rejected": -3.4486756324768066,
+      "step": 1069
+    },
+    {
+      "epoch": 1.9427792915531334,
+      "grad_norm": 6.471877574920654,
+      "learning_rate": 5.692720587938212e-06,
+      "logits/chosen": 0.05075512453913689,
+      "logits/rejected": 0.10243435204029083,
+      "logps/chosen": -95.43891143798828,
+      "logps/rejected": -106.59917449951172,
+      "loss": 2.0455,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5541107654571533,
+      "rewards/margins": 0.7671210765838623,
+      "rewards/rejected": -3.3212318420410156,
+      "step": 1070
+    },
+    {
+      "epoch": 1.9445958219800181,
+      "grad_norm": 2.8575029373168945,
+      "learning_rate": 5.686300304416725e-06,
+      "logits/chosen": 0.00602412223815918,
+      "logits/rejected": 0.0314582996070385,
+      "logps/chosen": -86.1510238647461,
+      "logps/rejected": -97.31475830078125,
+      "loss": 1.4992,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.817497491836548,
+      "rewards/margins": 1.1453136205673218,
+      "rewards/rejected": -3.9628114700317383,
+      "step": 1071
+    },
+    {
+      "epoch": 1.9464123524069028,
+      "grad_norm": 2.725543737411499,
+      "learning_rate": 5.67987964099648e-06,
+      "logits/chosen": 0.04529724270105362,
+      "logits/rejected": 0.06778516620397568,
+      "logps/chosen": -92.77943420410156,
+      "logps/rejected": -104.79008483886719,
+      "loss": 2.0695,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.81473970413208,
+      "rewards/margins": 0.7794060111045837,
+      "rewards/rejected": -3.5941455364227295,
+      "step": 1072
+    },
+    {
+      "epoch": 1.9482288828337875,
+      "grad_norm": 3.011185884475708,
+      "learning_rate": 5.673458610770327e-06,
+      "logits/chosen": -0.03981207311153412,
+      "logits/rejected": 0.045006051659584045,
+      "logps/chosen": -86.32022857666016,
+      "logps/rejected": -105.84201049804688,
+      "loss": 1.7233,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.8359944820404053,
+      "rewards/margins": 1.0611416101455688,
+      "rewards/rejected": -3.8971357345581055,
+      "step": 1073
+    },
+    {
+      "epoch": 1.9500454132606722,
+      "grad_norm": 3.054245710372925,
+      "learning_rate": 5.667037226831865e-06,
+      "logits/chosen": 0.028957592323422432,
+      "logits/rejected": 0.09366678446531296,
+      "logps/chosen": -86.51834869384766,
+      "logps/rejected": -107.35171508789062,
+      "loss": 2.0319,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -3.000333070755005,
+      "rewards/margins": 1.069792628288269,
+      "rewards/rejected": -4.070126056671143,
+      "step": 1074
+    },
+    {
+      "epoch": 1.9518619436875566,
+      "grad_norm": 2.7530715465545654,
+      "learning_rate": 5.660615502275413e-06,
+      "logits/chosen": 0.004540876019746065,
+      "logits/rejected": 0.049830369651317596,
+      "logps/chosen": -90.24873352050781,
+      "logps/rejected": -104.93528747558594,
+      "loss": 1.7892,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.7523155212402344,
+      "rewards/margins": 1.0245788097381592,
+      "rewards/rejected": -3.7768945693969727,
+      "step": 1075
+    },
+    {
+      "epoch": 1.9536784741144415,
+      "grad_norm": 2.7152786254882812,
+      "learning_rate": 5.654193450195986e-06,
+      "logits/chosen": 0.1072491779923439,
+      "logits/rejected": 0.07327581942081451,
+      "logps/chosen": -90.66412353515625,
+      "logps/rejected": -88.46973419189453,
+      "loss": 2.3361,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.8444371223449707,
+      "rewards/margins": 0.5099895596504211,
+      "rewards/rejected": -3.354426622390747,
+      "step": 1076
+    },
+    {
+      "epoch": 1.955495004541326,
+      "grad_norm": 3.480952501296997,
+      "learning_rate": 5.647771083689266e-06,
+      "logits/chosen": 0.04179378226399422,
+      "logits/rejected": 0.1350129097700119,
+      "logps/chosen": -78.02560424804688,
+      "logps/rejected": -93.27855682373047,
+      "loss": 2.1204,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -3.1416115760803223,
+      "rewards/margins": 0.7916675806045532,
+      "rewards/rejected": -3.933278799057007,
+      "step": 1077
+    },
+    {
+      "epoch": 1.9573115349682109,
+      "grad_norm": 3.130934953689575,
+      "learning_rate": 5.641348415851578e-06,
+      "logits/chosen": 0.054842621088027954,
+      "logits/rejected": 0.08140455186367035,
+      "logps/chosen": -87.1384048461914,
+      "logps/rejected": -92.8744888305664,
+      "loss": 2.3019,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -3.00537109375,
+      "rewards/margins": 0.7598800659179688,
+      "rewards/rejected": -3.765251398086548,
+      "step": 1078
+    },
+    {
+      "epoch": 1.9591280653950953,
+      "grad_norm": 3.1723556518554688,
+      "learning_rate": 5.634925459779858e-06,
+      "logits/chosen": 0.05505143851041794,
+      "logits/rejected": 0.05938095226883888,
+      "logps/chosen": -97.43035125732422,
+      "logps/rejected": -111.23731994628906,
+      "loss": 1.8076,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.8198184967041016,
+      "rewards/margins": 1.150580883026123,
+      "rewards/rejected": -3.9703996181488037,
+      "step": 1079
+    },
+    {
+      "epoch": 1.96094459582198,
+      "grad_norm": 2.666764497756958,
+      "learning_rate": 5.6285022285716325e-06,
+      "logits/chosen": 0.08428171277046204,
+      "logits/rejected": 0.05373971536755562,
+      "logps/chosen": -86.85575866699219,
+      "logps/rejected": -94.35645294189453,
+      "loss": 1.6456,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.844283103942871,
+      "rewards/margins": 1.0076135396957397,
+      "rewards/rejected": -3.8518965244293213,
+      "step": 1080
+    },
+    {
+      "epoch": 1.9627611262488647,
+      "grad_norm": 2.6962080001831055,
+      "learning_rate": 5.622078735324991e-06,
+      "logits/chosen": 0.11438596993684769,
+      "logits/rejected": 0.09278252720832825,
+      "logps/chosen": -87.12863159179688,
+      "logps/rejected": -92.41490173339844,
+      "loss": 2.4464,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -3.2357699871063232,
+      "rewards/margins": 0.6380851864814758,
+      "rewards/rejected": -3.8738555908203125,
+      "step": 1081
+    },
+    {
+      "epoch": 1.9645776566757494,
+      "grad_norm": 3.0744364261627197,
+      "learning_rate": 5.615654993138551e-06,
+      "logits/chosen": 0.07801391184329987,
+      "logits/rejected": 0.08932562172412872,
+      "logps/chosen": -79.82630920410156,
+      "logps/rejected": -91.24948120117188,
+      "loss": 2.2637,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.83866548538208,
+      "rewards/margins": 0.880230188369751,
+      "rewards/rejected": -3.718895673751831,
+      "step": 1082
+    },
+    {
+      "epoch": 1.966394187102634,
+      "grad_norm": 2.861772298812866,
+      "learning_rate": 5.6092310151114416e-06,
+      "logits/chosen": 0.07315719872713089,
+      "logits/rejected": 0.10782472044229507,
+      "logps/chosen": -95.10383605957031,
+      "logps/rejected": -109.89063262939453,
+      "loss": 1.9737,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.9198198318481445,
+      "rewards/margins": 1.027293086051941,
+      "rewards/rejected": -3.947113037109375,
+      "step": 1083
+    },
+    {
+      "epoch": 1.9682107175295185,
+      "grad_norm": 2.773005485534668,
+      "learning_rate": 5.602806814343275e-06,
+      "logits/chosen": 0.1266084760427475,
+      "logits/rejected": 0.1481492519378662,
+      "logps/chosen": -88.75833129882812,
+      "logps/rejected": -97.42185974121094,
+      "loss": 1.7235,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.7945189476013184,
+      "rewards/margins": 0.9863830208778381,
+      "rewards/rejected": -3.780902147293091,
+      "step": 1084
+    },
+    {
+      "epoch": 1.9700272479564034,
+      "grad_norm": 2.5498030185699463,
+      "learning_rate": 5.596382403934112e-06,
+      "logits/chosen": 0.03134167939424515,
+      "logits/rejected": 0.0482235848903656,
+      "logps/chosen": -89.30046844482422,
+      "logps/rejected": -104.6332015991211,
+      "loss": 1.7914,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.870337724685669,
+      "rewards/margins": 1.177640438079834,
+      "rewards/rejected": -4.047978401184082,
+      "step": 1085
+    },
+    {
+      "epoch": 1.9718437783832878,
+      "grad_norm": 2.9151346683502197,
+      "learning_rate": 5.589957796984445e-06,
+      "logits/chosen": 0.10600137710571289,
+      "logits/rejected": 0.027712268754839897,
+      "logps/chosen": -95.32709503173828,
+      "logps/rejected": -100.17862701416016,
+      "loss": 1.7396,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.8309273719787598,
+      "rewards/margins": 0.9780685901641846,
+      "rewards/rejected": -3.8089957237243652,
+      "step": 1086
+    },
+    {
+      "epoch": 1.9736603088101727,
+      "grad_norm": 2.6284403800964355,
+      "learning_rate": 5.583533006595169e-06,
+      "logits/chosen": 0.028934508562088013,
+      "logits/rejected": 0.14620086550712585,
+      "logps/chosen": -82.0864486694336,
+      "logps/rejected": -103.23451232910156,
+      "loss": 1.8953,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -3.100543975830078,
+      "rewards/margins": 1.002597689628601,
+      "rewards/rejected": -4.103141784667969,
+      "step": 1087
+    },
+    {
+      "epoch": 1.9754768392370572,
+      "grad_norm": 3.4660964012145996,
+      "learning_rate": 5.577108045867548e-06,
+      "logits/chosen": 0.014269035309553146,
+      "logits/rejected": 0.01597430743277073,
+      "logps/chosen": -80.68024444580078,
+      "logps/rejected": -92.7439956665039,
+      "loss": 1.7933,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.932746410369873,
+      "rewards/margins": 0.8568322062492371,
+      "rewards/rejected": -3.789578914642334,
+      "step": 1088
+    },
+    {
+      "epoch": 1.9772933696639419,
+      "grad_norm": 2.9386420249938965,
+      "learning_rate": 5.570682927903194e-06,
+      "logits/chosen": 0.17257757484912872,
+      "logits/rejected": 0.13612157106399536,
+      "logps/chosen": -81.04582977294922,
+      "logps/rejected": -84.23763275146484,
+      "loss": 2.1234,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.903170585632324,
+      "rewards/margins": 0.7135775089263916,
+      "rewards/rejected": -3.616748094558716,
+      "step": 1089
+    },
+    {
+      "epoch": 1.9791099000908265,
+      "grad_norm": 2.753761053085327,
+      "learning_rate": 5.564257665804044e-06,
+      "logits/chosen": 0.01994311809539795,
+      "logits/rejected": 0.058431755751371384,
+      "logps/chosen": -85.41407775878906,
+      "logps/rejected": -101.63247680664062,
+      "loss": 1.6846,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -3.0370702743530273,
+      "rewards/margins": 1.2280018329620361,
+      "rewards/rejected": -4.265071868896484,
+      "step": 1090
+    },
+    {
+      "epoch": 1.9809264305177112,
+      "grad_norm": 2.6164751052856445,
+      "learning_rate": 5.557832272672325e-06,
+      "logits/chosen": 0.01375737227499485,
+      "logits/rejected": 0.054067350924015045,
+      "logps/chosen": -76.8437271118164,
+      "logps/rejected": -87.47756958007812,
+      "loss": 1.8208,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4862310886383057,
+      "rewards/margins": 1.013425588607788,
+      "rewards/rejected": -3.4996564388275146,
+      "step": 1091
+    },
+    {
+      "epoch": 1.982742960944596,
+      "grad_norm": 2.4975688457489014,
+      "learning_rate": 5.551406761610531e-06,
+      "logits/chosen": 0.020210791379213333,
+      "logits/rejected": 0.044297587126493454,
+      "logps/chosen": -81.97608184814453,
+      "logps/rejected": -101.89766693115234,
+      "loss": 1.8884,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.924888849258423,
+      "rewards/margins": 1.1990878582000732,
+      "rewards/rejected": -4.123976230621338,
+      "step": 1092
+    },
+    {
+      "epoch": 1.9845594913714804,
+      "grad_norm": 2.7430906295776367,
+      "learning_rate": 5.544981145721401e-06,
+      "logits/chosen": 0.016364507377147675,
+      "logits/rejected": 0.014219507575035095,
+      "logps/chosen": -84.45879364013672,
+      "logps/rejected": -95.0400619506836,
+      "loss": 1.7552,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.800570249557495,
+      "rewards/margins": 1.0338683128356934,
+      "rewards/rejected": -3.8344383239746094,
+      "step": 1093
+    },
+    {
+      "epoch": 1.9863760217983653,
+      "grad_norm": 2.5156731605529785,
+      "learning_rate": 5.538555438107883e-06,
+      "logits/chosen": 0.12248750030994415,
+      "logits/rejected": 0.19048044085502625,
+      "logps/chosen": -86.73338317871094,
+      "logps/rejected": -100.41438293457031,
+      "loss": 1.7055,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.8860414028167725,
+      "rewards/margins": 1.0440818071365356,
+      "rewards/rejected": -3.9301233291625977,
+      "step": 1094
+    },
+    {
+      "epoch": 1.9881925522252497,
+      "grad_norm": 3.9679181575775146,
+      "learning_rate": 5.532129651873112e-06,
+      "logits/chosen": 0.07864493131637573,
+      "logits/rejected": 0.11612318456172943,
+      "logps/chosen": -82.85931396484375,
+      "logps/rejected": -99.84012603759766,
+      "loss": 1.8296,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -3.08736252784729,
+      "rewards/margins": 1.1726363897323608,
+      "rewards/rejected": -4.259998798370361,
+      "step": 1095
+    },
+    {
+      "epoch": 1.9900090826521344,
+      "grad_norm": 2.9341135025024414,
+      "learning_rate": 5.5257038001203885e-06,
+      "logits/chosen": 0.05851954594254494,
+      "logits/rejected": 0.0659928172826767,
+      "logps/chosen": -87.36480712890625,
+      "logps/rejected": -98.98983001708984,
+      "loss": 1.9843,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.759835958480835,
+      "rewards/margins": 0.979812502861023,
+      "rewards/rejected": -3.7396485805511475,
+      "step": 1096
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.153270461680845e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-1096/training_args.bin b/checkpoint-1096/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-1096/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-1096/zero_to_fp32.py b/checkpoint-1096/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-1096/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-137/README.md b/checkpoint-137/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-137/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-137/adapter_config.json b/checkpoint-137/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-137/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-137/adapter_model.safetensors b/checkpoint-137/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..773905a95d1ffb635544bb85b02798833973306b
--- /dev/null
+++ b/checkpoint-137/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08f983e5fac190502f6e69c40823f52ccc319a197e1d7ea52e0dd25b734bd60f
+size 207244392
diff --git a/checkpoint-137/latest b/checkpoint-137/latest
new file mode 100644
index 0000000000000000000000000000000000000000..13fc4bbf20aed3a527939f85d0e7e526c3503d13
--- /dev/null
+++ b/checkpoint-137/latest
@@ -0,0 +1 @@
+global_step137
\ No newline at end of file
diff --git a/checkpoint-137/rng_state_0.pth b/checkpoint-137/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b6473612e41c5cfd6973c2e71fa5f3ad2b2bcad1
--- /dev/null
+++ b/checkpoint-137/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984
diff --git a/checkpoint-137/rng_state_1.pth b/checkpoint-137/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..8506e00431b6ac7067699c0ea4f59adb6fa0ba20
--- /dev/null
+++ b/checkpoint-137/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984
diff --git a/checkpoint-137/rng_state_2.pth b/checkpoint-137/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..ea499e285c97cca07fedd34662c3d4ab44ff6f47
--- /dev/null
+++ b/checkpoint-137/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984
diff --git a/checkpoint-137/rng_state_3.pth b/checkpoint-137/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..aeb38f92f106ac3f08bae4f82179a8a12243bccb
--- /dev/null
+++ b/checkpoint-137/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984
diff --git a/checkpoint-137/rng_state_4.pth b/checkpoint-137/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..9d5856cb7a3f15092fa5593507022316916f648e
--- /dev/null
+++ b/checkpoint-137/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984
diff --git a/checkpoint-137/rng_state_5.pth b/checkpoint-137/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b824ee24d256695aad4a69a62d8e7125f51a17f2
--- /dev/null
+++ b/checkpoint-137/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984
diff --git a/checkpoint-137/rng_state_6.pth b/checkpoint-137/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a9fd0364bb8f1a8e91eca45be5e1b6672b4d9afd
--- /dev/null
+++ b/checkpoint-137/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984
diff --git a/checkpoint-137/rng_state_7.pth b/checkpoint-137/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..4e80125fd18efcb1097384319888b699f4dce7e7
--- /dev/null
+++ b/checkpoint-137/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984
diff --git a/checkpoint-137/special_tokens_map.json b/checkpoint-137/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-137/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-137/tokenizer.json b/checkpoint-137/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-137/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-137/tokenizer_config.json b/checkpoint-137/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-137/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-137/trainer_state.json b/checkpoint-137/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4311f8055b9c839e162c509a128b53eeabe6b72b
--- /dev/null
+++ b/checkpoint-137/trainer_state.json
@@ -0,0 +1,2088 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.2488646684831971,
+  "eval_steps": 500,
+  "global_step": 137,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 6.440843988697088e+16,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-137/training_args.bin b/checkpoint-137/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-137/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-137/zero_to_fp32.py b/checkpoint-137/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-137/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-274/README.md b/checkpoint-274/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-274/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-274/adapter_config.json b/checkpoint-274/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-274/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-274/adapter_model.safetensors b/checkpoint-274/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..da1060d02e20328d7bc8e1cb287898f19b57e818
--- /dev/null
+++ b/checkpoint-274/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2b7b19a3b1b92b8e7b15bb559e98ca1a17da196d7e4b133c152b40f86424376
+size 207244392
diff --git a/checkpoint-274/latest b/checkpoint-274/latest
new file mode 100644
index 0000000000000000000000000000000000000000..f21caafe112d9847a078b929c7cc2570f86f134d
--- /dev/null
+++ b/checkpoint-274/latest
@@ -0,0 +1 @@
+global_step274
\ No newline at end of file
diff --git a/checkpoint-274/rng_state_0.pth b/checkpoint-274/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b6473612e41c5cfd6973c2e71fa5f3ad2b2bcad1
--- /dev/null
+++ b/checkpoint-274/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984
diff --git a/checkpoint-274/rng_state_1.pth b/checkpoint-274/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..8506e00431b6ac7067699c0ea4f59adb6fa0ba20
--- /dev/null
+++ b/checkpoint-274/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984
diff --git a/checkpoint-274/rng_state_2.pth b/checkpoint-274/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..ea499e285c97cca07fedd34662c3d4ab44ff6f47
--- /dev/null
+++ b/checkpoint-274/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984
diff --git a/checkpoint-274/rng_state_3.pth b/checkpoint-274/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..aeb38f92f106ac3f08bae4f82179a8a12243bccb
--- /dev/null
+++ b/checkpoint-274/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984
diff --git a/checkpoint-274/rng_state_4.pth b/checkpoint-274/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..9d5856cb7a3f15092fa5593507022316916f648e
--- /dev/null
+++ b/checkpoint-274/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984
diff --git a/checkpoint-274/rng_state_5.pth b/checkpoint-274/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b824ee24d256695aad4a69a62d8e7125f51a17f2
--- /dev/null
+++ b/checkpoint-274/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984
diff --git a/checkpoint-274/rng_state_6.pth b/checkpoint-274/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a9fd0364bb8f1a8e91eca45be5e1b6672b4d9afd
--- /dev/null
+++ b/checkpoint-274/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984
diff --git a/checkpoint-274/rng_state_7.pth b/checkpoint-274/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..4e80125fd18efcb1097384319888b699f4dce7e7
--- /dev/null
+++ b/checkpoint-274/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984
diff --git a/checkpoint-274/special_tokens_map.json b/checkpoint-274/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-274/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-274/tokenizer.json b/checkpoint-274/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-274/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-274/tokenizer_config.json b/checkpoint-274/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-274/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-274/trainer_state.json b/checkpoint-274/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..423a100bab5e201ab5c0454856c0cac872bee929
--- /dev/null
+++ b/checkpoint-274/trainer_state.json
@@ -0,0 +1,4143 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.4977293369663942,
+  "eval_steps": 500,
+  "global_step": 274,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.299919692365824e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-274/training_args.bin b/checkpoint-274/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-274/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-274/zero_to_fp32.py b/checkpoint-274/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-274/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-411/README.md b/checkpoint-411/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-411/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-411/adapter_config.json b/checkpoint-411/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-411/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-411/adapter_model.safetensors b/checkpoint-411/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..39fe096d777dcc0ee0ca7728240c83b3259a9fd2
--- /dev/null
+++ b/checkpoint-411/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be1b84f1ff3fbc70d0f97b39e95d136fbf2193aa564349685d47caf345724722
+size 207244392
diff --git a/checkpoint-411/latest b/checkpoint-411/latest
new file mode 100644
index 0000000000000000000000000000000000000000..a1ea2f1ce0ba500dd708a484f2be3288502c1558
--- /dev/null
+++ b/checkpoint-411/latest
@@ -0,0 +1 @@
+global_step411
\ No newline at end of file
diff --git a/checkpoint-411/rng_state_0.pth b/checkpoint-411/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b6473612e41c5cfd6973c2e71fa5f3ad2b2bcad1
--- /dev/null
+++ b/checkpoint-411/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984
diff --git a/checkpoint-411/rng_state_1.pth b/checkpoint-411/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..8506e00431b6ac7067699c0ea4f59adb6fa0ba20
--- /dev/null
+++ b/checkpoint-411/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984
diff --git a/checkpoint-411/rng_state_2.pth b/checkpoint-411/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..ea499e285c97cca07fedd34662c3d4ab44ff6f47
--- /dev/null
+++ b/checkpoint-411/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984
diff --git a/checkpoint-411/rng_state_3.pth b/checkpoint-411/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..aeb38f92f106ac3f08bae4f82179a8a12243bccb
--- /dev/null
+++ b/checkpoint-411/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984
diff --git a/checkpoint-411/rng_state_4.pth b/checkpoint-411/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..9d5856cb7a3f15092fa5593507022316916f648e
--- /dev/null
+++ b/checkpoint-411/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984
diff --git a/checkpoint-411/rng_state_5.pth b/checkpoint-411/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b824ee24d256695aad4a69a62d8e7125f51a17f2
--- /dev/null
+++ b/checkpoint-411/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984
diff --git a/checkpoint-411/rng_state_6.pth b/checkpoint-411/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a9fd0364bb8f1a8e91eca45be5e1b6672b4d9afd
--- /dev/null
+++ b/checkpoint-411/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984
diff --git a/checkpoint-411/rng_state_7.pth b/checkpoint-411/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..4e80125fd18efcb1097384319888b699f4dce7e7
--- /dev/null
+++ b/checkpoint-411/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984
diff --git a/checkpoint-411/special_tokens_map.json b/checkpoint-411/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-411/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-411/tokenizer.json b/checkpoint-411/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-411/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-411/tokenizer_config.json b/checkpoint-411/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-411/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-411/trainer_state.json b/checkpoint-411/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5584c7a8e6ae3a4cef9e845ef4ecc136251209c5
--- /dev/null
+++ b/checkpoint-411/trainer_state.json
@@ -0,0 +1,6198 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.7465940054495913,
+  "eval_steps": 500,
+  "global_step": 411,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    },
+    {
+      "epoch": 0.49954586739327883,
+      "grad_norm": 2.1769776344299316,
+      "learning_rate": 9.65745789630079e-06,
+      "logits/chosen": 0.11112834513187408,
+      "logits/rejected": 0.10816515237092972,
+      "logps/chosen": -83.5405044555664,
+      "logps/rejected": -83.06329345703125,
+      "loss": 3.0939,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.142024040222168,
+      "rewards/margins": 0.1820124089717865,
+      "rewards/rejected": -2.3240363597869873,
+      "step": 275
+    },
+    {
+      "epoch": 0.5013623978201635,
+      "grad_norm": 1.7329221963882446,
+      "learning_rate": 9.654994540260396e-06,
+      "logits/chosen": 0.0653618574142456,
+      "logits/rejected": 0.08004368096590042,
+      "logps/chosen": -80.83209228515625,
+      "logps/rejected": -82.97142028808594,
+      "loss": 2.7161,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.851841688156128,
+      "rewards/margins": 0.2948168218135834,
+      "rewards/rejected": -2.146658420562744,
+      "step": 276
+    },
+    {
+      "epoch": 0.5031789282470481,
+      "grad_norm": 1.592657208442688,
+      "learning_rate": 9.65252271146268e-06,
+      "logits/chosen": 0.09880789369344711,
+      "logits/rejected": 0.14229761064052582,
+      "logps/chosen": -67.1727294921875,
+      "logps/rejected": -75.03417205810547,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8714643716812134,
+      "rewards/margins": 0.37937116622924805,
+      "rewards/rejected": -2.250835657119751,
+      "step": 277
+    },
+    {
+      "epoch": 0.5049954586739328,
+      "grad_norm": 1.6487712860107422,
+      "learning_rate": 9.650042414948133e-06,
+      "logits/chosen": 0.13465353846549988,
+      "logits/rejected": 0.12865689396858215,
+      "logps/chosen": -76.4417724609375,
+      "logps/rejected": -78.6947021484375,
+      "loss": 2.6085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.875291347503662,
+      "rewards/margins": 0.35175251960754395,
+      "rewards/rejected": -2.227043867111206,
+      "step": 278
+    },
+    {
+      "epoch": 0.5068119891008175,
+      "grad_norm": 1.6523009538650513,
+      "learning_rate": 9.64755365577451e-06,
+      "logits/chosen": 0.04238567873835564,
+      "logits/rejected": 0.07994347810745239,
+      "logps/chosen": -76.92097473144531,
+      "logps/rejected": -83.20886993408203,
+      "loss": 2.3156,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9759610891342163,
+      "rewards/margins": 0.5398927330970764,
+      "rewards/rejected": -2.5158536434173584,
+      "step": 279
+    },
+    {
+      "epoch": 0.508628519527702,
+      "grad_norm": 1.7999261617660522,
+      "learning_rate": 9.645056439016827e-06,
+      "logits/chosen": 0.07349395751953125,
+      "logits/rejected": 0.07518415153026581,
+      "logps/chosen": -77.56079864501953,
+      "logps/rejected": -84.97645568847656,
+      "loss": 2.8085,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9408115148544312,
+      "rewards/margins": 0.3470652997493744,
+      "rewards/rejected": -2.287877082824707,
+      "step": 280
+    },
+    {
+      "epoch": 0.5104450499545867,
+      "grad_norm": 1.5684200525283813,
+      "learning_rate": 9.642550769767342e-06,
+      "logits/chosen": 0.16188879311084747,
+      "logits/rejected": 0.12772323191165924,
+      "logps/chosen": -89.29315185546875,
+      "logps/rejected": -94.35065460205078,
+      "loss": 2.2314,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8958841562271118,
+      "rewards/margins": 0.5102108716964722,
+      "rewards/rejected": -2.406095027923584,
+      "step": 281
+    },
+    {
+      "epoch": 0.5122615803814714,
+      "grad_norm": 1.7878178358078003,
+      "learning_rate": 9.640036653135548e-06,
+      "logits/chosen": 0.060573749244213104,
+      "logits/rejected": 0.13457715511322021,
+      "logps/chosen": -68.9404525756836,
+      "logps/rejected": -74.77693176269531,
+      "loss": 2.4359,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.889530897140503,
+      "rewards/margins": 0.45911547541618347,
+      "rewards/rejected": -2.348646402359009,
+      "step": 282
+    },
+    {
+      "epoch": 0.5140781108083561,
+      "grad_norm": 1.4741288423538208,
+      "learning_rate": 9.637514094248172e-06,
+      "logits/chosen": 0.10433132946491241,
+      "logits/rejected": 0.1439315229654312,
+      "logps/chosen": -74.89447021484375,
+      "logps/rejected": -86.58551788330078,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9090807437896729,
+      "rewards/margins": 0.6733560562133789,
+      "rewards/rejected": -2.5824368000030518,
+      "step": 283
+    },
+    {
+      "epoch": 0.5158946412352406,
+      "grad_norm": 1.2864018678665161,
+      "learning_rate": 9.634983098249146e-06,
+      "logits/chosen": 0.10891089588403702,
+      "logits/rejected": 0.11755162477493286,
+      "logps/chosen": -66.90985107421875,
+      "logps/rejected": -71.8125228881836,
+      "loss": 2.1538,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.7779250144958496,
+      "rewards/margins": 0.5625147223472595,
+      "rewards/rejected": -2.340439796447754,
+      "step": 284
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 1.465747594833374,
+      "learning_rate": 9.632443670299616e-06,
+      "logits/chosen": 0.08224496245384216,
+      "logits/rejected": 0.12130744755268097,
+      "logps/chosen": -75.4281997680664,
+      "logps/rejected": -85.0781021118164,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7217226028442383,
+      "rewards/margins": 0.4750482439994812,
+      "rewards/rejected": -2.196770668029785,
+      "step": 285
+    },
+    {
+      "epoch": 0.51952770208901,
+      "grad_norm": 1.492859959602356,
+      "learning_rate": 9.629895815577915e-06,
+      "logits/chosen": 0.06619664281606674,
+      "logits/rejected": 0.13152630627155304,
+      "logps/chosen": -96.65383911132812,
+      "logps/rejected": -111.93521881103516,
+      "loss": 2.2831,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7611618041992188,
+      "rewards/margins": 0.5886337757110596,
+      "rewards/rejected": -2.349795341491699,
+      "step": 286
+    },
+    {
+      "epoch": 0.5213442325158947,
+      "grad_norm": 1.5534065961837769,
+      "learning_rate": 9.627339539279564e-06,
+      "logits/chosen": 0.06637927144765854,
+      "logits/rejected": 0.09107412397861481,
+      "logps/chosen": -71.92534637451172,
+      "logps/rejected": -82.98391723632812,
+      "loss": 2.5101,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7867075204849243,
+      "rewards/margins": 0.39493298530578613,
+      "rewards/rejected": -2.181640625,
+      "step": 287
+    },
+    {
+      "epoch": 0.5231607629427792,
+      "grad_norm": 1.6778221130371094,
+      "learning_rate": 9.624774846617254e-06,
+      "logits/chosen": 0.14700329303741455,
+      "logits/rejected": 0.12778782844543457,
+      "logps/chosen": -65.2364730834961,
+      "logps/rejected": -74.39017486572266,
+      "loss": 2.5236,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6942007541656494,
+      "rewards/margins": 0.35334131121635437,
+      "rewards/rejected": -2.047542095184326,
+      "step": 288
+    },
+    {
+      "epoch": 0.5249772933696639,
+      "grad_norm": 1.462215542793274,
+      "learning_rate": 9.622201742820839e-06,
+      "logits/chosen": 0.11467991769313812,
+      "logits/rejected": 0.1028795838356018,
+      "logps/chosen": -67.85939025878906,
+      "logps/rejected": -74.3462142944336,
+      "loss": 2.3891,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7194863557815552,
+      "rewards/margins": 0.3846975564956665,
+      "rewards/rejected": -2.1041836738586426,
+      "step": 289
+    },
+    {
+      "epoch": 0.5267938237965486,
+      "grad_norm": 1.7591735124588013,
+      "learning_rate": 9.619620233137328e-06,
+      "logits/chosen": 0.08407986164093018,
+      "logits/rejected": 0.12201236933469772,
+      "logps/chosen": -82.57340240478516,
+      "logps/rejected": -86.94192504882812,
+      "loss": 2.4569,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8536537885665894,
+      "rewards/margins": 0.38806283473968506,
+      "rewards/rejected": -2.2417166233062744,
+      "step": 290
+    },
+    {
+      "epoch": 0.5286103542234333,
+      "grad_norm": 1.5584774017333984,
+      "learning_rate": 9.617030322830868e-06,
+      "logits/chosen": 0.1126309484243393,
+      "logits/rejected": 0.08907752484083176,
+      "logps/chosen": -73.36901092529297,
+      "logps/rejected": -76.65506744384766,
+      "loss": 2.4021,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8091729879379272,
+      "rewards/margins": 0.3603411316871643,
+      "rewards/rejected": -2.1695141792297363,
+      "step": 291
+    },
+    {
+      "epoch": 0.5304268846503178,
+      "grad_norm": 1.487938642501831,
+      "learning_rate": 9.614432017182736e-06,
+      "logits/chosen": 0.09620118141174316,
+      "logits/rejected": 0.08514149487018585,
+      "logps/chosen": -88.66265106201172,
+      "logps/rejected": -92.26091766357422,
+      "loss": 2.3636,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.5980674028396606,
+      "rewards/margins": 0.4888036549091339,
+      "rewards/rejected": -2.0868711471557617,
+      "step": 292
+    },
+    {
+      "epoch": 0.5322434150772025,
+      "grad_norm": 1.445408582687378,
+      "learning_rate": 9.611825321491331e-06,
+      "logits/chosen": 0.10014252364635468,
+      "logits/rejected": 0.12857215106487274,
+      "logps/chosen": -82.46649932861328,
+      "logps/rejected": -92.84205627441406,
+      "loss": 2.343,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4620139598846436,
+      "rewards/margins": 0.5616316199302673,
+      "rewards/rejected": -2.0236456394195557,
+      "step": 293
+    },
+    {
+      "epoch": 0.5340599455040872,
+      "grad_norm": 1.262193202972412,
+      "learning_rate": 9.609210241072158e-06,
+      "logits/chosen": 0.09868282079696655,
+      "logits/rejected": 0.16074486076831818,
+      "logps/chosen": -67.0407485961914,
+      "logps/rejected": -82.36201477050781,
+      "loss": 2.1163,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5362765789031982,
+      "rewards/margins": 0.7091963887214661,
+      "rewards/rejected": -2.2454731464385986,
+      "step": 294
+    },
+    {
+      "epoch": 0.5358764759309719,
+      "grad_norm": 1.601891279220581,
+      "learning_rate": 9.606586781257822e-06,
+      "logits/chosen": 0.1236480325460434,
+      "logits/rejected": 0.12187935411930084,
+      "logps/chosen": -78.2726058959961,
+      "logps/rejected": -84.49217987060547,
+      "loss": 2.5075,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.779215931892395,
+      "rewards/margins": 0.38987410068511963,
+      "rewards/rejected": -2.1690900325775146,
+      "step": 295
+    },
+    {
+      "epoch": 0.5376930063578564,
+      "grad_norm": 1.4211548566818237,
+      "learning_rate": 9.603954947398016e-06,
+      "logits/chosen": 0.18375667929649353,
+      "logits/rejected": 0.21945703029632568,
+      "logps/chosen": -70.29829406738281,
+      "logps/rejected": -78.42284393310547,
+      "loss": 2.1959,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.7448753118515015,
+      "rewards/margins": 0.5106840133666992,
+      "rewards/rejected": -2.2555594444274902,
+      "step": 296
+    },
+    {
+      "epoch": 0.5395095367847411,
+      "grad_norm": 1.458617925643921,
+      "learning_rate": 9.601314744859504e-06,
+      "logits/chosen": 0.07360847294330597,
+      "logits/rejected": 0.1596718430519104,
+      "logps/chosen": -72.28546142578125,
+      "logps/rejected": -87.13935089111328,
+      "loss": 2.3538,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8007893562316895,
+      "rewards/margins": 0.5433144569396973,
+      "rewards/rejected": -2.3441038131713867,
+      "step": 297
+    },
+    {
+      "epoch": 0.5413260672116258,
+      "grad_norm": 1.6526339054107666,
+      "learning_rate": 9.598666179026123e-06,
+      "logits/chosen": 0.12013350427150726,
+      "logits/rejected": 0.10237178206443787,
+      "logps/chosen": -83.06564331054688,
+      "logps/rejected": -85.60771942138672,
+      "loss": 2.4531,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.911988615989685,
+      "rewards/margins": 0.35885563492774963,
+      "rewards/rejected": -2.2708444595336914,
+      "step": 298
+    },
+    {
+      "epoch": 0.5431425976385105,
+      "grad_norm": 1.9808340072631836,
+      "learning_rate": 9.596009255298755e-06,
+      "logits/chosen": 0.062342576682567596,
+      "logits/rejected": 0.09053834527730942,
+      "logps/chosen": -96.33489990234375,
+      "logps/rejected": -93.43024444580078,
+      "loss": 2.8549,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.845227599143982,
+      "rewards/margins": 0.22291362285614014,
+      "rewards/rejected": -2.068141222000122,
+      "step": 299
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 1.6276236772537231,
+      "learning_rate": 9.593343979095334e-06,
+      "logits/chosen": 0.2073422521352768,
+      "logits/rejected": 0.13107003271579742,
+      "logps/chosen": -72.45758056640625,
+      "logps/rejected": -77.42770385742188,
+      "loss": 2.4644,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9407453536987305,
+      "rewards/margins": 0.3134301006793976,
+      "rewards/rejected": -2.2541751861572266,
+      "step": 300
+    },
+    {
+      "epoch": 0.5467756584922797,
+      "grad_norm": 1.5001753568649292,
+      "learning_rate": 9.590670355850819e-06,
+      "logits/chosen": 0.15097060799598694,
+      "logits/rejected": 0.16254279017448425,
+      "logps/chosen": -70.44524383544922,
+      "logps/rejected": -70.64158630371094,
+      "loss": 2.4159,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.6921097040176392,
+      "rewards/margins": 0.3336741626262665,
+      "rewards/rejected": -2.0257837772369385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5485921889191644,
+      "grad_norm": 1.8360233306884766,
+      "learning_rate": 9.587988391017198e-06,
+      "logits/chosen": 0.14593760669231415,
+      "logits/rejected": 0.12311654537916183,
+      "logps/chosen": -78.32576751708984,
+      "logps/rejected": -88.25840759277344,
+      "loss": 2.5826,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.6591644287109375,
+      "rewards/margins": 0.5071319341659546,
+      "rewards/rejected": -2.1662964820861816,
+      "step": 302
+    },
+    {
+      "epoch": 0.5504087193460491,
+      "grad_norm": 1.4822838306427002,
+      "learning_rate": 9.585298090063459e-06,
+      "logits/chosen": 0.20818498730659485,
+      "logits/rejected": 0.16436657309532166,
+      "logps/chosen": -67.59429931640625,
+      "logps/rejected": -77.14763641357422,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6798239946365356,
+      "rewards/margins": 0.5954271554946899,
+      "rewards/rejected": -2.2752511501312256,
+      "step": 303
+    },
+    {
+      "epoch": 0.5522252497729337,
+      "grad_norm": 1.6118305921554565,
+      "learning_rate": 9.582599458475598e-06,
+      "logits/chosen": 0.0827561467885971,
+      "logits/rejected": 0.09151773154735565,
+      "logps/chosen": -73.28964233398438,
+      "logps/rejected": -76.90730285644531,
+      "loss": 2.3792,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.8056440353393555,
+      "rewards/margins": 0.39551618695259094,
+      "rewards/rejected": -2.201160192489624,
+      "step": 304
+    },
+    {
+      "epoch": 0.5540417801998183,
+      "grad_norm": 1.4609856605529785,
+      "learning_rate": 9.579892501756593e-06,
+      "logits/chosen": 0.05332394689321518,
+      "logits/rejected": 0.1590057611465454,
+      "logps/chosen": -77.51653289794922,
+      "logps/rejected": -95.51261138916016,
+      "loss": 2.2105,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.740310549736023,
+      "rewards/margins": 0.6771562695503235,
+      "rewards/rejected": -2.4174671173095703,
+      "step": 305
+    },
+    {
+      "epoch": 0.555858310626703,
+      "grad_norm": 1.7181960344314575,
+      "learning_rate": 9.5771772254264e-06,
+      "logits/chosen": 0.05865276977419853,
+      "logits/rejected": 0.03500773757696152,
+      "logps/chosen": -83.72650146484375,
+      "logps/rejected": -93.69804382324219,
+      "loss": 2.1687,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6025928258895874,
+      "rewards/margins": 0.7339029312133789,
+      "rewards/rejected": -2.3364956378936768,
+      "step": 306
+    },
+    {
+      "epoch": 0.5576748410535877,
+      "grad_norm": 1.497698187828064,
+      "learning_rate": 9.57445363502194e-06,
+      "logits/chosen": 0.1225174218416214,
+      "logits/rejected": 0.09580346196889877,
+      "logps/chosen": -76.27323150634766,
+      "logps/rejected": -76.90716552734375,
+      "loss": 2.2116,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.715336799621582,
+      "rewards/margins": 0.4648086428642273,
+      "rewards/rejected": -2.180145263671875,
+      "step": 307
+    },
+    {
+      "epoch": 0.5594913714804723,
+      "grad_norm": 1.3957064151763916,
+      "learning_rate": 9.571721736097089e-06,
+      "logits/chosen": 0.07596514374017715,
+      "logits/rejected": 0.14832191169261932,
+      "logps/chosen": -84.65274810791016,
+      "logps/rejected": -105.76422119140625,
+      "loss": 1.9942,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.6643587350845337,
+      "rewards/margins": 0.7884883284568787,
+      "rewards/rejected": -2.4528470039367676,
+      "step": 308
+    },
+    {
+      "epoch": 0.5613079019073569,
+      "grad_norm": 1.6824349164962769,
+      "learning_rate": 9.568981534222664e-06,
+      "logits/chosen": 0.049865882843732834,
+      "logits/rejected": 0.023348212242126465,
+      "logps/chosen": -80.3697509765625,
+      "logps/rejected": -86.07074737548828,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.009049654006958,
+      "rewards/margins": 0.5470253825187683,
+      "rewards/rejected": -2.556075096130371,
+      "step": 309
+    },
+    {
+      "epoch": 0.5631244323342416,
+      "grad_norm": 1.8752086162567139,
+      "learning_rate": 9.566233034986413e-06,
+      "logits/chosen": 0.07250591367483139,
+      "logits/rejected": 0.12809628248214722,
+      "logps/chosen": -74.84794616699219,
+      "logps/rejected": -81.79740905761719,
+      "loss": 2.6286,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.75887131690979,
+      "rewards/margins": 0.33979111909866333,
+      "rewards/rejected": -2.0986623764038086,
+      "step": 310
+    },
+    {
+      "epoch": 0.5649409627611263,
+      "grad_norm": 1.7387233972549438,
+      "learning_rate": 9.563476243993008e-06,
+      "logits/chosen": 0.130618155002594,
+      "logits/rejected": 0.12952059507369995,
+      "logps/chosen": -80.75495910644531,
+      "logps/rejected": -90.4281234741211,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.836004614830017,
+      "rewards/margins": 0.4418962597846985,
+      "rewards/rejected": -2.2779006958007812,
+      "step": 311
+    },
+    {
+      "epoch": 0.5667574931880109,
+      "grad_norm": 1.5147444009780884,
+      "learning_rate": 9.56071116686402e-06,
+      "logits/chosen": 0.10193713754415512,
+      "logits/rejected": 0.22481802105903625,
+      "logps/chosen": -73.21926879882812,
+      "logps/rejected": -81.27864837646484,
+      "loss": 2.578,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6358015537261963,
+      "rewards/margins": 0.38279739022254944,
+      "rewards/rejected": -2.018598794937134,
+      "step": 312
+    },
+    {
+      "epoch": 0.5685740236148955,
+      "grad_norm": 1.6253665685653687,
+      "learning_rate": 9.557937809237927e-06,
+      "logits/chosen": 0.09468917548656464,
+      "logits/rejected": 0.09415112435817719,
+      "logps/chosen": -80.62995147705078,
+      "logps/rejected": -86.65946197509766,
+      "loss": 2.3854,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8635404109954834,
+      "rewards/margins": 0.4438764154911041,
+      "rewards/rejected": -2.3074169158935547,
+      "step": 313
+    },
+    {
+      "epoch": 0.5703905540417802,
+      "grad_norm": 1.7893344163894653,
+      "learning_rate": 9.555156176770087e-06,
+      "logits/chosen": 0.15863659977912903,
+      "logits/rejected": 0.09485571831464767,
+      "logps/chosen": -76.10442352294922,
+      "logps/rejected": -72.73162078857422,
+      "loss": 2.6795,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.935206413269043,
+      "rewards/margins": 0.22100940346717834,
+      "rewards/rejected": -2.1562156677246094,
+      "step": 314
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 1.696327805519104,
+      "learning_rate": 9.552366275132733e-06,
+      "logits/chosen": 0.07012113183736801,
+      "logits/rejected": 0.13891686499118805,
+      "logps/chosen": -78.08012390136719,
+      "logps/rejected": -83.05044555664062,
+      "loss": 2.6181,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7869051694869995,
+      "rewards/margins": 0.23424415290355682,
+      "rewards/rejected": -2.0211493968963623,
+      "step": 315
+    },
+    {
+      "epoch": 0.5740236148955495,
+      "grad_norm": 1.6044729948043823,
+      "learning_rate": 9.54956811001496e-06,
+      "logits/chosen": 0.11298641562461853,
+      "logits/rejected": 0.2595204710960388,
+      "logps/chosen": -75.5998306274414,
+      "logps/rejected": -85.31849670410156,
+      "loss": 2.5849,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.5757699012756348,
+      "rewards/margins": 0.35041162371635437,
+      "rewards/rejected": -1.9261815547943115,
+      "step": 316
+    },
+    {
+      "epoch": 0.5758401453224341,
+      "grad_norm": 1.7399548292160034,
+      "learning_rate": 9.546761687122715e-06,
+      "logits/chosen": 0.14933931827545166,
+      "logits/rejected": 0.1868455857038498,
+      "logps/chosen": -68.05965423583984,
+      "logps/rejected": -82.37442779541016,
+      "loss": 2.3208,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8474600315093994,
+      "rewards/margins": 0.5766161680221558,
+      "rewards/rejected": -2.4240763187408447,
+      "step": 317
+    },
+    {
+      "epoch": 0.5776566757493188,
+      "grad_norm": 1.597495675086975,
+      "learning_rate": 9.54394701217878e-06,
+      "logits/chosen": 0.04729313403367996,
+      "logits/rejected": 0.05815067142248154,
+      "logps/chosen": -97.36541748046875,
+      "logps/rejected": -104.02650451660156,
+      "loss": 2.2106,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8631618022918701,
+      "rewards/margins": 0.6027428507804871,
+      "rewards/rejected": -2.465904474258423,
+      "step": 318
+    },
+    {
+      "epoch": 0.5794732061762035,
+      "grad_norm": 1.8646149635314941,
+      "learning_rate": 9.541124090922771e-06,
+      "logits/chosen": 0.1399674415588379,
+      "logits/rejected": 0.11107950657606125,
+      "logps/chosen": -82.39740753173828,
+      "logps/rejected": -76.80902099609375,
+      "loss": 2.6435,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.7670300006866455,
+      "rewards/margins": 0.15345275402069092,
+      "rewards/rejected": -1.920482873916626,
+      "step": 319
+    },
+    {
+      "epoch": 0.5812897366030881,
+      "grad_norm": 1.502447247505188,
+      "learning_rate": 9.538292929111114e-06,
+      "logits/chosen": 0.10672347247600555,
+      "logits/rejected": 0.08290571719408035,
+      "logps/chosen": -74.72683715820312,
+      "logps/rejected": -86.37158203125,
+      "loss": 2.2137,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8041073083877563,
+      "rewards/margins": 0.596272885799408,
+      "rewards/rejected": -2.4003803730010986,
+      "step": 320
+    },
+    {
+      "epoch": 0.5831062670299727,
+      "grad_norm": 1.7287745475769043,
+      "learning_rate": 9.535453532517039e-06,
+      "logits/chosen": 0.13742896914482117,
+      "logits/rejected": 0.07607734203338623,
+      "logps/chosen": -86.82475280761719,
+      "logps/rejected": -77.8934555053711,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7918696403503418,
+      "rewards/margins": 0.2591831088066101,
+      "rewards/rejected": -2.0510525703430176,
+      "step": 321
+    },
+    {
+      "epoch": 0.5849227974568574,
+      "grad_norm": 1.5914596319198608,
+      "learning_rate": 9.532605906930575e-06,
+      "logits/chosen": 0.11301672458648682,
+      "logits/rejected": 0.1944851577281952,
+      "logps/chosen": -68.9898452758789,
+      "logps/rejected": -73.17451477050781,
+      "loss": 2.4565,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6999026536941528,
+      "rewards/margins": 0.30462026596069336,
+      "rewards/rejected": -2.0045228004455566,
+      "step": 322
+    },
+    {
+      "epoch": 0.5867393278837421,
+      "grad_norm": 1.6072031259536743,
+      "learning_rate": 9.529750058158522e-06,
+      "logits/chosen": 0.07092760503292084,
+      "logits/rejected": 0.08035591244697571,
+      "logps/chosen": -81.5555419921875,
+      "logps/rejected": -81.93560028076172,
+      "loss": 2.2729,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6834638118743896,
+      "rewards/margins": 0.48882579803466797,
+      "rewards/rejected": -2.1722893714904785,
+      "step": 323
+    },
+    {
+      "epoch": 0.5885558583106267,
+      "grad_norm": 1.6433221101760864,
+      "learning_rate": 9.526885992024453e-06,
+      "logits/chosen": 0.13823899626731873,
+      "logits/rejected": 0.13610433042049408,
+      "logps/chosen": -78.9625244140625,
+      "logps/rejected": -83.2666015625,
+      "loss": 2.2075,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6939194202423096,
+      "rewards/margins": 0.5755473971366882,
+      "rewards/rejected": -2.2694668769836426,
+      "step": 324
+    },
+    {
+      "epoch": 0.5903723887375113,
+      "grad_norm": 1.7674918174743652,
+      "learning_rate": 9.524013714368702e-06,
+      "logits/chosen": 0.1704932153224945,
+      "logits/rejected": 0.1073535829782486,
+      "logps/chosen": -72.47895050048828,
+      "logps/rejected": -73.0809326171875,
+      "loss": 2.5473,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8390110731124878,
+      "rewards/margins": 0.34467947483062744,
+      "rewards/rejected": -2.1836905479431152,
+      "step": 325
+    },
+    {
+      "epoch": 0.592188919164396,
+      "grad_norm": 1.8219250440597534,
+      "learning_rate": 9.521133231048338e-06,
+      "logits/chosen": 0.07941028475761414,
+      "logits/rejected": 0.13045310974121094,
+      "logps/chosen": -80.73494720458984,
+      "logps/rejected": -91.58990478515625,
+      "loss": 2.4866,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8594518899917603,
+      "rewards/margins": 0.4554288685321808,
+      "rewards/rejected": -2.314880847930908,
+      "step": 326
+    },
+    {
+      "epoch": 0.5940054495912807,
+      "grad_norm": 1.3753328323364258,
+      "learning_rate": 9.51824454793717e-06,
+      "logits/chosen": 0.08879546821117401,
+      "logits/rejected": 0.04693777486681938,
+      "logps/chosen": -82.9569091796875,
+      "logps/rejected": -91.44571685791016,
+      "loss": 2.2602,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.549119472503662,
+      "rewards/margins": 0.5145683288574219,
+      "rewards/rejected": -2.063688039779663,
+      "step": 327
+    },
+    {
+      "epoch": 0.5958219800181653,
+      "grad_norm": 1.4296562671661377,
+      "learning_rate": 9.515347670925728e-06,
+      "logits/chosen": 0.15614314377307892,
+      "logits/rejected": 0.1598319411277771,
+      "logps/chosen": -75.60345458984375,
+      "logps/rejected": -80.81770324707031,
+      "loss": 2.3946,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.556883454322815,
+      "rewards/margins": 0.4525718092918396,
+      "rewards/rejected": -2.0094552040100098,
+      "step": 328
+    },
+    {
+      "epoch": 0.59763851044505,
+      "grad_norm": 1.7932195663452148,
+      "learning_rate": 9.512442605921245e-06,
+      "logits/chosen": 0.05863601714372635,
+      "logits/rejected": 0.0788806602358818,
+      "logps/chosen": -75.95397186279297,
+      "logps/rejected": -84.50177001953125,
+      "loss": 2.5085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9159862995147705,
+      "rewards/margins": 0.4600725769996643,
+      "rewards/rejected": -2.376059055328369,
+      "step": 329
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 1.6558383703231812,
+      "learning_rate": 9.509529358847655e-06,
+      "logits/chosen": 0.08205496519804001,
+      "logits/rejected": 0.13091425597667694,
+      "logps/chosen": -82.73474884033203,
+      "logps/rejected": -94.02273559570312,
+      "loss": 2.4419,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.083285093307495,
+      "rewards/margins": 0.5300815105438232,
+      "rewards/rejected": -2.6133666038513184,
+      "step": 330
+    },
+    {
+      "epoch": 0.6012715712988193,
+      "grad_norm": 1.7888520956039429,
+      "learning_rate": 9.506607935645579e-06,
+      "logits/chosen": 0.11793217062950134,
+      "logits/rejected": 0.19568441808223724,
+      "logps/chosen": -79.86200714111328,
+      "logps/rejected": -89.25471496582031,
+      "loss": 2.4273,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9273384809494019,
+      "rewards/margins": 0.5115458965301514,
+      "rewards/rejected": -2.4388844966888428,
+      "step": 331
+    },
+    {
+      "epoch": 0.6030881017257039,
+      "grad_norm": 2.0436160564422607,
+      "learning_rate": 9.503678342272306e-06,
+      "logits/chosen": 0.043473344296216965,
+      "logits/rejected": 0.15554016828536987,
+      "logps/chosen": -75.2901611328125,
+      "logps/rejected": -83.20555877685547,
+      "loss": 2.6941,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.725769281387329,
+      "rewards/margins": 0.2789250314235687,
+      "rewards/rejected": -2.0046942234039307,
+      "step": 332
+    },
+    {
+      "epoch": 0.6049046321525886,
+      "grad_norm": 1.665578842163086,
+      "learning_rate": 9.500740584701785e-06,
+      "logits/chosen": 0.17119848728179932,
+      "logits/rejected": 0.14128939807415009,
+      "logps/chosen": -83.1322021484375,
+      "logps/rejected": -93.38603210449219,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.8975954055786133,
+      "rewards/margins": 0.7810046672821045,
+      "rewards/rejected": -2.6786000728607178,
+      "step": 333
+    },
+    {
+      "epoch": 0.6067211625794732,
+      "grad_norm": 1.9258419275283813,
+      "learning_rate": 9.497794668924617e-06,
+      "logits/chosen": 0.028591612353920937,
+      "logits/rejected": 0.08894480764865875,
+      "logps/chosen": -78.12230682373047,
+      "logps/rejected": -90.00184631347656,
+      "loss": 2.4006,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8113291263580322,
+      "rewards/margins": 0.5277000665664673,
+      "rewards/rejected": -2.339029550552368,
+      "step": 334
+    },
+    {
+      "epoch": 0.6085376930063578,
+      "grad_norm": 1.4878523349761963,
+      "learning_rate": 9.494840600948038e-06,
+      "logits/chosen": 0.03548199310898781,
+      "logits/rejected": 0.09170147776603699,
+      "logps/chosen": -73.15607452392578,
+      "logps/rejected": -83.77317810058594,
+      "loss": 2.3301,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6781296730041504,
+      "rewards/margins": 0.4735565185546875,
+      "rewards/rejected": -2.151685953140259,
+      "step": 335
+    },
+    {
+      "epoch": 0.6103542234332425,
+      "grad_norm": 1.8042774200439453,
+      "learning_rate": 9.491878386795906e-06,
+      "logits/chosen": 0.046132348477840424,
+      "logits/rejected": 0.0721711814403534,
+      "logps/chosen": -81.01045227050781,
+      "logps/rejected": -89.28679656982422,
+      "loss": 2.5593,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.012852191925049,
+      "rewards/margins": 0.44549623131752014,
+      "rewards/rejected": -2.458348512649536,
+      "step": 336
+    },
+    {
+      "epoch": 0.6121707538601272,
+      "grad_norm": 2.048952102661133,
+      "learning_rate": 9.488908032508691e-06,
+      "logits/chosen": 0.10774732381105423,
+      "logits/rejected": 0.11935572326183319,
+      "logps/chosen": -91.25210571289062,
+      "logps/rejected": -90.42224884033203,
+      "loss": 2.8282,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0127718448638916,
+      "rewards/margins": 0.16889013350009918,
+      "rewards/rejected": -2.18166184425354,
+      "step": 337
+    },
+    {
+      "epoch": 0.6139872842870118,
+      "grad_norm": 1.6447219848632812,
+      "learning_rate": 9.485929544143462e-06,
+      "logits/chosen": 0.12260966747999191,
+      "logits/rejected": 0.1517799198627472,
+      "logps/chosen": -72.49166870117188,
+      "logps/rejected": -82.3083724975586,
+      "loss": 2.4029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.624878168106079,
+      "rewards/margins": 0.5487539768218994,
+      "rewards/rejected": -2.1736321449279785,
+      "step": 338
+    },
+    {
+      "epoch": 0.6158038147138964,
+      "grad_norm": 1.599530577659607,
+      "learning_rate": 9.482942927773876e-06,
+      "logits/chosen": 0.09051798284053802,
+      "logits/rejected": 0.11202570050954819,
+      "logps/chosen": -84.42019653320312,
+      "logps/rejected": -92.61677551269531,
+      "loss": 2.2269,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6424546241760254,
+      "rewards/margins": 0.5715836882591248,
+      "rewards/rejected": -2.214038133621216,
+      "step": 339
+    },
+    {
+      "epoch": 0.6176203451407811,
+      "grad_norm": 1.5411081314086914,
+      "learning_rate": 9.479948189490164e-06,
+      "logits/chosen": 0.06563656777143478,
+      "logits/rejected": 0.17349205911159515,
+      "logps/chosen": -70.98219299316406,
+      "logps/rejected": -78.88846588134766,
+      "loss": 2.3328,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.824373722076416,
+      "rewards/margins": 0.46550822257995605,
+      "rewards/rejected": -2.289881706237793,
+      "step": 340
+    },
+    {
+      "epoch": 0.6194368755676658,
+      "grad_norm": 1.8855030536651611,
+      "learning_rate": 9.476945335399122e-06,
+      "logits/chosen": 0.11399642378091812,
+      "logits/rejected": 0.12690946459770203,
+      "logps/chosen": -85.86933898925781,
+      "logps/rejected": -89.75601196289062,
+      "loss": 2.4269,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9560539722442627,
+      "rewards/margins": 0.38427016139030457,
+      "rewards/rejected": -2.3403239250183105,
+      "step": 341
+    },
+    {
+      "epoch": 0.6212534059945504,
+      "grad_norm": 1.9973480701446533,
+      "learning_rate": 9.473934371624087e-06,
+      "logits/chosen": 0.029105912894010544,
+      "logits/rejected": 0.12763622403144836,
+      "logps/chosen": -80.68119049072266,
+      "logps/rejected": -93.90747833251953,
+      "loss": 2.4709,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.906477928161621,
+      "rewards/margins": 0.6422742009162903,
+      "rewards/rejected": -2.5487518310546875,
+      "step": 342
+    },
+    {
+      "epoch": 0.623069936421435,
+      "grad_norm": 1.7752224206924438,
+      "learning_rate": 9.47091530430494e-06,
+      "logits/chosen": 0.14096824824810028,
+      "logits/rejected": 0.1774113029241562,
+      "logps/chosen": -71.80216217041016,
+      "logps/rejected": -79.65301513671875,
+      "loss": 2.2678,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.5735952854156494,
+      "rewards/margins": 0.5823659300804138,
+      "rewards/rejected": -2.155961036682129,
+      "step": 343
+    },
+    {
+      "epoch": 0.6248864668483197,
+      "grad_norm": 1.855377197265625,
+      "learning_rate": 9.467888139598086e-06,
+      "logits/chosen": 0.10637074708938599,
+      "logits/rejected": 0.0798158049583435,
+      "logps/chosen": -79.01347351074219,
+      "logps/rejected": -80.77239990234375,
+      "loss": 2.7867,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9598251581192017,
+      "rewards/margins": 0.3184160590171814,
+      "rewards/rejected": -2.2782411575317383,
+      "step": 344
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 1.423040747642517,
+      "learning_rate": 9.464852883676441e-06,
+      "logits/chosen": 0.1304859220981598,
+      "logits/rejected": 0.1759231686592102,
+      "logps/chosen": -74.09733581542969,
+      "logps/rejected": -89.27589416503906,
+      "loss": 2.2777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7697277069091797,
+      "rewards/margins": 0.635611355304718,
+      "rewards/rejected": -2.405339002609253,
+      "step": 345
+    },
+    {
+      "epoch": 0.628519527702089,
+      "grad_norm": 1.6246038675308228,
+      "learning_rate": 9.461809542729421e-06,
+      "logits/chosen": 0.03750050812959671,
+      "logits/rejected": 0.10337980091571808,
+      "logps/chosen": -81.89110565185547,
+      "logps/rejected": -95.36811828613281,
+      "loss": 2.0438,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7299752235412598,
+      "rewards/margins": 0.7404756546020508,
+      "rewards/rejected": -2.4704508781433105,
+      "step": 346
+    },
+    {
+      "epoch": 0.6303360581289736,
+      "grad_norm": 1.6618752479553223,
+      "learning_rate": 9.458758122962926e-06,
+      "logits/chosen": 0.05359608680009842,
+      "logits/rejected": 0.10455545783042908,
+      "logps/chosen": -85.34078216552734,
+      "logps/rejected": -90.46200561523438,
+      "loss": 2.3299,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9331045150756836,
+      "rewards/margins": 0.4755082130432129,
+      "rewards/rejected": -2.4086129665374756,
+      "step": 347
+    },
+    {
+      "epoch": 0.6321525885558583,
+      "grad_norm": 1.5805696249008179,
+      "learning_rate": 9.455698630599332e-06,
+      "logits/chosen": 0.10048776119947433,
+      "logits/rejected": 0.1140337884426117,
+      "logps/chosen": -80.25875854492188,
+      "logps/rejected": -94.44998168945312,
+      "loss": 2.3439,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8421945571899414,
+      "rewards/margins": 0.5977468490600586,
+      "rewards/rejected": -2.43994140625,
+      "step": 348
+    },
+    {
+      "epoch": 0.633969118982743,
+      "grad_norm": 1.37093985080719,
+      "learning_rate": 9.452631071877478e-06,
+      "logits/chosen": 0.11764326691627502,
+      "logits/rejected": 0.10735289752483368,
+      "logps/chosen": -72.02367401123047,
+      "logps/rejected": -77.9522705078125,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.6527085304260254,
+      "rewards/margins": 0.6065118908882141,
+      "rewards/rejected": -2.259220600128174,
+      "step": 349
+    },
+    {
+      "epoch": 0.6357856494096276,
+      "grad_norm": 2.0214192867279053,
+      "learning_rate": 9.449555453052652e-06,
+      "logits/chosen": 0.13177426159381866,
+      "logits/rejected": 0.12408209592103958,
+      "logps/chosen": -76.62931823730469,
+      "logps/rejected": -81.16517639160156,
+      "loss": 2.9175,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.9575048685073853,
+      "rewards/margins": 0.21441945433616638,
+      "rewards/rejected": -2.171924352645874,
+      "step": 350
+    },
+    {
+      "epoch": 0.6376021798365122,
+      "grad_norm": 1.6041687726974487,
+      "learning_rate": 9.446471780396573e-06,
+      "logits/chosen": 0.18574532866477966,
+      "logits/rejected": 0.22683225572109222,
+      "logps/chosen": -73.95652770996094,
+      "logps/rejected": -80.29524993896484,
+      "loss": 2.3833,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8684748411178589,
+      "rewards/margins": 0.37268272042274475,
+      "rewards/rejected": -2.2411575317382812,
+      "step": 351
+    },
+    {
+      "epoch": 0.6394187102633969,
+      "grad_norm": 1.7410838603973389,
+      "learning_rate": 9.443380060197387e-06,
+      "logits/chosen": 0.07876043766736984,
+      "logits/rejected": 0.14103996753692627,
+      "logps/chosen": -74.64432525634766,
+      "logps/rejected": -83.6999740600586,
+      "loss": 2.5558,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.029916763305664,
+      "rewards/margins": 0.4088842272758484,
+      "rewards/rejected": -2.438800811767578,
+      "step": 352
+    },
+    {
+      "epoch": 0.6412352406902816,
+      "grad_norm": 1.87971031665802,
+      "learning_rate": 9.440280298759653e-06,
+      "logits/chosen": 0.13997013866901398,
+      "logits/rejected": 0.1322249174118042,
+      "logps/chosen": -88.78776550292969,
+      "logps/rejected": -88.22732543945312,
+      "loss": 2.6547,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9983258247375488,
+      "rewards/margins": 0.2458382397890091,
+      "rewards/rejected": -2.244164228439331,
+      "step": 353
+    },
+    {
+      "epoch": 0.6430517711171662,
+      "grad_norm": 1.7056363821029663,
+      "learning_rate": 9.437172502404318e-06,
+      "logits/chosen": 0.07248476892709732,
+      "logits/rejected": 0.13019773364067078,
+      "logps/chosen": -78.4591064453125,
+      "logps/rejected": -80.49126434326172,
+      "loss": 2.3809,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9740041494369507,
+      "rewards/margins": 0.4853326082229614,
+      "rewards/rejected": -2.459336757659912,
+      "step": 354
+    },
+    {
+      "epoch": 0.6448683015440508,
+      "grad_norm": 1.435718059539795,
+      "learning_rate": 9.434056677468726e-06,
+      "logits/chosen": 0.09164869040250778,
+      "logits/rejected": 0.14243285357952118,
+      "logps/chosen": -77.83367156982422,
+      "logps/rejected": -86.18836212158203,
+      "loss": 2.006,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.7355122566223145,
+      "rewards/margins": 0.6869419813156128,
+      "rewards/rejected": -2.422454357147217,
+      "step": 355
+    },
+    {
+      "epoch": 0.6466848319709355,
+      "grad_norm": 1.6553188562393188,
+      "learning_rate": 9.430932830306587e-06,
+      "logits/chosen": 0.04967673122882843,
+      "logits/rejected": 0.1456151008605957,
+      "logps/chosen": -73.50204467773438,
+      "logps/rejected": -85.31363677978516,
+      "loss": 2.4388,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.978384256362915,
+      "rewards/margins": 0.48918718099594116,
+      "rewards/rejected": -2.467571496963501,
+      "step": 356
+    },
+    {
+      "epoch": 0.6485013623978202,
+      "grad_norm": 1.8443480730056763,
+      "learning_rate": 9.427800967287963e-06,
+      "logits/chosen": 0.06455090641975403,
+      "logits/rejected": 0.12608012557029724,
+      "logps/chosen": -75.49232482910156,
+      "logps/rejected": -77.49136352539062,
+      "loss": 2.706,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.8730812072753906,
+      "rewards/margins": 0.2989741861820221,
+      "rewards/rejected": -2.17205548286438,
+      "step": 357
+    },
+    {
+      "epoch": 0.6503178928247049,
+      "grad_norm": 1.6912306547164917,
+      "learning_rate": 9.424661094799273e-06,
+      "logits/chosen": 0.11878645420074463,
+      "logits/rejected": 0.13628609478473663,
+      "logps/chosen": -73.09883880615234,
+      "logps/rejected": -84.19624328613281,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.928421139717102,
+      "rewards/margins": 0.5017052292823792,
+      "rewards/rejected": -2.430126190185547,
+      "step": 358
+    },
+    {
+      "epoch": 0.6521344232515894,
+      "grad_norm": 1.3643461465835571,
+      "learning_rate": 9.421513219243262e-06,
+      "logits/chosen": 0.07683826237916946,
+      "logits/rejected": 0.15765298902988434,
+      "logps/chosen": -76.0871353149414,
+      "logps/rejected": -97.47781372070312,
+      "loss": 1.7844,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.8300602436065674,
+      "rewards/margins": 0.911116361618042,
+      "rewards/rejected": -2.7411766052246094,
+      "step": 359
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 1.6327749490737915,
+      "learning_rate": 9.418357347038999e-06,
+      "logits/chosen": 0.1078951433300972,
+      "logits/rejected": 0.12233921140432358,
+      "logps/chosen": -75.89913177490234,
+      "logps/rejected": -78.4587173461914,
+      "loss": 2.4976,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.7992844581604004,
+      "rewards/margins": 0.311392605304718,
+      "rewards/rejected": -2.1106772422790527,
+      "step": 360
+    },
+    {
+      "epoch": 0.6557674841053588,
+      "grad_norm": 1.463025689125061,
+      "learning_rate": 9.415193484621852e-06,
+      "logits/chosen": 0.057331383228302,
+      "logits/rejected": 0.1563551127910614,
+      "logps/chosen": -80.7889633178711,
+      "logps/rejected": -93.21602630615234,
+      "loss": 2.1398,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.744189739227295,
+      "rewards/margins": 0.576555073261261,
+      "rewards/rejected": -2.320744752883911,
+      "step": 361
+    },
+    {
+      "epoch": 0.6575840145322435,
+      "grad_norm": 1.743695616722107,
+      "learning_rate": 9.412021638443491e-06,
+      "logits/chosen": 0.03781123086810112,
+      "logits/rejected": 0.154897540807724,
+      "logps/chosen": -77.72047424316406,
+      "logps/rejected": -88.80599975585938,
+      "loss": 2.2361,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9849369525909424,
+      "rewards/margins": 0.6562294363975525,
+      "rewards/rejected": -2.6411664485931396,
+      "step": 362
+    },
+    {
+      "epoch": 0.659400544959128,
+      "grad_norm": 1.5648279190063477,
+      "learning_rate": 9.408841814971862e-06,
+      "logits/chosen": 0.10401102900505066,
+      "logits/rejected": 0.12699122726917267,
+      "logps/chosen": -83.10671997070312,
+      "logps/rejected": -86.8663101196289,
+      "loss": 2.4009,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7048468589782715,
+      "rewards/margins": 0.4090174734592438,
+      "rewards/rejected": -2.1138644218444824,
+      "step": 363
+    },
+    {
+      "epoch": 0.6612170753860127,
+      "grad_norm": 1.3385239839553833,
+      "learning_rate": 9.405654020691178e-06,
+      "logits/chosen": 0.06100422143936157,
+      "logits/rejected": 0.10872650146484375,
+      "logps/chosen": -76.72837829589844,
+      "logps/rejected": -87.63470458984375,
+      "loss": 2.141,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.747226595878601,
+      "rewards/margins": 0.615075945854187,
+      "rewards/rejected": -2.362302780151367,
+      "step": 364
+    },
+    {
+      "epoch": 0.6630336058128974,
+      "grad_norm": 1.4060734510421753,
+      "learning_rate": 9.402458262101906e-06,
+      "logits/chosen": 0.10638861358165741,
+      "logits/rejected": 0.16951681673526764,
+      "logps/chosen": -77.35757446289062,
+      "logps/rejected": -92.03012084960938,
+      "loss": 2.0511,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8359463214874268,
+      "rewards/margins": 0.7444422841072083,
+      "rewards/rejected": -2.5803885459899902,
+      "step": 365
+    },
+    {
+      "epoch": 0.6648501362397821,
+      "grad_norm": 1.4371014833450317,
+      "learning_rate": 9.399254545720757e-06,
+      "logits/chosen": 0.0383220911026001,
+      "logits/rejected": 0.05856206640601158,
+      "logps/chosen": -82.40064239501953,
+      "logps/rejected": -92.1117935180664,
+      "loss": 2.1756,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8580785989761353,
+      "rewards/margins": 0.6475991010665894,
+      "rewards/rejected": -2.5056777000427246,
+      "step": 366
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.6673945188522339,
+      "learning_rate": 9.396042878080661e-06,
+      "logits/chosen": 0.14657820761203766,
+      "logits/rejected": 0.19021111726760864,
+      "logps/chosen": -73.56724548339844,
+      "logps/rejected": -78.85279846191406,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9079951047897339,
+      "rewards/margins": 0.3709837794303894,
+      "rewards/rejected": -2.2789790630340576,
+      "step": 367
+    },
+    {
+      "epoch": 0.6684831970935513,
+      "grad_norm": 1.7402448654174805,
+      "learning_rate": 9.392823265730775e-06,
+      "logits/chosen": 0.16515754163265228,
+      "logits/rejected": 0.12460774928331375,
+      "logps/chosen": -69.89266204833984,
+      "logps/rejected": -74.53837585449219,
+      "loss": 2.4262,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.0627622604370117,
+      "rewards/margins": 0.5317320823669434,
+      "rewards/rejected": -2.594494342803955,
+      "step": 368
+    },
+    {
+      "epoch": 0.670299727520436,
+      "grad_norm": 1.5290364027023315,
+      "learning_rate": 9.389595715236446e-06,
+      "logits/chosen": 0.0954434722661972,
+      "logits/rejected": 0.16226956248283386,
+      "logps/chosen": -77.17019653320312,
+      "logps/rejected": -91.07398986816406,
+      "loss": 2.2511,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.241170644760132,
+      "rewards/margins": 0.6319360136985779,
+      "rewards/rejected": -2.8731067180633545,
+      "step": 369
+    },
+    {
+      "epoch": 0.6721162579473207,
+      "grad_norm": 1.4633573293685913,
+      "learning_rate": 9.386360233179206e-06,
+      "logits/chosen": 0.06517557799816132,
+      "logits/rejected": 0.07250035554170609,
+      "logps/chosen": -79.23770141601562,
+      "logps/rejected": -88.50403594970703,
+      "loss": 2.3858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.751698613166809,
+      "rewards/margins": 0.49529415369033813,
+      "rewards/rejected": -2.246993064880371,
+      "step": 370
+    },
+    {
+      "epoch": 0.6739327883742052,
+      "grad_norm": 2.007215976715088,
+      "learning_rate": 9.383116826156775e-06,
+      "logits/chosen": 0.13584929704666138,
+      "logits/rejected": 0.10950647294521332,
+      "logps/chosen": -80.75408935546875,
+      "logps/rejected": -79.3903579711914,
+      "loss": 2.6967,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.2421724796295166,
+      "rewards/margins": 0.34328368306159973,
+      "rewards/rejected": -2.585456132888794,
+      "step": 371
+    },
+    {
+      "epoch": 0.6757493188010899,
+      "grad_norm": 1.7323477268218994,
+      "learning_rate": 9.37986550078302e-06,
+      "logits/chosen": 0.011626070365309715,
+      "logits/rejected": 0.06207559257745743,
+      "logps/chosen": -75.51494598388672,
+      "logps/rejected": -85.83218383789062,
+      "loss": 2.4608,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9569413661956787,
+      "rewards/margins": 0.5160585641860962,
+      "rewards/rejected": -2.4729998111724854,
+      "step": 372
+    },
+    {
+      "epoch": 0.6775658492279746,
+      "grad_norm": 1.6365752220153809,
+      "learning_rate": 9.376606263687959e-06,
+      "logits/chosen": 0.10213632136583328,
+      "logits/rejected": 0.08355780690908432,
+      "logps/chosen": -79.01370239257812,
+      "logps/rejected": -81.9417953491211,
+      "loss": 2.4701,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.064467668533325,
+      "rewards/margins": 0.4157037138938904,
+      "rewards/rejected": -2.4801712036132812,
+      "step": 373
+    },
+    {
+      "epoch": 0.6793823796548593,
+      "grad_norm": 1.5918922424316406,
+      "learning_rate": 9.373339121517748e-06,
+      "logits/chosen": 0.09486684203147888,
+      "logits/rejected": 0.07311725616455078,
+      "logps/chosen": -82.15351867675781,
+      "logps/rejected": -91.7690658569336,
+      "loss": 2.1297,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0153114795684814,
+      "rewards/margins": 0.669613242149353,
+      "rewards/rejected": -2.684924840927124,
+      "step": 374
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 1.6511566638946533,
+      "learning_rate": 9.370064080934654e-06,
+      "logits/chosen": 0.1406637728214264,
+      "logits/rejected": 0.20172299444675446,
+      "logps/chosen": -69.41023254394531,
+      "logps/rejected": -74.65138244628906,
+      "loss": 2.4704,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.061182737350464,
+      "rewards/margins": 0.3090137243270874,
+      "rewards/rejected": -2.3701963424682617,
+      "step": 375
+    },
+    {
+      "epoch": 0.6830154405086285,
+      "grad_norm": 1.770624041557312,
+      "learning_rate": 9.366781148617056e-06,
+      "logits/chosen": 0.10267248749732971,
+      "logits/rejected": 0.07900385558605194,
+      "logps/chosen": -77.27003479003906,
+      "logps/rejected": -88.09040832519531,
+      "loss": 2.2813,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.132927656173706,
+      "rewards/margins": 0.6629016399383545,
+      "rewards/rejected": -2.7958290576934814,
+      "step": 376
+    },
+    {
+      "epoch": 0.6848319709355132,
+      "grad_norm": 1.5172028541564941,
+      "learning_rate": 9.363490331259426e-06,
+      "logits/chosen": 0.05240853130817413,
+      "logits/rejected": 0.10445387661457062,
+      "logps/chosen": -75.5218276977539,
+      "logps/rejected": -85.49366760253906,
+      "loss": 2.1442,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9302880764007568,
+      "rewards/margins": 0.5776917338371277,
+      "rewards/rejected": -2.5079798698425293,
+      "step": 377
+    },
+    {
+      "epoch": 0.6866485013623979,
+      "grad_norm": 1.6540950536727905,
+      "learning_rate": 9.360191635572313e-06,
+      "logits/chosen": 0.14195458590984344,
+      "logits/rejected": 0.0978107899427414,
+      "logps/chosen": -85.00608825683594,
+      "logps/rejected": -85.1456298828125,
+      "loss": 2.423,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8765841722488403,
+      "rewards/margins": 0.49441370368003845,
+      "rewards/rejected": -2.370997905731201,
+      "step": 378
+    },
+    {
+      "epoch": 0.6884650317892824,
+      "grad_norm": 1.6463801860809326,
+      "learning_rate": 9.356885068282334e-06,
+      "logits/chosen": 0.13197994232177734,
+      "logits/rejected": 0.09924699366092682,
+      "logps/chosen": -86.94219970703125,
+      "logps/rejected": -86.85704803466797,
+      "loss": 2.4801,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9674016237258911,
+      "rewards/margins": 0.4530283510684967,
+      "rewards/rejected": -2.4204299449920654,
+      "step": 379
+    },
+    {
+      "epoch": 0.6902815622161671,
+      "grad_norm": 1.5428895950317383,
+      "learning_rate": 9.353570636132151e-06,
+      "logits/chosen": 0.10434838384389877,
+      "logits/rejected": 0.12619757652282715,
+      "logps/chosen": -78.56607818603516,
+      "logps/rejected": -86.33320617675781,
+      "loss": 2.2711,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9667280912399292,
+      "rewards/margins": 0.6350463628768921,
+      "rewards/rejected": -2.601774215698242,
+      "step": 380
+    },
+    {
+      "epoch": 0.6920980926430518,
+      "grad_norm": 1.6211251020431519,
+      "learning_rate": 9.350248345880471e-06,
+      "logits/chosen": 0.14081251621246338,
+      "logits/rejected": 0.1453506350517273,
+      "logps/chosen": -74.38184356689453,
+      "logps/rejected": -80.61524200439453,
+      "loss": 2.3863,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.0189671516418457,
+      "rewards/margins": 0.4528267979621887,
+      "rewards/rejected": -2.4717938899993896,
+      "step": 381
+    },
+    {
+      "epoch": 0.6939146230699365,
+      "grad_norm": 2.6346240043640137,
+      "learning_rate": 9.346918204302022e-06,
+      "logits/chosen": 0.0519830696284771,
+      "logits/rejected": 0.04284125566482544,
+      "logps/chosen": -82.38153839111328,
+      "logps/rejected": -86.35488891601562,
+      "loss": 2.5434,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9824391603469849,
+      "rewards/margins": 0.4156208634376526,
+      "rewards/rejected": -2.3980600833892822,
+      "step": 382
+    },
+    {
+      "epoch": 0.695731153496821,
+      "grad_norm": 1.4762026071548462,
+      "learning_rate": 9.343580218187544e-06,
+      "logits/chosen": 0.02196469157934189,
+      "logits/rejected": 0.04618150740861893,
+      "logps/chosen": -72.66182708740234,
+      "logps/rejected": -88.68983459472656,
+      "loss": 2.1699,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9227774143218994,
+      "rewards/margins": 0.9200201630592346,
+      "rewards/rejected": -2.8427975177764893,
+      "step": 383
+    },
+    {
+      "epoch": 0.6975476839237057,
+      "grad_norm": 1.9675711393356323,
+      "learning_rate": 9.340234394343768e-06,
+      "logits/chosen": 0.12956203520298004,
+      "logits/rejected": 0.16079677641391754,
+      "logps/chosen": -87.38492584228516,
+      "logps/rejected": -87.80662536621094,
+      "loss": 2.3138,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.144991397857666,
+      "rewards/margins": 0.4893474578857422,
+      "rewards/rejected": -2.634338855743408,
+      "step": 384
+    },
+    {
+      "epoch": 0.6993642143505904,
+      "grad_norm": 1.9116860628128052,
+      "learning_rate": 9.336880739593415e-06,
+      "logits/chosen": 0.06013559550046921,
+      "logits/rejected": 0.09889352321624756,
+      "logps/chosen": -86.55718231201172,
+      "logps/rejected": -91.01994323730469,
+      "loss": 2.4864,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9528815746307373,
+      "rewards/margins": 0.44100096821784973,
+      "rewards/rejected": -2.3938825130462646,
+      "step": 385
+    },
+    {
+      "epoch": 0.701180744777475,
+      "grad_norm": 2.0582289695739746,
+      "learning_rate": 9.33351926077517e-06,
+      "logits/chosen": 0.11538423597812653,
+      "logits/rejected": 0.17568480968475342,
+      "logps/chosen": -76.24241638183594,
+      "logps/rejected": -87.1357650756836,
+      "loss": 2.6878,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.102504253387451,
+      "rewards/margins": 0.3671773076057434,
+      "rewards/rejected": -2.469681739807129,
+      "step": 386
+    },
+    {
+      "epoch": 0.7029972752043597,
+      "grad_norm": 1.6962640285491943,
+      "learning_rate": 9.330149964743674e-06,
+      "logits/chosen": 0.12425235658884048,
+      "logits/rejected": 0.1852879822254181,
+      "logps/chosen": -82.67887115478516,
+      "logps/rejected": -98.66133117675781,
+      "loss": 2.343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.182525634765625,
+      "rewards/margins": 0.6602460145950317,
+      "rewards/rejected": -2.842771530151367,
+      "step": 387
+    },
+    {
+      "epoch": 0.7048138056312443,
+      "grad_norm": 1.6582266092300415,
+      "learning_rate": 9.326772858369506e-06,
+      "logits/chosen": 0.14438432455062866,
+      "logits/rejected": 0.0922938883304596,
+      "logps/chosen": -78.87794494628906,
+      "logps/rejected": -84.929443359375,
+      "loss": 2.3036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.988234281539917,
+      "rewards/margins": 0.4737466275691986,
+      "rewards/rejected": -2.4619810581207275,
+      "step": 388
+    },
+    {
+      "epoch": 0.706630336058129,
+      "grad_norm": 1.6689919233322144,
+      "learning_rate": 9.323387948539176e-06,
+      "logits/chosen": 0.1282673329114914,
+      "logits/rejected": 0.13633723556995392,
+      "logps/chosen": -70.22606658935547,
+      "logps/rejected": -79.26126861572266,
+      "loss": 2.241,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9162399768829346,
+      "rewards/margins": 0.5345122814178467,
+      "rewards/rejected": -2.4507524967193604,
+      "step": 389
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 1.5861207246780396,
+      "learning_rate": 9.319995242155102e-06,
+      "logits/chosen": 0.13024169206619263,
+      "logits/rejected": 0.14390775561332703,
+      "logps/chosen": -87.00923919677734,
+      "logps/rejected": -91.24594116210938,
+      "loss": 2.284,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9127002954483032,
+      "rewards/margins": 0.5265774130821228,
+      "rewards/rejected": -2.4392776489257812,
+      "step": 390
+    },
+    {
+      "epoch": 0.7102633969118983,
+      "grad_norm": 2.0118792057037354,
+      "learning_rate": 9.316594746135608e-06,
+      "logits/chosen": 0.14648675918579102,
+      "logits/rejected": 0.11786539107561111,
+      "logps/chosen": -82.68962097167969,
+      "logps/rejected": -88.26126861572266,
+      "loss": 2.7281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9463748931884766,
+      "rewards/margins": 0.20282456278800964,
+      "rewards/rejected": -2.1491994857788086,
+      "step": 391
+    },
+    {
+      "epoch": 0.7120799273387829,
+      "grad_norm": 1.8176332712173462,
+      "learning_rate": 9.313186467414892e-06,
+      "logits/chosen": 0.12084021419286728,
+      "logits/rejected": 0.1433698982000351,
+      "logps/chosen": -73.46707153320312,
+      "logps/rejected": -82.46024322509766,
+      "loss": 2.5247,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0802624225616455,
+      "rewards/margins": 0.41601985692977905,
+      "rewards/rejected": -2.4962823390960693,
+      "step": 392
+    },
+    {
+      "epoch": 0.7138964577656676,
+      "grad_norm": 1.7207351922988892,
+      "learning_rate": 9.30977041294303e-06,
+      "logits/chosen": 0.11983273923397064,
+      "logits/rejected": 0.16106371581554413,
+      "logps/chosen": -76.09751892089844,
+      "logps/rejected": -85.20323944091797,
+      "loss": 2.5284,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9018669128417969,
+      "rewards/margins": 0.4894491136074066,
+      "rewards/rejected": -2.3913159370422363,
+      "step": 393
+    },
+    {
+      "epoch": 0.7157129881925522,
+      "grad_norm": 1.6489367485046387,
+      "learning_rate": 9.306346589685956e-06,
+      "logits/chosen": 0.14845696091651917,
+      "logits/rejected": 0.10356368124485016,
+      "logps/chosen": -83.94526672363281,
+      "logps/rejected": -86.53378295898438,
+      "loss": 2.3394,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9089899063110352,
+      "rewards/margins": 0.4619474411010742,
+      "rewards/rejected": -2.3709375858306885,
+      "step": 394
+    },
+    {
+      "epoch": 0.7175295186194369,
+      "grad_norm": 1.792077898979187,
+      "learning_rate": 9.302915004625435e-06,
+      "logits/chosen": 0.06936248391866684,
+      "logits/rejected": 0.13349927961826324,
+      "logps/chosen": -91.82075500488281,
+      "logps/rejected": -98.7391357421875,
+      "loss": 2.3875,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9371806383132935,
+      "rewards/margins": 0.6781354546546936,
+      "rewards/rejected": -2.6153156757354736,
+      "step": 395
+    },
+    {
+      "epoch": 0.7193460490463215,
+      "grad_norm": 2.049694538116455,
+      "learning_rate": 9.29947566475907e-06,
+      "logits/chosen": 0.08474650233983994,
+      "logits/rejected": 0.13380834460258484,
+      "logps/chosen": -84.59528350830078,
+      "logps/rejected": -96.24267578125,
+      "loss": 2.7001,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9375882148742676,
+      "rewards/margins": 0.4607384204864502,
+      "rewards/rejected": -2.3983266353607178,
+      "step": 396
+    },
+    {
+      "epoch": 0.7211625794732062,
+      "grad_norm": 1.9235875606536865,
+      "learning_rate": 9.296028577100271e-06,
+      "logits/chosen": 0.06255945563316345,
+      "logits/rejected": 0.17074811458587646,
+      "logps/chosen": -68.99498748779297,
+      "logps/rejected": -84.45899963378906,
+      "loss": 2.347,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9344629049301147,
+      "rewards/margins": 0.688963770866394,
+      "rewards/rejected": -2.6234264373779297,
+      "step": 397
+    },
+    {
+      "epoch": 0.7229791099000908,
+      "grad_norm": 1.6617276668548584,
+      "learning_rate": 9.292573748678254e-06,
+      "logits/chosen": 0.11962493509054184,
+      "logits/rejected": 0.11647717654705048,
+      "logps/chosen": -85.3626708984375,
+      "logps/rejected": -95.70562744140625,
+      "loss": 2.2229,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.8968594074249268,
+      "rewards/margins": 0.6918852925300598,
+      "rewards/rejected": -2.588744640350342,
+      "step": 398
+    },
+    {
+      "epoch": 0.7247956403269755,
+      "grad_norm": 1.4650444984436035,
+      "learning_rate": 9.289111186538013e-06,
+      "logits/chosen": 0.06390775740146637,
+      "logits/rejected": 0.11076060682535172,
+      "logps/chosen": -69.9189453125,
+      "logps/rejected": -81.5618667602539,
+      "loss": 2.2271,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.857581377029419,
+      "rewards/margins": 0.5540103316307068,
+      "rewards/rejected": -2.4115917682647705,
+      "step": 399
+    },
+    {
+      "epoch": 0.7266121707538601,
+      "grad_norm": 1.5728726387023926,
+      "learning_rate": 9.285640897740316e-06,
+      "logits/chosen": 0.08816932141780853,
+      "logits/rejected": 0.1330798715353012,
+      "logps/chosen": -78.04735565185547,
+      "logps/rejected": -87.89193725585938,
+      "loss": 2.3595,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.823161244392395,
+      "rewards/margins": 0.4568220376968384,
+      "rewards/rejected": -2.2799830436706543,
+      "step": 400
+    },
+    {
+      "epoch": 0.7284287011807448,
+      "grad_norm": 1.6291779279708862,
+      "learning_rate": 9.282162889361686e-06,
+      "logits/chosen": 0.17718347907066345,
+      "logits/rejected": 0.20603135228157043,
+      "logps/chosen": -67.99884033203125,
+      "logps/rejected": -75.93153381347656,
+      "loss": 2.3002,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.671633243560791,
+      "rewards/margins": 0.45543500781059265,
+      "rewards/rejected": -2.127068281173706,
+      "step": 401
+    },
+    {
+      "epoch": 0.7302452316076294,
+      "grad_norm": 1.5066276788711548,
+      "learning_rate": 9.278677168494388e-06,
+      "logits/chosen": 0.16840124130249023,
+      "logits/rejected": 0.17102479934692383,
+      "logps/chosen": -78.1556625366211,
+      "logps/rejected": -84.91495513916016,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9279460906982422,
+      "rewards/margins": 0.6659258008003235,
+      "rewards/rejected": -2.593871831893921,
+      "step": 402
+    },
+    {
+      "epoch": 0.7320617620345141,
+      "grad_norm": 1.71995210647583,
+      "learning_rate": 9.275183742246412e-06,
+      "logits/chosen": -0.0005522281862795353,
+      "logits/rejected": 0.1412215530872345,
+      "logps/chosen": -73.37832641601562,
+      "logps/rejected": -92.71226501464844,
+      "loss": 2.3315,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6880759000778198,
+      "rewards/margins": 0.5844776630401611,
+      "rewards/rejected": -2.2725534439086914,
+      "step": 403
+    },
+    {
+      "epoch": 0.7338782924613987,
+      "grad_norm": 1.676321029663086,
+      "learning_rate": 9.271682617741466e-06,
+      "logits/chosen": 0.15319044888019562,
+      "logits/rejected": 0.14503052830696106,
+      "logps/chosen": -75.85136413574219,
+      "logps/rejected": -82.61698150634766,
+      "loss": 2.3528,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.819403052330017,
+      "rewards/margins": 0.3919827342033386,
+      "rewards/rejected": -2.211385726928711,
+      "step": 404
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 1.4811294078826904,
+      "learning_rate": 9.268173802118949e-06,
+      "logits/chosen": 0.07122528553009033,
+      "logits/rejected": 0.12146291136741638,
+      "logps/chosen": -76.78916931152344,
+      "logps/rejected": -85.33805084228516,
+      "loss": 2.1505,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9886473417282104,
+      "rewards/margins": 0.5911746025085449,
+      "rewards/rejected": -2.579822301864624,
+      "step": 405
+    },
+    {
+      "epoch": 0.737511353315168,
+      "grad_norm": 1.629601001739502,
+      "learning_rate": 9.264657302533947e-06,
+      "logits/chosen": 0.11274963617324829,
+      "logits/rejected": 0.0651293620467186,
+      "logps/chosen": -82.86985778808594,
+      "logps/rejected": -88.68803405761719,
+      "loss": 2.1873,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8538885116577148,
+      "rewards/margins": 0.6726438403129578,
+      "rewards/rejected": -2.5265324115753174,
+      "step": 406
+    },
+    {
+      "epoch": 0.7393278837420527,
+      "grad_norm": 1.5687483549118042,
+      "learning_rate": 9.261133126157218e-06,
+      "logits/chosen": 0.06054290384054184,
+      "logits/rejected": 0.1543959081172943,
+      "logps/chosen": -68.5821304321289,
+      "logps/rejected": -80.544921875,
+      "loss": 2.3587,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7550634145736694,
+      "rewards/margins": 0.46927812695503235,
+      "rewards/rejected": -2.224341630935669,
+      "step": 407
+    },
+    {
+      "epoch": 0.7411444141689373,
+      "grad_norm": 1.849908471107483,
+      "learning_rate": 9.257601280175167e-06,
+      "logits/chosen": 0.15177126228809357,
+      "logits/rejected": 0.1485670655965805,
+      "logps/chosen": -72.81680297851562,
+      "logps/rejected": -80.58684539794922,
+      "loss": 2.444,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9739688634872437,
+      "rewards/margins": 0.6166737079620361,
+      "rewards/rejected": -2.5906424522399902,
+      "step": 408
+    },
+    {
+      "epoch": 0.742960944595822,
+      "grad_norm": 1.779342770576477,
+      "learning_rate": 9.254061771789847e-06,
+      "logits/chosen": 0.05093669891357422,
+      "logits/rejected": 0.1048060953617096,
+      "logps/chosen": -84.48141479492188,
+      "logps/rejected": -91.49285125732422,
+      "loss": 2.5324,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9952929019927979,
+      "rewards/margins": 0.3900759220123291,
+      "rewards/rejected": -2.385368585586548,
+      "step": 409
+    },
+    {
+      "epoch": 0.7447774750227066,
+      "grad_norm": 1.71133291721344,
+      "learning_rate": 9.25051460821893e-06,
+      "logits/chosen": 0.1308833658695221,
+      "logits/rejected": 0.08994461596012115,
+      "logps/chosen": -74.09046936035156,
+      "logps/rejected": -80.80438995361328,
+      "loss": 2.1895,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8457674980163574,
+      "rewards/margins": 0.5731581449508667,
+      "rewards/rejected": -2.4189257621765137,
+      "step": 410
+    },
+    {
+      "epoch": 0.7465940054495913,
+      "grad_norm": 1.6980112791061401,
+      "learning_rate": 9.2469597966957e-06,
+      "logits/chosen": 0.05856658145785332,
+      "logits/rejected": 0.08371179550886154,
+      "logps/chosen": -78.58245849609375,
+      "logps/rejected": -88.30054473876953,
+      "loss": 2.46,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9803493022918701,
+      "rewards/margins": 0.6115507483482361,
+      "rewards/rejected": -2.591899871826172,
+      "step": 411
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.9534665846725018e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-411/training_args.bin b/checkpoint-411/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-411/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-411/zero_to_fp32.py b/checkpoint-411/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-411/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-548/README.md b/checkpoint-548/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-548/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-548/adapter_config.json b/checkpoint-548/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-548/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-548/adapter_model.safetensors b/checkpoint-548/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..9ba1c3720ea70ec4360f6a448fe619c4aca7093f
--- /dev/null
+++ b/checkpoint-548/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95ceafc4bed69eef0d2d4335029b6da40b7f3a2103964721eeb2fb8a92140d74
+size 207244392
diff --git a/checkpoint-548/latest b/checkpoint-548/latest
new file mode 100644
index 0000000000000000000000000000000000000000..4de5b28f19e81a3c7caa8e8e1dac6f5caf01a9ae
--- /dev/null
+++ b/checkpoint-548/latest
@@ -0,0 +1 @@
+global_step548
\ No newline at end of file
diff --git a/checkpoint-548/rng_state_0.pth b/checkpoint-548/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b6473612e41c5cfd6973c2e71fa5f3ad2b2bcad1
--- /dev/null
+++ b/checkpoint-548/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984
diff --git a/checkpoint-548/rng_state_1.pth b/checkpoint-548/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..8506e00431b6ac7067699c0ea4f59adb6fa0ba20
--- /dev/null
+++ b/checkpoint-548/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984
diff --git a/checkpoint-548/rng_state_2.pth b/checkpoint-548/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..ea499e285c97cca07fedd34662c3d4ab44ff6f47
--- /dev/null
+++ b/checkpoint-548/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984
diff --git a/checkpoint-548/rng_state_3.pth b/checkpoint-548/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..aeb38f92f106ac3f08bae4f82179a8a12243bccb
--- /dev/null
+++ b/checkpoint-548/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984
diff --git a/checkpoint-548/rng_state_4.pth b/checkpoint-548/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..9d5856cb7a3f15092fa5593507022316916f648e
--- /dev/null
+++ b/checkpoint-548/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984
diff --git a/checkpoint-548/rng_state_5.pth b/checkpoint-548/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b824ee24d256695aad4a69a62d8e7125f51a17f2
--- /dev/null
+++ b/checkpoint-548/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984
diff --git a/checkpoint-548/rng_state_6.pth b/checkpoint-548/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a9fd0364bb8f1a8e91eca45be5e1b6672b4d9afd
--- /dev/null
+++ b/checkpoint-548/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984
diff --git a/checkpoint-548/rng_state_7.pth b/checkpoint-548/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..4e80125fd18efcb1097384319888b699f4dce7e7
--- /dev/null
+++ b/checkpoint-548/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984
diff --git a/checkpoint-548/special_tokens_map.json b/checkpoint-548/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-548/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-548/tokenizer.json b/checkpoint-548/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-548/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-548/tokenizer_config.json b/checkpoint-548/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-548/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-548/trainer_state.json b/checkpoint-548/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4b4e18e7017fb30cd14292c55e468a39fee17a3e
--- /dev/null
+++ b/checkpoint-548/trainer_state.json
@@ -0,0 +1,8253 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9954586739327884,
+  "eval_steps": 500,
+  "global_step": 548,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    },
+    {
+      "epoch": 0.49954586739327883,
+      "grad_norm": 2.1769776344299316,
+      "learning_rate": 9.65745789630079e-06,
+      "logits/chosen": 0.11112834513187408,
+      "logits/rejected": 0.10816515237092972,
+      "logps/chosen": -83.5405044555664,
+      "logps/rejected": -83.06329345703125,
+      "loss": 3.0939,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.142024040222168,
+      "rewards/margins": 0.1820124089717865,
+      "rewards/rejected": -2.3240363597869873,
+      "step": 275
+    },
+    {
+      "epoch": 0.5013623978201635,
+      "grad_norm": 1.7329221963882446,
+      "learning_rate": 9.654994540260396e-06,
+      "logits/chosen": 0.0653618574142456,
+      "logits/rejected": 0.08004368096590042,
+      "logps/chosen": -80.83209228515625,
+      "logps/rejected": -82.97142028808594,
+      "loss": 2.7161,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.851841688156128,
+      "rewards/margins": 0.2948168218135834,
+      "rewards/rejected": -2.146658420562744,
+      "step": 276
+    },
+    {
+      "epoch": 0.5031789282470481,
+      "grad_norm": 1.592657208442688,
+      "learning_rate": 9.65252271146268e-06,
+      "logits/chosen": 0.09880789369344711,
+      "logits/rejected": 0.14229761064052582,
+      "logps/chosen": -67.1727294921875,
+      "logps/rejected": -75.03417205810547,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8714643716812134,
+      "rewards/margins": 0.37937116622924805,
+      "rewards/rejected": -2.250835657119751,
+      "step": 277
+    },
+    {
+      "epoch": 0.5049954586739328,
+      "grad_norm": 1.6487712860107422,
+      "learning_rate": 9.650042414948133e-06,
+      "logits/chosen": 0.13465353846549988,
+      "logits/rejected": 0.12865689396858215,
+      "logps/chosen": -76.4417724609375,
+      "logps/rejected": -78.6947021484375,
+      "loss": 2.6085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.875291347503662,
+      "rewards/margins": 0.35175251960754395,
+      "rewards/rejected": -2.227043867111206,
+      "step": 278
+    },
+    {
+      "epoch": 0.5068119891008175,
+      "grad_norm": 1.6523009538650513,
+      "learning_rate": 9.64755365577451e-06,
+      "logits/chosen": 0.04238567873835564,
+      "logits/rejected": 0.07994347810745239,
+      "logps/chosen": -76.92097473144531,
+      "logps/rejected": -83.20886993408203,
+      "loss": 2.3156,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9759610891342163,
+      "rewards/margins": 0.5398927330970764,
+      "rewards/rejected": -2.5158536434173584,
+      "step": 279
+    },
+    {
+      "epoch": 0.508628519527702,
+      "grad_norm": 1.7999261617660522,
+      "learning_rate": 9.645056439016827e-06,
+      "logits/chosen": 0.07349395751953125,
+      "logits/rejected": 0.07518415153026581,
+      "logps/chosen": -77.56079864501953,
+      "logps/rejected": -84.97645568847656,
+      "loss": 2.8085,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9408115148544312,
+      "rewards/margins": 0.3470652997493744,
+      "rewards/rejected": -2.287877082824707,
+      "step": 280
+    },
+    {
+      "epoch": 0.5104450499545867,
+      "grad_norm": 1.5684200525283813,
+      "learning_rate": 9.642550769767342e-06,
+      "logits/chosen": 0.16188879311084747,
+      "logits/rejected": 0.12772323191165924,
+      "logps/chosen": -89.29315185546875,
+      "logps/rejected": -94.35065460205078,
+      "loss": 2.2314,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8958841562271118,
+      "rewards/margins": 0.5102108716964722,
+      "rewards/rejected": -2.406095027923584,
+      "step": 281
+    },
+    {
+      "epoch": 0.5122615803814714,
+      "grad_norm": 1.7878178358078003,
+      "learning_rate": 9.640036653135548e-06,
+      "logits/chosen": 0.060573749244213104,
+      "logits/rejected": 0.13457715511322021,
+      "logps/chosen": -68.9404525756836,
+      "logps/rejected": -74.77693176269531,
+      "loss": 2.4359,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.889530897140503,
+      "rewards/margins": 0.45911547541618347,
+      "rewards/rejected": -2.348646402359009,
+      "step": 282
+    },
+    {
+      "epoch": 0.5140781108083561,
+      "grad_norm": 1.4741288423538208,
+      "learning_rate": 9.637514094248172e-06,
+      "logits/chosen": 0.10433132946491241,
+      "logits/rejected": 0.1439315229654312,
+      "logps/chosen": -74.89447021484375,
+      "logps/rejected": -86.58551788330078,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9090807437896729,
+      "rewards/margins": 0.6733560562133789,
+      "rewards/rejected": -2.5824368000030518,
+      "step": 283
+    },
+    {
+      "epoch": 0.5158946412352406,
+      "grad_norm": 1.2864018678665161,
+      "learning_rate": 9.634983098249146e-06,
+      "logits/chosen": 0.10891089588403702,
+      "logits/rejected": 0.11755162477493286,
+      "logps/chosen": -66.90985107421875,
+      "logps/rejected": -71.8125228881836,
+      "loss": 2.1538,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.7779250144958496,
+      "rewards/margins": 0.5625147223472595,
+      "rewards/rejected": -2.340439796447754,
+      "step": 284
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 1.465747594833374,
+      "learning_rate": 9.632443670299616e-06,
+      "logits/chosen": 0.08224496245384216,
+      "logits/rejected": 0.12130744755268097,
+      "logps/chosen": -75.4281997680664,
+      "logps/rejected": -85.0781021118164,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7217226028442383,
+      "rewards/margins": 0.4750482439994812,
+      "rewards/rejected": -2.196770668029785,
+      "step": 285
+    },
+    {
+      "epoch": 0.51952770208901,
+      "grad_norm": 1.492859959602356,
+      "learning_rate": 9.629895815577915e-06,
+      "logits/chosen": 0.06619664281606674,
+      "logits/rejected": 0.13152630627155304,
+      "logps/chosen": -96.65383911132812,
+      "logps/rejected": -111.93521881103516,
+      "loss": 2.2831,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7611618041992188,
+      "rewards/margins": 0.5886337757110596,
+      "rewards/rejected": -2.349795341491699,
+      "step": 286
+    },
+    {
+      "epoch": 0.5213442325158947,
+      "grad_norm": 1.5534065961837769,
+      "learning_rate": 9.627339539279564e-06,
+      "logits/chosen": 0.06637927144765854,
+      "logits/rejected": 0.09107412397861481,
+      "logps/chosen": -71.92534637451172,
+      "logps/rejected": -82.98391723632812,
+      "loss": 2.5101,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7867075204849243,
+      "rewards/margins": 0.39493298530578613,
+      "rewards/rejected": -2.181640625,
+      "step": 287
+    },
+    {
+      "epoch": 0.5231607629427792,
+      "grad_norm": 1.6778221130371094,
+      "learning_rate": 9.624774846617254e-06,
+      "logits/chosen": 0.14700329303741455,
+      "logits/rejected": 0.12778782844543457,
+      "logps/chosen": -65.2364730834961,
+      "logps/rejected": -74.39017486572266,
+      "loss": 2.5236,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6942007541656494,
+      "rewards/margins": 0.35334131121635437,
+      "rewards/rejected": -2.047542095184326,
+      "step": 288
+    },
+    {
+      "epoch": 0.5249772933696639,
+      "grad_norm": 1.462215542793274,
+      "learning_rate": 9.622201742820839e-06,
+      "logits/chosen": 0.11467991769313812,
+      "logits/rejected": 0.1028795838356018,
+      "logps/chosen": -67.85939025878906,
+      "logps/rejected": -74.3462142944336,
+      "loss": 2.3891,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7194863557815552,
+      "rewards/margins": 0.3846975564956665,
+      "rewards/rejected": -2.1041836738586426,
+      "step": 289
+    },
+    {
+      "epoch": 0.5267938237965486,
+      "grad_norm": 1.7591735124588013,
+      "learning_rate": 9.619620233137328e-06,
+      "logits/chosen": 0.08407986164093018,
+      "logits/rejected": 0.12201236933469772,
+      "logps/chosen": -82.57340240478516,
+      "logps/rejected": -86.94192504882812,
+      "loss": 2.4569,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8536537885665894,
+      "rewards/margins": 0.38806283473968506,
+      "rewards/rejected": -2.2417166233062744,
+      "step": 290
+    },
+    {
+      "epoch": 0.5286103542234333,
+      "grad_norm": 1.5584774017333984,
+      "learning_rate": 9.617030322830868e-06,
+      "logits/chosen": 0.1126309484243393,
+      "logits/rejected": 0.08907752484083176,
+      "logps/chosen": -73.36901092529297,
+      "logps/rejected": -76.65506744384766,
+      "loss": 2.4021,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8091729879379272,
+      "rewards/margins": 0.3603411316871643,
+      "rewards/rejected": -2.1695141792297363,
+      "step": 291
+    },
+    {
+      "epoch": 0.5304268846503178,
+      "grad_norm": 1.487938642501831,
+      "learning_rate": 9.614432017182736e-06,
+      "logits/chosen": 0.09620118141174316,
+      "logits/rejected": 0.08514149487018585,
+      "logps/chosen": -88.66265106201172,
+      "logps/rejected": -92.26091766357422,
+      "loss": 2.3636,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.5980674028396606,
+      "rewards/margins": 0.4888036549091339,
+      "rewards/rejected": -2.0868711471557617,
+      "step": 292
+    },
+    {
+      "epoch": 0.5322434150772025,
+      "grad_norm": 1.445408582687378,
+      "learning_rate": 9.611825321491331e-06,
+      "logits/chosen": 0.10014252364635468,
+      "logits/rejected": 0.12857215106487274,
+      "logps/chosen": -82.46649932861328,
+      "logps/rejected": -92.84205627441406,
+      "loss": 2.343,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4620139598846436,
+      "rewards/margins": 0.5616316199302673,
+      "rewards/rejected": -2.0236456394195557,
+      "step": 293
+    },
+    {
+      "epoch": 0.5340599455040872,
+      "grad_norm": 1.262193202972412,
+      "learning_rate": 9.609210241072158e-06,
+      "logits/chosen": 0.09868282079696655,
+      "logits/rejected": 0.16074486076831818,
+      "logps/chosen": -67.0407485961914,
+      "logps/rejected": -82.36201477050781,
+      "loss": 2.1163,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5362765789031982,
+      "rewards/margins": 0.7091963887214661,
+      "rewards/rejected": -2.2454731464385986,
+      "step": 294
+    },
+    {
+      "epoch": 0.5358764759309719,
+      "grad_norm": 1.601891279220581,
+      "learning_rate": 9.606586781257822e-06,
+      "logits/chosen": 0.1236480325460434,
+      "logits/rejected": 0.12187935411930084,
+      "logps/chosen": -78.2726058959961,
+      "logps/rejected": -84.49217987060547,
+      "loss": 2.5075,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.779215931892395,
+      "rewards/margins": 0.38987410068511963,
+      "rewards/rejected": -2.1690900325775146,
+      "step": 295
+    },
+    {
+      "epoch": 0.5376930063578564,
+      "grad_norm": 1.4211548566818237,
+      "learning_rate": 9.603954947398016e-06,
+      "logits/chosen": 0.18375667929649353,
+      "logits/rejected": 0.21945703029632568,
+      "logps/chosen": -70.29829406738281,
+      "logps/rejected": -78.42284393310547,
+      "loss": 2.1959,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.7448753118515015,
+      "rewards/margins": 0.5106840133666992,
+      "rewards/rejected": -2.2555594444274902,
+      "step": 296
+    },
+    {
+      "epoch": 0.5395095367847411,
+      "grad_norm": 1.458617925643921,
+      "learning_rate": 9.601314744859504e-06,
+      "logits/chosen": 0.07360847294330597,
+      "logits/rejected": 0.1596718430519104,
+      "logps/chosen": -72.28546142578125,
+      "logps/rejected": -87.13935089111328,
+      "loss": 2.3538,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8007893562316895,
+      "rewards/margins": 0.5433144569396973,
+      "rewards/rejected": -2.3441038131713867,
+      "step": 297
+    },
+    {
+      "epoch": 0.5413260672116258,
+      "grad_norm": 1.6526339054107666,
+      "learning_rate": 9.598666179026123e-06,
+      "logits/chosen": 0.12013350427150726,
+      "logits/rejected": 0.10237178206443787,
+      "logps/chosen": -83.06564331054688,
+      "logps/rejected": -85.60771942138672,
+      "loss": 2.4531,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.911988615989685,
+      "rewards/margins": 0.35885563492774963,
+      "rewards/rejected": -2.2708444595336914,
+      "step": 298
+    },
+    {
+      "epoch": 0.5431425976385105,
+      "grad_norm": 1.9808340072631836,
+      "learning_rate": 9.596009255298755e-06,
+      "logits/chosen": 0.062342576682567596,
+      "logits/rejected": 0.09053834527730942,
+      "logps/chosen": -96.33489990234375,
+      "logps/rejected": -93.43024444580078,
+      "loss": 2.8549,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.845227599143982,
+      "rewards/margins": 0.22291362285614014,
+      "rewards/rejected": -2.068141222000122,
+      "step": 299
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 1.6276236772537231,
+      "learning_rate": 9.593343979095334e-06,
+      "logits/chosen": 0.2073422521352768,
+      "logits/rejected": 0.13107003271579742,
+      "logps/chosen": -72.45758056640625,
+      "logps/rejected": -77.42770385742188,
+      "loss": 2.4644,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9407453536987305,
+      "rewards/margins": 0.3134301006793976,
+      "rewards/rejected": -2.2541751861572266,
+      "step": 300
+    },
+    {
+      "epoch": 0.5467756584922797,
+      "grad_norm": 1.5001753568649292,
+      "learning_rate": 9.590670355850819e-06,
+      "logits/chosen": 0.15097060799598694,
+      "logits/rejected": 0.16254279017448425,
+      "logps/chosen": -70.44524383544922,
+      "logps/rejected": -70.64158630371094,
+      "loss": 2.4159,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.6921097040176392,
+      "rewards/margins": 0.3336741626262665,
+      "rewards/rejected": -2.0257837772369385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5485921889191644,
+      "grad_norm": 1.8360233306884766,
+      "learning_rate": 9.587988391017198e-06,
+      "logits/chosen": 0.14593760669231415,
+      "logits/rejected": 0.12311654537916183,
+      "logps/chosen": -78.32576751708984,
+      "logps/rejected": -88.25840759277344,
+      "loss": 2.5826,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.6591644287109375,
+      "rewards/margins": 0.5071319341659546,
+      "rewards/rejected": -2.1662964820861816,
+      "step": 302
+    },
+    {
+      "epoch": 0.5504087193460491,
+      "grad_norm": 1.4822838306427002,
+      "learning_rate": 9.585298090063459e-06,
+      "logits/chosen": 0.20818498730659485,
+      "logits/rejected": 0.16436657309532166,
+      "logps/chosen": -67.59429931640625,
+      "logps/rejected": -77.14763641357422,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6798239946365356,
+      "rewards/margins": 0.5954271554946899,
+      "rewards/rejected": -2.2752511501312256,
+      "step": 303
+    },
+    {
+      "epoch": 0.5522252497729337,
+      "grad_norm": 1.6118305921554565,
+      "learning_rate": 9.582599458475598e-06,
+      "logits/chosen": 0.0827561467885971,
+      "logits/rejected": 0.09151773154735565,
+      "logps/chosen": -73.28964233398438,
+      "logps/rejected": -76.90730285644531,
+      "loss": 2.3792,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.8056440353393555,
+      "rewards/margins": 0.39551618695259094,
+      "rewards/rejected": -2.201160192489624,
+      "step": 304
+    },
+    {
+      "epoch": 0.5540417801998183,
+      "grad_norm": 1.4609856605529785,
+      "learning_rate": 9.579892501756593e-06,
+      "logits/chosen": 0.05332394689321518,
+      "logits/rejected": 0.1590057611465454,
+      "logps/chosen": -77.51653289794922,
+      "logps/rejected": -95.51261138916016,
+      "loss": 2.2105,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.740310549736023,
+      "rewards/margins": 0.6771562695503235,
+      "rewards/rejected": -2.4174671173095703,
+      "step": 305
+    },
+    {
+      "epoch": 0.555858310626703,
+      "grad_norm": 1.7181960344314575,
+      "learning_rate": 9.5771772254264e-06,
+      "logits/chosen": 0.05865276977419853,
+      "logits/rejected": 0.03500773757696152,
+      "logps/chosen": -83.72650146484375,
+      "logps/rejected": -93.69804382324219,
+      "loss": 2.1687,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6025928258895874,
+      "rewards/margins": 0.7339029312133789,
+      "rewards/rejected": -2.3364956378936768,
+      "step": 306
+    },
+    {
+      "epoch": 0.5576748410535877,
+      "grad_norm": 1.497698187828064,
+      "learning_rate": 9.57445363502194e-06,
+      "logits/chosen": 0.1225174218416214,
+      "logits/rejected": 0.09580346196889877,
+      "logps/chosen": -76.27323150634766,
+      "logps/rejected": -76.90716552734375,
+      "loss": 2.2116,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.715336799621582,
+      "rewards/margins": 0.4648086428642273,
+      "rewards/rejected": -2.180145263671875,
+      "step": 307
+    },
+    {
+      "epoch": 0.5594913714804723,
+      "grad_norm": 1.3957064151763916,
+      "learning_rate": 9.571721736097089e-06,
+      "logits/chosen": 0.07596514374017715,
+      "logits/rejected": 0.14832191169261932,
+      "logps/chosen": -84.65274810791016,
+      "logps/rejected": -105.76422119140625,
+      "loss": 1.9942,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.6643587350845337,
+      "rewards/margins": 0.7884883284568787,
+      "rewards/rejected": -2.4528470039367676,
+      "step": 308
+    },
+    {
+      "epoch": 0.5613079019073569,
+      "grad_norm": 1.6824349164962769,
+      "learning_rate": 9.568981534222664e-06,
+      "logits/chosen": 0.049865882843732834,
+      "logits/rejected": 0.023348212242126465,
+      "logps/chosen": -80.3697509765625,
+      "logps/rejected": -86.07074737548828,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.009049654006958,
+      "rewards/margins": 0.5470253825187683,
+      "rewards/rejected": -2.556075096130371,
+      "step": 309
+    },
+    {
+      "epoch": 0.5631244323342416,
+      "grad_norm": 1.8752086162567139,
+      "learning_rate": 9.566233034986413e-06,
+      "logits/chosen": 0.07250591367483139,
+      "logits/rejected": 0.12809628248214722,
+      "logps/chosen": -74.84794616699219,
+      "logps/rejected": -81.79740905761719,
+      "loss": 2.6286,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.75887131690979,
+      "rewards/margins": 0.33979111909866333,
+      "rewards/rejected": -2.0986623764038086,
+      "step": 310
+    },
+    {
+      "epoch": 0.5649409627611263,
+      "grad_norm": 1.7387233972549438,
+      "learning_rate": 9.563476243993008e-06,
+      "logits/chosen": 0.130618155002594,
+      "logits/rejected": 0.12952059507369995,
+      "logps/chosen": -80.75495910644531,
+      "logps/rejected": -90.4281234741211,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.836004614830017,
+      "rewards/margins": 0.4418962597846985,
+      "rewards/rejected": -2.2779006958007812,
+      "step": 311
+    },
+    {
+      "epoch": 0.5667574931880109,
+      "grad_norm": 1.5147444009780884,
+      "learning_rate": 9.56071116686402e-06,
+      "logits/chosen": 0.10193713754415512,
+      "logits/rejected": 0.22481802105903625,
+      "logps/chosen": -73.21926879882812,
+      "logps/rejected": -81.27864837646484,
+      "loss": 2.578,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6358015537261963,
+      "rewards/margins": 0.38279739022254944,
+      "rewards/rejected": -2.018598794937134,
+      "step": 312
+    },
+    {
+      "epoch": 0.5685740236148955,
+      "grad_norm": 1.6253665685653687,
+      "learning_rate": 9.557937809237927e-06,
+      "logits/chosen": 0.09468917548656464,
+      "logits/rejected": 0.09415112435817719,
+      "logps/chosen": -80.62995147705078,
+      "logps/rejected": -86.65946197509766,
+      "loss": 2.3854,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8635404109954834,
+      "rewards/margins": 0.4438764154911041,
+      "rewards/rejected": -2.3074169158935547,
+      "step": 313
+    },
+    {
+      "epoch": 0.5703905540417802,
+      "grad_norm": 1.7893344163894653,
+      "learning_rate": 9.555156176770087e-06,
+      "logits/chosen": 0.15863659977912903,
+      "logits/rejected": 0.09485571831464767,
+      "logps/chosen": -76.10442352294922,
+      "logps/rejected": -72.73162078857422,
+      "loss": 2.6795,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.935206413269043,
+      "rewards/margins": 0.22100940346717834,
+      "rewards/rejected": -2.1562156677246094,
+      "step": 314
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 1.696327805519104,
+      "learning_rate": 9.552366275132733e-06,
+      "logits/chosen": 0.07012113183736801,
+      "logits/rejected": 0.13891686499118805,
+      "logps/chosen": -78.08012390136719,
+      "logps/rejected": -83.05044555664062,
+      "loss": 2.6181,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7869051694869995,
+      "rewards/margins": 0.23424415290355682,
+      "rewards/rejected": -2.0211493968963623,
+      "step": 315
+    },
+    {
+      "epoch": 0.5740236148955495,
+      "grad_norm": 1.6044729948043823,
+      "learning_rate": 9.54956811001496e-06,
+      "logits/chosen": 0.11298641562461853,
+      "logits/rejected": 0.2595204710960388,
+      "logps/chosen": -75.5998306274414,
+      "logps/rejected": -85.31849670410156,
+      "loss": 2.5849,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.5757699012756348,
+      "rewards/margins": 0.35041162371635437,
+      "rewards/rejected": -1.9261815547943115,
+      "step": 316
+    },
+    {
+      "epoch": 0.5758401453224341,
+      "grad_norm": 1.7399548292160034,
+      "learning_rate": 9.546761687122715e-06,
+      "logits/chosen": 0.14933931827545166,
+      "logits/rejected": 0.1868455857038498,
+      "logps/chosen": -68.05965423583984,
+      "logps/rejected": -82.37442779541016,
+      "loss": 2.3208,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8474600315093994,
+      "rewards/margins": 0.5766161680221558,
+      "rewards/rejected": -2.4240763187408447,
+      "step": 317
+    },
+    {
+      "epoch": 0.5776566757493188,
+      "grad_norm": 1.597495675086975,
+      "learning_rate": 9.54394701217878e-06,
+      "logits/chosen": 0.04729313403367996,
+      "logits/rejected": 0.05815067142248154,
+      "logps/chosen": -97.36541748046875,
+      "logps/rejected": -104.02650451660156,
+      "loss": 2.2106,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8631618022918701,
+      "rewards/margins": 0.6027428507804871,
+      "rewards/rejected": -2.465904474258423,
+      "step": 318
+    },
+    {
+      "epoch": 0.5794732061762035,
+      "grad_norm": 1.8646149635314941,
+      "learning_rate": 9.541124090922771e-06,
+      "logits/chosen": 0.1399674415588379,
+      "logits/rejected": 0.11107950657606125,
+      "logps/chosen": -82.39740753173828,
+      "logps/rejected": -76.80902099609375,
+      "loss": 2.6435,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.7670300006866455,
+      "rewards/margins": 0.15345275402069092,
+      "rewards/rejected": -1.920482873916626,
+      "step": 319
+    },
+    {
+      "epoch": 0.5812897366030881,
+      "grad_norm": 1.502447247505188,
+      "learning_rate": 9.538292929111114e-06,
+      "logits/chosen": 0.10672347247600555,
+      "logits/rejected": 0.08290571719408035,
+      "logps/chosen": -74.72683715820312,
+      "logps/rejected": -86.37158203125,
+      "loss": 2.2137,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8041073083877563,
+      "rewards/margins": 0.596272885799408,
+      "rewards/rejected": -2.4003803730010986,
+      "step": 320
+    },
+    {
+      "epoch": 0.5831062670299727,
+      "grad_norm": 1.7287745475769043,
+      "learning_rate": 9.535453532517039e-06,
+      "logits/chosen": 0.13742896914482117,
+      "logits/rejected": 0.07607734203338623,
+      "logps/chosen": -86.82475280761719,
+      "logps/rejected": -77.8934555053711,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7918696403503418,
+      "rewards/margins": 0.2591831088066101,
+      "rewards/rejected": -2.0510525703430176,
+      "step": 321
+    },
+    {
+      "epoch": 0.5849227974568574,
+      "grad_norm": 1.5914596319198608,
+      "learning_rate": 9.532605906930575e-06,
+      "logits/chosen": 0.11301672458648682,
+      "logits/rejected": 0.1944851577281952,
+      "logps/chosen": -68.9898452758789,
+      "logps/rejected": -73.17451477050781,
+      "loss": 2.4565,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6999026536941528,
+      "rewards/margins": 0.30462026596069336,
+      "rewards/rejected": -2.0045228004455566,
+      "step": 322
+    },
+    {
+      "epoch": 0.5867393278837421,
+      "grad_norm": 1.6072031259536743,
+      "learning_rate": 9.529750058158522e-06,
+      "logits/chosen": 0.07092760503292084,
+      "logits/rejected": 0.08035591244697571,
+      "logps/chosen": -81.5555419921875,
+      "logps/rejected": -81.93560028076172,
+      "loss": 2.2729,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6834638118743896,
+      "rewards/margins": 0.48882579803466797,
+      "rewards/rejected": -2.1722893714904785,
+      "step": 323
+    },
+    {
+      "epoch": 0.5885558583106267,
+      "grad_norm": 1.6433221101760864,
+      "learning_rate": 9.526885992024453e-06,
+      "logits/chosen": 0.13823899626731873,
+      "logits/rejected": 0.13610433042049408,
+      "logps/chosen": -78.9625244140625,
+      "logps/rejected": -83.2666015625,
+      "loss": 2.2075,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6939194202423096,
+      "rewards/margins": 0.5755473971366882,
+      "rewards/rejected": -2.2694668769836426,
+      "step": 324
+    },
+    {
+      "epoch": 0.5903723887375113,
+      "grad_norm": 1.7674918174743652,
+      "learning_rate": 9.524013714368702e-06,
+      "logits/chosen": 0.1704932153224945,
+      "logits/rejected": 0.1073535829782486,
+      "logps/chosen": -72.47895050048828,
+      "logps/rejected": -73.0809326171875,
+      "loss": 2.5473,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8390110731124878,
+      "rewards/margins": 0.34467947483062744,
+      "rewards/rejected": -2.1836905479431152,
+      "step": 325
+    },
+    {
+      "epoch": 0.592188919164396,
+      "grad_norm": 1.8219250440597534,
+      "learning_rate": 9.521133231048338e-06,
+      "logits/chosen": 0.07941028475761414,
+      "logits/rejected": 0.13045310974121094,
+      "logps/chosen": -80.73494720458984,
+      "logps/rejected": -91.58990478515625,
+      "loss": 2.4866,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8594518899917603,
+      "rewards/margins": 0.4554288685321808,
+      "rewards/rejected": -2.314880847930908,
+      "step": 326
+    },
+    {
+      "epoch": 0.5940054495912807,
+      "grad_norm": 1.3753328323364258,
+      "learning_rate": 9.51824454793717e-06,
+      "logits/chosen": 0.08879546821117401,
+      "logits/rejected": 0.04693777486681938,
+      "logps/chosen": -82.9569091796875,
+      "logps/rejected": -91.44571685791016,
+      "loss": 2.2602,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.549119472503662,
+      "rewards/margins": 0.5145683288574219,
+      "rewards/rejected": -2.063688039779663,
+      "step": 327
+    },
+    {
+      "epoch": 0.5958219800181653,
+      "grad_norm": 1.4296562671661377,
+      "learning_rate": 9.515347670925728e-06,
+      "logits/chosen": 0.15614314377307892,
+      "logits/rejected": 0.1598319411277771,
+      "logps/chosen": -75.60345458984375,
+      "logps/rejected": -80.81770324707031,
+      "loss": 2.3946,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.556883454322815,
+      "rewards/margins": 0.4525718092918396,
+      "rewards/rejected": -2.0094552040100098,
+      "step": 328
+    },
+    {
+      "epoch": 0.59763851044505,
+      "grad_norm": 1.7932195663452148,
+      "learning_rate": 9.512442605921245e-06,
+      "logits/chosen": 0.05863601714372635,
+      "logits/rejected": 0.0788806602358818,
+      "logps/chosen": -75.95397186279297,
+      "logps/rejected": -84.50177001953125,
+      "loss": 2.5085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9159862995147705,
+      "rewards/margins": 0.4600725769996643,
+      "rewards/rejected": -2.376059055328369,
+      "step": 329
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 1.6558383703231812,
+      "learning_rate": 9.509529358847655e-06,
+      "logits/chosen": 0.08205496519804001,
+      "logits/rejected": 0.13091425597667694,
+      "logps/chosen": -82.73474884033203,
+      "logps/rejected": -94.02273559570312,
+      "loss": 2.4419,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.083285093307495,
+      "rewards/margins": 0.5300815105438232,
+      "rewards/rejected": -2.6133666038513184,
+      "step": 330
+    },
+    {
+      "epoch": 0.6012715712988193,
+      "grad_norm": 1.7888520956039429,
+      "learning_rate": 9.506607935645579e-06,
+      "logits/chosen": 0.11793217062950134,
+      "logits/rejected": 0.19568441808223724,
+      "logps/chosen": -79.86200714111328,
+      "logps/rejected": -89.25471496582031,
+      "loss": 2.4273,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9273384809494019,
+      "rewards/margins": 0.5115458965301514,
+      "rewards/rejected": -2.4388844966888428,
+      "step": 331
+    },
+    {
+      "epoch": 0.6030881017257039,
+      "grad_norm": 2.0436160564422607,
+      "learning_rate": 9.503678342272306e-06,
+      "logits/chosen": 0.043473344296216965,
+      "logits/rejected": 0.15554016828536987,
+      "logps/chosen": -75.2901611328125,
+      "logps/rejected": -83.20555877685547,
+      "loss": 2.6941,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.725769281387329,
+      "rewards/margins": 0.2789250314235687,
+      "rewards/rejected": -2.0046942234039307,
+      "step": 332
+    },
+    {
+      "epoch": 0.6049046321525886,
+      "grad_norm": 1.665578842163086,
+      "learning_rate": 9.500740584701785e-06,
+      "logits/chosen": 0.17119848728179932,
+      "logits/rejected": 0.14128939807415009,
+      "logps/chosen": -83.1322021484375,
+      "logps/rejected": -93.38603210449219,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.8975954055786133,
+      "rewards/margins": 0.7810046672821045,
+      "rewards/rejected": -2.6786000728607178,
+      "step": 333
+    },
+    {
+      "epoch": 0.6067211625794732,
+      "grad_norm": 1.9258419275283813,
+      "learning_rate": 9.497794668924617e-06,
+      "logits/chosen": 0.028591612353920937,
+      "logits/rejected": 0.08894480764865875,
+      "logps/chosen": -78.12230682373047,
+      "logps/rejected": -90.00184631347656,
+      "loss": 2.4006,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8113291263580322,
+      "rewards/margins": 0.5277000665664673,
+      "rewards/rejected": -2.339029550552368,
+      "step": 334
+    },
+    {
+      "epoch": 0.6085376930063578,
+      "grad_norm": 1.4878523349761963,
+      "learning_rate": 9.494840600948038e-06,
+      "logits/chosen": 0.03548199310898781,
+      "logits/rejected": 0.09170147776603699,
+      "logps/chosen": -73.15607452392578,
+      "logps/rejected": -83.77317810058594,
+      "loss": 2.3301,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6781296730041504,
+      "rewards/margins": 0.4735565185546875,
+      "rewards/rejected": -2.151685953140259,
+      "step": 335
+    },
+    {
+      "epoch": 0.6103542234332425,
+      "grad_norm": 1.8042774200439453,
+      "learning_rate": 9.491878386795906e-06,
+      "logits/chosen": 0.046132348477840424,
+      "logits/rejected": 0.0721711814403534,
+      "logps/chosen": -81.01045227050781,
+      "logps/rejected": -89.28679656982422,
+      "loss": 2.5593,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.012852191925049,
+      "rewards/margins": 0.44549623131752014,
+      "rewards/rejected": -2.458348512649536,
+      "step": 336
+    },
+    {
+      "epoch": 0.6121707538601272,
+      "grad_norm": 2.048952102661133,
+      "learning_rate": 9.488908032508691e-06,
+      "logits/chosen": 0.10774732381105423,
+      "logits/rejected": 0.11935572326183319,
+      "logps/chosen": -91.25210571289062,
+      "logps/rejected": -90.42224884033203,
+      "loss": 2.8282,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0127718448638916,
+      "rewards/margins": 0.16889013350009918,
+      "rewards/rejected": -2.18166184425354,
+      "step": 337
+    },
+    {
+      "epoch": 0.6139872842870118,
+      "grad_norm": 1.6447219848632812,
+      "learning_rate": 9.485929544143462e-06,
+      "logits/chosen": 0.12260966747999191,
+      "logits/rejected": 0.1517799198627472,
+      "logps/chosen": -72.49166870117188,
+      "logps/rejected": -82.3083724975586,
+      "loss": 2.4029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.624878168106079,
+      "rewards/margins": 0.5487539768218994,
+      "rewards/rejected": -2.1736321449279785,
+      "step": 338
+    },
+    {
+      "epoch": 0.6158038147138964,
+      "grad_norm": 1.599530577659607,
+      "learning_rate": 9.482942927773876e-06,
+      "logits/chosen": 0.09051798284053802,
+      "logits/rejected": 0.11202570050954819,
+      "logps/chosen": -84.42019653320312,
+      "logps/rejected": -92.61677551269531,
+      "loss": 2.2269,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6424546241760254,
+      "rewards/margins": 0.5715836882591248,
+      "rewards/rejected": -2.214038133621216,
+      "step": 339
+    },
+    {
+      "epoch": 0.6176203451407811,
+      "grad_norm": 1.5411081314086914,
+      "learning_rate": 9.479948189490164e-06,
+      "logits/chosen": 0.06563656777143478,
+      "logits/rejected": 0.17349205911159515,
+      "logps/chosen": -70.98219299316406,
+      "logps/rejected": -78.88846588134766,
+      "loss": 2.3328,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.824373722076416,
+      "rewards/margins": 0.46550822257995605,
+      "rewards/rejected": -2.289881706237793,
+      "step": 340
+    },
+    {
+      "epoch": 0.6194368755676658,
+      "grad_norm": 1.8855030536651611,
+      "learning_rate": 9.476945335399122e-06,
+      "logits/chosen": 0.11399642378091812,
+      "logits/rejected": 0.12690946459770203,
+      "logps/chosen": -85.86933898925781,
+      "logps/rejected": -89.75601196289062,
+      "loss": 2.4269,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9560539722442627,
+      "rewards/margins": 0.38427016139030457,
+      "rewards/rejected": -2.3403239250183105,
+      "step": 341
+    },
+    {
+      "epoch": 0.6212534059945504,
+      "grad_norm": 1.9973480701446533,
+      "learning_rate": 9.473934371624087e-06,
+      "logits/chosen": 0.029105912894010544,
+      "logits/rejected": 0.12763622403144836,
+      "logps/chosen": -80.68119049072266,
+      "logps/rejected": -93.90747833251953,
+      "loss": 2.4709,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.906477928161621,
+      "rewards/margins": 0.6422742009162903,
+      "rewards/rejected": -2.5487518310546875,
+      "step": 342
+    },
+    {
+      "epoch": 0.623069936421435,
+      "grad_norm": 1.7752224206924438,
+      "learning_rate": 9.47091530430494e-06,
+      "logits/chosen": 0.14096824824810028,
+      "logits/rejected": 0.1774113029241562,
+      "logps/chosen": -71.80216217041016,
+      "logps/rejected": -79.65301513671875,
+      "loss": 2.2678,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.5735952854156494,
+      "rewards/margins": 0.5823659300804138,
+      "rewards/rejected": -2.155961036682129,
+      "step": 343
+    },
+    {
+      "epoch": 0.6248864668483197,
+      "grad_norm": 1.855377197265625,
+      "learning_rate": 9.467888139598086e-06,
+      "logits/chosen": 0.10637074708938599,
+      "logits/rejected": 0.0798158049583435,
+      "logps/chosen": -79.01347351074219,
+      "logps/rejected": -80.77239990234375,
+      "loss": 2.7867,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9598251581192017,
+      "rewards/margins": 0.3184160590171814,
+      "rewards/rejected": -2.2782411575317383,
+      "step": 344
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 1.423040747642517,
+      "learning_rate": 9.464852883676441e-06,
+      "logits/chosen": 0.1304859220981598,
+      "logits/rejected": 0.1759231686592102,
+      "logps/chosen": -74.09733581542969,
+      "logps/rejected": -89.27589416503906,
+      "loss": 2.2777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7697277069091797,
+      "rewards/margins": 0.635611355304718,
+      "rewards/rejected": -2.405339002609253,
+      "step": 345
+    },
+    {
+      "epoch": 0.628519527702089,
+      "grad_norm": 1.6246038675308228,
+      "learning_rate": 9.461809542729421e-06,
+      "logits/chosen": 0.03750050812959671,
+      "logits/rejected": 0.10337980091571808,
+      "logps/chosen": -81.89110565185547,
+      "logps/rejected": -95.36811828613281,
+      "loss": 2.0438,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7299752235412598,
+      "rewards/margins": 0.7404756546020508,
+      "rewards/rejected": -2.4704508781433105,
+      "step": 346
+    },
+    {
+      "epoch": 0.6303360581289736,
+      "grad_norm": 1.6618752479553223,
+      "learning_rate": 9.458758122962926e-06,
+      "logits/chosen": 0.05359608680009842,
+      "logits/rejected": 0.10455545783042908,
+      "logps/chosen": -85.34078216552734,
+      "logps/rejected": -90.46200561523438,
+      "loss": 2.3299,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9331045150756836,
+      "rewards/margins": 0.4755082130432129,
+      "rewards/rejected": -2.4086129665374756,
+      "step": 347
+    },
+    {
+      "epoch": 0.6321525885558583,
+      "grad_norm": 1.5805696249008179,
+      "learning_rate": 9.455698630599332e-06,
+      "logits/chosen": 0.10048776119947433,
+      "logits/rejected": 0.1140337884426117,
+      "logps/chosen": -80.25875854492188,
+      "logps/rejected": -94.44998168945312,
+      "loss": 2.3439,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8421945571899414,
+      "rewards/margins": 0.5977468490600586,
+      "rewards/rejected": -2.43994140625,
+      "step": 348
+    },
+    {
+      "epoch": 0.633969118982743,
+      "grad_norm": 1.37093985080719,
+      "learning_rate": 9.452631071877478e-06,
+      "logits/chosen": 0.11764326691627502,
+      "logits/rejected": 0.10735289752483368,
+      "logps/chosen": -72.02367401123047,
+      "logps/rejected": -77.9522705078125,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.6527085304260254,
+      "rewards/margins": 0.6065118908882141,
+      "rewards/rejected": -2.259220600128174,
+      "step": 349
+    },
+    {
+      "epoch": 0.6357856494096276,
+      "grad_norm": 2.0214192867279053,
+      "learning_rate": 9.449555453052652e-06,
+      "logits/chosen": 0.13177426159381866,
+      "logits/rejected": 0.12408209592103958,
+      "logps/chosen": -76.62931823730469,
+      "logps/rejected": -81.16517639160156,
+      "loss": 2.9175,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.9575048685073853,
+      "rewards/margins": 0.21441945433616638,
+      "rewards/rejected": -2.171924352645874,
+      "step": 350
+    },
+    {
+      "epoch": 0.6376021798365122,
+      "grad_norm": 1.6041687726974487,
+      "learning_rate": 9.446471780396573e-06,
+      "logits/chosen": 0.18574532866477966,
+      "logits/rejected": 0.22683225572109222,
+      "logps/chosen": -73.95652770996094,
+      "logps/rejected": -80.29524993896484,
+      "loss": 2.3833,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8684748411178589,
+      "rewards/margins": 0.37268272042274475,
+      "rewards/rejected": -2.2411575317382812,
+      "step": 351
+    },
+    {
+      "epoch": 0.6394187102633969,
+      "grad_norm": 1.7410838603973389,
+      "learning_rate": 9.443380060197387e-06,
+      "logits/chosen": 0.07876043766736984,
+      "logits/rejected": 0.14103996753692627,
+      "logps/chosen": -74.64432525634766,
+      "logps/rejected": -83.6999740600586,
+      "loss": 2.5558,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.029916763305664,
+      "rewards/margins": 0.4088842272758484,
+      "rewards/rejected": -2.438800811767578,
+      "step": 352
+    },
+    {
+      "epoch": 0.6412352406902816,
+      "grad_norm": 1.87971031665802,
+      "learning_rate": 9.440280298759653e-06,
+      "logits/chosen": 0.13997013866901398,
+      "logits/rejected": 0.1322249174118042,
+      "logps/chosen": -88.78776550292969,
+      "logps/rejected": -88.22732543945312,
+      "loss": 2.6547,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9983258247375488,
+      "rewards/margins": 0.2458382397890091,
+      "rewards/rejected": -2.244164228439331,
+      "step": 353
+    },
+    {
+      "epoch": 0.6430517711171662,
+      "grad_norm": 1.7056363821029663,
+      "learning_rate": 9.437172502404318e-06,
+      "logits/chosen": 0.07248476892709732,
+      "logits/rejected": 0.13019773364067078,
+      "logps/chosen": -78.4591064453125,
+      "logps/rejected": -80.49126434326172,
+      "loss": 2.3809,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9740041494369507,
+      "rewards/margins": 0.4853326082229614,
+      "rewards/rejected": -2.459336757659912,
+      "step": 354
+    },
+    {
+      "epoch": 0.6448683015440508,
+      "grad_norm": 1.435718059539795,
+      "learning_rate": 9.434056677468726e-06,
+      "logits/chosen": 0.09164869040250778,
+      "logits/rejected": 0.14243285357952118,
+      "logps/chosen": -77.83367156982422,
+      "logps/rejected": -86.18836212158203,
+      "loss": 2.006,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.7355122566223145,
+      "rewards/margins": 0.6869419813156128,
+      "rewards/rejected": -2.422454357147217,
+      "step": 355
+    },
+    {
+      "epoch": 0.6466848319709355,
+      "grad_norm": 1.6553188562393188,
+      "learning_rate": 9.430932830306587e-06,
+      "logits/chosen": 0.04967673122882843,
+      "logits/rejected": 0.1456151008605957,
+      "logps/chosen": -73.50204467773438,
+      "logps/rejected": -85.31363677978516,
+      "loss": 2.4388,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.978384256362915,
+      "rewards/margins": 0.48918718099594116,
+      "rewards/rejected": -2.467571496963501,
+      "step": 356
+    },
+    {
+      "epoch": 0.6485013623978202,
+      "grad_norm": 1.8443480730056763,
+      "learning_rate": 9.427800967287963e-06,
+      "logits/chosen": 0.06455090641975403,
+      "logits/rejected": 0.12608012557029724,
+      "logps/chosen": -75.49232482910156,
+      "logps/rejected": -77.49136352539062,
+      "loss": 2.706,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.8730812072753906,
+      "rewards/margins": 0.2989741861820221,
+      "rewards/rejected": -2.17205548286438,
+      "step": 357
+    },
+    {
+      "epoch": 0.6503178928247049,
+      "grad_norm": 1.6912306547164917,
+      "learning_rate": 9.424661094799273e-06,
+      "logits/chosen": 0.11878645420074463,
+      "logits/rejected": 0.13628609478473663,
+      "logps/chosen": -73.09883880615234,
+      "logps/rejected": -84.19624328613281,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.928421139717102,
+      "rewards/margins": 0.5017052292823792,
+      "rewards/rejected": -2.430126190185547,
+      "step": 358
+    },
+    {
+      "epoch": 0.6521344232515894,
+      "grad_norm": 1.3643461465835571,
+      "learning_rate": 9.421513219243262e-06,
+      "logits/chosen": 0.07683826237916946,
+      "logits/rejected": 0.15765298902988434,
+      "logps/chosen": -76.0871353149414,
+      "logps/rejected": -97.47781372070312,
+      "loss": 1.7844,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.8300602436065674,
+      "rewards/margins": 0.911116361618042,
+      "rewards/rejected": -2.7411766052246094,
+      "step": 359
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 1.6327749490737915,
+      "learning_rate": 9.418357347038999e-06,
+      "logits/chosen": 0.1078951433300972,
+      "logits/rejected": 0.12233921140432358,
+      "logps/chosen": -75.89913177490234,
+      "logps/rejected": -78.4587173461914,
+      "loss": 2.4976,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.7992844581604004,
+      "rewards/margins": 0.311392605304718,
+      "rewards/rejected": -2.1106772422790527,
+      "step": 360
+    },
+    {
+      "epoch": 0.6557674841053588,
+      "grad_norm": 1.463025689125061,
+      "learning_rate": 9.415193484621852e-06,
+      "logits/chosen": 0.057331383228302,
+      "logits/rejected": 0.1563551127910614,
+      "logps/chosen": -80.7889633178711,
+      "logps/rejected": -93.21602630615234,
+      "loss": 2.1398,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.744189739227295,
+      "rewards/margins": 0.576555073261261,
+      "rewards/rejected": -2.320744752883911,
+      "step": 361
+    },
+    {
+      "epoch": 0.6575840145322435,
+      "grad_norm": 1.743695616722107,
+      "learning_rate": 9.412021638443491e-06,
+      "logits/chosen": 0.03781123086810112,
+      "logits/rejected": 0.154897540807724,
+      "logps/chosen": -77.72047424316406,
+      "logps/rejected": -88.80599975585938,
+      "loss": 2.2361,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9849369525909424,
+      "rewards/margins": 0.6562294363975525,
+      "rewards/rejected": -2.6411664485931396,
+      "step": 362
+    },
+    {
+      "epoch": 0.659400544959128,
+      "grad_norm": 1.5648279190063477,
+      "learning_rate": 9.408841814971862e-06,
+      "logits/chosen": 0.10401102900505066,
+      "logits/rejected": 0.12699122726917267,
+      "logps/chosen": -83.10671997070312,
+      "logps/rejected": -86.8663101196289,
+      "loss": 2.4009,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7048468589782715,
+      "rewards/margins": 0.4090174734592438,
+      "rewards/rejected": -2.1138644218444824,
+      "step": 363
+    },
+    {
+      "epoch": 0.6612170753860127,
+      "grad_norm": 1.3385239839553833,
+      "learning_rate": 9.405654020691178e-06,
+      "logits/chosen": 0.06100422143936157,
+      "logits/rejected": 0.10872650146484375,
+      "logps/chosen": -76.72837829589844,
+      "logps/rejected": -87.63470458984375,
+      "loss": 2.141,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.747226595878601,
+      "rewards/margins": 0.615075945854187,
+      "rewards/rejected": -2.362302780151367,
+      "step": 364
+    },
+    {
+      "epoch": 0.6630336058128974,
+      "grad_norm": 1.4060734510421753,
+      "learning_rate": 9.402458262101906e-06,
+      "logits/chosen": 0.10638861358165741,
+      "logits/rejected": 0.16951681673526764,
+      "logps/chosen": -77.35757446289062,
+      "logps/rejected": -92.03012084960938,
+      "loss": 2.0511,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8359463214874268,
+      "rewards/margins": 0.7444422841072083,
+      "rewards/rejected": -2.5803885459899902,
+      "step": 365
+    },
+    {
+      "epoch": 0.6648501362397821,
+      "grad_norm": 1.4371014833450317,
+      "learning_rate": 9.399254545720757e-06,
+      "logits/chosen": 0.0383220911026001,
+      "logits/rejected": 0.05856206640601158,
+      "logps/chosen": -82.40064239501953,
+      "logps/rejected": -92.1117935180664,
+      "loss": 2.1756,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8580785989761353,
+      "rewards/margins": 0.6475991010665894,
+      "rewards/rejected": -2.5056777000427246,
+      "step": 366
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.6673945188522339,
+      "learning_rate": 9.396042878080661e-06,
+      "logits/chosen": 0.14657820761203766,
+      "logits/rejected": 0.19021111726760864,
+      "logps/chosen": -73.56724548339844,
+      "logps/rejected": -78.85279846191406,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9079951047897339,
+      "rewards/margins": 0.3709837794303894,
+      "rewards/rejected": -2.2789790630340576,
+      "step": 367
+    },
+    {
+      "epoch": 0.6684831970935513,
+      "grad_norm": 1.7402448654174805,
+      "learning_rate": 9.392823265730775e-06,
+      "logits/chosen": 0.16515754163265228,
+      "logits/rejected": 0.12460774928331375,
+      "logps/chosen": -69.89266204833984,
+      "logps/rejected": -74.53837585449219,
+      "loss": 2.4262,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.0627622604370117,
+      "rewards/margins": 0.5317320823669434,
+      "rewards/rejected": -2.594494342803955,
+      "step": 368
+    },
+    {
+      "epoch": 0.670299727520436,
+      "grad_norm": 1.5290364027023315,
+      "learning_rate": 9.389595715236446e-06,
+      "logits/chosen": 0.0954434722661972,
+      "logits/rejected": 0.16226956248283386,
+      "logps/chosen": -77.17019653320312,
+      "logps/rejected": -91.07398986816406,
+      "loss": 2.2511,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.241170644760132,
+      "rewards/margins": 0.6319360136985779,
+      "rewards/rejected": -2.8731067180633545,
+      "step": 369
+    },
+    {
+      "epoch": 0.6721162579473207,
+      "grad_norm": 1.4633573293685913,
+      "learning_rate": 9.386360233179206e-06,
+      "logits/chosen": 0.06517557799816132,
+      "logits/rejected": 0.07250035554170609,
+      "logps/chosen": -79.23770141601562,
+      "logps/rejected": -88.50403594970703,
+      "loss": 2.3858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.751698613166809,
+      "rewards/margins": 0.49529415369033813,
+      "rewards/rejected": -2.246993064880371,
+      "step": 370
+    },
+    {
+      "epoch": 0.6739327883742052,
+      "grad_norm": 2.007215976715088,
+      "learning_rate": 9.383116826156775e-06,
+      "logits/chosen": 0.13584929704666138,
+      "logits/rejected": 0.10950647294521332,
+      "logps/chosen": -80.75408935546875,
+      "logps/rejected": -79.3903579711914,
+      "loss": 2.6967,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.2421724796295166,
+      "rewards/margins": 0.34328368306159973,
+      "rewards/rejected": -2.585456132888794,
+      "step": 371
+    },
+    {
+      "epoch": 0.6757493188010899,
+      "grad_norm": 1.7323477268218994,
+      "learning_rate": 9.37986550078302e-06,
+      "logits/chosen": 0.011626070365309715,
+      "logits/rejected": 0.06207559257745743,
+      "logps/chosen": -75.51494598388672,
+      "logps/rejected": -85.83218383789062,
+      "loss": 2.4608,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9569413661956787,
+      "rewards/margins": 0.5160585641860962,
+      "rewards/rejected": -2.4729998111724854,
+      "step": 372
+    },
+    {
+      "epoch": 0.6775658492279746,
+      "grad_norm": 1.6365752220153809,
+      "learning_rate": 9.376606263687959e-06,
+      "logits/chosen": 0.10213632136583328,
+      "logits/rejected": 0.08355780690908432,
+      "logps/chosen": -79.01370239257812,
+      "logps/rejected": -81.9417953491211,
+      "loss": 2.4701,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.064467668533325,
+      "rewards/margins": 0.4157037138938904,
+      "rewards/rejected": -2.4801712036132812,
+      "step": 373
+    },
+    {
+      "epoch": 0.6793823796548593,
+      "grad_norm": 1.5918922424316406,
+      "learning_rate": 9.373339121517748e-06,
+      "logits/chosen": 0.09486684203147888,
+      "logits/rejected": 0.07311725616455078,
+      "logps/chosen": -82.15351867675781,
+      "logps/rejected": -91.7690658569336,
+      "loss": 2.1297,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0153114795684814,
+      "rewards/margins": 0.669613242149353,
+      "rewards/rejected": -2.684924840927124,
+      "step": 374
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 1.6511566638946533,
+      "learning_rate": 9.370064080934654e-06,
+      "logits/chosen": 0.1406637728214264,
+      "logits/rejected": 0.20172299444675446,
+      "logps/chosen": -69.41023254394531,
+      "logps/rejected": -74.65138244628906,
+      "loss": 2.4704,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.061182737350464,
+      "rewards/margins": 0.3090137243270874,
+      "rewards/rejected": -2.3701963424682617,
+      "step": 375
+    },
+    {
+      "epoch": 0.6830154405086285,
+      "grad_norm": 1.770624041557312,
+      "learning_rate": 9.366781148617056e-06,
+      "logits/chosen": 0.10267248749732971,
+      "logits/rejected": 0.07900385558605194,
+      "logps/chosen": -77.27003479003906,
+      "logps/rejected": -88.09040832519531,
+      "loss": 2.2813,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.132927656173706,
+      "rewards/margins": 0.6629016399383545,
+      "rewards/rejected": -2.7958290576934814,
+      "step": 376
+    },
+    {
+      "epoch": 0.6848319709355132,
+      "grad_norm": 1.5172028541564941,
+      "learning_rate": 9.363490331259426e-06,
+      "logits/chosen": 0.05240853130817413,
+      "logits/rejected": 0.10445387661457062,
+      "logps/chosen": -75.5218276977539,
+      "logps/rejected": -85.49366760253906,
+      "loss": 2.1442,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9302880764007568,
+      "rewards/margins": 0.5776917338371277,
+      "rewards/rejected": -2.5079798698425293,
+      "step": 377
+    },
+    {
+      "epoch": 0.6866485013623979,
+      "grad_norm": 1.6540950536727905,
+      "learning_rate": 9.360191635572313e-06,
+      "logits/chosen": 0.14195458590984344,
+      "logits/rejected": 0.0978107899427414,
+      "logps/chosen": -85.00608825683594,
+      "logps/rejected": -85.1456298828125,
+      "loss": 2.423,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8765841722488403,
+      "rewards/margins": 0.49441370368003845,
+      "rewards/rejected": -2.370997905731201,
+      "step": 378
+    },
+    {
+      "epoch": 0.6884650317892824,
+      "grad_norm": 1.6463801860809326,
+      "learning_rate": 9.356885068282334e-06,
+      "logits/chosen": 0.13197994232177734,
+      "logits/rejected": 0.09924699366092682,
+      "logps/chosen": -86.94219970703125,
+      "logps/rejected": -86.85704803466797,
+      "loss": 2.4801,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9674016237258911,
+      "rewards/margins": 0.4530283510684967,
+      "rewards/rejected": -2.4204299449920654,
+      "step": 379
+    },
+    {
+      "epoch": 0.6902815622161671,
+      "grad_norm": 1.5428895950317383,
+      "learning_rate": 9.353570636132151e-06,
+      "logits/chosen": 0.10434838384389877,
+      "logits/rejected": 0.12619757652282715,
+      "logps/chosen": -78.56607818603516,
+      "logps/rejected": -86.33320617675781,
+      "loss": 2.2711,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9667280912399292,
+      "rewards/margins": 0.6350463628768921,
+      "rewards/rejected": -2.601774215698242,
+      "step": 380
+    },
+    {
+      "epoch": 0.6920980926430518,
+      "grad_norm": 1.6211251020431519,
+      "learning_rate": 9.350248345880471e-06,
+      "logits/chosen": 0.14081251621246338,
+      "logits/rejected": 0.1453506350517273,
+      "logps/chosen": -74.38184356689453,
+      "logps/rejected": -80.61524200439453,
+      "loss": 2.3863,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.0189671516418457,
+      "rewards/margins": 0.4528267979621887,
+      "rewards/rejected": -2.4717938899993896,
+      "step": 381
+    },
+    {
+      "epoch": 0.6939146230699365,
+      "grad_norm": 2.6346240043640137,
+      "learning_rate": 9.346918204302022e-06,
+      "logits/chosen": 0.0519830696284771,
+      "logits/rejected": 0.04284125566482544,
+      "logps/chosen": -82.38153839111328,
+      "logps/rejected": -86.35488891601562,
+      "loss": 2.5434,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9824391603469849,
+      "rewards/margins": 0.4156208634376526,
+      "rewards/rejected": -2.3980600833892822,
+      "step": 382
+    },
+    {
+      "epoch": 0.695731153496821,
+      "grad_norm": 1.4762026071548462,
+      "learning_rate": 9.343580218187544e-06,
+      "logits/chosen": 0.02196469157934189,
+      "logits/rejected": 0.04618150740861893,
+      "logps/chosen": -72.66182708740234,
+      "logps/rejected": -88.68983459472656,
+      "loss": 2.1699,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9227774143218994,
+      "rewards/margins": 0.9200201630592346,
+      "rewards/rejected": -2.8427975177764893,
+      "step": 383
+    },
+    {
+      "epoch": 0.6975476839237057,
+      "grad_norm": 1.9675711393356323,
+      "learning_rate": 9.340234394343768e-06,
+      "logits/chosen": 0.12956203520298004,
+      "logits/rejected": 0.16079677641391754,
+      "logps/chosen": -87.38492584228516,
+      "logps/rejected": -87.80662536621094,
+      "loss": 2.3138,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.144991397857666,
+      "rewards/margins": 0.4893474578857422,
+      "rewards/rejected": -2.634338855743408,
+      "step": 384
+    },
+    {
+      "epoch": 0.6993642143505904,
+      "grad_norm": 1.9116860628128052,
+      "learning_rate": 9.336880739593415e-06,
+      "logits/chosen": 0.06013559550046921,
+      "logits/rejected": 0.09889352321624756,
+      "logps/chosen": -86.55718231201172,
+      "logps/rejected": -91.01994323730469,
+      "loss": 2.4864,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9528815746307373,
+      "rewards/margins": 0.44100096821784973,
+      "rewards/rejected": -2.3938825130462646,
+      "step": 385
+    },
+    {
+      "epoch": 0.701180744777475,
+      "grad_norm": 2.0582289695739746,
+      "learning_rate": 9.33351926077517e-06,
+      "logits/chosen": 0.11538423597812653,
+      "logits/rejected": 0.17568480968475342,
+      "logps/chosen": -76.24241638183594,
+      "logps/rejected": -87.1357650756836,
+      "loss": 2.6878,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.102504253387451,
+      "rewards/margins": 0.3671773076057434,
+      "rewards/rejected": -2.469681739807129,
+      "step": 386
+    },
+    {
+      "epoch": 0.7029972752043597,
+      "grad_norm": 1.6962640285491943,
+      "learning_rate": 9.330149964743674e-06,
+      "logits/chosen": 0.12425235658884048,
+      "logits/rejected": 0.1852879822254181,
+      "logps/chosen": -82.67887115478516,
+      "logps/rejected": -98.66133117675781,
+      "loss": 2.343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.182525634765625,
+      "rewards/margins": 0.6602460145950317,
+      "rewards/rejected": -2.842771530151367,
+      "step": 387
+    },
+    {
+      "epoch": 0.7048138056312443,
+      "grad_norm": 1.6582266092300415,
+      "learning_rate": 9.326772858369506e-06,
+      "logits/chosen": 0.14438432455062866,
+      "logits/rejected": 0.0922938883304596,
+      "logps/chosen": -78.87794494628906,
+      "logps/rejected": -84.929443359375,
+      "loss": 2.3036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.988234281539917,
+      "rewards/margins": 0.4737466275691986,
+      "rewards/rejected": -2.4619810581207275,
+      "step": 388
+    },
+    {
+      "epoch": 0.706630336058129,
+      "grad_norm": 1.6689919233322144,
+      "learning_rate": 9.323387948539176e-06,
+      "logits/chosen": 0.1282673329114914,
+      "logits/rejected": 0.13633723556995392,
+      "logps/chosen": -70.22606658935547,
+      "logps/rejected": -79.26126861572266,
+      "loss": 2.241,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9162399768829346,
+      "rewards/margins": 0.5345122814178467,
+      "rewards/rejected": -2.4507524967193604,
+      "step": 389
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 1.5861207246780396,
+      "learning_rate": 9.319995242155102e-06,
+      "logits/chosen": 0.13024169206619263,
+      "logits/rejected": 0.14390775561332703,
+      "logps/chosen": -87.00923919677734,
+      "logps/rejected": -91.24594116210938,
+      "loss": 2.284,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9127002954483032,
+      "rewards/margins": 0.5265774130821228,
+      "rewards/rejected": -2.4392776489257812,
+      "step": 390
+    },
+    {
+      "epoch": 0.7102633969118983,
+      "grad_norm": 2.0118792057037354,
+      "learning_rate": 9.316594746135608e-06,
+      "logits/chosen": 0.14648675918579102,
+      "logits/rejected": 0.11786539107561111,
+      "logps/chosen": -82.68962097167969,
+      "logps/rejected": -88.26126861572266,
+      "loss": 2.7281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9463748931884766,
+      "rewards/margins": 0.20282456278800964,
+      "rewards/rejected": -2.1491994857788086,
+      "step": 391
+    },
+    {
+      "epoch": 0.7120799273387829,
+      "grad_norm": 1.8176332712173462,
+      "learning_rate": 9.313186467414892e-06,
+      "logits/chosen": 0.12084021419286728,
+      "logits/rejected": 0.1433698982000351,
+      "logps/chosen": -73.46707153320312,
+      "logps/rejected": -82.46024322509766,
+      "loss": 2.5247,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0802624225616455,
+      "rewards/margins": 0.41601985692977905,
+      "rewards/rejected": -2.4962823390960693,
+      "step": 392
+    },
+    {
+      "epoch": 0.7138964577656676,
+      "grad_norm": 1.7207351922988892,
+      "learning_rate": 9.30977041294303e-06,
+      "logits/chosen": 0.11983273923397064,
+      "logits/rejected": 0.16106371581554413,
+      "logps/chosen": -76.09751892089844,
+      "logps/rejected": -85.20323944091797,
+      "loss": 2.5284,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9018669128417969,
+      "rewards/margins": 0.4894491136074066,
+      "rewards/rejected": -2.3913159370422363,
+      "step": 393
+    },
+    {
+      "epoch": 0.7157129881925522,
+      "grad_norm": 1.6489367485046387,
+      "learning_rate": 9.306346589685956e-06,
+      "logits/chosen": 0.14845696091651917,
+      "logits/rejected": 0.10356368124485016,
+      "logps/chosen": -83.94526672363281,
+      "logps/rejected": -86.53378295898438,
+      "loss": 2.3394,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9089899063110352,
+      "rewards/margins": 0.4619474411010742,
+      "rewards/rejected": -2.3709375858306885,
+      "step": 394
+    },
+    {
+      "epoch": 0.7175295186194369,
+      "grad_norm": 1.792077898979187,
+      "learning_rate": 9.302915004625435e-06,
+      "logits/chosen": 0.06936248391866684,
+      "logits/rejected": 0.13349927961826324,
+      "logps/chosen": -91.82075500488281,
+      "logps/rejected": -98.7391357421875,
+      "loss": 2.3875,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9371806383132935,
+      "rewards/margins": 0.6781354546546936,
+      "rewards/rejected": -2.6153156757354736,
+      "step": 395
+    },
+    {
+      "epoch": 0.7193460490463215,
+      "grad_norm": 2.049694538116455,
+      "learning_rate": 9.29947566475907e-06,
+      "logits/chosen": 0.08474650233983994,
+      "logits/rejected": 0.13380834460258484,
+      "logps/chosen": -84.59528350830078,
+      "logps/rejected": -96.24267578125,
+      "loss": 2.7001,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9375882148742676,
+      "rewards/margins": 0.4607384204864502,
+      "rewards/rejected": -2.3983266353607178,
+      "step": 396
+    },
+    {
+      "epoch": 0.7211625794732062,
+      "grad_norm": 1.9235875606536865,
+      "learning_rate": 9.296028577100271e-06,
+      "logits/chosen": 0.06255945563316345,
+      "logits/rejected": 0.17074811458587646,
+      "logps/chosen": -68.99498748779297,
+      "logps/rejected": -84.45899963378906,
+      "loss": 2.347,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9344629049301147,
+      "rewards/margins": 0.688963770866394,
+      "rewards/rejected": -2.6234264373779297,
+      "step": 397
+    },
+    {
+      "epoch": 0.7229791099000908,
+      "grad_norm": 1.6617276668548584,
+      "learning_rate": 9.292573748678254e-06,
+      "logits/chosen": 0.11962493509054184,
+      "logits/rejected": 0.11647717654705048,
+      "logps/chosen": -85.3626708984375,
+      "logps/rejected": -95.70562744140625,
+      "loss": 2.2229,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.8968594074249268,
+      "rewards/margins": 0.6918852925300598,
+      "rewards/rejected": -2.588744640350342,
+      "step": 398
+    },
+    {
+      "epoch": 0.7247956403269755,
+      "grad_norm": 1.4650444984436035,
+      "learning_rate": 9.289111186538013e-06,
+      "logits/chosen": 0.06390775740146637,
+      "logits/rejected": 0.11076060682535172,
+      "logps/chosen": -69.9189453125,
+      "logps/rejected": -81.5618667602539,
+      "loss": 2.2271,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.857581377029419,
+      "rewards/margins": 0.5540103316307068,
+      "rewards/rejected": -2.4115917682647705,
+      "step": 399
+    },
+    {
+      "epoch": 0.7266121707538601,
+      "grad_norm": 1.5728726387023926,
+      "learning_rate": 9.285640897740316e-06,
+      "logits/chosen": 0.08816932141780853,
+      "logits/rejected": 0.1330798715353012,
+      "logps/chosen": -78.04735565185547,
+      "logps/rejected": -87.89193725585938,
+      "loss": 2.3595,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.823161244392395,
+      "rewards/margins": 0.4568220376968384,
+      "rewards/rejected": -2.2799830436706543,
+      "step": 400
+    },
+    {
+      "epoch": 0.7284287011807448,
+      "grad_norm": 1.6291779279708862,
+      "learning_rate": 9.282162889361686e-06,
+      "logits/chosen": 0.17718347907066345,
+      "logits/rejected": 0.20603135228157043,
+      "logps/chosen": -67.99884033203125,
+      "logps/rejected": -75.93153381347656,
+      "loss": 2.3002,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.671633243560791,
+      "rewards/margins": 0.45543500781059265,
+      "rewards/rejected": -2.127068281173706,
+      "step": 401
+    },
+    {
+      "epoch": 0.7302452316076294,
+      "grad_norm": 1.5066276788711548,
+      "learning_rate": 9.278677168494388e-06,
+      "logits/chosen": 0.16840124130249023,
+      "logits/rejected": 0.17102479934692383,
+      "logps/chosen": -78.1556625366211,
+      "logps/rejected": -84.91495513916016,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9279460906982422,
+      "rewards/margins": 0.6659258008003235,
+      "rewards/rejected": -2.593871831893921,
+      "step": 402
+    },
+    {
+      "epoch": 0.7320617620345141,
+      "grad_norm": 1.71995210647583,
+      "learning_rate": 9.275183742246412e-06,
+      "logits/chosen": -0.0005522281862795353,
+      "logits/rejected": 0.1412215530872345,
+      "logps/chosen": -73.37832641601562,
+      "logps/rejected": -92.71226501464844,
+      "loss": 2.3315,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6880759000778198,
+      "rewards/margins": 0.5844776630401611,
+      "rewards/rejected": -2.2725534439086914,
+      "step": 403
+    },
+    {
+      "epoch": 0.7338782924613987,
+      "grad_norm": 1.676321029663086,
+      "learning_rate": 9.271682617741466e-06,
+      "logits/chosen": 0.15319044888019562,
+      "logits/rejected": 0.14503052830696106,
+      "logps/chosen": -75.85136413574219,
+      "logps/rejected": -82.61698150634766,
+      "loss": 2.3528,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.819403052330017,
+      "rewards/margins": 0.3919827342033386,
+      "rewards/rejected": -2.211385726928711,
+      "step": 404
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 1.4811294078826904,
+      "learning_rate": 9.268173802118949e-06,
+      "logits/chosen": 0.07122528553009033,
+      "logits/rejected": 0.12146291136741638,
+      "logps/chosen": -76.78916931152344,
+      "logps/rejected": -85.33805084228516,
+      "loss": 2.1505,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9886473417282104,
+      "rewards/margins": 0.5911746025085449,
+      "rewards/rejected": -2.579822301864624,
+      "step": 405
+    },
+    {
+      "epoch": 0.737511353315168,
+      "grad_norm": 1.629601001739502,
+      "learning_rate": 9.264657302533947e-06,
+      "logits/chosen": 0.11274963617324829,
+      "logits/rejected": 0.0651293620467186,
+      "logps/chosen": -82.86985778808594,
+      "logps/rejected": -88.68803405761719,
+      "loss": 2.1873,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8538885116577148,
+      "rewards/margins": 0.6726438403129578,
+      "rewards/rejected": -2.5265324115753174,
+      "step": 406
+    },
+    {
+      "epoch": 0.7393278837420527,
+      "grad_norm": 1.5687483549118042,
+      "learning_rate": 9.261133126157218e-06,
+      "logits/chosen": 0.06054290384054184,
+      "logits/rejected": 0.1543959081172943,
+      "logps/chosen": -68.5821304321289,
+      "logps/rejected": -80.544921875,
+      "loss": 2.3587,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7550634145736694,
+      "rewards/margins": 0.46927812695503235,
+      "rewards/rejected": -2.224341630935669,
+      "step": 407
+    },
+    {
+      "epoch": 0.7411444141689373,
+      "grad_norm": 1.849908471107483,
+      "learning_rate": 9.257601280175167e-06,
+      "logits/chosen": 0.15177126228809357,
+      "logits/rejected": 0.1485670655965805,
+      "logps/chosen": -72.81680297851562,
+      "logps/rejected": -80.58684539794922,
+      "loss": 2.444,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9739688634872437,
+      "rewards/margins": 0.6166737079620361,
+      "rewards/rejected": -2.5906424522399902,
+      "step": 408
+    },
+    {
+      "epoch": 0.742960944595822,
+      "grad_norm": 1.779342770576477,
+      "learning_rate": 9.254061771789847e-06,
+      "logits/chosen": 0.05093669891357422,
+      "logits/rejected": 0.1048060953617096,
+      "logps/chosen": -84.48141479492188,
+      "logps/rejected": -91.49285125732422,
+      "loss": 2.5324,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9952929019927979,
+      "rewards/margins": 0.3900759220123291,
+      "rewards/rejected": -2.385368585586548,
+      "step": 409
+    },
+    {
+      "epoch": 0.7447774750227066,
+      "grad_norm": 1.71133291721344,
+      "learning_rate": 9.25051460821893e-06,
+      "logits/chosen": 0.1308833658695221,
+      "logits/rejected": 0.08994461596012115,
+      "logps/chosen": -74.09046936035156,
+      "logps/rejected": -80.80438995361328,
+      "loss": 2.1895,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8457674980163574,
+      "rewards/margins": 0.5731581449508667,
+      "rewards/rejected": -2.4189257621765137,
+      "step": 410
+    },
+    {
+      "epoch": 0.7465940054495913,
+      "grad_norm": 1.6980112791061401,
+      "learning_rate": 9.2469597966957e-06,
+      "logits/chosen": 0.05856658145785332,
+      "logits/rejected": 0.08371179550886154,
+      "logps/chosen": -78.58245849609375,
+      "logps/rejected": -88.30054473876953,
+      "loss": 2.46,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9803493022918701,
+      "rewards/margins": 0.6115507483482361,
+      "rewards/rejected": -2.591899871826172,
+      "step": 411
+    },
+    {
+      "epoch": 0.7484105358764759,
+      "grad_norm": 1.8350476026535034,
+      "learning_rate": 9.243397344469037e-06,
+      "logits/chosen": 0.05285171419382095,
+      "logits/rejected": 0.13240401446819305,
+      "logps/chosen": -73.79881286621094,
+      "logps/rejected": -86.56880187988281,
+      "loss": 2.2929,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.060075283050537,
+      "rewards/margins": 0.54813152551651,
+      "rewards/rejected": -2.6082065105438232,
+      "step": 412
+    },
+    {
+      "epoch": 0.7502270663033606,
+      "grad_norm": 1.4403142929077148,
+      "learning_rate": 9.239827258803402e-06,
+      "logits/chosen": 0.07327421009540558,
+      "logits/rejected": 0.18387822806835175,
+      "logps/chosen": -68.3246078491211,
+      "logps/rejected": -81.33067321777344,
+      "loss": 1.8794,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8658640384674072,
+      "rewards/margins": 0.7600168585777283,
+      "rewards/rejected": -2.625880718231201,
+      "step": 413
+    },
+    {
+      "epoch": 0.7520435967302452,
+      "grad_norm": 1.883155345916748,
+      "learning_rate": 9.23624954697882e-06,
+      "logits/chosen": 0.10411994159221649,
+      "logits/rejected": 0.06234448403120041,
+      "logps/chosen": -78.31743621826172,
+      "logps/rejected": -80.0100326538086,
+      "loss": 2.5646,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.934099555015564,
+      "rewards/margins": 0.28737103939056396,
+      "rewards/rejected": -2.221470832824707,
+      "step": 414
+    },
+    {
+      "epoch": 0.7538601271571299,
+      "grad_norm": 1.5926934480667114,
+      "learning_rate": 9.232664216290868e-06,
+      "logits/chosen": 0.09589770436286926,
+      "logits/rejected": 0.041689179837703705,
+      "logps/chosen": -77.89552307128906,
+      "logps/rejected": -84.02517700195312,
+      "loss": 2.2284,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.787453532218933,
+      "rewards/margins": 0.49444466829299927,
+      "rewards/rejected": -2.281898260116577,
+      "step": 415
+    },
+    {
+      "epoch": 0.7556766575840145,
+      "grad_norm": 1.511832356452942,
+      "learning_rate": 9.229071274050663e-06,
+      "logits/chosen": 0.07688678801059723,
+      "logits/rejected": 0.14106512069702148,
+      "logps/chosen": -78.53334045410156,
+      "logps/rejected": -86.92496490478516,
+      "loss": 2.2211,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0686798095703125,
+      "rewards/margins": 0.5378819704055786,
+      "rewards/rejected": -2.6065618991851807,
+      "step": 416
+    },
+    {
+      "epoch": 0.7574931880108992,
+      "grad_norm": 1.7858117818832397,
+      "learning_rate": 9.225470727584835e-06,
+      "logits/chosen": 0.14854730665683746,
+      "logits/rejected": 0.06465649604797363,
+      "logps/chosen": -87.31832122802734,
+      "logps/rejected": -79.3746566772461,
+      "loss": 2.4446,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9429931640625,
+      "rewards/margins": 0.38861486315727234,
+      "rewards/rejected": -2.3316078186035156,
+      "step": 417
+    },
+    {
+      "epoch": 0.7593097184377838,
+      "grad_norm": 1.9230687618255615,
+      "learning_rate": 9.221862584235527e-06,
+      "logits/chosen": 0.07233145087957382,
+      "logits/rejected": 0.13825634121894836,
+      "logps/chosen": -81.53103637695312,
+      "logps/rejected": -90.54696655273438,
+      "loss": 2.7171,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.065859794616699,
+      "rewards/margins": 0.30683329701423645,
+      "rewards/rejected": -2.3726933002471924,
+      "step": 418
+    },
+    {
+      "epoch": 0.7611262488646685,
+      "grad_norm": 2.063098192214966,
+      "learning_rate": 9.218246851360374e-06,
+      "logits/chosen": 0.10360075533390045,
+      "logits/rejected": 0.14677830040454865,
+      "logps/chosen": -85.08120727539062,
+      "logps/rejected": -89.35393524169922,
+      "loss": 2.4292,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.300565719604492,
+      "rewards/margins": 0.4277467727661133,
+      "rewards/rejected": -2.7283124923706055,
+      "step": 419
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 2.4728786945343018,
+      "learning_rate": 9.214623536332483e-06,
+      "logits/chosen": 0.08115474879741669,
+      "logits/rejected": 0.07649822533130646,
+      "logps/chosen": -83.79824829101562,
+      "logps/rejected": -87.47264099121094,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.0838184356689453,
+      "rewards/margins": 0.41770505905151367,
+      "rewards/rejected": -2.501523494720459,
+      "step": 420
+    },
+    {
+      "epoch": 0.7647593097184378,
+      "grad_norm": 1.7078697681427002,
+      "learning_rate": 9.210992646540425e-06,
+      "logits/chosen": 0.08155008405447006,
+      "logits/rejected": 0.08723931759595871,
+      "logps/chosen": -74.4200210571289,
+      "logps/rejected": -87.49038696289062,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9076974391937256,
+      "rewards/margins": 0.6807244420051575,
+      "rewards/rejected": -2.5884220600128174,
+      "step": 421
+    },
+    {
+      "epoch": 0.7665758401453224,
+      "grad_norm": 1.5644993782043457,
+      "learning_rate": 9.207354189388214e-06,
+      "logits/chosen": 0.16107802093029022,
+      "logits/rejected": 0.15951679646968842,
+      "logps/chosen": -74.85030364990234,
+      "logps/rejected": -81.25472259521484,
+      "loss": 2.2389,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0082590579986572,
+      "rewards/margins": 0.5728211402893066,
+      "rewards/rejected": -2.581080198287964,
+      "step": 422
+    },
+    {
+      "epoch": 0.7683923705722071,
+      "grad_norm": 1.6519557237625122,
+      "learning_rate": 9.203708172295299e-06,
+      "logits/chosen": 0.08725707978010178,
+      "logits/rejected": 0.11393023282289505,
+      "logps/chosen": -79.56071472167969,
+      "logps/rejected": -90.27279663085938,
+      "loss": 2.2853,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9312469959259033,
+      "rewards/margins": 0.6229196786880493,
+      "rewards/rejected": -2.554166793823242,
+      "step": 423
+    },
+    {
+      "epoch": 0.7702089009990918,
+      "grad_norm": 1.668648600578308,
+      "learning_rate": 9.200054602696544e-06,
+      "logits/chosen": 0.1369432657957077,
+      "logits/rejected": 0.08549812436103821,
+      "logps/chosen": -80.69041442871094,
+      "logps/rejected": -85.47969055175781,
+      "loss": 2.3022,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7405030727386475,
+      "rewards/margins": 0.4421282112598419,
+      "rewards/rejected": -2.182631492614746,
+      "step": 424
+    },
+    {
+      "epoch": 0.7720254314259763,
+      "grad_norm": 1.7834622859954834,
+      "learning_rate": 9.196393488042213e-06,
+      "logits/chosen": 0.06559979915618896,
+      "logits/rejected": 0.05396304652094841,
+      "logps/chosen": -76.51896667480469,
+      "logps/rejected": -88.19717407226562,
+      "loss": 2.5216,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.050562620162964,
+      "rewards/margins": 0.5811472535133362,
+      "rewards/rejected": -2.631709575653076,
+      "step": 425
+    },
+    {
+      "epoch": 0.773841961852861,
+      "grad_norm": 1.5034058094024658,
+      "learning_rate": 9.192724835797956e-06,
+      "logits/chosen": 0.10820607095956802,
+      "logits/rejected": 0.14427171647548676,
+      "logps/chosen": -81.0079116821289,
+      "logps/rejected": -95.6849136352539,
+      "loss": 2.0472,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.19939923286438,
+      "rewards/margins": 0.8476728796958923,
+      "rewards/rejected": -3.047071933746338,
+      "step": 426
+    },
+    {
+      "epoch": 0.7756584922797457,
+      "grad_norm": 1.7205973863601685,
+      "learning_rate": 9.189048653444796e-06,
+      "logits/chosen": 0.08522580564022064,
+      "logits/rejected": 0.09135682880878448,
+      "logps/chosen": -77.4021224975586,
+      "logps/rejected": -85.47643280029297,
+      "loss": 2.2618,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.002034902572632,
+      "rewards/margins": 0.5705874562263489,
+      "rewards/rejected": -2.572622776031494,
+      "step": 427
+    },
+    {
+      "epoch": 0.7774750227066304,
+      "grad_norm": 2.131197214126587,
+      "learning_rate": 9.185364948479109e-06,
+      "logits/chosen": 0.05514610558748245,
+      "logits/rejected": 0.0597330704331398,
+      "logps/chosen": -86.83938598632812,
+      "logps/rejected": -94.71990966796875,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.088331699371338,
+      "rewards/margins": 0.47913864254951477,
+      "rewards/rejected": -2.567470073699951,
+      "step": 428
+    },
+    {
+      "epoch": 0.779291553133515,
+      "grad_norm": 3.0228030681610107,
+      "learning_rate": 9.181673728412605e-06,
+      "logits/chosen": 0.09502460807561874,
+      "logits/rejected": 0.10438862442970276,
+      "logps/chosen": -77.20478820800781,
+      "logps/rejected": -80.07164764404297,
+      "loss": 2.5924,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.107954978942871,
+      "rewards/margins": 0.4795181155204773,
+      "rewards/rejected": -2.5874733924865723,
+      "step": 429
+    },
+    {
+      "epoch": 0.7811080835603996,
+      "grad_norm": 2.630490779876709,
+      "learning_rate": 9.17797500077233e-06,
+      "logits/chosen": 0.053769052028656006,
+      "logits/rejected": 0.1357191652059555,
+      "logps/chosen": -74.01277160644531,
+      "logps/rejected": -86.05644989013672,
+      "loss": 2.6208,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0179696083068848,
+      "rewards/margins": 0.3262583613395691,
+      "rewards/rejected": -2.3442280292510986,
+      "step": 430
+    },
+    {
+      "epoch": 0.7829246139872843,
+      "grad_norm": 1.7841664552688599,
+      "learning_rate": 9.174268773100636e-06,
+      "logits/chosen": 0.09468546509742737,
+      "logits/rejected": 0.16804733872413635,
+      "logps/chosen": -80.07379913330078,
+      "logps/rejected": -87.17518615722656,
+      "loss": 2.4676,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.27384877204895,
+      "rewards/margins": 0.4068894386291504,
+      "rewards/rejected": -2.6807379722595215,
+      "step": 431
+    },
+    {
+      "epoch": 0.784741144414169,
+      "grad_norm": 1.7030746936798096,
+      "learning_rate": 9.170555052955158e-06,
+      "logits/chosen": 0.10446357727050781,
+      "logits/rejected": 0.1235620528459549,
+      "logps/chosen": -89.81584930419922,
+      "logps/rejected": -94.43669128417969,
+      "loss": 2.3366,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.98858642578125,
+      "rewards/margins": 0.5402320623397827,
+      "rewards/rejected": -2.5288188457489014,
+      "step": 432
+    },
+    {
+      "epoch": 0.7865576748410535,
+      "grad_norm": 1.4258465766906738,
+      "learning_rate": 9.166833847908825e-06,
+      "logits/chosen": 0.10827025771141052,
+      "logits/rejected": 0.10875076055526733,
+      "logps/chosen": -69.29400634765625,
+      "logps/rejected": -80.72003936767578,
+      "loss": 2.2501,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9478650093078613,
+      "rewards/margins": 0.5940475463867188,
+      "rewards/rejected": -2.54191255569458,
+      "step": 433
+    },
+    {
+      "epoch": 0.7883742052679382,
+      "grad_norm": 1.8069814443588257,
+      "learning_rate": 9.163105165549819e-06,
+      "logits/chosen": 0.08856840431690216,
+      "logits/rejected": 0.07843751460313797,
+      "logps/chosen": -79.93275451660156,
+      "logps/rejected": -83.75475311279297,
+      "loss": 2.5214,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0315518379211426,
+      "rewards/margins": 0.45457565784454346,
+      "rewards/rejected": -2.4861276149749756,
+      "step": 434
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 1.7881503105163574,
+      "learning_rate": 9.159369013481574e-06,
+      "logits/chosen": 0.08166562020778656,
+      "logits/rejected": 0.08240097761154175,
+      "logps/chosen": -92.98298645019531,
+      "logps/rejected": -96.15763092041016,
+      "loss": 2.053,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4435133934020996,
+      "rewards/margins": 0.7140956521034241,
+      "rewards/rejected": -3.157609462738037,
+      "step": 435
+    },
+    {
+      "epoch": 0.7920072661217076,
+      "grad_norm": 1.7574496269226074,
+      "learning_rate": 9.155625399322754e-06,
+      "logits/chosen": 0.08114133030176163,
+      "logits/rejected": 0.11279730498790741,
+      "logps/chosen": -74.07096862792969,
+      "logps/rejected": -80.94386291503906,
+      "loss": 2.4289,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2497410774230957,
+      "rewards/margins": 0.4424628019332886,
+      "rewards/rejected": -2.6922037601470947,
+      "step": 436
+    },
+    {
+      "epoch": 0.7938237965485921,
+      "grad_norm": 1.4814083576202393,
+      "learning_rate": 9.15187433070724e-06,
+      "logits/chosen": 0.026073571294546127,
+      "logits/rejected": 0.1382063925266266,
+      "logps/chosen": -72.39479064941406,
+      "logps/rejected": -86.76889038085938,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9130635261535645,
+      "rewards/margins": 0.6362709999084473,
+      "rewards/rejected": -2.549334764480591,
+      "step": 437
+    },
+    {
+      "epoch": 0.7956403269754768,
+      "grad_norm": 1.6441129446029663,
+      "learning_rate": 9.148115815284113e-06,
+      "logits/chosen": 0.07432619482278824,
+      "logits/rejected": 0.10215617716312408,
+      "logps/chosen": -77.05310821533203,
+      "logps/rejected": -80.81717681884766,
+      "loss": 2.338,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9012683629989624,
+      "rewards/margins": 0.43776583671569824,
+      "rewards/rejected": -2.339034080505371,
+      "step": 438
+    },
+    {
+      "epoch": 0.7974568574023615,
+      "grad_norm": 1.8839560747146606,
+      "learning_rate": 9.144349860717643e-06,
+      "logits/chosen": 0.07576426863670349,
+      "logits/rejected": 0.08868670463562012,
+      "logps/chosen": -74.13190460205078,
+      "logps/rejected": -80.07232666015625,
+      "loss": 2.3485,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9554617404937744,
+      "rewards/margins": 0.531810998916626,
+      "rewards/rejected": -2.4872727394104004,
+      "step": 439
+    },
+    {
+      "epoch": 0.7992733878292462,
+      "grad_norm": 1.826811671257019,
+      "learning_rate": 9.140576474687263e-06,
+      "logits/chosen": 0.09120994061231613,
+      "logits/rejected": 0.08987519890069962,
+      "logps/chosen": -75.1444320678711,
+      "logps/rejected": -78.58987426757812,
+      "loss": 2.6513,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158261299133301,
+      "rewards/margins": 0.28016677498817444,
+      "rewards/rejected": -2.4384284019470215,
+      "step": 440
+    },
+    {
+      "epoch": 0.8010899182561307,
+      "grad_norm": 1.8647096157073975,
+      "learning_rate": 9.13679566488757e-06,
+      "logits/chosen": 0.12438184767961502,
+      "logits/rejected": 0.13951222598552704,
+      "logps/chosen": -83.79389953613281,
+      "logps/rejected": -90.33230590820312,
+      "loss": 2.3956,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0185930728912354,
+      "rewards/margins": 0.43685024976730347,
+      "rewards/rejected": -2.4554433822631836,
+      "step": 441
+    },
+    {
+      "epoch": 0.8029064486830154,
+      "grad_norm": 1.7610801458358765,
+      "learning_rate": 9.133007439028288e-06,
+      "logits/chosen": 0.08218151330947876,
+      "logits/rejected": 0.072816863656044,
+      "logps/chosen": -86.35633087158203,
+      "logps/rejected": -84.08189392089844,
+      "loss": 2.243,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8836033344268799,
+      "rewards/margins": 0.516740083694458,
+      "rewards/rejected": -2.400343418121338,
+      "step": 442
+    },
+    {
+      "epoch": 0.8047229791099001,
+      "grad_norm": 1.6049703359603882,
+      "learning_rate": 9.129211804834271e-06,
+      "logits/chosen": 0.059631846845149994,
+      "logits/rejected": 0.09744230657815933,
+      "logps/chosen": -76.90062713623047,
+      "logps/rejected": -82.29356384277344,
+      "loss": 2.2574,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9436204433441162,
+      "rewards/margins": 0.47444695234298706,
+      "rewards/rejected": -2.418067216873169,
+      "step": 443
+    },
+    {
+      "epoch": 0.8065395095367848,
+      "grad_norm": 1.9819791316986084,
+      "learning_rate": 9.12540877004548e-06,
+      "logits/chosen": 0.07371871173381805,
+      "logits/rejected": 0.058843065053224564,
+      "logps/chosen": -78.42516326904297,
+      "logps/rejected": -82.23822784423828,
+      "loss": 2.6959,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.273909330368042,
+      "rewards/margins": 0.3514346182346344,
+      "rewards/rejected": -2.6253440380096436,
+      "step": 444
+    },
+    {
+      "epoch": 0.8083560399636693,
+      "grad_norm": 1.9540653228759766,
+      "learning_rate": 9.12159834241696e-06,
+      "logits/chosen": 0.13155074417591095,
+      "logits/rejected": 0.11775672435760498,
+      "logps/chosen": -69.17151641845703,
+      "logps/rejected": -77.9190902709961,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.073258638381958,
+      "rewards/margins": 0.5955159664154053,
+      "rewards/rejected": -2.6687746047973633,
+      "step": 445
+    },
+    {
+      "epoch": 0.810172570390554,
+      "grad_norm": 1.9634339809417725,
+      "learning_rate": 9.117780529718843e-06,
+      "logits/chosen": 0.05265370383858681,
+      "logits/rejected": 0.09651105850934982,
+      "logps/chosen": -78.77818298339844,
+      "logps/rejected": -84.63584899902344,
+      "loss": 2.2157,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.084688186645508,
+      "rewards/margins": 0.541061282157898,
+      "rewards/rejected": -2.625749349594116,
+      "step": 446
+    },
+    {
+      "epoch": 0.8119891008174387,
+      "grad_norm": 1.9086997509002686,
+      "learning_rate": 9.113955339736309e-06,
+      "logits/chosen": 0.08473093807697296,
+      "logits/rejected": 0.047919195145368576,
+      "logps/chosen": -83.44226837158203,
+      "logps/rejected": -87.24287414550781,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9542481899261475,
+      "rewards/margins": 0.4577428698539734,
+      "rewards/rejected": -2.4119908809661865,
+      "step": 447
+    },
+    {
+      "epoch": 0.8138056312443234,
+      "grad_norm": 1.6967483758926392,
+      "learning_rate": 9.11012278026959e-06,
+      "logits/chosen": 0.062431350350379944,
+      "logits/rejected": 0.10980932414531708,
+      "logps/chosen": -90.98541259765625,
+      "logps/rejected": -102.51634979248047,
+      "loss": 2.3339,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1866049766540527,
+      "rewards/margins": 0.6549904346466064,
+      "rewards/rejected": -2.841595411300659,
+      "step": 448
+    },
+    {
+      "epoch": 0.815622161671208,
+      "grad_norm": 2.295504331588745,
+      "learning_rate": 9.106282859133936e-06,
+      "logits/chosen": 0.04481218010187149,
+      "logits/rejected": 0.002619542181491852,
+      "logps/chosen": -81.81380462646484,
+      "logps/rejected": -80.60113525390625,
+      "loss": 2.8324,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.1727848052978516,
+      "rewards/margins": 0.2777697443962097,
+      "rewards/rejected": -2.450554609298706,
+      "step": 449
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 1.7951968908309937,
+      "learning_rate": 9.102435584159623e-06,
+      "logits/chosen": 0.029065577313303947,
+      "logits/rejected": 0.11651361733675003,
+      "logps/chosen": -78.08231353759766,
+      "logps/rejected": -88.59742736816406,
+      "loss": 2.5562,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.2762022018432617,
+      "rewards/margins": 0.4410257041454315,
+      "rewards/rejected": -2.7172276973724365,
+      "step": 450
+    },
+    {
+      "epoch": 0.8192552225249773,
+      "grad_norm": 1.6301803588867188,
+      "learning_rate": 9.098580963191908e-06,
+      "logits/chosen": 0.06471782922744751,
+      "logits/rejected": 0.11030608415603638,
+      "logps/chosen": -78.3387451171875,
+      "logps/rejected": -91.82540130615234,
+      "loss": 2.1826,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.885233998298645,
+      "rewards/margins": 0.6971657872200012,
+      "rewards/rejected": -2.582399845123291,
+      "step": 451
+    },
+    {
+      "epoch": 0.821071752951862,
+      "grad_norm": 1.705962061882019,
+      "learning_rate": 9.094719004091039e-06,
+      "logits/chosen": 0.12591060996055603,
+      "logits/rejected": 0.1107010543346405,
+      "logps/chosen": -76.82783508300781,
+      "logps/rejected": -84.23365020751953,
+      "loss": 2.2843,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.04795241355896,
+      "rewards/margins": 0.5103356242179871,
+      "rewards/rejected": -2.558288335800171,
+      "step": 452
+    },
+    {
+      "epoch": 0.8228882833787466,
+      "grad_norm": 1.6535886526107788,
+      "learning_rate": 9.090849714732217e-06,
+      "logits/chosen": 0.11322569847106934,
+      "logits/rejected": 0.13092264533042908,
+      "logps/chosen": -80.16563415527344,
+      "logps/rejected": -82.96965026855469,
+      "loss": 2.3556,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9074076414108276,
+      "rewards/margins": 0.39431121945381165,
+      "rewards/rejected": -2.3017189502716064,
+      "step": 453
+    },
+    {
+      "epoch": 0.8247048138056312,
+      "grad_norm": 1.7194557189941406,
+      "learning_rate": 9.086973103005602e-06,
+      "logits/chosen": 0.1069367378950119,
+      "logits/rejected": 0.06237747147679329,
+      "logps/chosen": -72.59986877441406,
+      "logps/rejected": -77.62615203857422,
+      "loss": 2.2768,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.023085355758667,
+      "rewards/margins": 0.6052648425102234,
+      "rewards/rejected": -2.6283504962921143,
+      "step": 454
+    },
+    {
+      "epoch": 0.8265213442325159,
+      "grad_norm": 1.7921446561813354,
+      "learning_rate": 9.08308917681628e-06,
+      "logits/chosen": 0.12843580543994904,
+      "logits/rejected": 0.08002308011054993,
+      "logps/chosen": -80.6107177734375,
+      "logps/rejected": -80.92134857177734,
+      "loss": 2.3881,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9404933452606201,
+      "rewards/margins": 0.37478840351104736,
+      "rewards/rejected": -2.315281629562378,
+      "step": 455
+    },
+    {
+      "epoch": 0.8283378746594006,
+      "grad_norm": 2.3353819847106934,
+      "learning_rate": 9.07919794408425e-06,
+      "logits/chosen": 0.05452323704957962,
+      "logits/rejected": 0.12883196771144867,
+      "logps/chosen": -83.56001281738281,
+      "logps/rejected": -95.23291015625,
+      "loss": 2.8503,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.2597994804382324,
+      "rewards/margins": 0.4092620015144348,
+      "rewards/rejected": -2.6690614223480225,
+      "step": 456
+    },
+    {
+      "epoch": 0.8301544050862852,
+      "grad_norm": 1.6829804182052612,
+      "learning_rate": 9.075299412744417e-06,
+      "logits/chosen": 0.1223950982093811,
+      "logits/rejected": 0.10596577078104019,
+      "logps/chosen": -77.75115203857422,
+      "logps/rejected": -83.63490295410156,
+      "loss": 2.1639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.026258945465088,
+      "rewards/margins": 0.6107547879219055,
+      "rewards/rejected": -2.6370139122009277,
+      "step": 457
+    },
+    {
+      "epoch": 0.8319709355131698,
+      "grad_norm": 2.053755283355713,
+      "learning_rate": 9.07139359074656e-06,
+      "logits/chosen": 0.08021271228790283,
+      "logits/rejected": 0.11617676913738251,
+      "logps/chosen": -75.00261688232422,
+      "logps/rejected": -87.25222778320312,
+      "loss": 2.4048,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3207621574401855,
+      "rewards/margins": 0.5956254005432129,
+      "rewards/rejected": -2.9163873195648193,
+      "step": 458
+    },
+    {
+      "epoch": 0.8337874659400545,
+      "grad_norm": 2.1006064414978027,
+      "learning_rate": 9.067480486055333e-06,
+      "logits/chosen": 0.1115679144859314,
+      "logits/rejected": 0.08167213946580887,
+      "logps/chosen": -81.78447723388672,
+      "logps/rejected": -82.21318817138672,
+      "loss": 2.8725,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.165053367614746,
+      "rewards/margins": 0.2777022123336792,
+      "rewards/rejected": -2.442755699157715,
+      "step": 459
+    },
+    {
+      "epoch": 0.8356039963669392,
+      "grad_norm": 1.6339340209960938,
+      "learning_rate": 9.063560106650238e-06,
+      "logits/chosen": 0.0966312363743782,
+      "logits/rejected": 0.09943810105323792,
+      "logps/chosen": -71.0906982421875,
+      "logps/rejected": -76.22811126708984,
+      "loss": 2.2922,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.0007636547088623,
+      "rewards/margins": 0.47423097491264343,
+      "rewards/rejected": -2.474994659423828,
+      "step": 460
+    },
+    {
+      "epoch": 0.8374205267938238,
+      "grad_norm": 1.8504972457885742,
+      "learning_rate": 9.059632460525613e-06,
+      "logits/chosen": 0.09517084062099457,
+      "logits/rejected": 0.14127981662750244,
+      "logps/chosen": -79.06185913085938,
+      "logps/rejected": -89.11494445800781,
+      "loss": 2.3288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0016026496887207,
+      "rewards/margins": 0.5282158255577087,
+      "rewards/rejected": -2.529818534851074,
+      "step": 461
+    },
+    {
+      "epoch": 0.8392370572207084,
+      "grad_norm": 1.8151304721832275,
+      "learning_rate": 9.055697555690607e-06,
+      "logits/chosen": 0.08868349343538284,
+      "logits/rejected": 0.14090785384178162,
+      "logps/chosen": -77.26272583007812,
+      "logps/rejected": -85.14620971679688,
+      "loss": 2.3694,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.087087869644165,
+      "rewards/margins": 0.5032299757003784,
+      "rewards/rejected": -2.590317964553833,
+      "step": 462
+    },
+    {
+      "epoch": 0.8410535876475931,
+      "grad_norm": 1.8596280813217163,
+      "learning_rate": 9.051755400169182e-06,
+      "logits/chosen": 0.06272133439779282,
+      "logits/rejected": 0.12342572212219238,
+      "logps/chosen": -75.46456909179688,
+      "logps/rejected": -81.9759292602539,
+      "loss": 2.5167,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0463178157806396,
+      "rewards/margins": 0.3146142363548279,
+      "rewards/rejected": -2.3609323501586914,
+      "step": 463
+    },
+    {
+      "epoch": 0.8428701180744778,
+      "grad_norm": 1.7658417224884033,
+      "learning_rate": 9.047806002000075e-06,
+      "logits/chosen": -0.007259421981871128,
+      "logits/rejected": 0.09611339122056961,
+      "logps/chosen": -73.88439178466797,
+      "logps/rejected": -80.77063751220703,
+      "loss": 2.4938,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9998183250427246,
+      "rewards/margins": 0.4115346670150757,
+      "rewards/rejected": -2.4113528728485107,
+      "step": 464
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 1.6596916913986206,
+      "learning_rate": 9.043849369236799e-06,
+      "logits/chosen": 0.03223409131169319,
+      "logits/rejected": 0.05010119825601578,
+      "logps/chosen": -66.32562255859375,
+      "logps/rejected": -69.36509704589844,
+      "loss": 2.5881,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.909837245941162,
+      "rewards/margins": 0.2923263907432556,
+      "rewards/rejected": -2.2021636962890625,
+      "step": 465
+    },
+    {
+      "epoch": 0.846503178928247,
+      "grad_norm": 1.6832820177078247,
+      "learning_rate": 9.039885509947616e-06,
+      "logits/chosen": 0.06287454813718796,
+      "logits/rejected": 0.12746769189834595,
+      "logps/chosen": -76.6998519897461,
+      "logps/rejected": -87.36117553710938,
+      "loss": 2.3914,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.029526710510254,
+      "rewards/margins": 0.48344865441322327,
+      "rewards/rejected": -2.5129752159118652,
+      "step": 466
+    },
+    {
+      "epoch": 0.8483197093551317,
+      "grad_norm": 1.6972366571426392,
+      "learning_rate": 9.035914432215527e-06,
+      "logits/chosen": 0.1179896891117096,
+      "logits/rejected": 0.14000467956066132,
+      "logps/chosen": -75.3287353515625,
+      "logps/rejected": -78.60227966308594,
+      "loss": 2.4194,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.088515043258667,
+      "rewards/margins": 0.42544490098953247,
+      "rewards/rejected": -2.513960123062134,
+      "step": 467
+    },
+    {
+      "epoch": 0.8501362397820164,
+      "grad_norm": 2.5104973316192627,
+      "learning_rate": 9.031936144138247e-06,
+      "logits/chosen": 0.08423591405153275,
+      "logits/rejected": 0.1387752890586853,
+      "logps/chosen": -78.6176528930664,
+      "logps/rejected": -83.52703094482422,
+      "loss": 2.7262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1089184284210205,
+      "rewards/margins": 0.44402289390563965,
+      "rewards/rejected": -2.55294132232666,
+      "step": 468
+    },
+    {
+      "epoch": 0.851952770208901,
+      "grad_norm": 1.944931149482727,
+      "learning_rate": 9.027950653828202e-06,
+      "logits/chosen": 0.12252221256494522,
+      "logits/rejected": 0.14959384500980377,
+      "logps/chosen": -72.63184356689453,
+      "logps/rejected": -70.89137268066406,
+      "loss": 2.7323,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9612714052200317,
+      "rewards/margins": 0.24262896180152893,
+      "rewards/rejected": -2.2039003372192383,
+      "step": 469
+    },
+    {
+      "epoch": 0.8537693006357856,
+      "grad_norm": 1.6699979305267334,
+      "learning_rate": 9.0239579694125e-06,
+      "logits/chosen": 0.11827167868614197,
+      "logits/rejected": 0.11933022737503052,
+      "logps/chosen": -70.45188903808594,
+      "logps/rejected": -76.2336654663086,
+      "loss": 2.457,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.854758381843567,
+      "rewards/margins": 0.47020024061203003,
+      "rewards/rejected": -2.3249588012695312,
+      "step": 470
+    },
+    {
+      "epoch": 0.8555858310626703,
+      "grad_norm": 1.5790536403656006,
+      "learning_rate": 9.019958099032919e-06,
+      "logits/chosen": 0.04856352508068085,
+      "logits/rejected": 0.11418008804321289,
+      "logps/chosen": -75.7929458618164,
+      "logps/rejected": -86.55818176269531,
+      "loss": 2.2809,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.003899335861206,
+      "rewards/margins": 0.5960561037063599,
+      "rewards/rejected": -2.5999553203582764,
+      "step": 471
+    },
+    {
+      "epoch": 0.857402361489555,
+      "grad_norm": 1.3594144582748413,
+      "learning_rate": 9.015951050845891e-06,
+      "logits/chosen": 0.10341258347034454,
+      "logits/rejected": 0.12394269555807114,
+      "logps/chosen": -78.8187255859375,
+      "logps/rejected": -84.04349517822266,
+      "loss": 1.9536,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9372440576553345,
+      "rewards/margins": 0.6742948889732361,
+      "rewards/rejected": -2.611538887023926,
+      "step": 472
+    },
+    {
+      "epoch": 0.8592188919164396,
+      "grad_norm": 1.7592689990997314,
+      "learning_rate": 9.011936833022485e-06,
+      "logits/chosen": 0.10970023274421692,
+      "logits/rejected": 0.09040558338165283,
+      "logps/chosen": -80.59556579589844,
+      "logps/rejected": -84.2996826171875,
+      "loss": 2.5658,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.935511827468872,
+      "rewards/margins": 0.35643547773361206,
+      "rewards/rejected": -2.29194712638855,
+      "step": 473
+    },
+    {
+      "epoch": 0.8610354223433242,
+      "grad_norm": 1.6418087482452393,
+      "learning_rate": 9.00791545374839e-06,
+      "logits/chosen": 0.16755647957324982,
+      "logits/rejected": 0.1745520681142807,
+      "logps/chosen": -82.31010437011719,
+      "logps/rejected": -85.66128540039062,
+      "loss": 2.1854,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.9519617557525635,
+      "rewards/margins": 0.5059336423873901,
+      "rewards/rejected": -2.457895278930664,
+      "step": 474
+    },
+    {
+      "epoch": 0.8628519527702089,
+      "grad_norm": 1.4616092443466187,
+      "learning_rate": 9.0038869212239e-06,
+      "logits/chosen": 0.03978392109274864,
+      "logits/rejected": 0.1235240027308464,
+      "logps/chosen": -77.78251647949219,
+      "logps/rejected": -85.66397094726562,
+      "loss": 2.1995,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8721034526824951,
+      "rewards/margins": 0.45375847816467285,
+      "rewards/rejected": -2.325861930847168,
+      "step": 475
+    },
+    {
+      "epoch": 0.8646684831970936,
+      "grad_norm": 1.838180422782898,
+      "learning_rate": 8.99985124366389e-06,
+      "logits/chosen": 0.08277668058872223,
+      "logits/rejected": 0.035900432616472244,
+      "logps/chosen": -83.23967742919922,
+      "logps/rejected": -83.16535949707031,
+      "loss": 2.7347,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.2339279651641846,
+      "rewards/margins": 0.16711921989917755,
+      "rewards/rejected": -2.4010472297668457,
+      "step": 476
+    },
+    {
+      "epoch": 0.8664850136239782,
+      "grad_norm": 1.6746636629104614,
+      "learning_rate": 8.995808429297815e-06,
+      "logits/chosen": 0.14687396585941315,
+      "logits/rejected": 0.09570137411355972,
+      "logps/chosen": -78.79898071289062,
+      "logps/rejected": -77.9619369506836,
+      "loss": 2.4449,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8334993124008179,
+      "rewards/margins": 0.3442971706390381,
+      "rewards/rejected": -2.1777963638305664,
+      "step": 477
+    },
+    {
+      "epoch": 0.8683015440508629,
+      "grad_norm": 1.7042535543441772,
+      "learning_rate": 8.991758486369675e-06,
+      "logits/chosen": 0.047281138598918915,
+      "logits/rejected": -0.023666318506002426,
+      "logps/chosen": -83.798583984375,
+      "logps/rejected": -91.88795471191406,
+      "loss": 2.3072,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9728612899780273,
+      "rewards/margins": 0.5659723281860352,
+      "rewards/rejected": -2.5388338565826416,
+      "step": 478
+    },
+    {
+      "epoch": 0.8701180744777475,
+      "grad_norm": 1.5552817583084106,
+      "learning_rate": 8.987701423138007e-06,
+      "logits/chosen": 0.0772751122713089,
+      "logits/rejected": 0.08708472549915314,
+      "logps/chosen": -72.42391204833984,
+      "logps/rejected": -78.93070220947266,
+      "loss": 2.2815,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9803509712219238,
+      "rewards/margins": 0.5386512279510498,
+      "rewards/rejected": -2.5190021991729736,
+      "step": 479
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.6372044086456299,
+      "learning_rate": 8.983637247875872e-06,
+      "logits/chosen": -0.03352706879377365,
+      "logits/rejected": 0.01581352949142456,
+      "logps/chosen": -81.29138946533203,
+      "logps/rejected": -89.1744155883789,
+      "loss": 2.09,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9441754817962646,
+      "rewards/margins": 0.6137264370918274,
+      "rewards/rejected": -2.5579018592834473,
+      "step": 480
+    },
+    {
+      "epoch": 0.8737511353315168,
+      "grad_norm": 1.6375274658203125,
+      "learning_rate": 8.979565968870831e-06,
+      "logits/chosen": 0.07507088780403137,
+      "logits/rejected": 0.1289597451686859,
+      "logps/chosen": -74.78109741210938,
+      "logps/rejected": -84.55204772949219,
+      "loss": 2.4566,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.9041385650634766,
+      "rewards/margins": 0.42814433574676514,
+      "rewards/rejected": -2.332282781600952,
+      "step": 481
+    },
+    {
+      "epoch": 0.8755676657584015,
+      "grad_norm": 1.5601258277893066,
+      "learning_rate": 8.975487594424927e-06,
+      "logits/chosen": 0.012452262453734875,
+      "logits/rejected": 0.06930352002382278,
+      "logps/chosen": -83.01705169677734,
+      "logps/rejected": -92.38286590576172,
+      "loss": 2.2736,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.013011932373047,
+      "rewards/margins": 0.5085774064064026,
+      "rewards/rejected": -2.5215890407562256,
+      "step": 482
+    },
+    {
+      "epoch": 0.8773841961852861,
+      "grad_norm": 1.8685195446014404,
+      "learning_rate": 8.971402132854677e-06,
+      "logits/chosen": 0.09415426105260849,
+      "logits/rejected": 0.12719042599201202,
+      "logps/chosen": -79.85440063476562,
+      "logps/rejected": -82.80875396728516,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.114530086517334,
+      "rewards/margins": 0.46328115463256836,
+      "rewards/rejected": -2.5778112411499023,
+      "step": 483
+    },
+    {
+      "epoch": 0.8792007266121707,
+      "grad_norm": 1.8812834024429321,
+      "learning_rate": 8.967309592491052e-06,
+      "logits/chosen": 0.13547101616859436,
+      "logits/rejected": 0.11872326582670212,
+      "logps/chosen": -74.42122650146484,
+      "logps/rejected": -81.9700927734375,
+      "loss": 2.5016,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.1828627586364746,
+      "rewards/margins": 0.44510167837142944,
+      "rewards/rejected": -2.627964496612549,
+      "step": 484
+    },
+    {
+      "epoch": 0.8810172570390554,
+      "grad_norm": 1.510517954826355,
+      "learning_rate": 8.963209981679451e-06,
+      "logits/chosen": 0.03274242952466011,
+      "logits/rejected": 0.07188954204320908,
+      "logps/chosen": -82.55563354492188,
+      "logps/rejected": -100.33609771728516,
+      "loss": 1.9177,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9487042427062988,
+      "rewards/margins": 0.8189151287078857,
+      "rewards/rejected": -2.7676191329956055,
+      "step": 485
+    },
+    {
+      "epoch": 0.8828337874659401,
+      "grad_norm": 1.5973646640777588,
+      "learning_rate": 8.959103308779696e-06,
+      "logits/chosen": 0.0365552082657814,
+      "logits/rejected": 0.051129020750522614,
+      "logps/chosen": -72.4110107421875,
+      "logps/rejected": -83.10322570800781,
+      "loss": 2.1547,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9375314712524414,
+      "rewards/margins": 0.6991211175918579,
+      "rewards/rejected": -2.6366524696350098,
+      "step": 486
+    },
+    {
+      "epoch": 0.8846503178928247,
+      "grad_norm": 1.6707032918930054,
+      "learning_rate": 8.954989582166009e-06,
+      "logits/chosen": 0.0635693296790123,
+      "logits/rejected": 0.009854275733232498,
+      "logps/chosen": -87.12749481201172,
+      "logps/rejected": -88.9491195678711,
+      "loss": 2.2043,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9257593154907227,
+      "rewards/margins": 0.5429801344871521,
+      "rewards/rejected": -2.4687397480010986,
+      "step": 487
+    },
+    {
+      "epoch": 0.8864668483197093,
+      "grad_norm": 1.4422773122787476,
+      "learning_rate": 8.95086881022699e-06,
+      "logits/chosen": -0.024460218846797943,
+      "logits/rejected": 0.14137038588523865,
+      "logps/chosen": -74.60330200195312,
+      "logps/rejected": -95.01730346679688,
+      "loss": 2.0246,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.054253339767456,
+      "rewards/margins": 0.8260326981544495,
+      "rewards/rejected": -2.88028621673584,
+      "step": 488
+    },
+    {
+      "epoch": 0.888283378746594,
+      "grad_norm": 1.8121236562728882,
+      "learning_rate": 8.946741001365616e-06,
+      "logits/chosen": 0.05854415148496628,
+      "logits/rejected": 0.13758361339569092,
+      "logps/chosen": -74.46614074707031,
+      "logps/rejected": -83.96118927001953,
+      "loss": 2.5066,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.105980157852173,
+      "rewards/margins": 0.41478782892227173,
+      "rewards/rejected": -2.520768165588379,
+      "step": 489
+    },
+    {
+      "epoch": 0.8900999091734787,
+      "grad_norm": 2.023728370666504,
+      "learning_rate": 8.942606163999205e-06,
+      "logits/chosen": 0.04619833081960678,
+      "logits/rejected": 0.09330146014690399,
+      "logps/chosen": -75.35594940185547,
+      "logps/rejected": -89.18836975097656,
+      "loss": 2.2655,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9396051168441772,
+      "rewards/margins": 0.6726698875427246,
+      "rewards/rejected": -2.6122751235961914,
+      "step": 490
+    },
+    {
+      "epoch": 0.8919164396003633,
+      "grad_norm": 1.559735894203186,
+      "learning_rate": 8.938464306559412e-06,
+      "logits/chosen": 0.093504399061203,
+      "logits/rejected": 0.09722252935171127,
+      "logps/chosen": -80.77140808105469,
+      "logps/rejected": -87.86373901367188,
+      "loss": 2.1676,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.1073269844055176,
+      "rewards/margins": 0.5740070939064026,
+      "rewards/rejected": -2.6813340187072754,
+      "step": 491
+    },
+    {
+      "epoch": 0.8937329700272479,
+      "grad_norm": 1.6771574020385742,
+      "learning_rate": 8.934315437492203e-06,
+      "logits/chosen": 0.06368491798639297,
+      "logits/rejected": 0.04900998994708061,
+      "logps/chosen": -78.2313003540039,
+      "logps/rejected": -92.83306121826172,
+      "loss": 2.3435,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.110978364944458,
+      "rewards/margins": 0.6751725673675537,
+      "rewards/rejected": -2.786151170730591,
+      "step": 492
+    },
+    {
+      "epoch": 0.8955495004541326,
+      "grad_norm": 1.651183843612671,
+      "learning_rate": 8.930159565257846e-06,
+      "logits/chosen": 0.08393608778715134,
+      "logits/rejected": 0.0475153923034668,
+      "logps/chosen": -77.8431396484375,
+      "logps/rejected": -88.03681945800781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0274362564086914,
+      "rewards/margins": 0.8075671195983887,
+      "rewards/rejected": -2.835003137588501,
+      "step": 493
+    },
+    {
+      "epoch": 0.8973660308810173,
+      "grad_norm": 2.262662172317505,
+      "learning_rate": 8.925996698330887e-06,
+      "logits/chosen": 0.014226208440959454,
+      "logits/rejected": 0.09588178992271423,
+      "logps/chosen": -88.56185913085938,
+      "logps/rejected": -95.1131591796875,
+      "loss": 2.5853,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4549288749694824,
+      "rewards/margins": 0.5275144577026367,
+      "rewards/rejected": -2.982443332672119,
+      "step": 494
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 1.6070181131362915,
+      "learning_rate": 8.92182684520014e-06,
+      "logits/chosen": 0.11615607142448425,
+      "logits/rejected": 0.16645964980125427,
+      "logps/chosen": -78.97079467773438,
+      "logps/rejected": -90.49947357177734,
+      "loss": 2.1072,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.175706386566162,
+      "rewards/margins": 0.7435849905014038,
+      "rewards/rejected": -2.9192914962768555,
+      "step": 495
+    },
+    {
+      "epoch": 0.9009990917347865,
+      "grad_norm": 1.8343369960784912,
+      "learning_rate": 8.917650014368658e-06,
+      "logits/chosen": 0.12583515048027039,
+      "logits/rejected": 0.07308925688266754,
+      "logps/chosen": -83.38356018066406,
+      "logps/rejected": -90.73155975341797,
+      "loss": 2.3116,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3124871253967285,
+      "rewards/margins": 0.5869801640510559,
+      "rewards/rejected": -2.8994674682617188,
+      "step": 496
+    },
+    {
+      "epoch": 0.9028156221616712,
+      "grad_norm": 1.7175630331039429,
+      "learning_rate": 8.913466214353728e-06,
+      "logits/chosen": 0.14971572160720825,
+      "logits/rejected": 0.17372727394104004,
+      "logps/chosen": -86.66060638427734,
+      "logps/rejected": -94.42353820800781,
+      "loss": 2.3541,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.943577766418457,
+      "rewards/margins": 0.5955328345298767,
+      "rewards/rejected": -2.5391108989715576,
+      "step": 497
+    },
+    {
+      "epoch": 0.9046321525885559,
+      "grad_norm": 1.754708170890808,
+      "learning_rate": 8.909275453686845e-06,
+      "logits/chosen": 0.07779194414615631,
+      "logits/rejected": 0.09394712746143341,
+      "logps/chosen": -71.64151000976562,
+      "logps/rejected": -78.53131103515625,
+      "loss": 2.0436,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0102486610412598,
+      "rewards/margins": 0.6668623685836792,
+      "rewards/rejected": -2.6771109104156494,
+      "step": 498
+    },
+    {
+      "epoch": 0.9064486830154405,
+      "grad_norm": 1.9303216934204102,
+      "learning_rate": 8.905077740913701e-06,
+      "logits/chosen": 0.15853236615657806,
+      "logits/rejected": 0.11452697217464447,
+      "logps/chosen": -81.46524047851562,
+      "logps/rejected": -89.1716537475586,
+      "loss": 2.2477,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.264355182647705,
+      "rewards/margins": 0.6477693915367126,
+      "rewards/rejected": -2.9121243953704834,
+      "step": 499
+    },
+    {
+      "epoch": 0.9082652134423251,
+      "grad_norm": 2.1294379234313965,
+      "learning_rate": 8.900873084594164e-06,
+      "logits/chosen": 0.029480352997779846,
+      "logits/rejected": 0.09827219694852829,
+      "logps/chosen": -95.93231964111328,
+      "logps/rejected": -106.21436309814453,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.3455631732940674,
+      "rewards/margins": 0.5977038145065308,
+      "rewards/rejected": -2.9432668685913086,
+      "step": 500
+    },
+    {
+      "epoch": 0.9100817438692098,
+      "grad_norm": 1.7347708940505981,
+      "learning_rate": 8.896661493302258e-06,
+      "logits/chosen": 0.08531993627548218,
+      "logits/rejected": 0.14993277192115784,
+      "logps/chosen": -79.09967041015625,
+      "logps/rejected": -90.50935363769531,
+      "loss": 2.2969,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9908722639083862,
+      "rewards/margins": 0.5710910558700562,
+      "rewards/rejected": -2.5619633197784424,
+      "step": 501
+    },
+    {
+      "epoch": 0.9118982742960945,
+      "grad_norm": 2.065664529800415,
+      "learning_rate": 8.892442975626152e-06,
+      "logits/chosen": 0.18976512551307678,
+      "logits/rejected": 0.13936059176921844,
+      "logps/chosen": -82.80276489257812,
+      "logps/rejected": -84.57154083251953,
+      "loss": 2.5968,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4863569736480713,
+      "rewards/margins": 0.3651903569698334,
+      "rewards/rejected": -2.8515477180480957,
+      "step": 502
+    },
+    {
+      "epoch": 0.9137148047229791,
+      "grad_norm": 2.054955005645752,
+      "learning_rate": 8.888217540168139e-06,
+      "logits/chosen": 0.01604822278022766,
+      "logits/rejected": 0.1370609700679779,
+      "logps/chosen": -74.58015441894531,
+      "logps/rejected": -96.80826568603516,
+      "loss": 2.169,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.397087574005127,
+      "rewards/margins": 0.9368999004364014,
+      "rewards/rejected": -3.333987236022949,
+      "step": 503
+    },
+    {
+      "epoch": 0.9155313351498637,
+      "grad_norm": 2.179532527923584,
+      "learning_rate": 8.883985195544617e-06,
+      "logits/chosen": 0.03752445429563522,
+      "logits/rejected": 0.024428365752100945,
+      "logps/chosen": -83.81852722167969,
+      "logps/rejected": -88.1905288696289,
+      "loss": 2.5361,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.239609956741333,
+      "rewards/margins": 0.4188295900821686,
+      "rewards/rejected": -2.6584396362304688,
+      "step": 504
+    },
+    {
+      "epoch": 0.9173478655767484,
+      "grad_norm": 1.9171282052993774,
+      "learning_rate": 8.879745950386075e-06,
+      "logits/chosen": 0.17986616492271423,
+      "logits/rejected": 0.20335282385349274,
+      "logps/chosen": -72.679931640625,
+      "logps/rejected": -71.41309356689453,
+      "loss": 2.5452,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.261596918106079,
+      "rewards/margins": 0.3737923502922058,
+      "rewards/rejected": -2.6353893280029297,
+      "step": 505
+    },
+    {
+      "epoch": 0.9191643960036331,
+      "grad_norm": 2.3437929153442383,
+      "learning_rate": 8.87549981333707e-06,
+      "logits/chosen": 0.09010382741689682,
+      "logits/rejected": 0.040441811084747314,
+      "logps/chosen": -81.56139373779297,
+      "logps/rejected": -83.21439361572266,
+      "loss": 2.74,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2574033737182617,
+      "rewards/margins": 0.28719204664230347,
+      "rewards/rejected": -2.544595241546631,
+      "step": 506
+    },
+    {
+      "epoch": 0.9209809264305178,
+      "grad_norm": 2.4240329265594482,
+      "learning_rate": 8.871246793056215e-06,
+      "logits/chosen": 0.02888420596718788,
+      "logits/rejected": 0.11542786657810211,
+      "logps/chosen": -84.60144805908203,
+      "logps/rejected": -100.05682373046875,
+      "loss": 2.7052,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.863104820251465,
+      "rewards/margins": 0.41941094398498535,
+      "rewards/rejected": -3.28251576423645,
+      "step": 507
+    },
+    {
+      "epoch": 0.9227974568574023,
+      "grad_norm": 1.9247711896896362,
+      "learning_rate": 8.866986898216157e-06,
+      "logits/chosen": 0.06140238791704178,
+      "logits/rejected": 0.05623817816376686,
+      "logps/chosen": -77.40584564208984,
+      "logps/rejected": -86.71392822265625,
+      "loss": 2.2445,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4979231357574463,
+      "rewards/margins": 0.6161274313926697,
+      "rewards/rejected": -3.11405086517334,
+      "step": 508
+    },
+    {
+      "epoch": 0.924613987284287,
+      "grad_norm": 1.854956030845642,
+      "learning_rate": 8.862720137503568e-06,
+      "logits/chosen": 0.06755004823207855,
+      "logits/rejected": 0.03762562945485115,
+      "logps/chosen": -86.03059387207031,
+      "logps/rejected": -90.12970733642578,
+      "loss": 2.1459,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3017396926879883,
+      "rewards/margins": 0.6131142973899841,
+      "rewards/rejected": -2.914853572845459,
+      "step": 509
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.8044302463531494,
+      "learning_rate": 8.858446519619113e-06,
+      "logits/chosen": 0.08446178585290909,
+      "logits/rejected": 0.08818987011909485,
+      "logps/chosen": -77.58124542236328,
+      "logps/rejected": -85.02752685546875,
+      "loss": 2.1311,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2378087043762207,
+      "rewards/margins": 0.7067832350730896,
+      "rewards/rejected": -2.944591999053955,
+      "step": 510
+    },
+    {
+      "epoch": 0.9282470481380564,
+      "grad_norm": 1.4898866415023804,
+      "learning_rate": 8.854166053277443e-06,
+      "logits/chosen": 0.1039918065071106,
+      "logits/rejected": 0.1396111100912094,
+      "logps/chosen": -87.59841918945312,
+      "logps/rejected": -101.05659484863281,
+      "loss": 2.1092,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1167173385620117,
+      "rewards/margins": 0.7099149227142334,
+      "rewards/rejected": -2.826632499694824,
+      "step": 511
+    },
+    {
+      "epoch": 0.9300635785649409,
+      "grad_norm": 1.8009706735610962,
+      "learning_rate": 8.849878747207175e-06,
+      "logits/chosen": 0.11744043976068497,
+      "logits/rejected": 0.07621707767248154,
+      "logps/chosen": -78.37222290039062,
+      "logps/rejected": -74.32958221435547,
+      "loss": 2.6399,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.4002914428710938,
+      "rewards/margins": 0.21068021655082703,
+      "rewards/rejected": -2.610971450805664,
+      "step": 512
+    },
+    {
+      "epoch": 0.9318801089918256,
+      "grad_norm": 1.6368399858474731,
+      "learning_rate": 8.845584610150871e-06,
+      "logits/chosen": 0.09938757866621017,
+      "logits/rejected": 0.12675486505031586,
+      "logps/chosen": -82.24690246582031,
+      "logps/rejected": -92.60913848876953,
+      "loss": 1.9078,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.0825133323669434,
+      "rewards/margins": 0.8327144384384155,
+      "rewards/rejected": -2.9152278900146484,
+      "step": 513
+    },
+    {
+      "epoch": 0.9336966394187103,
+      "grad_norm": 1.8117483854293823,
+      "learning_rate": 8.841283650865027e-06,
+      "logits/chosen": 0.1355382651090622,
+      "logits/rejected": 0.14869986474514008,
+      "logps/chosen": -85.03866577148438,
+      "logps/rejected": -90.55638885498047,
+      "loss": 2.1243,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3315932750701904,
+      "rewards/margins": 0.6421911120414734,
+      "rewards/rejected": -2.9737846851348877,
+      "step": 514
+    },
+    {
+      "epoch": 0.935513169845595,
+      "grad_norm": 1.785927414894104,
+      "learning_rate": 8.836975878120046e-06,
+      "logits/chosen": 0.11557039618492126,
+      "logits/rejected": 0.11716655641794205,
+      "logps/chosen": -83.99951171875,
+      "logps/rejected": -92.36119079589844,
+      "loss": 2.4256,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.439833402633667,
+      "rewards/margins": 0.38893458247184753,
+      "rewards/rejected": -2.8287675380706787,
+      "step": 515
+    },
+    {
+      "epoch": 0.9373297002724795,
+      "grad_norm": 3.526102066040039,
+      "learning_rate": 8.832661300700228e-06,
+      "logits/chosen": 0.10778439044952393,
+      "logits/rejected": 0.18004637956619263,
+      "logps/chosen": -84.20158386230469,
+      "logps/rejected": -91.33497619628906,
+      "loss": 2.3568,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2032201290130615,
+      "rewards/margins": 0.5167797207832336,
+      "rewards/rejected": -2.7200000286102295,
+      "step": 516
+    },
+    {
+      "epoch": 0.9391462306993642,
+      "grad_norm": 1.7708752155303955,
+      "learning_rate": 8.828339927403745e-06,
+      "logits/chosen": 0.0862819030880928,
+      "logits/rejected": 0.16991934180259705,
+      "logps/chosen": -87.4437026977539,
+      "logps/rejected": -96.60844421386719,
+      "loss": 2.3041,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.230825185775757,
+      "rewards/margins": 0.5197271108627319,
+      "rewards/rejected": -2.7505524158477783,
+      "step": 517
+    },
+    {
+      "epoch": 0.9409627611262489,
+      "grad_norm": 2.1336302757263184,
+      "learning_rate": 8.824011767042631e-06,
+      "logits/chosen": 0.12070942670106888,
+      "logits/rejected": 0.19643370807170868,
+      "logps/chosen": -75.97718811035156,
+      "logps/rejected": -83.26056671142578,
+      "loss": 2.3468,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3768177032470703,
+      "rewards/margins": 0.5026894211769104,
+      "rewards/rejected": -2.879507064819336,
+      "step": 518
+    },
+    {
+      "epoch": 0.9427792915531336,
+      "grad_norm": 2.6361515522003174,
+      "learning_rate": 8.819676828442758e-06,
+      "logits/chosen": 0.04481849446892738,
+      "logits/rejected": -0.011357773095369339,
+      "logps/chosen": -84.09026336669922,
+      "logps/rejected": -87.71636199951172,
+      "loss": 2.8427,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.4249284267425537,
+      "rewards/margins": 0.41216185688972473,
+      "rewards/rejected": -2.837090015411377,
+      "step": 519
+    },
+    {
+      "epoch": 0.9445958219800181,
+      "grad_norm": 2.1068665981292725,
+      "learning_rate": 8.815335120443822e-06,
+      "logits/chosen": 0.1595858335494995,
+      "logits/rejected": 0.14017102122306824,
+      "logps/chosen": -94.00933837890625,
+      "logps/rejected": -96.91961669921875,
+      "loss": 2.6559,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.4731483459472656,
+      "rewards/margins": 0.3141450881958008,
+      "rewards/rejected": -2.7872931957244873,
+      "step": 520
+    },
+    {
+      "epoch": 0.9464123524069028,
+      "grad_norm": 1.7320570945739746,
+      "learning_rate": 8.810986651899322e-06,
+      "logits/chosen": 0.10371927917003632,
+      "logits/rejected": 0.17988254129886627,
+      "logps/chosen": -78.32308197021484,
+      "logps/rejected": -88.74859619140625,
+      "loss": 2.1866,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3973731994628906,
+      "rewards/margins": 0.7199669480323792,
+      "rewards/rejected": -3.117340087890625,
+      "step": 521
+    },
+    {
+      "epoch": 0.9482288828337875,
+      "grad_norm": 1.6995161771774292,
+      "learning_rate": 8.80663143167654e-06,
+      "logits/chosen": 0.14279219508171082,
+      "logits/rejected": 0.1058904379606247,
+      "logps/chosen": -87.79972839355469,
+      "logps/rejected": -94.35545349121094,
+      "loss": 2.1217,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.979785680770874,
+      "rewards/margins": 0.6210839152336121,
+      "rewards/rejected": -2.600869655609131,
+      "step": 522
+    },
+    {
+      "epoch": 0.9500454132606722,
+      "grad_norm": 2.1860764026641846,
+      "learning_rate": 8.80226946865653e-06,
+      "logits/chosen": 0.11601082980632782,
+      "logits/rejected": 0.15693408250808716,
+      "logps/chosen": -88.86433410644531,
+      "logps/rejected": -92.43590545654297,
+      "loss": 2.5629,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.2875983715057373,
+      "rewards/margins": 0.39402255415916443,
+      "rewards/rejected": -2.6816210746765137,
+      "step": 523
+    },
+    {
+      "epoch": 0.9518619436875567,
+      "grad_norm": 1.829032063484192,
+      "learning_rate": 8.797900771734094e-06,
+      "logits/chosen": 0.12328370660543442,
+      "logits/rejected": 0.042389824986457825,
+      "logps/chosen": -77.34864807128906,
+      "logps/rejected": -78.61405944824219,
+      "loss": 2.5327,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2514185905456543,
+      "rewards/margins": 0.3621898591518402,
+      "rewards/rejected": -2.6136088371276855,
+      "step": 524
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 2.235381841659546,
+      "learning_rate": 8.793525349817765e-06,
+      "logits/chosen": 0.12495981156826019,
+      "logits/rejected": 0.13291439414024353,
+      "logps/chosen": -76.7927474975586,
+      "logps/rejected": -86.1208267211914,
+      "loss": 2.5188,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3426284790039062,
+      "rewards/margins": 0.46077418327331543,
+      "rewards/rejected": -2.8034026622772217,
+      "step": 525
+    },
+    {
+      "epoch": 0.9554950045413261,
+      "grad_norm": 1.8071073293685913,
+      "learning_rate": 8.78914321182979e-06,
+      "logits/chosen": 0.13614074885845184,
+      "logits/rejected": 0.1179000735282898,
+      "logps/chosen": -74.44164276123047,
+      "logps/rejected": -77.69221496582031,
+      "loss": 2.6082,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0528411865234375,
+      "rewards/margins": 0.29493850469589233,
+      "rewards/rejected": -2.3477797508239746,
+      "step": 526
+    },
+    {
+      "epoch": 0.9573115349682108,
+      "grad_norm": 1.8623837232589722,
+      "learning_rate": 8.784754366706115e-06,
+      "logits/chosen": 0.02768833190202713,
+      "logits/rejected": 0.09584817290306091,
+      "logps/chosen": -77.13914489746094,
+      "logps/rejected": -86.53968048095703,
+      "loss": 2.2244,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.199755907058716,
+      "rewards/margins": 0.6271941661834717,
+      "rewards/rejected": -2.8269503116607666,
+      "step": 527
+    },
+    {
+      "epoch": 0.9591280653950953,
+      "grad_norm": 1.5299558639526367,
+      "learning_rate": 8.780358823396352e-06,
+      "logits/chosen": 0.11699292808771133,
+      "logits/rejected": 0.10016686469316483,
+      "logps/chosen": -86.1636962890625,
+      "logps/rejected": -87.16045379638672,
+      "loss": 2.0645,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.163857936859131,
+      "rewards/margins": 0.697452187538147,
+      "rewards/rejected": -2.8613100051879883,
+      "step": 528
+    },
+    {
+      "epoch": 0.96094459582198,
+      "grad_norm": 1.6236025094985962,
+      "learning_rate": 8.775956590863785e-06,
+      "logits/chosen": 0.11622033268213272,
+      "logits/rejected": 0.06629584729671478,
+      "logps/chosen": -76.42916870117188,
+      "logps/rejected": -80.24000549316406,
+      "loss": 1.8904,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.102374315261841,
+      "rewards/margins": 0.7230857014656067,
+      "rewards/rejected": -2.825460195541382,
+      "step": 529
+    },
+    {
+      "epoch": 0.9627611262488647,
+      "grad_norm": 1.6119427680969238,
+      "learning_rate": 8.771547678085332e-06,
+      "logits/chosen": 0.07009143382310867,
+      "logits/rejected": 0.16355563700199127,
+      "logps/chosen": -72.83528137207031,
+      "logps/rejected": -87.72441101074219,
+      "loss": 2.0136,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1697566509246826,
+      "rewards/margins": 0.8312336206436157,
+      "rewards/rejected": -3.000990390777588,
+      "step": 530
+    },
+    {
+      "epoch": 0.9645776566757494,
+      "grad_norm": 2.0324013233184814,
+      "learning_rate": 8.767132094051534e-06,
+      "logits/chosen": 0.14837129414081573,
+      "logits/rejected": 0.0700330063700676,
+      "logps/chosen": -84.65350341796875,
+      "logps/rejected": -91.84485626220703,
+      "loss": 2.2551,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5343966484069824,
+      "rewards/margins": 0.6457280516624451,
+      "rewards/rejected": -3.180124521255493,
+      "step": 531
+    },
+    {
+      "epoch": 0.9663941871026339,
+      "grad_norm": 1.4526044130325317,
+      "learning_rate": 8.762709847766532e-06,
+      "logits/chosen": 0.013620391488075256,
+      "logits/rejected": 0.10376289486885071,
+      "logps/chosen": -70.80522155761719,
+      "logps/rejected": -84.26114654541016,
+      "loss": 1.8607,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.249685287475586,
+      "rewards/margins": 0.8104004859924316,
+      "rewards/rejected": -3.0600855350494385,
+      "step": 532
+    },
+    {
+      "epoch": 0.9682107175295186,
+      "grad_norm": 2.0312576293945312,
+      "learning_rate": 8.758280948248059e-06,
+      "logits/chosen": 0.0965358167886734,
+      "logits/rejected": 0.10093791037797928,
+      "logps/chosen": -81.83265686035156,
+      "logps/rejected": -88.52181243896484,
+      "loss": 2.6128,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4901068210601807,
+      "rewards/margins": 0.30889907479286194,
+      "rewards/rejected": -2.7990057468414307,
+      "step": 533
+    },
+    {
+      "epoch": 0.9700272479564033,
+      "grad_norm": 2.547189712524414,
+      "learning_rate": 8.753845404527413e-06,
+      "logits/chosen": 0.020555848255753517,
+      "logits/rejected": 0.0945032387971878,
+      "logps/chosen": -72.16207122802734,
+      "logps/rejected": -86.27986907958984,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.148913860321045,
+      "rewards/margins": 0.7298768162727356,
+      "rewards/rejected": -2.8787906169891357,
+      "step": 534
+    },
+    {
+      "epoch": 0.971843778383288,
+      "grad_norm": 1.996645450592041,
+      "learning_rate": 8.74940322564944e-06,
+      "logits/chosen": 0.0345986932516098,
+      "logits/rejected": 0.08187264204025269,
+      "logps/chosen": -78.76763916015625,
+      "logps/rejected": -91.45555114746094,
+      "loss": 2.4587,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.4772286415100098,
+      "rewards/margins": 0.4609605073928833,
+      "rewards/rejected": -2.9381890296936035,
+      "step": 535
+    },
+    {
+      "epoch": 0.9736603088101726,
+      "grad_norm": 1.8145204782485962,
+      "learning_rate": 8.744954420672514e-06,
+      "logits/chosen": 0.03380711376667023,
+      "logits/rejected": 0.09039584547281265,
+      "logps/chosen": -79.83094787597656,
+      "logps/rejected": -90.87137603759766,
+      "loss": 2.3589,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.1433308124542236,
+      "rewards/margins": 0.5286959409713745,
+      "rewards/rejected": -2.6720268726348877,
+      "step": 536
+    },
+    {
+      "epoch": 0.9754768392370572,
+      "grad_norm": 1.9745460748672485,
+      "learning_rate": 8.740498998668523e-06,
+      "logits/chosen": 0.07233883440494537,
+      "logits/rejected": 0.17097605764865875,
+      "logps/chosen": -72.3458251953125,
+      "logps/rejected": -83.47319793701172,
+      "loss": 2.3469,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199948310852051,
+      "rewards/margins": 0.47777751088142395,
+      "rewards/rejected": -2.6777257919311523,
+      "step": 537
+    },
+    {
+      "epoch": 0.9772933696639419,
+      "grad_norm": 1.665255069732666,
+      "learning_rate": 8.736036968722851e-06,
+      "logits/chosen": 0.09596951305866241,
+      "logits/rejected": 0.06625551730394363,
+      "logps/chosen": -80.11689758300781,
+      "logps/rejected": -88.4573974609375,
+      "loss": 2.1889,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1831865310668945,
+      "rewards/margins": 0.6106584668159485,
+      "rewards/rejected": -2.7938451766967773,
+      "step": 538
+    },
+    {
+      "epoch": 0.9791099000908265,
+      "grad_norm": 1.5587108135223389,
+      "learning_rate": 8.73156833993435e-06,
+      "logits/chosen": 0.07950340211391449,
+      "logits/rejected": 0.0903201699256897,
+      "logps/chosen": -78.9182357788086,
+      "logps/rejected": -92.290283203125,
+      "loss": 2.048,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.12032413482666,
+      "rewards/margins": 0.7009568214416504,
+      "rewards/rejected": -2.8212809562683105,
+      "step": 539
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 1.7177670001983643,
+      "learning_rate": 8.727093121415338e-06,
+      "logits/chosen": 0.06052964925765991,
+      "logits/rejected": 0.14043518900871277,
+      "logps/chosen": -79.04943084716797,
+      "logps/rejected": -88.881591796875,
+      "loss": 2.2089,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2680561542510986,
+      "rewards/margins": 0.5380589962005615,
+      "rewards/rejected": -2.806114912033081,
+      "step": 540
+    },
+    {
+      "epoch": 0.9827429609445958,
+      "grad_norm": 1.7630436420440674,
+      "learning_rate": 8.722611322291558e-06,
+      "logits/chosen": 0.02032918483018875,
+      "logits/rejected": 0.11728623509407043,
+      "logps/chosen": -78.1409912109375,
+      "logps/rejected": -89.3685073852539,
+      "loss": 2.3303,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.356957197189331,
+      "rewards/margins": 0.5139226317405701,
+      "rewards/rejected": -2.870880126953125,
+      "step": 541
+    },
+    {
+      "epoch": 0.9845594913714805,
+      "grad_norm": 1.852418065071106,
+      "learning_rate": 8.718122951702183e-06,
+      "logits/chosen": 0.0439969077706337,
+      "logits/rejected": 0.16671347618103027,
+      "logps/chosen": -69.96346282958984,
+      "logps/rejected": -85.24410247802734,
+      "loss": 2.1393,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3295838832855225,
+      "rewards/margins": 0.6544574499130249,
+      "rewards/rejected": -2.984041213989258,
+      "step": 542
+    },
+    {
+      "epoch": 0.9863760217983651,
+      "grad_norm": 1.9028044939041138,
+      "learning_rate": 8.713628018799782e-06,
+      "logits/chosen": 0.04875154793262482,
+      "logits/rejected": -0.009283583611249924,
+      "logps/chosen": -75.80145263671875,
+      "logps/rejected": -78.19249725341797,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.241586208343506,
+      "rewards/margins": 0.4672737419605255,
+      "rewards/rejected": -2.708860158920288,
+      "step": 543
+    },
+    {
+      "epoch": 0.9881925522252498,
+      "grad_norm": 1.5648458003997803,
+      "learning_rate": 8.709126532750304e-06,
+      "logits/chosen": 0.02982347458600998,
+      "logits/rejected": 0.07835812866687775,
+      "logps/chosen": -75.35140991210938,
+      "logps/rejected": -91.11531829833984,
+      "loss": 2.0945,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.062386989593506,
+      "rewards/margins": 0.6832989454269409,
+      "rewards/rejected": -2.7456860542297363,
+      "step": 544
+    },
+    {
+      "epoch": 0.9900090826521344,
+      "grad_norm": 1.6214131116867065,
+      "learning_rate": 8.704618502733064e-06,
+      "logits/chosen": 0.09551770985126495,
+      "logits/rejected": 0.11310344189405441,
+      "logps/chosen": -80.15961456298828,
+      "logps/rejected": -93.38126373291016,
+      "loss": 2.1642,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.255985736846924,
+      "rewards/margins": 0.7712126970291138,
+      "rewards/rejected": -3.027198553085327,
+      "step": 545
+    },
+    {
+      "epoch": 0.9918256130790191,
+      "grad_norm": 1.762953519821167,
+      "learning_rate": 8.70010393794072e-06,
+      "logits/chosen": 0.0971008688211441,
+      "logits/rejected": 0.16590046882629395,
+      "logps/chosen": -86.24911499023438,
+      "logps/rejected": -102.59831237792969,
+      "loss": 2.0951,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3542394638061523,
+      "rewards/margins": 0.7857018709182739,
+      "rewards/rejected": -3.139941453933716,
+      "step": 546
+    },
+    {
+      "epoch": 0.9936421435059037,
+      "grad_norm": 1.6033780574798584,
+      "learning_rate": 8.695582847579254e-06,
+      "logits/chosen": 0.15147003531455994,
+      "logits/rejected": 0.13297875225543976,
+      "logps/chosen": -78.5172348022461,
+      "logps/rejected": -86.49285888671875,
+      "loss": 2.1101,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2025821208953857,
+      "rewards/margins": 0.7471145391464233,
+      "rewards/rejected": -2.9496963024139404,
+      "step": 547
+    },
+    {
+      "epoch": 0.9954586739327884,
+      "grad_norm": 1.4905331134796143,
+      "learning_rate": 8.69105524086796e-06,
+      "logits/chosen": 0.07559307664632797,
+      "logits/rejected": 0.09286531805992126,
+      "logps/chosen": -74.92215728759766,
+      "logps/rejected": -78.23614501953125,
+      "loss": 2.2745,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.24694561958313,
+      "rewards/margins": 0.5010396838188171,
+      "rewards/rejected": -2.747985363006592,
+      "step": 548
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.5850121523350733e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-548/training_args.bin b/checkpoint-548/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-548/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-548/zero_to_fp32.py b/checkpoint-548/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-548/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-685/README.md b/checkpoint-685/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-685/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-685/adapter_config.json b/checkpoint-685/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-685/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-685/adapter_model.safetensors b/checkpoint-685/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..dc56bbad0bc86fea0ac4e56c4f053f7ffd8a96a1
--- /dev/null
+++ b/checkpoint-685/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4fa335202f0556d741fec1f4d2305d2af475a0932de05191ff6d6c814d453f3c
+size 207244392
diff --git a/checkpoint-685/latest b/checkpoint-685/latest
new file mode 100644
index 0000000000000000000000000000000000000000..32b7f894d10e5e12f7ef9cea66d082aaff9baad6
--- /dev/null
+++ b/checkpoint-685/latest
@@ -0,0 +1 @@
+global_step684
\ No newline at end of file
diff --git a/checkpoint-685/rng_state_0.pth b/checkpoint-685/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b346349ce12dd5a17d4b91ed2a5722bb52550950
--- /dev/null
+++ b/checkpoint-685/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
+size 15984
diff --git a/checkpoint-685/rng_state_1.pth b/checkpoint-685/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..68f3c6994456cb8d0592a5375d99503c8924b1c4
--- /dev/null
+++ b/checkpoint-685/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73
+size 15984
diff --git a/checkpoint-685/rng_state_2.pth b/checkpoint-685/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..be044f6ceeed587d30e80c2f72d5aa19fdc9947b
--- /dev/null
+++ b/checkpoint-685/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b
+size 15984
diff --git a/checkpoint-685/rng_state_3.pth b/checkpoint-685/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..fc825249656a9b858782542bd3f4386250f1dfe0
--- /dev/null
+++ b/checkpoint-685/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc
+size 15984
diff --git a/checkpoint-685/rng_state_4.pth b/checkpoint-685/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d30f52a44be563c152ae09db6ae934da6da0d3ed
--- /dev/null
+++ b/checkpoint-685/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972
+size 15984
diff --git a/checkpoint-685/rng_state_5.pth b/checkpoint-685/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..c8715d27ab23ae545d58039cf949cc44ecc1da5e
--- /dev/null
+++ b/checkpoint-685/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991
+size 15984
diff --git a/checkpoint-685/rng_state_6.pth b/checkpoint-685/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1ed791b6ef76eadf0b0c55a5733411771e2ae027
--- /dev/null
+++ b/checkpoint-685/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa
+size 15984
diff --git a/checkpoint-685/rng_state_7.pth b/checkpoint-685/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..800c3bbbc5edf7db01a8316069d439c5fb8d8c30
--- /dev/null
+++ b/checkpoint-685/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773
+size 15984
diff --git a/checkpoint-685/special_tokens_map.json b/checkpoint-685/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-685/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-685/tokenizer.json b/checkpoint-685/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-685/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-685/tokenizer_config.json b/checkpoint-685/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-685/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-685/trainer_state.json b/checkpoint-685/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a4e4d5eaf07e46436ac74840f440e627a4bbf46e
--- /dev/null
+++ b/checkpoint-685/trainer_state.json
@@ -0,0 +1,10308 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.2434150772025432,
+  "eval_steps": 500,
+  "global_step": 685,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    },
+    {
+      "epoch": 0.49954586739327883,
+      "grad_norm": 2.1769776344299316,
+      "learning_rate": 9.65745789630079e-06,
+      "logits/chosen": 0.11112834513187408,
+      "logits/rejected": 0.10816515237092972,
+      "logps/chosen": -83.5405044555664,
+      "logps/rejected": -83.06329345703125,
+      "loss": 3.0939,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.142024040222168,
+      "rewards/margins": 0.1820124089717865,
+      "rewards/rejected": -2.3240363597869873,
+      "step": 275
+    },
+    {
+      "epoch": 0.5013623978201635,
+      "grad_norm": 1.7329221963882446,
+      "learning_rate": 9.654994540260396e-06,
+      "logits/chosen": 0.0653618574142456,
+      "logits/rejected": 0.08004368096590042,
+      "logps/chosen": -80.83209228515625,
+      "logps/rejected": -82.97142028808594,
+      "loss": 2.7161,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.851841688156128,
+      "rewards/margins": 0.2948168218135834,
+      "rewards/rejected": -2.146658420562744,
+      "step": 276
+    },
+    {
+      "epoch": 0.5031789282470481,
+      "grad_norm": 1.592657208442688,
+      "learning_rate": 9.65252271146268e-06,
+      "logits/chosen": 0.09880789369344711,
+      "logits/rejected": 0.14229761064052582,
+      "logps/chosen": -67.1727294921875,
+      "logps/rejected": -75.03417205810547,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8714643716812134,
+      "rewards/margins": 0.37937116622924805,
+      "rewards/rejected": -2.250835657119751,
+      "step": 277
+    },
+    {
+      "epoch": 0.5049954586739328,
+      "grad_norm": 1.6487712860107422,
+      "learning_rate": 9.650042414948133e-06,
+      "logits/chosen": 0.13465353846549988,
+      "logits/rejected": 0.12865689396858215,
+      "logps/chosen": -76.4417724609375,
+      "logps/rejected": -78.6947021484375,
+      "loss": 2.6085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.875291347503662,
+      "rewards/margins": 0.35175251960754395,
+      "rewards/rejected": -2.227043867111206,
+      "step": 278
+    },
+    {
+      "epoch": 0.5068119891008175,
+      "grad_norm": 1.6523009538650513,
+      "learning_rate": 9.64755365577451e-06,
+      "logits/chosen": 0.04238567873835564,
+      "logits/rejected": 0.07994347810745239,
+      "logps/chosen": -76.92097473144531,
+      "logps/rejected": -83.20886993408203,
+      "loss": 2.3156,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9759610891342163,
+      "rewards/margins": 0.5398927330970764,
+      "rewards/rejected": -2.5158536434173584,
+      "step": 279
+    },
+    {
+      "epoch": 0.508628519527702,
+      "grad_norm": 1.7999261617660522,
+      "learning_rate": 9.645056439016827e-06,
+      "logits/chosen": 0.07349395751953125,
+      "logits/rejected": 0.07518415153026581,
+      "logps/chosen": -77.56079864501953,
+      "logps/rejected": -84.97645568847656,
+      "loss": 2.8085,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9408115148544312,
+      "rewards/margins": 0.3470652997493744,
+      "rewards/rejected": -2.287877082824707,
+      "step": 280
+    },
+    {
+      "epoch": 0.5104450499545867,
+      "grad_norm": 1.5684200525283813,
+      "learning_rate": 9.642550769767342e-06,
+      "logits/chosen": 0.16188879311084747,
+      "logits/rejected": 0.12772323191165924,
+      "logps/chosen": -89.29315185546875,
+      "logps/rejected": -94.35065460205078,
+      "loss": 2.2314,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8958841562271118,
+      "rewards/margins": 0.5102108716964722,
+      "rewards/rejected": -2.406095027923584,
+      "step": 281
+    },
+    {
+      "epoch": 0.5122615803814714,
+      "grad_norm": 1.7878178358078003,
+      "learning_rate": 9.640036653135548e-06,
+      "logits/chosen": 0.060573749244213104,
+      "logits/rejected": 0.13457715511322021,
+      "logps/chosen": -68.9404525756836,
+      "logps/rejected": -74.77693176269531,
+      "loss": 2.4359,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.889530897140503,
+      "rewards/margins": 0.45911547541618347,
+      "rewards/rejected": -2.348646402359009,
+      "step": 282
+    },
+    {
+      "epoch": 0.5140781108083561,
+      "grad_norm": 1.4741288423538208,
+      "learning_rate": 9.637514094248172e-06,
+      "logits/chosen": 0.10433132946491241,
+      "logits/rejected": 0.1439315229654312,
+      "logps/chosen": -74.89447021484375,
+      "logps/rejected": -86.58551788330078,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9090807437896729,
+      "rewards/margins": 0.6733560562133789,
+      "rewards/rejected": -2.5824368000030518,
+      "step": 283
+    },
+    {
+      "epoch": 0.5158946412352406,
+      "grad_norm": 1.2864018678665161,
+      "learning_rate": 9.634983098249146e-06,
+      "logits/chosen": 0.10891089588403702,
+      "logits/rejected": 0.11755162477493286,
+      "logps/chosen": -66.90985107421875,
+      "logps/rejected": -71.8125228881836,
+      "loss": 2.1538,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.7779250144958496,
+      "rewards/margins": 0.5625147223472595,
+      "rewards/rejected": -2.340439796447754,
+      "step": 284
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 1.465747594833374,
+      "learning_rate": 9.632443670299616e-06,
+      "logits/chosen": 0.08224496245384216,
+      "logits/rejected": 0.12130744755268097,
+      "logps/chosen": -75.4281997680664,
+      "logps/rejected": -85.0781021118164,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7217226028442383,
+      "rewards/margins": 0.4750482439994812,
+      "rewards/rejected": -2.196770668029785,
+      "step": 285
+    },
+    {
+      "epoch": 0.51952770208901,
+      "grad_norm": 1.492859959602356,
+      "learning_rate": 9.629895815577915e-06,
+      "logits/chosen": 0.06619664281606674,
+      "logits/rejected": 0.13152630627155304,
+      "logps/chosen": -96.65383911132812,
+      "logps/rejected": -111.93521881103516,
+      "loss": 2.2831,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7611618041992188,
+      "rewards/margins": 0.5886337757110596,
+      "rewards/rejected": -2.349795341491699,
+      "step": 286
+    },
+    {
+      "epoch": 0.5213442325158947,
+      "grad_norm": 1.5534065961837769,
+      "learning_rate": 9.627339539279564e-06,
+      "logits/chosen": 0.06637927144765854,
+      "logits/rejected": 0.09107412397861481,
+      "logps/chosen": -71.92534637451172,
+      "logps/rejected": -82.98391723632812,
+      "loss": 2.5101,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7867075204849243,
+      "rewards/margins": 0.39493298530578613,
+      "rewards/rejected": -2.181640625,
+      "step": 287
+    },
+    {
+      "epoch": 0.5231607629427792,
+      "grad_norm": 1.6778221130371094,
+      "learning_rate": 9.624774846617254e-06,
+      "logits/chosen": 0.14700329303741455,
+      "logits/rejected": 0.12778782844543457,
+      "logps/chosen": -65.2364730834961,
+      "logps/rejected": -74.39017486572266,
+      "loss": 2.5236,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6942007541656494,
+      "rewards/margins": 0.35334131121635437,
+      "rewards/rejected": -2.047542095184326,
+      "step": 288
+    },
+    {
+      "epoch": 0.5249772933696639,
+      "grad_norm": 1.462215542793274,
+      "learning_rate": 9.622201742820839e-06,
+      "logits/chosen": 0.11467991769313812,
+      "logits/rejected": 0.1028795838356018,
+      "logps/chosen": -67.85939025878906,
+      "logps/rejected": -74.3462142944336,
+      "loss": 2.3891,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7194863557815552,
+      "rewards/margins": 0.3846975564956665,
+      "rewards/rejected": -2.1041836738586426,
+      "step": 289
+    },
+    {
+      "epoch": 0.5267938237965486,
+      "grad_norm": 1.7591735124588013,
+      "learning_rate": 9.619620233137328e-06,
+      "logits/chosen": 0.08407986164093018,
+      "logits/rejected": 0.12201236933469772,
+      "logps/chosen": -82.57340240478516,
+      "logps/rejected": -86.94192504882812,
+      "loss": 2.4569,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8536537885665894,
+      "rewards/margins": 0.38806283473968506,
+      "rewards/rejected": -2.2417166233062744,
+      "step": 290
+    },
+    {
+      "epoch": 0.5286103542234333,
+      "grad_norm": 1.5584774017333984,
+      "learning_rate": 9.617030322830868e-06,
+      "logits/chosen": 0.1126309484243393,
+      "logits/rejected": 0.08907752484083176,
+      "logps/chosen": -73.36901092529297,
+      "logps/rejected": -76.65506744384766,
+      "loss": 2.4021,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8091729879379272,
+      "rewards/margins": 0.3603411316871643,
+      "rewards/rejected": -2.1695141792297363,
+      "step": 291
+    },
+    {
+      "epoch": 0.5304268846503178,
+      "grad_norm": 1.487938642501831,
+      "learning_rate": 9.614432017182736e-06,
+      "logits/chosen": 0.09620118141174316,
+      "logits/rejected": 0.08514149487018585,
+      "logps/chosen": -88.66265106201172,
+      "logps/rejected": -92.26091766357422,
+      "loss": 2.3636,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.5980674028396606,
+      "rewards/margins": 0.4888036549091339,
+      "rewards/rejected": -2.0868711471557617,
+      "step": 292
+    },
+    {
+      "epoch": 0.5322434150772025,
+      "grad_norm": 1.445408582687378,
+      "learning_rate": 9.611825321491331e-06,
+      "logits/chosen": 0.10014252364635468,
+      "logits/rejected": 0.12857215106487274,
+      "logps/chosen": -82.46649932861328,
+      "logps/rejected": -92.84205627441406,
+      "loss": 2.343,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4620139598846436,
+      "rewards/margins": 0.5616316199302673,
+      "rewards/rejected": -2.0236456394195557,
+      "step": 293
+    },
+    {
+      "epoch": 0.5340599455040872,
+      "grad_norm": 1.262193202972412,
+      "learning_rate": 9.609210241072158e-06,
+      "logits/chosen": 0.09868282079696655,
+      "logits/rejected": 0.16074486076831818,
+      "logps/chosen": -67.0407485961914,
+      "logps/rejected": -82.36201477050781,
+      "loss": 2.1163,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5362765789031982,
+      "rewards/margins": 0.7091963887214661,
+      "rewards/rejected": -2.2454731464385986,
+      "step": 294
+    },
+    {
+      "epoch": 0.5358764759309719,
+      "grad_norm": 1.601891279220581,
+      "learning_rate": 9.606586781257822e-06,
+      "logits/chosen": 0.1236480325460434,
+      "logits/rejected": 0.12187935411930084,
+      "logps/chosen": -78.2726058959961,
+      "logps/rejected": -84.49217987060547,
+      "loss": 2.5075,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.779215931892395,
+      "rewards/margins": 0.38987410068511963,
+      "rewards/rejected": -2.1690900325775146,
+      "step": 295
+    },
+    {
+      "epoch": 0.5376930063578564,
+      "grad_norm": 1.4211548566818237,
+      "learning_rate": 9.603954947398016e-06,
+      "logits/chosen": 0.18375667929649353,
+      "logits/rejected": 0.21945703029632568,
+      "logps/chosen": -70.29829406738281,
+      "logps/rejected": -78.42284393310547,
+      "loss": 2.1959,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.7448753118515015,
+      "rewards/margins": 0.5106840133666992,
+      "rewards/rejected": -2.2555594444274902,
+      "step": 296
+    },
+    {
+      "epoch": 0.5395095367847411,
+      "grad_norm": 1.458617925643921,
+      "learning_rate": 9.601314744859504e-06,
+      "logits/chosen": 0.07360847294330597,
+      "logits/rejected": 0.1596718430519104,
+      "logps/chosen": -72.28546142578125,
+      "logps/rejected": -87.13935089111328,
+      "loss": 2.3538,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8007893562316895,
+      "rewards/margins": 0.5433144569396973,
+      "rewards/rejected": -2.3441038131713867,
+      "step": 297
+    },
+    {
+      "epoch": 0.5413260672116258,
+      "grad_norm": 1.6526339054107666,
+      "learning_rate": 9.598666179026123e-06,
+      "logits/chosen": 0.12013350427150726,
+      "logits/rejected": 0.10237178206443787,
+      "logps/chosen": -83.06564331054688,
+      "logps/rejected": -85.60771942138672,
+      "loss": 2.4531,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.911988615989685,
+      "rewards/margins": 0.35885563492774963,
+      "rewards/rejected": -2.2708444595336914,
+      "step": 298
+    },
+    {
+      "epoch": 0.5431425976385105,
+      "grad_norm": 1.9808340072631836,
+      "learning_rate": 9.596009255298755e-06,
+      "logits/chosen": 0.062342576682567596,
+      "logits/rejected": 0.09053834527730942,
+      "logps/chosen": -96.33489990234375,
+      "logps/rejected": -93.43024444580078,
+      "loss": 2.8549,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.845227599143982,
+      "rewards/margins": 0.22291362285614014,
+      "rewards/rejected": -2.068141222000122,
+      "step": 299
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 1.6276236772537231,
+      "learning_rate": 9.593343979095334e-06,
+      "logits/chosen": 0.2073422521352768,
+      "logits/rejected": 0.13107003271579742,
+      "logps/chosen": -72.45758056640625,
+      "logps/rejected": -77.42770385742188,
+      "loss": 2.4644,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9407453536987305,
+      "rewards/margins": 0.3134301006793976,
+      "rewards/rejected": -2.2541751861572266,
+      "step": 300
+    },
+    {
+      "epoch": 0.5467756584922797,
+      "grad_norm": 1.5001753568649292,
+      "learning_rate": 9.590670355850819e-06,
+      "logits/chosen": 0.15097060799598694,
+      "logits/rejected": 0.16254279017448425,
+      "logps/chosen": -70.44524383544922,
+      "logps/rejected": -70.64158630371094,
+      "loss": 2.4159,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.6921097040176392,
+      "rewards/margins": 0.3336741626262665,
+      "rewards/rejected": -2.0257837772369385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5485921889191644,
+      "grad_norm": 1.8360233306884766,
+      "learning_rate": 9.587988391017198e-06,
+      "logits/chosen": 0.14593760669231415,
+      "logits/rejected": 0.12311654537916183,
+      "logps/chosen": -78.32576751708984,
+      "logps/rejected": -88.25840759277344,
+      "loss": 2.5826,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.6591644287109375,
+      "rewards/margins": 0.5071319341659546,
+      "rewards/rejected": -2.1662964820861816,
+      "step": 302
+    },
+    {
+      "epoch": 0.5504087193460491,
+      "grad_norm": 1.4822838306427002,
+      "learning_rate": 9.585298090063459e-06,
+      "logits/chosen": 0.20818498730659485,
+      "logits/rejected": 0.16436657309532166,
+      "logps/chosen": -67.59429931640625,
+      "logps/rejected": -77.14763641357422,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6798239946365356,
+      "rewards/margins": 0.5954271554946899,
+      "rewards/rejected": -2.2752511501312256,
+      "step": 303
+    },
+    {
+      "epoch": 0.5522252497729337,
+      "grad_norm": 1.6118305921554565,
+      "learning_rate": 9.582599458475598e-06,
+      "logits/chosen": 0.0827561467885971,
+      "logits/rejected": 0.09151773154735565,
+      "logps/chosen": -73.28964233398438,
+      "logps/rejected": -76.90730285644531,
+      "loss": 2.3792,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.8056440353393555,
+      "rewards/margins": 0.39551618695259094,
+      "rewards/rejected": -2.201160192489624,
+      "step": 304
+    },
+    {
+      "epoch": 0.5540417801998183,
+      "grad_norm": 1.4609856605529785,
+      "learning_rate": 9.579892501756593e-06,
+      "logits/chosen": 0.05332394689321518,
+      "logits/rejected": 0.1590057611465454,
+      "logps/chosen": -77.51653289794922,
+      "logps/rejected": -95.51261138916016,
+      "loss": 2.2105,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.740310549736023,
+      "rewards/margins": 0.6771562695503235,
+      "rewards/rejected": -2.4174671173095703,
+      "step": 305
+    },
+    {
+      "epoch": 0.555858310626703,
+      "grad_norm": 1.7181960344314575,
+      "learning_rate": 9.5771772254264e-06,
+      "logits/chosen": 0.05865276977419853,
+      "logits/rejected": 0.03500773757696152,
+      "logps/chosen": -83.72650146484375,
+      "logps/rejected": -93.69804382324219,
+      "loss": 2.1687,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6025928258895874,
+      "rewards/margins": 0.7339029312133789,
+      "rewards/rejected": -2.3364956378936768,
+      "step": 306
+    },
+    {
+      "epoch": 0.5576748410535877,
+      "grad_norm": 1.497698187828064,
+      "learning_rate": 9.57445363502194e-06,
+      "logits/chosen": 0.1225174218416214,
+      "logits/rejected": 0.09580346196889877,
+      "logps/chosen": -76.27323150634766,
+      "logps/rejected": -76.90716552734375,
+      "loss": 2.2116,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.715336799621582,
+      "rewards/margins": 0.4648086428642273,
+      "rewards/rejected": -2.180145263671875,
+      "step": 307
+    },
+    {
+      "epoch": 0.5594913714804723,
+      "grad_norm": 1.3957064151763916,
+      "learning_rate": 9.571721736097089e-06,
+      "logits/chosen": 0.07596514374017715,
+      "logits/rejected": 0.14832191169261932,
+      "logps/chosen": -84.65274810791016,
+      "logps/rejected": -105.76422119140625,
+      "loss": 1.9942,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.6643587350845337,
+      "rewards/margins": 0.7884883284568787,
+      "rewards/rejected": -2.4528470039367676,
+      "step": 308
+    },
+    {
+      "epoch": 0.5613079019073569,
+      "grad_norm": 1.6824349164962769,
+      "learning_rate": 9.568981534222664e-06,
+      "logits/chosen": 0.049865882843732834,
+      "logits/rejected": 0.023348212242126465,
+      "logps/chosen": -80.3697509765625,
+      "logps/rejected": -86.07074737548828,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.009049654006958,
+      "rewards/margins": 0.5470253825187683,
+      "rewards/rejected": -2.556075096130371,
+      "step": 309
+    },
+    {
+      "epoch": 0.5631244323342416,
+      "grad_norm": 1.8752086162567139,
+      "learning_rate": 9.566233034986413e-06,
+      "logits/chosen": 0.07250591367483139,
+      "logits/rejected": 0.12809628248214722,
+      "logps/chosen": -74.84794616699219,
+      "logps/rejected": -81.79740905761719,
+      "loss": 2.6286,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.75887131690979,
+      "rewards/margins": 0.33979111909866333,
+      "rewards/rejected": -2.0986623764038086,
+      "step": 310
+    },
+    {
+      "epoch": 0.5649409627611263,
+      "grad_norm": 1.7387233972549438,
+      "learning_rate": 9.563476243993008e-06,
+      "logits/chosen": 0.130618155002594,
+      "logits/rejected": 0.12952059507369995,
+      "logps/chosen": -80.75495910644531,
+      "logps/rejected": -90.4281234741211,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.836004614830017,
+      "rewards/margins": 0.4418962597846985,
+      "rewards/rejected": -2.2779006958007812,
+      "step": 311
+    },
+    {
+      "epoch": 0.5667574931880109,
+      "grad_norm": 1.5147444009780884,
+      "learning_rate": 9.56071116686402e-06,
+      "logits/chosen": 0.10193713754415512,
+      "logits/rejected": 0.22481802105903625,
+      "logps/chosen": -73.21926879882812,
+      "logps/rejected": -81.27864837646484,
+      "loss": 2.578,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6358015537261963,
+      "rewards/margins": 0.38279739022254944,
+      "rewards/rejected": -2.018598794937134,
+      "step": 312
+    },
+    {
+      "epoch": 0.5685740236148955,
+      "grad_norm": 1.6253665685653687,
+      "learning_rate": 9.557937809237927e-06,
+      "logits/chosen": 0.09468917548656464,
+      "logits/rejected": 0.09415112435817719,
+      "logps/chosen": -80.62995147705078,
+      "logps/rejected": -86.65946197509766,
+      "loss": 2.3854,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8635404109954834,
+      "rewards/margins": 0.4438764154911041,
+      "rewards/rejected": -2.3074169158935547,
+      "step": 313
+    },
+    {
+      "epoch": 0.5703905540417802,
+      "grad_norm": 1.7893344163894653,
+      "learning_rate": 9.555156176770087e-06,
+      "logits/chosen": 0.15863659977912903,
+      "logits/rejected": 0.09485571831464767,
+      "logps/chosen": -76.10442352294922,
+      "logps/rejected": -72.73162078857422,
+      "loss": 2.6795,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.935206413269043,
+      "rewards/margins": 0.22100940346717834,
+      "rewards/rejected": -2.1562156677246094,
+      "step": 314
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 1.696327805519104,
+      "learning_rate": 9.552366275132733e-06,
+      "logits/chosen": 0.07012113183736801,
+      "logits/rejected": 0.13891686499118805,
+      "logps/chosen": -78.08012390136719,
+      "logps/rejected": -83.05044555664062,
+      "loss": 2.6181,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7869051694869995,
+      "rewards/margins": 0.23424415290355682,
+      "rewards/rejected": -2.0211493968963623,
+      "step": 315
+    },
+    {
+      "epoch": 0.5740236148955495,
+      "grad_norm": 1.6044729948043823,
+      "learning_rate": 9.54956811001496e-06,
+      "logits/chosen": 0.11298641562461853,
+      "logits/rejected": 0.2595204710960388,
+      "logps/chosen": -75.5998306274414,
+      "logps/rejected": -85.31849670410156,
+      "loss": 2.5849,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.5757699012756348,
+      "rewards/margins": 0.35041162371635437,
+      "rewards/rejected": -1.9261815547943115,
+      "step": 316
+    },
+    {
+      "epoch": 0.5758401453224341,
+      "grad_norm": 1.7399548292160034,
+      "learning_rate": 9.546761687122715e-06,
+      "logits/chosen": 0.14933931827545166,
+      "logits/rejected": 0.1868455857038498,
+      "logps/chosen": -68.05965423583984,
+      "logps/rejected": -82.37442779541016,
+      "loss": 2.3208,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8474600315093994,
+      "rewards/margins": 0.5766161680221558,
+      "rewards/rejected": -2.4240763187408447,
+      "step": 317
+    },
+    {
+      "epoch": 0.5776566757493188,
+      "grad_norm": 1.597495675086975,
+      "learning_rate": 9.54394701217878e-06,
+      "logits/chosen": 0.04729313403367996,
+      "logits/rejected": 0.05815067142248154,
+      "logps/chosen": -97.36541748046875,
+      "logps/rejected": -104.02650451660156,
+      "loss": 2.2106,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8631618022918701,
+      "rewards/margins": 0.6027428507804871,
+      "rewards/rejected": -2.465904474258423,
+      "step": 318
+    },
+    {
+      "epoch": 0.5794732061762035,
+      "grad_norm": 1.8646149635314941,
+      "learning_rate": 9.541124090922771e-06,
+      "logits/chosen": 0.1399674415588379,
+      "logits/rejected": 0.11107950657606125,
+      "logps/chosen": -82.39740753173828,
+      "logps/rejected": -76.80902099609375,
+      "loss": 2.6435,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.7670300006866455,
+      "rewards/margins": 0.15345275402069092,
+      "rewards/rejected": -1.920482873916626,
+      "step": 319
+    },
+    {
+      "epoch": 0.5812897366030881,
+      "grad_norm": 1.502447247505188,
+      "learning_rate": 9.538292929111114e-06,
+      "logits/chosen": 0.10672347247600555,
+      "logits/rejected": 0.08290571719408035,
+      "logps/chosen": -74.72683715820312,
+      "logps/rejected": -86.37158203125,
+      "loss": 2.2137,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8041073083877563,
+      "rewards/margins": 0.596272885799408,
+      "rewards/rejected": -2.4003803730010986,
+      "step": 320
+    },
+    {
+      "epoch": 0.5831062670299727,
+      "grad_norm": 1.7287745475769043,
+      "learning_rate": 9.535453532517039e-06,
+      "logits/chosen": 0.13742896914482117,
+      "logits/rejected": 0.07607734203338623,
+      "logps/chosen": -86.82475280761719,
+      "logps/rejected": -77.8934555053711,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7918696403503418,
+      "rewards/margins": 0.2591831088066101,
+      "rewards/rejected": -2.0510525703430176,
+      "step": 321
+    },
+    {
+      "epoch": 0.5849227974568574,
+      "grad_norm": 1.5914596319198608,
+      "learning_rate": 9.532605906930575e-06,
+      "logits/chosen": 0.11301672458648682,
+      "logits/rejected": 0.1944851577281952,
+      "logps/chosen": -68.9898452758789,
+      "logps/rejected": -73.17451477050781,
+      "loss": 2.4565,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6999026536941528,
+      "rewards/margins": 0.30462026596069336,
+      "rewards/rejected": -2.0045228004455566,
+      "step": 322
+    },
+    {
+      "epoch": 0.5867393278837421,
+      "grad_norm": 1.6072031259536743,
+      "learning_rate": 9.529750058158522e-06,
+      "logits/chosen": 0.07092760503292084,
+      "logits/rejected": 0.08035591244697571,
+      "logps/chosen": -81.5555419921875,
+      "logps/rejected": -81.93560028076172,
+      "loss": 2.2729,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6834638118743896,
+      "rewards/margins": 0.48882579803466797,
+      "rewards/rejected": -2.1722893714904785,
+      "step": 323
+    },
+    {
+      "epoch": 0.5885558583106267,
+      "grad_norm": 1.6433221101760864,
+      "learning_rate": 9.526885992024453e-06,
+      "logits/chosen": 0.13823899626731873,
+      "logits/rejected": 0.13610433042049408,
+      "logps/chosen": -78.9625244140625,
+      "logps/rejected": -83.2666015625,
+      "loss": 2.2075,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6939194202423096,
+      "rewards/margins": 0.5755473971366882,
+      "rewards/rejected": -2.2694668769836426,
+      "step": 324
+    },
+    {
+      "epoch": 0.5903723887375113,
+      "grad_norm": 1.7674918174743652,
+      "learning_rate": 9.524013714368702e-06,
+      "logits/chosen": 0.1704932153224945,
+      "logits/rejected": 0.1073535829782486,
+      "logps/chosen": -72.47895050048828,
+      "logps/rejected": -73.0809326171875,
+      "loss": 2.5473,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8390110731124878,
+      "rewards/margins": 0.34467947483062744,
+      "rewards/rejected": -2.1836905479431152,
+      "step": 325
+    },
+    {
+      "epoch": 0.592188919164396,
+      "grad_norm": 1.8219250440597534,
+      "learning_rate": 9.521133231048338e-06,
+      "logits/chosen": 0.07941028475761414,
+      "logits/rejected": 0.13045310974121094,
+      "logps/chosen": -80.73494720458984,
+      "logps/rejected": -91.58990478515625,
+      "loss": 2.4866,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8594518899917603,
+      "rewards/margins": 0.4554288685321808,
+      "rewards/rejected": -2.314880847930908,
+      "step": 326
+    },
+    {
+      "epoch": 0.5940054495912807,
+      "grad_norm": 1.3753328323364258,
+      "learning_rate": 9.51824454793717e-06,
+      "logits/chosen": 0.08879546821117401,
+      "logits/rejected": 0.04693777486681938,
+      "logps/chosen": -82.9569091796875,
+      "logps/rejected": -91.44571685791016,
+      "loss": 2.2602,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.549119472503662,
+      "rewards/margins": 0.5145683288574219,
+      "rewards/rejected": -2.063688039779663,
+      "step": 327
+    },
+    {
+      "epoch": 0.5958219800181653,
+      "grad_norm": 1.4296562671661377,
+      "learning_rate": 9.515347670925728e-06,
+      "logits/chosen": 0.15614314377307892,
+      "logits/rejected": 0.1598319411277771,
+      "logps/chosen": -75.60345458984375,
+      "logps/rejected": -80.81770324707031,
+      "loss": 2.3946,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.556883454322815,
+      "rewards/margins": 0.4525718092918396,
+      "rewards/rejected": -2.0094552040100098,
+      "step": 328
+    },
+    {
+      "epoch": 0.59763851044505,
+      "grad_norm": 1.7932195663452148,
+      "learning_rate": 9.512442605921245e-06,
+      "logits/chosen": 0.05863601714372635,
+      "logits/rejected": 0.0788806602358818,
+      "logps/chosen": -75.95397186279297,
+      "logps/rejected": -84.50177001953125,
+      "loss": 2.5085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9159862995147705,
+      "rewards/margins": 0.4600725769996643,
+      "rewards/rejected": -2.376059055328369,
+      "step": 329
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 1.6558383703231812,
+      "learning_rate": 9.509529358847655e-06,
+      "logits/chosen": 0.08205496519804001,
+      "logits/rejected": 0.13091425597667694,
+      "logps/chosen": -82.73474884033203,
+      "logps/rejected": -94.02273559570312,
+      "loss": 2.4419,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.083285093307495,
+      "rewards/margins": 0.5300815105438232,
+      "rewards/rejected": -2.6133666038513184,
+      "step": 330
+    },
+    {
+      "epoch": 0.6012715712988193,
+      "grad_norm": 1.7888520956039429,
+      "learning_rate": 9.506607935645579e-06,
+      "logits/chosen": 0.11793217062950134,
+      "logits/rejected": 0.19568441808223724,
+      "logps/chosen": -79.86200714111328,
+      "logps/rejected": -89.25471496582031,
+      "loss": 2.4273,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9273384809494019,
+      "rewards/margins": 0.5115458965301514,
+      "rewards/rejected": -2.4388844966888428,
+      "step": 331
+    },
+    {
+      "epoch": 0.6030881017257039,
+      "grad_norm": 2.0436160564422607,
+      "learning_rate": 9.503678342272306e-06,
+      "logits/chosen": 0.043473344296216965,
+      "logits/rejected": 0.15554016828536987,
+      "logps/chosen": -75.2901611328125,
+      "logps/rejected": -83.20555877685547,
+      "loss": 2.6941,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.725769281387329,
+      "rewards/margins": 0.2789250314235687,
+      "rewards/rejected": -2.0046942234039307,
+      "step": 332
+    },
+    {
+      "epoch": 0.6049046321525886,
+      "grad_norm": 1.665578842163086,
+      "learning_rate": 9.500740584701785e-06,
+      "logits/chosen": 0.17119848728179932,
+      "logits/rejected": 0.14128939807415009,
+      "logps/chosen": -83.1322021484375,
+      "logps/rejected": -93.38603210449219,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.8975954055786133,
+      "rewards/margins": 0.7810046672821045,
+      "rewards/rejected": -2.6786000728607178,
+      "step": 333
+    },
+    {
+      "epoch": 0.6067211625794732,
+      "grad_norm": 1.9258419275283813,
+      "learning_rate": 9.497794668924617e-06,
+      "logits/chosen": 0.028591612353920937,
+      "logits/rejected": 0.08894480764865875,
+      "logps/chosen": -78.12230682373047,
+      "logps/rejected": -90.00184631347656,
+      "loss": 2.4006,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8113291263580322,
+      "rewards/margins": 0.5277000665664673,
+      "rewards/rejected": -2.339029550552368,
+      "step": 334
+    },
+    {
+      "epoch": 0.6085376930063578,
+      "grad_norm": 1.4878523349761963,
+      "learning_rate": 9.494840600948038e-06,
+      "logits/chosen": 0.03548199310898781,
+      "logits/rejected": 0.09170147776603699,
+      "logps/chosen": -73.15607452392578,
+      "logps/rejected": -83.77317810058594,
+      "loss": 2.3301,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6781296730041504,
+      "rewards/margins": 0.4735565185546875,
+      "rewards/rejected": -2.151685953140259,
+      "step": 335
+    },
+    {
+      "epoch": 0.6103542234332425,
+      "grad_norm": 1.8042774200439453,
+      "learning_rate": 9.491878386795906e-06,
+      "logits/chosen": 0.046132348477840424,
+      "logits/rejected": 0.0721711814403534,
+      "logps/chosen": -81.01045227050781,
+      "logps/rejected": -89.28679656982422,
+      "loss": 2.5593,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.012852191925049,
+      "rewards/margins": 0.44549623131752014,
+      "rewards/rejected": -2.458348512649536,
+      "step": 336
+    },
+    {
+      "epoch": 0.6121707538601272,
+      "grad_norm": 2.048952102661133,
+      "learning_rate": 9.488908032508691e-06,
+      "logits/chosen": 0.10774732381105423,
+      "logits/rejected": 0.11935572326183319,
+      "logps/chosen": -91.25210571289062,
+      "logps/rejected": -90.42224884033203,
+      "loss": 2.8282,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0127718448638916,
+      "rewards/margins": 0.16889013350009918,
+      "rewards/rejected": -2.18166184425354,
+      "step": 337
+    },
+    {
+      "epoch": 0.6139872842870118,
+      "grad_norm": 1.6447219848632812,
+      "learning_rate": 9.485929544143462e-06,
+      "logits/chosen": 0.12260966747999191,
+      "logits/rejected": 0.1517799198627472,
+      "logps/chosen": -72.49166870117188,
+      "logps/rejected": -82.3083724975586,
+      "loss": 2.4029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.624878168106079,
+      "rewards/margins": 0.5487539768218994,
+      "rewards/rejected": -2.1736321449279785,
+      "step": 338
+    },
+    {
+      "epoch": 0.6158038147138964,
+      "grad_norm": 1.599530577659607,
+      "learning_rate": 9.482942927773876e-06,
+      "logits/chosen": 0.09051798284053802,
+      "logits/rejected": 0.11202570050954819,
+      "logps/chosen": -84.42019653320312,
+      "logps/rejected": -92.61677551269531,
+      "loss": 2.2269,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6424546241760254,
+      "rewards/margins": 0.5715836882591248,
+      "rewards/rejected": -2.214038133621216,
+      "step": 339
+    },
+    {
+      "epoch": 0.6176203451407811,
+      "grad_norm": 1.5411081314086914,
+      "learning_rate": 9.479948189490164e-06,
+      "logits/chosen": 0.06563656777143478,
+      "logits/rejected": 0.17349205911159515,
+      "logps/chosen": -70.98219299316406,
+      "logps/rejected": -78.88846588134766,
+      "loss": 2.3328,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.824373722076416,
+      "rewards/margins": 0.46550822257995605,
+      "rewards/rejected": -2.289881706237793,
+      "step": 340
+    },
+    {
+      "epoch": 0.6194368755676658,
+      "grad_norm": 1.8855030536651611,
+      "learning_rate": 9.476945335399122e-06,
+      "logits/chosen": 0.11399642378091812,
+      "logits/rejected": 0.12690946459770203,
+      "logps/chosen": -85.86933898925781,
+      "logps/rejected": -89.75601196289062,
+      "loss": 2.4269,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9560539722442627,
+      "rewards/margins": 0.38427016139030457,
+      "rewards/rejected": -2.3403239250183105,
+      "step": 341
+    },
+    {
+      "epoch": 0.6212534059945504,
+      "grad_norm": 1.9973480701446533,
+      "learning_rate": 9.473934371624087e-06,
+      "logits/chosen": 0.029105912894010544,
+      "logits/rejected": 0.12763622403144836,
+      "logps/chosen": -80.68119049072266,
+      "logps/rejected": -93.90747833251953,
+      "loss": 2.4709,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.906477928161621,
+      "rewards/margins": 0.6422742009162903,
+      "rewards/rejected": -2.5487518310546875,
+      "step": 342
+    },
+    {
+      "epoch": 0.623069936421435,
+      "grad_norm": 1.7752224206924438,
+      "learning_rate": 9.47091530430494e-06,
+      "logits/chosen": 0.14096824824810028,
+      "logits/rejected": 0.1774113029241562,
+      "logps/chosen": -71.80216217041016,
+      "logps/rejected": -79.65301513671875,
+      "loss": 2.2678,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.5735952854156494,
+      "rewards/margins": 0.5823659300804138,
+      "rewards/rejected": -2.155961036682129,
+      "step": 343
+    },
+    {
+      "epoch": 0.6248864668483197,
+      "grad_norm": 1.855377197265625,
+      "learning_rate": 9.467888139598086e-06,
+      "logits/chosen": 0.10637074708938599,
+      "logits/rejected": 0.0798158049583435,
+      "logps/chosen": -79.01347351074219,
+      "logps/rejected": -80.77239990234375,
+      "loss": 2.7867,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9598251581192017,
+      "rewards/margins": 0.3184160590171814,
+      "rewards/rejected": -2.2782411575317383,
+      "step": 344
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 1.423040747642517,
+      "learning_rate": 9.464852883676441e-06,
+      "logits/chosen": 0.1304859220981598,
+      "logits/rejected": 0.1759231686592102,
+      "logps/chosen": -74.09733581542969,
+      "logps/rejected": -89.27589416503906,
+      "loss": 2.2777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7697277069091797,
+      "rewards/margins": 0.635611355304718,
+      "rewards/rejected": -2.405339002609253,
+      "step": 345
+    },
+    {
+      "epoch": 0.628519527702089,
+      "grad_norm": 1.6246038675308228,
+      "learning_rate": 9.461809542729421e-06,
+      "logits/chosen": 0.03750050812959671,
+      "logits/rejected": 0.10337980091571808,
+      "logps/chosen": -81.89110565185547,
+      "logps/rejected": -95.36811828613281,
+      "loss": 2.0438,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7299752235412598,
+      "rewards/margins": 0.7404756546020508,
+      "rewards/rejected": -2.4704508781433105,
+      "step": 346
+    },
+    {
+      "epoch": 0.6303360581289736,
+      "grad_norm": 1.6618752479553223,
+      "learning_rate": 9.458758122962926e-06,
+      "logits/chosen": 0.05359608680009842,
+      "logits/rejected": 0.10455545783042908,
+      "logps/chosen": -85.34078216552734,
+      "logps/rejected": -90.46200561523438,
+      "loss": 2.3299,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9331045150756836,
+      "rewards/margins": 0.4755082130432129,
+      "rewards/rejected": -2.4086129665374756,
+      "step": 347
+    },
+    {
+      "epoch": 0.6321525885558583,
+      "grad_norm": 1.5805696249008179,
+      "learning_rate": 9.455698630599332e-06,
+      "logits/chosen": 0.10048776119947433,
+      "logits/rejected": 0.1140337884426117,
+      "logps/chosen": -80.25875854492188,
+      "logps/rejected": -94.44998168945312,
+      "loss": 2.3439,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8421945571899414,
+      "rewards/margins": 0.5977468490600586,
+      "rewards/rejected": -2.43994140625,
+      "step": 348
+    },
+    {
+      "epoch": 0.633969118982743,
+      "grad_norm": 1.37093985080719,
+      "learning_rate": 9.452631071877478e-06,
+      "logits/chosen": 0.11764326691627502,
+      "logits/rejected": 0.10735289752483368,
+      "logps/chosen": -72.02367401123047,
+      "logps/rejected": -77.9522705078125,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.6527085304260254,
+      "rewards/margins": 0.6065118908882141,
+      "rewards/rejected": -2.259220600128174,
+      "step": 349
+    },
+    {
+      "epoch": 0.6357856494096276,
+      "grad_norm": 2.0214192867279053,
+      "learning_rate": 9.449555453052652e-06,
+      "logits/chosen": 0.13177426159381866,
+      "logits/rejected": 0.12408209592103958,
+      "logps/chosen": -76.62931823730469,
+      "logps/rejected": -81.16517639160156,
+      "loss": 2.9175,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.9575048685073853,
+      "rewards/margins": 0.21441945433616638,
+      "rewards/rejected": -2.171924352645874,
+      "step": 350
+    },
+    {
+      "epoch": 0.6376021798365122,
+      "grad_norm": 1.6041687726974487,
+      "learning_rate": 9.446471780396573e-06,
+      "logits/chosen": 0.18574532866477966,
+      "logits/rejected": 0.22683225572109222,
+      "logps/chosen": -73.95652770996094,
+      "logps/rejected": -80.29524993896484,
+      "loss": 2.3833,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8684748411178589,
+      "rewards/margins": 0.37268272042274475,
+      "rewards/rejected": -2.2411575317382812,
+      "step": 351
+    },
+    {
+      "epoch": 0.6394187102633969,
+      "grad_norm": 1.7410838603973389,
+      "learning_rate": 9.443380060197387e-06,
+      "logits/chosen": 0.07876043766736984,
+      "logits/rejected": 0.14103996753692627,
+      "logps/chosen": -74.64432525634766,
+      "logps/rejected": -83.6999740600586,
+      "loss": 2.5558,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.029916763305664,
+      "rewards/margins": 0.4088842272758484,
+      "rewards/rejected": -2.438800811767578,
+      "step": 352
+    },
+    {
+      "epoch": 0.6412352406902816,
+      "grad_norm": 1.87971031665802,
+      "learning_rate": 9.440280298759653e-06,
+      "logits/chosen": 0.13997013866901398,
+      "logits/rejected": 0.1322249174118042,
+      "logps/chosen": -88.78776550292969,
+      "logps/rejected": -88.22732543945312,
+      "loss": 2.6547,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9983258247375488,
+      "rewards/margins": 0.2458382397890091,
+      "rewards/rejected": -2.244164228439331,
+      "step": 353
+    },
+    {
+      "epoch": 0.6430517711171662,
+      "grad_norm": 1.7056363821029663,
+      "learning_rate": 9.437172502404318e-06,
+      "logits/chosen": 0.07248476892709732,
+      "logits/rejected": 0.13019773364067078,
+      "logps/chosen": -78.4591064453125,
+      "logps/rejected": -80.49126434326172,
+      "loss": 2.3809,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9740041494369507,
+      "rewards/margins": 0.4853326082229614,
+      "rewards/rejected": -2.459336757659912,
+      "step": 354
+    },
+    {
+      "epoch": 0.6448683015440508,
+      "grad_norm": 1.435718059539795,
+      "learning_rate": 9.434056677468726e-06,
+      "logits/chosen": 0.09164869040250778,
+      "logits/rejected": 0.14243285357952118,
+      "logps/chosen": -77.83367156982422,
+      "logps/rejected": -86.18836212158203,
+      "loss": 2.006,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.7355122566223145,
+      "rewards/margins": 0.6869419813156128,
+      "rewards/rejected": -2.422454357147217,
+      "step": 355
+    },
+    {
+      "epoch": 0.6466848319709355,
+      "grad_norm": 1.6553188562393188,
+      "learning_rate": 9.430932830306587e-06,
+      "logits/chosen": 0.04967673122882843,
+      "logits/rejected": 0.1456151008605957,
+      "logps/chosen": -73.50204467773438,
+      "logps/rejected": -85.31363677978516,
+      "loss": 2.4388,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.978384256362915,
+      "rewards/margins": 0.48918718099594116,
+      "rewards/rejected": -2.467571496963501,
+      "step": 356
+    },
+    {
+      "epoch": 0.6485013623978202,
+      "grad_norm": 1.8443480730056763,
+      "learning_rate": 9.427800967287963e-06,
+      "logits/chosen": 0.06455090641975403,
+      "logits/rejected": 0.12608012557029724,
+      "logps/chosen": -75.49232482910156,
+      "logps/rejected": -77.49136352539062,
+      "loss": 2.706,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.8730812072753906,
+      "rewards/margins": 0.2989741861820221,
+      "rewards/rejected": -2.17205548286438,
+      "step": 357
+    },
+    {
+      "epoch": 0.6503178928247049,
+      "grad_norm": 1.6912306547164917,
+      "learning_rate": 9.424661094799273e-06,
+      "logits/chosen": 0.11878645420074463,
+      "logits/rejected": 0.13628609478473663,
+      "logps/chosen": -73.09883880615234,
+      "logps/rejected": -84.19624328613281,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.928421139717102,
+      "rewards/margins": 0.5017052292823792,
+      "rewards/rejected": -2.430126190185547,
+      "step": 358
+    },
+    {
+      "epoch": 0.6521344232515894,
+      "grad_norm": 1.3643461465835571,
+      "learning_rate": 9.421513219243262e-06,
+      "logits/chosen": 0.07683826237916946,
+      "logits/rejected": 0.15765298902988434,
+      "logps/chosen": -76.0871353149414,
+      "logps/rejected": -97.47781372070312,
+      "loss": 1.7844,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.8300602436065674,
+      "rewards/margins": 0.911116361618042,
+      "rewards/rejected": -2.7411766052246094,
+      "step": 359
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 1.6327749490737915,
+      "learning_rate": 9.418357347038999e-06,
+      "logits/chosen": 0.1078951433300972,
+      "logits/rejected": 0.12233921140432358,
+      "logps/chosen": -75.89913177490234,
+      "logps/rejected": -78.4587173461914,
+      "loss": 2.4976,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.7992844581604004,
+      "rewards/margins": 0.311392605304718,
+      "rewards/rejected": -2.1106772422790527,
+      "step": 360
+    },
+    {
+      "epoch": 0.6557674841053588,
+      "grad_norm": 1.463025689125061,
+      "learning_rate": 9.415193484621852e-06,
+      "logits/chosen": 0.057331383228302,
+      "logits/rejected": 0.1563551127910614,
+      "logps/chosen": -80.7889633178711,
+      "logps/rejected": -93.21602630615234,
+      "loss": 2.1398,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.744189739227295,
+      "rewards/margins": 0.576555073261261,
+      "rewards/rejected": -2.320744752883911,
+      "step": 361
+    },
+    {
+      "epoch": 0.6575840145322435,
+      "grad_norm": 1.743695616722107,
+      "learning_rate": 9.412021638443491e-06,
+      "logits/chosen": 0.03781123086810112,
+      "logits/rejected": 0.154897540807724,
+      "logps/chosen": -77.72047424316406,
+      "logps/rejected": -88.80599975585938,
+      "loss": 2.2361,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9849369525909424,
+      "rewards/margins": 0.6562294363975525,
+      "rewards/rejected": -2.6411664485931396,
+      "step": 362
+    },
+    {
+      "epoch": 0.659400544959128,
+      "grad_norm": 1.5648279190063477,
+      "learning_rate": 9.408841814971862e-06,
+      "logits/chosen": 0.10401102900505066,
+      "logits/rejected": 0.12699122726917267,
+      "logps/chosen": -83.10671997070312,
+      "logps/rejected": -86.8663101196289,
+      "loss": 2.4009,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7048468589782715,
+      "rewards/margins": 0.4090174734592438,
+      "rewards/rejected": -2.1138644218444824,
+      "step": 363
+    },
+    {
+      "epoch": 0.6612170753860127,
+      "grad_norm": 1.3385239839553833,
+      "learning_rate": 9.405654020691178e-06,
+      "logits/chosen": 0.06100422143936157,
+      "logits/rejected": 0.10872650146484375,
+      "logps/chosen": -76.72837829589844,
+      "logps/rejected": -87.63470458984375,
+      "loss": 2.141,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.747226595878601,
+      "rewards/margins": 0.615075945854187,
+      "rewards/rejected": -2.362302780151367,
+      "step": 364
+    },
+    {
+      "epoch": 0.6630336058128974,
+      "grad_norm": 1.4060734510421753,
+      "learning_rate": 9.402458262101906e-06,
+      "logits/chosen": 0.10638861358165741,
+      "logits/rejected": 0.16951681673526764,
+      "logps/chosen": -77.35757446289062,
+      "logps/rejected": -92.03012084960938,
+      "loss": 2.0511,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8359463214874268,
+      "rewards/margins": 0.7444422841072083,
+      "rewards/rejected": -2.5803885459899902,
+      "step": 365
+    },
+    {
+      "epoch": 0.6648501362397821,
+      "grad_norm": 1.4371014833450317,
+      "learning_rate": 9.399254545720757e-06,
+      "logits/chosen": 0.0383220911026001,
+      "logits/rejected": 0.05856206640601158,
+      "logps/chosen": -82.40064239501953,
+      "logps/rejected": -92.1117935180664,
+      "loss": 2.1756,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8580785989761353,
+      "rewards/margins": 0.6475991010665894,
+      "rewards/rejected": -2.5056777000427246,
+      "step": 366
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.6673945188522339,
+      "learning_rate": 9.396042878080661e-06,
+      "logits/chosen": 0.14657820761203766,
+      "logits/rejected": 0.19021111726760864,
+      "logps/chosen": -73.56724548339844,
+      "logps/rejected": -78.85279846191406,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9079951047897339,
+      "rewards/margins": 0.3709837794303894,
+      "rewards/rejected": -2.2789790630340576,
+      "step": 367
+    },
+    {
+      "epoch": 0.6684831970935513,
+      "grad_norm": 1.7402448654174805,
+      "learning_rate": 9.392823265730775e-06,
+      "logits/chosen": 0.16515754163265228,
+      "logits/rejected": 0.12460774928331375,
+      "logps/chosen": -69.89266204833984,
+      "logps/rejected": -74.53837585449219,
+      "loss": 2.4262,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.0627622604370117,
+      "rewards/margins": 0.5317320823669434,
+      "rewards/rejected": -2.594494342803955,
+      "step": 368
+    },
+    {
+      "epoch": 0.670299727520436,
+      "grad_norm": 1.5290364027023315,
+      "learning_rate": 9.389595715236446e-06,
+      "logits/chosen": 0.0954434722661972,
+      "logits/rejected": 0.16226956248283386,
+      "logps/chosen": -77.17019653320312,
+      "logps/rejected": -91.07398986816406,
+      "loss": 2.2511,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.241170644760132,
+      "rewards/margins": 0.6319360136985779,
+      "rewards/rejected": -2.8731067180633545,
+      "step": 369
+    },
+    {
+      "epoch": 0.6721162579473207,
+      "grad_norm": 1.4633573293685913,
+      "learning_rate": 9.386360233179206e-06,
+      "logits/chosen": 0.06517557799816132,
+      "logits/rejected": 0.07250035554170609,
+      "logps/chosen": -79.23770141601562,
+      "logps/rejected": -88.50403594970703,
+      "loss": 2.3858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.751698613166809,
+      "rewards/margins": 0.49529415369033813,
+      "rewards/rejected": -2.246993064880371,
+      "step": 370
+    },
+    {
+      "epoch": 0.6739327883742052,
+      "grad_norm": 2.007215976715088,
+      "learning_rate": 9.383116826156775e-06,
+      "logits/chosen": 0.13584929704666138,
+      "logits/rejected": 0.10950647294521332,
+      "logps/chosen": -80.75408935546875,
+      "logps/rejected": -79.3903579711914,
+      "loss": 2.6967,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.2421724796295166,
+      "rewards/margins": 0.34328368306159973,
+      "rewards/rejected": -2.585456132888794,
+      "step": 371
+    },
+    {
+      "epoch": 0.6757493188010899,
+      "grad_norm": 1.7323477268218994,
+      "learning_rate": 9.37986550078302e-06,
+      "logits/chosen": 0.011626070365309715,
+      "logits/rejected": 0.06207559257745743,
+      "logps/chosen": -75.51494598388672,
+      "logps/rejected": -85.83218383789062,
+      "loss": 2.4608,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9569413661956787,
+      "rewards/margins": 0.5160585641860962,
+      "rewards/rejected": -2.4729998111724854,
+      "step": 372
+    },
+    {
+      "epoch": 0.6775658492279746,
+      "grad_norm": 1.6365752220153809,
+      "learning_rate": 9.376606263687959e-06,
+      "logits/chosen": 0.10213632136583328,
+      "logits/rejected": 0.08355780690908432,
+      "logps/chosen": -79.01370239257812,
+      "logps/rejected": -81.9417953491211,
+      "loss": 2.4701,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.064467668533325,
+      "rewards/margins": 0.4157037138938904,
+      "rewards/rejected": -2.4801712036132812,
+      "step": 373
+    },
+    {
+      "epoch": 0.6793823796548593,
+      "grad_norm": 1.5918922424316406,
+      "learning_rate": 9.373339121517748e-06,
+      "logits/chosen": 0.09486684203147888,
+      "logits/rejected": 0.07311725616455078,
+      "logps/chosen": -82.15351867675781,
+      "logps/rejected": -91.7690658569336,
+      "loss": 2.1297,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0153114795684814,
+      "rewards/margins": 0.669613242149353,
+      "rewards/rejected": -2.684924840927124,
+      "step": 374
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 1.6511566638946533,
+      "learning_rate": 9.370064080934654e-06,
+      "logits/chosen": 0.1406637728214264,
+      "logits/rejected": 0.20172299444675446,
+      "logps/chosen": -69.41023254394531,
+      "logps/rejected": -74.65138244628906,
+      "loss": 2.4704,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.061182737350464,
+      "rewards/margins": 0.3090137243270874,
+      "rewards/rejected": -2.3701963424682617,
+      "step": 375
+    },
+    {
+      "epoch": 0.6830154405086285,
+      "grad_norm": 1.770624041557312,
+      "learning_rate": 9.366781148617056e-06,
+      "logits/chosen": 0.10267248749732971,
+      "logits/rejected": 0.07900385558605194,
+      "logps/chosen": -77.27003479003906,
+      "logps/rejected": -88.09040832519531,
+      "loss": 2.2813,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.132927656173706,
+      "rewards/margins": 0.6629016399383545,
+      "rewards/rejected": -2.7958290576934814,
+      "step": 376
+    },
+    {
+      "epoch": 0.6848319709355132,
+      "grad_norm": 1.5172028541564941,
+      "learning_rate": 9.363490331259426e-06,
+      "logits/chosen": 0.05240853130817413,
+      "logits/rejected": 0.10445387661457062,
+      "logps/chosen": -75.5218276977539,
+      "logps/rejected": -85.49366760253906,
+      "loss": 2.1442,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9302880764007568,
+      "rewards/margins": 0.5776917338371277,
+      "rewards/rejected": -2.5079798698425293,
+      "step": 377
+    },
+    {
+      "epoch": 0.6866485013623979,
+      "grad_norm": 1.6540950536727905,
+      "learning_rate": 9.360191635572313e-06,
+      "logits/chosen": 0.14195458590984344,
+      "logits/rejected": 0.0978107899427414,
+      "logps/chosen": -85.00608825683594,
+      "logps/rejected": -85.1456298828125,
+      "loss": 2.423,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8765841722488403,
+      "rewards/margins": 0.49441370368003845,
+      "rewards/rejected": -2.370997905731201,
+      "step": 378
+    },
+    {
+      "epoch": 0.6884650317892824,
+      "grad_norm": 1.6463801860809326,
+      "learning_rate": 9.356885068282334e-06,
+      "logits/chosen": 0.13197994232177734,
+      "logits/rejected": 0.09924699366092682,
+      "logps/chosen": -86.94219970703125,
+      "logps/rejected": -86.85704803466797,
+      "loss": 2.4801,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9674016237258911,
+      "rewards/margins": 0.4530283510684967,
+      "rewards/rejected": -2.4204299449920654,
+      "step": 379
+    },
+    {
+      "epoch": 0.6902815622161671,
+      "grad_norm": 1.5428895950317383,
+      "learning_rate": 9.353570636132151e-06,
+      "logits/chosen": 0.10434838384389877,
+      "logits/rejected": 0.12619757652282715,
+      "logps/chosen": -78.56607818603516,
+      "logps/rejected": -86.33320617675781,
+      "loss": 2.2711,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9667280912399292,
+      "rewards/margins": 0.6350463628768921,
+      "rewards/rejected": -2.601774215698242,
+      "step": 380
+    },
+    {
+      "epoch": 0.6920980926430518,
+      "grad_norm": 1.6211251020431519,
+      "learning_rate": 9.350248345880471e-06,
+      "logits/chosen": 0.14081251621246338,
+      "logits/rejected": 0.1453506350517273,
+      "logps/chosen": -74.38184356689453,
+      "logps/rejected": -80.61524200439453,
+      "loss": 2.3863,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.0189671516418457,
+      "rewards/margins": 0.4528267979621887,
+      "rewards/rejected": -2.4717938899993896,
+      "step": 381
+    },
+    {
+      "epoch": 0.6939146230699365,
+      "grad_norm": 2.6346240043640137,
+      "learning_rate": 9.346918204302022e-06,
+      "logits/chosen": 0.0519830696284771,
+      "logits/rejected": 0.04284125566482544,
+      "logps/chosen": -82.38153839111328,
+      "logps/rejected": -86.35488891601562,
+      "loss": 2.5434,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9824391603469849,
+      "rewards/margins": 0.4156208634376526,
+      "rewards/rejected": -2.3980600833892822,
+      "step": 382
+    },
+    {
+      "epoch": 0.695731153496821,
+      "grad_norm": 1.4762026071548462,
+      "learning_rate": 9.343580218187544e-06,
+      "logits/chosen": 0.02196469157934189,
+      "logits/rejected": 0.04618150740861893,
+      "logps/chosen": -72.66182708740234,
+      "logps/rejected": -88.68983459472656,
+      "loss": 2.1699,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9227774143218994,
+      "rewards/margins": 0.9200201630592346,
+      "rewards/rejected": -2.8427975177764893,
+      "step": 383
+    },
+    {
+      "epoch": 0.6975476839237057,
+      "grad_norm": 1.9675711393356323,
+      "learning_rate": 9.340234394343768e-06,
+      "logits/chosen": 0.12956203520298004,
+      "logits/rejected": 0.16079677641391754,
+      "logps/chosen": -87.38492584228516,
+      "logps/rejected": -87.80662536621094,
+      "loss": 2.3138,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.144991397857666,
+      "rewards/margins": 0.4893474578857422,
+      "rewards/rejected": -2.634338855743408,
+      "step": 384
+    },
+    {
+      "epoch": 0.6993642143505904,
+      "grad_norm": 1.9116860628128052,
+      "learning_rate": 9.336880739593415e-06,
+      "logits/chosen": 0.06013559550046921,
+      "logits/rejected": 0.09889352321624756,
+      "logps/chosen": -86.55718231201172,
+      "logps/rejected": -91.01994323730469,
+      "loss": 2.4864,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9528815746307373,
+      "rewards/margins": 0.44100096821784973,
+      "rewards/rejected": -2.3938825130462646,
+      "step": 385
+    },
+    {
+      "epoch": 0.701180744777475,
+      "grad_norm": 2.0582289695739746,
+      "learning_rate": 9.33351926077517e-06,
+      "logits/chosen": 0.11538423597812653,
+      "logits/rejected": 0.17568480968475342,
+      "logps/chosen": -76.24241638183594,
+      "logps/rejected": -87.1357650756836,
+      "loss": 2.6878,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.102504253387451,
+      "rewards/margins": 0.3671773076057434,
+      "rewards/rejected": -2.469681739807129,
+      "step": 386
+    },
+    {
+      "epoch": 0.7029972752043597,
+      "grad_norm": 1.6962640285491943,
+      "learning_rate": 9.330149964743674e-06,
+      "logits/chosen": 0.12425235658884048,
+      "logits/rejected": 0.1852879822254181,
+      "logps/chosen": -82.67887115478516,
+      "logps/rejected": -98.66133117675781,
+      "loss": 2.343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.182525634765625,
+      "rewards/margins": 0.6602460145950317,
+      "rewards/rejected": -2.842771530151367,
+      "step": 387
+    },
+    {
+      "epoch": 0.7048138056312443,
+      "grad_norm": 1.6582266092300415,
+      "learning_rate": 9.326772858369506e-06,
+      "logits/chosen": 0.14438432455062866,
+      "logits/rejected": 0.0922938883304596,
+      "logps/chosen": -78.87794494628906,
+      "logps/rejected": -84.929443359375,
+      "loss": 2.3036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.988234281539917,
+      "rewards/margins": 0.4737466275691986,
+      "rewards/rejected": -2.4619810581207275,
+      "step": 388
+    },
+    {
+      "epoch": 0.706630336058129,
+      "grad_norm": 1.6689919233322144,
+      "learning_rate": 9.323387948539176e-06,
+      "logits/chosen": 0.1282673329114914,
+      "logits/rejected": 0.13633723556995392,
+      "logps/chosen": -70.22606658935547,
+      "logps/rejected": -79.26126861572266,
+      "loss": 2.241,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9162399768829346,
+      "rewards/margins": 0.5345122814178467,
+      "rewards/rejected": -2.4507524967193604,
+      "step": 389
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 1.5861207246780396,
+      "learning_rate": 9.319995242155102e-06,
+      "logits/chosen": 0.13024169206619263,
+      "logits/rejected": 0.14390775561332703,
+      "logps/chosen": -87.00923919677734,
+      "logps/rejected": -91.24594116210938,
+      "loss": 2.284,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9127002954483032,
+      "rewards/margins": 0.5265774130821228,
+      "rewards/rejected": -2.4392776489257812,
+      "step": 390
+    },
+    {
+      "epoch": 0.7102633969118983,
+      "grad_norm": 2.0118792057037354,
+      "learning_rate": 9.316594746135608e-06,
+      "logits/chosen": 0.14648675918579102,
+      "logits/rejected": 0.11786539107561111,
+      "logps/chosen": -82.68962097167969,
+      "logps/rejected": -88.26126861572266,
+      "loss": 2.7281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9463748931884766,
+      "rewards/margins": 0.20282456278800964,
+      "rewards/rejected": -2.1491994857788086,
+      "step": 391
+    },
+    {
+      "epoch": 0.7120799273387829,
+      "grad_norm": 1.8176332712173462,
+      "learning_rate": 9.313186467414892e-06,
+      "logits/chosen": 0.12084021419286728,
+      "logits/rejected": 0.1433698982000351,
+      "logps/chosen": -73.46707153320312,
+      "logps/rejected": -82.46024322509766,
+      "loss": 2.5247,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0802624225616455,
+      "rewards/margins": 0.41601985692977905,
+      "rewards/rejected": -2.4962823390960693,
+      "step": 392
+    },
+    {
+      "epoch": 0.7138964577656676,
+      "grad_norm": 1.7207351922988892,
+      "learning_rate": 9.30977041294303e-06,
+      "logits/chosen": 0.11983273923397064,
+      "logits/rejected": 0.16106371581554413,
+      "logps/chosen": -76.09751892089844,
+      "logps/rejected": -85.20323944091797,
+      "loss": 2.5284,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9018669128417969,
+      "rewards/margins": 0.4894491136074066,
+      "rewards/rejected": -2.3913159370422363,
+      "step": 393
+    },
+    {
+      "epoch": 0.7157129881925522,
+      "grad_norm": 1.6489367485046387,
+      "learning_rate": 9.306346589685956e-06,
+      "logits/chosen": 0.14845696091651917,
+      "logits/rejected": 0.10356368124485016,
+      "logps/chosen": -83.94526672363281,
+      "logps/rejected": -86.53378295898438,
+      "loss": 2.3394,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9089899063110352,
+      "rewards/margins": 0.4619474411010742,
+      "rewards/rejected": -2.3709375858306885,
+      "step": 394
+    },
+    {
+      "epoch": 0.7175295186194369,
+      "grad_norm": 1.792077898979187,
+      "learning_rate": 9.302915004625435e-06,
+      "logits/chosen": 0.06936248391866684,
+      "logits/rejected": 0.13349927961826324,
+      "logps/chosen": -91.82075500488281,
+      "logps/rejected": -98.7391357421875,
+      "loss": 2.3875,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9371806383132935,
+      "rewards/margins": 0.6781354546546936,
+      "rewards/rejected": -2.6153156757354736,
+      "step": 395
+    },
+    {
+      "epoch": 0.7193460490463215,
+      "grad_norm": 2.049694538116455,
+      "learning_rate": 9.29947566475907e-06,
+      "logits/chosen": 0.08474650233983994,
+      "logits/rejected": 0.13380834460258484,
+      "logps/chosen": -84.59528350830078,
+      "logps/rejected": -96.24267578125,
+      "loss": 2.7001,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9375882148742676,
+      "rewards/margins": 0.4607384204864502,
+      "rewards/rejected": -2.3983266353607178,
+      "step": 396
+    },
+    {
+      "epoch": 0.7211625794732062,
+      "grad_norm": 1.9235875606536865,
+      "learning_rate": 9.296028577100271e-06,
+      "logits/chosen": 0.06255945563316345,
+      "logits/rejected": 0.17074811458587646,
+      "logps/chosen": -68.99498748779297,
+      "logps/rejected": -84.45899963378906,
+      "loss": 2.347,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9344629049301147,
+      "rewards/margins": 0.688963770866394,
+      "rewards/rejected": -2.6234264373779297,
+      "step": 397
+    },
+    {
+      "epoch": 0.7229791099000908,
+      "grad_norm": 1.6617276668548584,
+      "learning_rate": 9.292573748678254e-06,
+      "logits/chosen": 0.11962493509054184,
+      "logits/rejected": 0.11647717654705048,
+      "logps/chosen": -85.3626708984375,
+      "logps/rejected": -95.70562744140625,
+      "loss": 2.2229,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.8968594074249268,
+      "rewards/margins": 0.6918852925300598,
+      "rewards/rejected": -2.588744640350342,
+      "step": 398
+    },
+    {
+      "epoch": 0.7247956403269755,
+      "grad_norm": 1.4650444984436035,
+      "learning_rate": 9.289111186538013e-06,
+      "logits/chosen": 0.06390775740146637,
+      "logits/rejected": 0.11076060682535172,
+      "logps/chosen": -69.9189453125,
+      "logps/rejected": -81.5618667602539,
+      "loss": 2.2271,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.857581377029419,
+      "rewards/margins": 0.5540103316307068,
+      "rewards/rejected": -2.4115917682647705,
+      "step": 399
+    },
+    {
+      "epoch": 0.7266121707538601,
+      "grad_norm": 1.5728726387023926,
+      "learning_rate": 9.285640897740316e-06,
+      "logits/chosen": 0.08816932141780853,
+      "logits/rejected": 0.1330798715353012,
+      "logps/chosen": -78.04735565185547,
+      "logps/rejected": -87.89193725585938,
+      "loss": 2.3595,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.823161244392395,
+      "rewards/margins": 0.4568220376968384,
+      "rewards/rejected": -2.2799830436706543,
+      "step": 400
+    },
+    {
+      "epoch": 0.7284287011807448,
+      "grad_norm": 1.6291779279708862,
+      "learning_rate": 9.282162889361686e-06,
+      "logits/chosen": 0.17718347907066345,
+      "logits/rejected": 0.20603135228157043,
+      "logps/chosen": -67.99884033203125,
+      "logps/rejected": -75.93153381347656,
+      "loss": 2.3002,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.671633243560791,
+      "rewards/margins": 0.45543500781059265,
+      "rewards/rejected": -2.127068281173706,
+      "step": 401
+    },
+    {
+      "epoch": 0.7302452316076294,
+      "grad_norm": 1.5066276788711548,
+      "learning_rate": 9.278677168494388e-06,
+      "logits/chosen": 0.16840124130249023,
+      "logits/rejected": 0.17102479934692383,
+      "logps/chosen": -78.1556625366211,
+      "logps/rejected": -84.91495513916016,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9279460906982422,
+      "rewards/margins": 0.6659258008003235,
+      "rewards/rejected": -2.593871831893921,
+      "step": 402
+    },
+    {
+      "epoch": 0.7320617620345141,
+      "grad_norm": 1.71995210647583,
+      "learning_rate": 9.275183742246412e-06,
+      "logits/chosen": -0.0005522281862795353,
+      "logits/rejected": 0.1412215530872345,
+      "logps/chosen": -73.37832641601562,
+      "logps/rejected": -92.71226501464844,
+      "loss": 2.3315,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6880759000778198,
+      "rewards/margins": 0.5844776630401611,
+      "rewards/rejected": -2.2725534439086914,
+      "step": 403
+    },
+    {
+      "epoch": 0.7338782924613987,
+      "grad_norm": 1.676321029663086,
+      "learning_rate": 9.271682617741466e-06,
+      "logits/chosen": 0.15319044888019562,
+      "logits/rejected": 0.14503052830696106,
+      "logps/chosen": -75.85136413574219,
+      "logps/rejected": -82.61698150634766,
+      "loss": 2.3528,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.819403052330017,
+      "rewards/margins": 0.3919827342033386,
+      "rewards/rejected": -2.211385726928711,
+      "step": 404
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 1.4811294078826904,
+      "learning_rate": 9.268173802118949e-06,
+      "logits/chosen": 0.07122528553009033,
+      "logits/rejected": 0.12146291136741638,
+      "logps/chosen": -76.78916931152344,
+      "logps/rejected": -85.33805084228516,
+      "loss": 2.1505,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9886473417282104,
+      "rewards/margins": 0.5911746025085449,
+      "rewards/rejected": -2.579822301864624,
+      "step": 405
+    },
+    {
+      "epoch": 0.737511353315168,
+      "grad_norm": 1.629601001739502,
+      "learning_rate": 9.264657302533947e-06,
+      "logits/chosen": 0.11274963617324829,
+      "logits/rejected": 0.0651293620467186,
+      "logps/chosen": -82.86985778808594,
+      "logps/rejected": -88.68803405761719,
+      "loss": 2.1873,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8538885116577148,
+      "rewards/margins": 0.6726438403129578,
+      "rewards/rejected": -2.5265324115753174,
+      "step": 406
+    },
+    {
+      "epoch": 0.7393278837420527,
+      "grad_norm": 1.5687483549118042,
+      "learning_rate": 9.261133126157218e-06,
+      "logits/chosen": 0.06054290384054184,
+      "logits/rejected": 0.1543959081172943,
+      "logps/chosen": -68.5821304321289,
+      "logps/rejected": -80.544921875,
+      "loss": 2.3587,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7550634145736694,
+      "rewards/margins": 0.46927812695503235,
+      "rewards/rejected": -2.224341630935669,
+      "step": 407
+    },
+    {
+      "epoch": 0.7411444141689373,
+      "grad_norm": 1.849908471107483,
+      "learning_rate": 9.257601280175167e-06,
+      "logits/chosen": 0.15177126228809357,
+      "logits/rejected": 0.1485670655965805,
+      "logps/chosen": -72.81680297851562,
+      "logps/rejected": -80.58684539794922,
+      "loss": 2.444,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9739688634872437,
+      "rewards/margins": 0.6166737079620361,
+      "rewards/rejected": -2.5906424522399902,
+      "step": 408
+    },
+    {
+      "epoch": 0.742960944595822,
+      "grad_norm": 1.779342770576477,
+      "learning_rate": 9.254061771789847e-06,
+      "logits/chosen": 0.05093669891357422,
+      "logits/rejected": 0.1048060953617096,
+      "logps/chosen": -84.48141479492188,
+      "logps/rejected": -91.49285125732422,
+      "loss": 2.5324,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9952929019927979,
+      "rewards/margins": 0.3900759220123291,
+      "rewards/rejected": -2.385368585586548,
+      "step": 409
+    },
+    {
+      "epoch": 0.7447774750227066,
+      "grad_norm": 1.71133291721344,
+      "learning_rate": 9.25051460821893e-06,
+      "logits/chosen": 0.1308833658695221,
+      "logits/rejected": 0.08994461596012115,
+      "logps/chosen": -74.09046936035156,
+      "logps/rejected": -80.80438995361328,
+      "loss": 2.1895,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8457674980163574,
+      "rewards/margins": 0.5731581449508667,
+      "rewards/rejected": -2.4189257621765137,
+      "step": 410
+    },
+    {
+      "epoch": 0.7465940054495913,
+      "grad_norm": 1.6980112791061401,
+      "learning_rate": 9.2469597966957e-06,
+      "logits/chosen": 0.05856658145785332,
+      "logits/rejected": 0.08371179550886154,
+      "logps/chosen": -78.58245849609375,
+      "logps/rejected": -88.30054473876953,
+      "loss": 2.46,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9803493022918701,
+      "rewards/margins": 0.6115507483482361,
+      "rewards/rejected": -2.591899871826172,
+      "step": 411
+    },
+    {
+      "epoch": 0.7484105358764759,
+      "grad_norm": 1.8350476026535034,
+      "learning_rate": 9.243397344469037e-06,
+      "logits/chosen": 0.05285171419382095,
+      "logits/rejected": 0.13240401446819305,
+      "logps/chosen": -73.79881286621094,
+      "logps/rejected": -86.56880187988281,
+      "loss": 2.2929,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.060075283050537,
+      "rewards/margins": 0.54813152551651,
+      "rewards/rejected": -2.6082065105438232,
+      "step": 412
+    },
+    {
+      "epoch": 0.7502270663033606,
+      "grad_norm": 1.4403142929077148,
+      "learning_rate": 9.239827258803402e-06,
+      "logits/chosen": 0.07327421009540558,
+      "logits/rejected": 0.18387822806835175,
+      "logps/chosen": -68.3246078491211,
+      "logps/rejected": -81.33067321777344,
+      "loss": 1.8794,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8658640384674072,
+      "rewards/margins": 0.7600168585777283,
+      "rewards/rejected": -2.625880718231201,
+      "step": 413
+    },
+    {
+      "epoch": 0.7520435967302452,
+      "grad_norm": 1.883155345916748,
+      "learning_rate": 9.23624954697882e-06,
+      "logits/chosen": 0.10411994159221649,
+      "logits/rejected": 0.06234448403120041,
+      "logps/chosen": -78.31743621826172,
+      "logps/rejected": -80.0100326538086,
+      "loss": 2.5646,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.934099555015564,
+      "rewards/margins": 0.28737103939056396,
+      "rewards/rejected": -2.221470832824707,
+      "step": 414
+    },
+    {
+      "epoch": 0.7538601271571299,
+      "grad_norm": 1.5926934480667114,
+      "learning_rate": 9.232664216290868e-06,
+      "logits/chosen": 0.09589770436286926,
+      "logits/rejected": 0.041689179837703705,
+      "logps/chosen": -77.89552307128906,
+      "logps/rejected": -84.02517700195312,
+      "loss": 2.2284,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.787453532218933,
+      "rewards/margins": 0.49444466829299927,
+      "rewards/rejected": -2.281898260116577,
+      "step": 415
+    },
+    {
+      "epoch": 0.7556766575840145,
+      "grad_norm": 1.511832356452942,
+      "learning_rate": 9.229071274050663e-06,
+      "logits/chosen": 0.07688678801059723,
+      "logits/rejected": 0.14106512069702148,
+      "logps/chosen": -78.53334045410156,
+      "logps/rejected": -86.92496490478516,
+      "loss": 2.2211,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0686798095703125,
+      "rewards/margins": 0.5378819704055786,
+      "rewards/rejected": -2.6065618991851807,
+      "step": 416
+    },
+    {
+      "epoch": 0.7574931880108992,
+      "grad_norm": 1.7858117818832397,
+      "learning_rate": 9.225470727584835e-06,
+      "logits/chosen": 0.14854730665683746,
+      "logits/rejected": 0.06465649604797363,
+      "logps/chosen": -87.31832122802734,
+      "logps/rejected": -79.3746566772461,
+      "loss": 2.4446,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9429931640625,
+      "rewards/margins": 0.38861486315727234,
+      "rewards/rejected": -2.3316078186035156,
+      "step": 417
+    },
+    {
+      "epoch": 0.7593097184377838,
+      "grad_norm": 1.9230687618255615,
+      "learning_rate": 9.221862584235527e-06,
+      "logits/chosen": 0.07233145087957382,
+      "logits/rejected": 0.13825634121894836,
+      "logps/chosen": -81.53103637695312,
+      "logps/rejected": -90.54696655273438,
+      "loss": 2.7171,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.065859794616699,
+      "rewards/margins": 0.30683329701423645,
+      "rewards/rejected": -2.3726933002471924,
+      "step": 418
+    },
+    {
+      "epoch": 0.7611262488646685,
+      "grad_norm": 2.063098192214966,
+      "learning_rate": 9.218246851360374e-06,
+      "logits/chosen": 0.10360075533390045,
+      "logits/rejected": 0.14677830040454865,
+      "logps/chosen": -85.08120727539062,
+      "logps/rejected": -89.35393524169922,
+      "loss": 2.4292,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.300565719604492,
+      "rewards/margins": 0.4277467727661133,
+      "rewards/rejected": -2.7283124923706055,
+      "step": 419
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 2.4728786945343018,
+      "learning_rate": 9.214623536332483e-06,
+      "logits/chosen": 0.08115474879741669,
+      "logits/rejected": 0.07649822533130646,
+      "logps/chosen": -83.79824829101562,
+      "logps/rejected": -87.47264099121094,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.0838184356689453,
+      "rewards/margins": 0.41770505905151367,
+      "rewards/rejected": -2.501523494720459,
+      "step": 420
+    },
+    {
+      "epoch": 0.7647593097184378,
+      "grad_norm": 1.7078697681427002,
+      "learning_rate": 9.210992646540425e-06,
+      "logits/chosen": 0.08155008405447006,
+      "logits/rejected": 0.08723931759595871,
+      "logps/chosen": -74.4200210571289,
+      "logps/rejected": -87.49038696289062,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9076974391937256,
+      "rewards/margins": 0.6807244420051575,
+      "rewards/rejected": -2.5884220600128174,
+      "step": 421
+    },
+    {
+      "epoch": 0.7665758401453224,
+      "grad_norm": 1.5644993782043457,
+      "learning_rate": 9.207354189388214e-06,
+      "logits/chosen": 0.16107802093029022,
+      "logits/rejected": 0.15951679646968842,
+      "logps/chosen": -74.85030364990234,
+      "logps/rejected": -81.25472259521484,
+      "loss": 2.2389,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0082590579986572,
+      "rewards/margins": 0.5728211402893066,
+      "rewards/rejected": -2.581080198287964,
+      "step": 422
+    },
+    {
+      "epoch": 0.7683923705722071,
+      "grad_norm": 1.6519557237625122,
+      "learning_rate": 9.203708172295299e-06,
+      "logits/chosen": 0.08725707978010178,
+      "logits/rejected": 0.11393023282289505,
+      "logps/chosen": -79.56071472167969,
+      "logps/rejected": -90.27279663085938,
+      "loss": 2.2853,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9312469959259033,
+      "rewards/margins": 0.6229196786880493,
+      "rewards/rejected": -2.554166793823242,
+      "step": 423
+    },
+    {
+      "epoch": 0.7702089009990918,
+      "grad_norm": 1.668648600578308,
+      "learning_rate": 9.200054602696544e-06,
+      "logits/chosen": 0.1369432657957077,
+      "logits/rejected": 0.08549812436103821,
+      "logps/chosen": -80.69041442871094,
+      "logps/rejected": -85.47969055175781,
+      "loss": 2.3022,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7405030727386475,
+      "rewards/margins": 0.4421282112598419,
+      "rewards/rejected": -2.182631492614746,
+      "step": 424
+    },
+    {
+      "epoch": 0.7720254314259763,
+      "grad_norm": 1.7834622859954834,
+      "learning_rate": 9.196393488042213e-06,
+      "logits/chosen": 0.06559979915618896,
+      "logits/rejected": 0.05396304652094841,
+      "logps/chosen": -76.51896667480469,
+      "logps/rejected": -88.19717407226562,
+      "loss": 2.5216,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.050562620162964,
+      "rewards/margins": 0.5811472535133362,
+      "rewards/rejected": -2.631709575653076,
+      "step": 425
+    },
+    {
+      "epoch": 0.773841961852861,
+      "grad_norm": 1.5034058094024658,
+      "learning_rate": 9.192724835797956e-06,
+      "logits/chosen": 0.10820607095956802,
+      "logits/rejected": 0.14427171647548676,
+      "logps/chosen": -81.0079116821289,
+      "logps/rejected": -95.6849136352539,
+      "loss": 2.0472,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.19939923286438,
+      "rewards/margins": 0.8476728796958923,
+      "rewards/rejected": -3.047071933746338,
+      "step": 426
+    },
+    {
+      "epoch": 0.7756584922797457,
+      "grad_norm": 1.7205973863601685,
+      "learning_rate": 9.189048653444796e-06,
+      "logits/chosen": 0.08522580564022064,
+      "logits/rejected": 0.09135682880878448,
+      "logps/chosen": -77.4021224975586,
+      "logps/rejected": -85.47643280029297,
+      "loss": 2.2618,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.002034902572632,
+      "rewards/margins": 0.5705874562263489,
+      "rewards/rejected": -2.572622776031494,
+      "step": 427
+    },
+    {
+      "epoch": 0.7774750227066304,
+      "grad_norm": 2.131197214126587,
+      "learning_rate": 9.185364948479109e-06,
+      "logits/chosen": 0.05514610558748245,
+      "logits/rejected": 0.0597330704331398,
+      "logps/chosen": -86.83938598632812,
+      "logps/rejected": -94.71990966796875,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.088331699371338,
+      "rewards/margins": 0.47913864254951477,
+      "rewards/rejected": -2.567470073699951,
+      "step": 428
+    },
+    {
+      "epoch": 0.779291553133515,
+      "grad_norm": 3.0228030681610107,
+      "learning_rate": 9.181673728412605e-06,
+      "logits/chosen": 0.09502460807561874,
+      "logits/rejected": 0.10438862442970276,
+      "logps/chosen": -77.20478820800781,
+      "logps/rejected": -80.07164764404297,
+      "loss": 2.5924,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.107954978942871,
+      "rewards/margins": 0.4795181155204773,
+      "rewards/rejected": -2.5874733924865723,
+      "step": 429
+    },
+    {
+      "epoch": 0.7811080835603996,
+      "grad_norm": 2.630490779876709,
+      "learning_rate": 9.17797500077233e-06,
+      "logits/chosen": 0.053769052028656006,
+      "logits/rejected": 0.1357191652059555,
+      "logps/chosen": -74.01277160644531,
+      "logps/rejected": -86.05644989013672,
+      "loss": 2.6208,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0179696083068848,
+      "rewards/margins": 0.3262583613395691,
+      "rewards/rejected": -2.3442280292510986,
+      "step": 430
+    },
+    {
+      "epoch": 0.7829246139872843,
+      "grad_norm": 1.7841664552688599,
+      "learning_rate": 9.174268773100636e-06,
+      "logits/chosen": 0.09468546509742737,
+      "logits/rejected": 0.16804733872413635,
+      "logps/chosen": -80.07379913330078,
+      "logps/rejected": -87.17518615722656,
+      "loss": 2.4676,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.27384877204895,
+      "rewards/margins": 0.4068894386291504,
+      "rewards/rejected": -2.6807379722595215,
+      "step": 431
+    },
+    {
+      "epoch": 0.784741144414169,
+      "grad_norm": 1.7030746936798096,
+      "learning_rate": 9.170555052955158e-06,
+      "logits/chosen": 0.10446357727050781,
+      "logits/rejected": 0.1235620528459549,
+      "logps/chosen": -89.81584930419922,
+      "logps/rejected": -94.43669128417969,
+      "loss": 2.3366,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.98858642578125,
+      "rewards/margins": 0.5402320623397827,
+      "rewards/rejected": -2.5288188457489014,
+      "step": 432
+    },
+    {
+      "epoch": 0.7865576748410535,
+      "grad_norm": 1.4258465766906738,
+      "learning_rate": 9.166833847908825e-06,
+      "logits/chosen": 0.10827025771141052,
+      "logits/rejected": 0.10875076055526733,
+      "logps/chosen": -69.29400634765625,
+      "logps/rejected": -80.72003936767578,
+      "loss": 2.2501,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9478650093078613,
+      "rewards/margins": 0.5940475463867188,
+      "rewards/rejected": -2.54191255569458,
+      "step": 433
+    },
+    {
+      "epoch": 0.7883742052679382,
+      "grad_norm": 1.8069814443588257,
+      "learning_rate": 9.163105165549819e-06,
+      "logits/chosen": 0.08856840431690216,
+      "logits/rejected": 0.07843751460313797,
+      "logps/chosen": -79.93275451660156,
+      "logps/rejected": -83.75475311279297,
+      "loss": 2.5214,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0315518379211426,
+      "rewards/margins": 0.45457565784454346,
+      "rewards/rejected": -2.4861276149749756,
+      "step": 434
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 1.7881503105163574,
+      "learning_rate": 9.159369013481574e-06,
+      "logits/chosen": 0.08166562020778656,
+      "logits/rejected": 0.08240097761154175,
+      "logps/chosen": -92.98298645019531,
+      "logps/rejected": -96.15763092041016,
+      "loss": 2.053,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4435133934020996,
+      "rewards/margins": 0.7140956521034241,
+      "rewards/rejected": -3.157609462738037,
+      "step": 435
+    },
+    {
+      "epoch": 0.7920072661217076,
+      "grad_norm": 1.7574496269226074,
+      "learning_rate": 9.155625399322754e-06,
+      "logits/chosen": 0.08114133030176163,
+      "logits/rejected": 0.11279730498790741,
+      "logps/chosen": -74.07096862792969,
+      "logps/rejected": -80.94386291503906,
+      "loss": 2.4289,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2497410774230957,
+      "rewards/margins": 0.4424628019332886,
+      "rewards/rejected": -2.6922037601470947,
+      "step": 436
+    },
+    {
+      "epoch": 0.7938237965485921,
+      "grad_norm": 1.4814083576202393,
+      "learning_rate": 9.15187433070724e-06,
+      "logits/chosen": 0.026073571294546127,
+      "logits/rejected": 0.1382063925266266,
+      "logps/chosen": -72.39479064941406,
+      "logps/rejected": -86.76889038085938,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9130635261535645,
+      "rewards/margins": 0.6362709999084473,
+      "rewards/rejected": -2.549334764480591,
+      "step": 437
+    },
+    {
+      "epoch": 0.7956403269754768,
+      "grad_norm": 1.6441129446029663,
+      "learning_rate": 9.148115815284113e-06,
+      "logits/chosen": 0.07432619482278824,
+      "logits/rejected": 0.10215617716312408,
+      "logps/chosen": -77.05310821533203,
+      "logps/rejected": -80.81717681884766,
+      "loss": 2.338,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9012683629989624,
+      "rewards/margins": 0.43776583671569824,
+      "rewards/rejected": -2.339034080505371,
+      "step": 438
+    },
+    {
+      "epoch": 0.7974568574023615,
+      "grad_norm": 1.8839560747146606,
+      "learning_rate": 9.144349860717643e-06,
+      "logits/chosen": 0.07576426863670349,
+      "logits/rejected": 0.08868670463562012,
+      "logps/chosen": -74.13190460205078,
+      "logps/rejected": -80.07232666015625,
+      "loss": 2.3485,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9554617404937744,
+      "rewards/margins": 0.531810998916626,
+      "rewards/rejected": -2.4872727394104004,
+      "step": 439
+    },
+    {
+      "epoch": 0.7992733878292462,
+      "grad_norm": 1.826811671257019,
+      "learning_rate": 9.140576474687263e-06,
+      "logits/chosen": 0.09120994061231613,
+      "logits/rejected": 0.08987519890069962,
+      "logps/chosen": -75.1444320678711,
+      "logps/rejected": -78.58987426757812,
+      "loss": 2.6513,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158261299133301,
+      "rewards/margins": 0.28016677498817444,
+      "rewards/rejected": -2.4384284019470215,
+      "step": 440
+    },
+    {
+      "epoch": 0.8010899182561307,
+      "grad_norm": 1.8647096157073975,
+      "learning_rate": 9.13679566488757e-06,
+      "logits/chosen": 0.12438184767961502,
+      "logits/rejected": 0.13951222598552704,
+      "logps/chosen": -83.79389953613281,
+      "logps/rejected": -90.33230590820312,
+      "loss": 2.3956,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0185930728912354,
+      "rewards/margins": 0.43685024976730347,
+      "rewards/rejected": -2.4554433822631836,
+      "step": 441
+    },
+    {
+      "epoch": 0.8029064486830154,
+      "grad_norm": 1.7610801458358765,
+      "learning_rate": 9.133007439028288e-06,
+      "logits/chosen": 0.08218151330947876,
+      "logits/rejected": 0.072816863656044,
+      "logps/chosen": -86.35633087158203,
+      "logps/rejected": -84.08189392089844,
+      "loss": 2.243,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8836033344268799,
+      "rewards/margins": 0.516740083694458,
+      "rewards/rejected": -2.400343418121338,
+      "step": 442
+    },
+    {
+      "epoch": 0.8047229791099001,
+      "grad_norm": 1.6049703359603882,
+      "learning_rate": 9.129211804834271e-06,
+      "logits/chosen": 0.059631846845149994,
+      "logits/rejected": 0.09744230657815933,
+      "logps/chosen": -76.90062713623047,
+      "logps/rejected": -82.29356384277344,
+      "loss": 2.2574,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9436204433441162,
+      "rewards/margins": 0.47444695234298706,
+      "rewards/rejected": -2.418067216873169,
+      "step": 443
+    },
+    {
+      "epoch": 0.8065395095367848,
+      "grad_norm": 1.9819791316986084,
+      "learning_rate": 9.12540877004548e-06,
+      "logits/chosen": 0.07371871173381805,
+      "logits/rejected": 0.058843065053224564,
+      "logps/chosen": -78.42516326904297,
+      "logps/rejected": -82.23822784423828,
+      "loss": 2.6959,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.273909330368042,
+      "rewards/margins": 0.3514346182346344,
+      "rewards/rejected": -2.6253440380096436,
+      "step": 444
+    },
+    {
+      "epoch": 0.8083560399636693,
+      "grad_norm": 1.9540653228759766,
+      "learning_rate": 9.12159834241696e-06,
+      "logits/chosen": 0.13155074417591095,
+      "logits/rejected": 0.11775672435760498,
+      "logps/chosen": -69.17151641845703,
+      "logps/rejected": -77.9190902709961,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.073258638381958,
+      "rewards/margins": 0.5955159664154053,
+      "rewards/rejected": -2.6687746047973633,
+      "step": 445
+    },
+    {
+      "epoch": 0.810172570390554,
+      "grad_norm": 1.9634339809417725,
+      "learning_rate": 9.117780529718843e-06,
+      "logits/chosen": 0.05265370383858681,
+      "logits/rejected": 0.09651105850934982,
+      "logps/chosen": -78.77818298339844,
+      "logps/rejected": -84.63584899902344,
+      "loss": 2.2157,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.084688186645508,
+      "rewards/margins": 0.541061282157898,
+      "rewards/rejected": -2.625749349594116,
+      "step": 446
+    },
+    {
+      "epoch": 0.8119891008174387,
+      "grad_norm": 1.9086997509002686,
+      "learning_rate": 9.113955339736309e-06,
+      "logits/chosen": 0.08473093807697296,
+      "logits/rejected": 0.047919195145368576,
+      "logps/chosen": -83.44226837158203,
+      "logps/rejected": -87.24287414550781,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9542481899261475,
+      "rewards/margins": 0.4577428698539734,
+      "rewards/rejected": -2.4119908809661865,
+      "step": 447
+    },
+    {
+      "epoch": 0.8138056312443234,
+      "grad_norm": 1.6967483758926392,
+      "learning_rate": 9.11012278026959e-06,
+      "logits/chosen": 0.062431350350379944,
+      "logits/rejected": 0.10980932414531708,
+      "logps/chosen": -90.98541259765625,
+      "logps/rejected": -102.51634979248047,
+      "loss": 2.3339,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1866049766540527,
+      "rewards/margins": 0.6549904346466064,
+      "rewards/rejected": -2.841595411300659,
+      "step": 448
+    },
+    {
+      "epoch": 0.815622161671208,
+      "grad_norm": 2.295504331588745,
+      "learning_rate": 9.106282859133936e-06,
+      "logits/chosen": 0.04481218010187149,
+      "logits/rejected": 0.002619542181491852,
+      "logps/chosen": -81.81380462646484,
+      "logps/rejected": -80.60113525390625,
+      "loss": 2.8324,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.1727848052978516,
+      "rewards/margins": 0.2777697443962097,
+      "rewards/rejected": -2.450554609298706,
+      "step": 449
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 1.7951968908309937,
+      "learning_rate": 9.102435584159623e-06,
+      "logits/chosen": 0.029065577313303947,
+      "logits/rejected": 0.11651361733675003,
+      "logps/chosen": -78.08231353759766,
+      "logps/rejected": -88.59742736816406,
+      "loss": 2.5562,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.2762022018432617,
+      "rewards/margins": 0.4410257041454315,
+      "rewards/rejected": -2.7172276973724365,
+      "step": 450
+    },
+    {
+      "epoch": 0.8192552225249773,
+      "grad_norm": 1.6301803588867188,
+      "learning_rate": 9.098580963191908e-06,
+      "logits/chosen": 0.06471782922744751,
+      "logits/rejected": 0.11030608415603638,
+      "logps/chosen": -78.3387451171875,
+      "logps/rejected": -91.82540130615234,
+      "loss": 2.1826,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.885233998298645,
+      "rewards/margins": 0.6971657872200012,
+      "rewards/rejected": -2.582399845123291,
+      "step": 451
+    },
+    {
+      "epoch": 0.821071752951862,
+      "grad_norm": 1.705962061882019,
+      "learning_rate": 9.094719004091039e-06,
+      "logits/chosen": 0.12591060996055603,
+      "logits/rejected": 0.1107010543346405,
+      "logps/chosen": -76.82783508300781,
+      "logps/rejected": -84.23365020751953,
+      "loss": 2.2843,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.04795241355896,
+      "rewards/margins": 0.5103356242179871,
+      "rewards/rejected": -2.558288335800171,
+      "step": 452
+    },
+    {
+      "epoch": 0.8228882833787466,
+      "grad_norm": 1.6535886526107788,
+      "learning_rate": 9.090849714732217e-06,
+      "logits/chosen": 0.11322569847106934,
+      "logits/rejected": 0.13092264533042908,
+      "logps/chosen": -80.16563415527344,
+      "logps/rejected": -82.96965026855469,
+      "loss": 2.3556,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9074076414108276,
+      "rewards/margins": 0.39431121945381165,
+      "rewards/rejected": -2.3017189502716064,
+      "step": 453
+    },
+    {
+      "epoch": 0.8247048138056312,
+      "grad_norm": 1.7194557189941406,
+      "learning_rate": 9.086973103005602e-06,
+      "logits/chosen": 0.1069367378950119,
+      "logits/rejected": 0.06237747147679329,
+      "logps/chosen": -72.59986877441406,
+      "logps/rejected": -77.62615203857422,
+      "loss": 2.2768,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.023085355758667,
+      "rewards/margins": 0.6052648425102234,
+      "rewards/rejected": -2.6283504962921143,
+      "step": 454
+    },
+    {
+      "epoch": 0.8265213442325159,
+      "grad_norm": 1.7921446561813354,
+      "learning_rate": 9.08308917681628e-06,
+      "logits/chosen": 0.12843580543994904,
+      "logits/rejected": 0.08002308011054993,
+      "logps/chosen": -80.6107177734375,
+      "logps/rejected": -80.92134857177734,
+      "loss": 2.3881,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9404933452606201,
+      "rewards/margins": 0.37478840351104736,
+      "rewards/rejected": -2.315281629562378,
+      "step": 455
+    },
+    {
+      "epoch": 0.8283378746594006,
+      "grad_norm": 2.3353819847106934,
+      "learning_rate": 9.07919794408425e-06,
+      "logits/chosen": 0.05452323704957962,
+      "logits/rejected": 0.12883196771144867,
+      "logps/chosen": -83.56001281738281,
+      "logps/rejected": -95.23291015625,
+      "loss": 2.8503,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.2597994804382324,
+      "rewards/margins": 0.4092620015144348,
+      "rewards/rejected": -2.6690614223480225,
+      "step": 456
+    },
+    {
+      "epoch": 0.8301544050862852,
+      "grad_norm": 1.6829804182052612,
+      "learning_rate": 9.075299412744417e-06,
+      "logits/chosen": 0.1223950982093811,
+      "logits/rejected": 0.10596577078104019,
+      "logps/chosen": -77.75115203857422,
+      "logps/rejected": -83.63490295410156,
+      "loss": 2.1639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.026258945465088,
+      "rewards/margins": 0.6107547879219055,
+      "rewards/rejected": -2.6370139122009277,
+      "step": 457
+    },
+    {
+      "epoch": 0.8319709355131698,
+      "grad_norm": 2.053755283355713,
+      "learning_rate": 9.07139359074656e-06,
+      "logits/chosen": 0.08021271228790283,
+      "logits/rejected": 0.11617676913738251,
+      "logps/chosen": -75.00261688232422,
+      "logps/rejected": -87.25222778320312,
+      "loss": 2.4048,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3207621574401855,
+      "rewards/margins": 0.5956254005432129,
+      "rewards/rejected": -2.9163873195648193,
+      "step": 458
+    },
+    {
+      "epoch": 0.8337874659400545,
+      "grad_norm": 2.1006064414978027,
+      "learning_rate": 9.067480486055333e-06,
+      "logits/chosen": 0.1115679144859314,
+      "logits/rejected": 0.08167213946580887,
+      "logps/chosen": -81.78447723388672,
+      "logps/rejected": -82.21318817138672,
+      "loss": 2.8725,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.165053367614746,
+      "rewards/margins": 0.2777022123336792,
+      "rewards/rejected": -2.442755699157715,
+      "step": 459
+    },
+    {
+      "epoch": 0.8356039963669392,
+      "grad_norm": 1.6339340209960938,
+      "learning_rate": 9.063560106650238e-06,
+      "logits/chosen": 0.0966312363743782,
+      "logits/rejected": 0.09943810105323792,
+      "logps/chosen": -71.0906982421875,
+      "logps/rejected": -76.22811126708984,
+      "loss": 2.2922,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.0007636547088623,
+      "rewards/margins": 0.47423097491264343,
+      "rewards/rejected": -2.474994659423828,
+      "step": 460
+    },
+    {
+      "epoch": 0.8374205267938238,
+      "grad_norm": 1.8504972457885742,
+      "learning_rate": 9.059632460525613e-06,
+      "logits/chosen": 0.09517084062099457,
+      "logits/rejected": 0.14127981662750244,
+      "logps/chosen": -79.06185913085938,
+      "logps/rejected": -89.11494445800781,
+      "loss": 2.3288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0016026496887207,
+      "rewards/margins": 0.5282158255577087,
+      "rewards/rejected": -2.529818534851074,
+      "step": 461
+    },
+    {
+      "epoch": 0.8392370572207084,
+      "grad_norm": 1.8151304721832275,
+      "learning_rate": 9.055697555690607e-06,
+      "logits/chosen": 0.08868349343538284,
+      "logits/rejected": 0.14090785384178162,
+      "logps/chosen": -77.26272583007812,
+      "logps/rejected": -85.14620971679688,
+      "loss": 2.3694,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.087087869644165,
+      "rewards/margins": 0.5032299757003784,
+      "rewards/rejected": -2.590317964553833,
+      "step": 462
+    },
+    {
+      "epoch": 0.8410535876475931,
+      "grad_norm": 1.8596280813217163,
+      "learning_rate": 9.051755400169182e-06,
+      "logits/chosen": 0.06272133439779282,
+      "logits/rejected": 0.12342572212219238,
+      "logps/chosen": -75.46456909179688,
+      "logps/rejected": -81.9759292602539,
+      "loss": 2.5167,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0463178157806396,
+      "rewards/margins": 0.3146142363548279,
+      "rewards/rejected": -2.3609323501586914,
+      "step": 463
+    },
+    {
+      "epoch": 0.8428701180744778,
+      "grad_norm": 1.7658417224884033,
+      "learning_rate": 9.047806002000075e-06,
+      "logits/chosen": -0.007259421981871128,
+      "logits/rejected": 0.09611339122056961,
+      "logps/chosen": -73.88439178466797,
+      "logps/rejected": -80.77063751220703,
+      "loss": 2.4938,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9998183250427246,
+      "rewards/margins": 0.4115346670150757,
+      "rewards/rejected": -2.4113528728485107,
+      "step": 464
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 1.6596916913986206,
+      "learning_rate": 9.043849369236799e-06,
+      "logits/chosen": 0.03223409131169319,
+      "logits/rejected": 0.05010119825601578,
+      "logps/chosen": -66.32562255859375,
+      "logps/rejected": -69.36509704589844,
+      "loss": 2.5881,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.909837245941162,
+      "rewards/margins": 0.2923263907432556,
+      "rewards/rejected": -2.2021636962890625,
+      "step": 465
+    },
+    {
+      "epoch": 0.846503178928247,
+      "grad_norm": 1.6832820177078247,
+      "learning_rate": 9.039885509947616e-06,
+      "logits/chosen": 0.06287454813718796,
+      "logits/rejected": 0.12746769189834595,
+      "logps/chosen": -76.6998519897461,
+      "logps/rejected": -87.36117553710938,
+      "loss": 2.3914,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.029526710510254,
+      "rewards/margins": 0.48344865441322327,
+      "rewards/rejected": -2.5129752159118652,
+      "step": 466
+    },
+    {
+      "epoch": 0.8483197093551317,
+      "grad_norm": 1.6972366571426392,
+      "learning_rate": 9.035914432215527e-06,
+      "logits/chosen": 0.1179896891117096,
+      "logits/rejected": 0.14000467956066132,
+      "logps/chosen": -75.3287353515625,
+      "logps/rejected": -78.60227966308594,
+      "loss": 2.4194,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.088515043258667,
+      "rewards/margins": 0.42544490098953247,
+      "rewards/rejected": -2.513960123062134,
+      "step": 467
+    },
+    {
+      "epoch": 0.8501362397820164,
+      "grad_norm": 2.5104973316192627,
+      "learning_rate": 9.031936144138247e-06,
+      "logits/chosen": 0.08423591405153275,
+      "logits/rejected": 0.1387752890586853,
+      "logps/chosen": -78.6176528930664,
+      "logps/rejected": -83.52703094482422,
+      "loss": 2.7262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1089184284210205,
+      "rewards/margins": 0.44402289390563965,
+      "rewards/rejected": -2.55294132232666,
+      "step": 468
+    },
+    {
+      "epoch": 0.851952770208901,
+      "grad_norm": 1.944931149482727,
+      "learning_rate": 9.027950653828202e-06,
+      "logits/chosen": 0.12252221256494522,
+      "logits/rejected": 0.14959384500980377,
+      "logps/chosen": -72.63184356689453,
+      "logps/rejected": -70.89137268066406,
+      "loss": 2.7323,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9612714052200317,
+      "rewards/margins": 0.24262896180152893,
+      "rewards/rejected": -2.2039003372192383,
+      "step": 469
+    },
+    {
+      "epoch": 0.8537693006357856,
+      "grad_norm": 1.6699979305267334,
+      "learning_rate": 9.0239579694125e-06,
+      "logits/chosen": 0.11827167868614197,
+      "logits/rejected": 0.11933022737503052,
+      "logps/chosen": -70.45188903808594,
+      "logps/rejected": -76.2336654663086,
+      "loss": 2.457,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.854758381843567,
+      "rewards/margins": 0.47020024061203003,
+      "rewards/rejected": -2.3249588012695312,
+      "step": 470
+    },
+    {
+      "epoch": 0.8555858310626703,
+      "grad_norm": 1.5790536403656006,
+      "learning_rate": 9.019958099032919e-06,
+      "logits/chosen": 0.04856352508068085,
+      "logits/rejected": 0.11418008804321289,
+      "logps/chosen": -75.7929458618164,
+      "logps/rejected": -86.55818176269531,
+      "loss": 2.2809,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.003899335861206,
+      "rewards/margins": 0.5960561037063599,
+      "rewards/rejected": -2.5999553203582764,
+      "step": 471
+    },
+    {
+      "epoch": 0.857402361489555,
+      "grad_norm": 1.3594144582748413,
+      "learning_rate": 9.015951050845891e-06,
+      "logits/chosen": 0.10341258347034454,
+      "logits/rejected": 0.12394269555807114,
+      "logps/chosen": -78.8187255859375,
+      "logps/rejected": -84.04349517822266,
+      "loss": 1.9536,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9372440576553345,
+      "rewards/margins": 0.6742948889732361,
+      "rewards/rejected": -2.611538887023926,
+      "step": 472
+    },
+    {
+      "epoch": 0.8592188919164396,
+      "grad_norm": 1.7592689990997314,
+      "learning_rate": 9.011936833022485e-06,
+      "logits/chosen": 0.10970023274421692,
+      "logits/rejected": 0.09040558338165283,
+      "logps/chosen": -80.59556579589844,
+      "logps/rejected": -84.2996826171875,
+      "loss": 2.5658,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.935511827468872,
+      "rewards/margins": 0.35643547773361206,
+      "rewards/rejected": -2.29194712638855,
+      "step": 473
+    },
+    {
+      "epoch": 0.8610354223433242,
+      "grad_norm": 1.6418087482452393,
+      "learning_rate": 9.00791545374839e-06,
+      "logits/chosen": 0.16755647957324982,
+      "logits/rejected": 0.1745520681142807,
+      "logps/chosen": -82.31010437011719,
+      "logps/rejected": -85.66128540039062,
+      "loss": 2.1854,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.9519617557525635,
+      "rewards/margins": 0.5059336423873901,
+      "rewards/rejected": -2.457895278930664,
+      "step": 474
+    },
+    {
+      "epoch": 0.8628519527702089,
+      "grad_norm": 1.4616092443466187,
+      "learning_rate": 9.0038869212239e-06,
+      "logits/chosen": 0.03978392109274864,
+      "logits/rejected": 0.1235240027308464,
+      "logps/chosen": -77.78251647949219,
+      "logps/rejected": -85.66397094726562,
+      "loss": 2.1995,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8721034526824951,
+      "rewards/margins": 0.45375847816467285,
+      "rewards/rejected": -2.325861930847168,
+      "step": 475
+    },
+    {
+      "epoch": 0.8646684831970936,
+      "grad_norm": 1.838180422782898,
+      "learning_rate": 8.99985124366389e-06,
+      "logits/chosen": 0.08277668058872223,
+      "logits/rejected": 0.035900432616472244,
+      "logps/chosen": -83.23967742919922,
+      "logps/rejected": -83.16535949707031,
+      "loss": 2.7347,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.2339279651641846,
+      "rewards/margins": 0.16711921989917755,
+      "rewards/rejected": -2.4010472297668457,
+      "step": 476
+    },
+    {
+      "epoch": 0.8664850136239782,
+      "grad_norm": 1.6746636629104614,
+      "learning_rate": 8.995808429297815e-06,
+      "logits/chosen": 0.14687396585941315,
+      "logits/rejected": 0.09570137411355972,
+      "logps/chosen": -78.79898071289062,
+      "logps/rejected": -77.9619369506836,
+      "loss": 2.4449,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8334993124008179,
+      "rewards/margins": 0.3442971706390381,
+      "rewards/rejected": -2.1777963638305664,
+      "step": 477
+    },
+    {
+      "epoch": 0.8683015440508629,
+      "grad_norm": 1.7042535543441772,
+      "learning_rate": 8.991758486369675e-06,
+      "logits/chosen": 0.047281138598918915,
+      "logits/rejected": -0.023666318506002426,
+      "logps/chosen": -83.798583984375,
+      "logps/rejected": -91.88795471191406,
+      "loss": 2.3072,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9728612899780273,
+      "rewards/margins": 0.5659723281860352,
+      "rewards/rejected": -2.5388338565826416,
+      "step": 478
+    },
+    {
+      "epoch": 0.8701180744777475,
+      "grad_norm": 1.5552817583084106,
+      "learning_rate": 8.987701423138007e-06,
+      "logits/chosen": 0.0772751122713089,
+      "logits/rejected": 0.08708472549915314,
+      "logps/chosen": -72.42391204833984,
+      "logps/rejected": -78.93070220947266,
+      "loss": 2.2815,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9803509712219238,
+      "rewards/margins": 0.5386512279510498,
+      "rewards/rejected": -2.5190021991729736,
+      "step": 479
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.6372044086456299,
+      "learning_rate": 8.983637247875872e-06,
+      "logits/chosen": -0.03352706879377365,
+      "logits/rejected": 0.01581352949142456,
+      "logps/chosen": -81.29138946533203,
+      "logps/rejected": -89.1744155883789,
+      "loss": 2.09,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9441754817962646,
+      "rewards/margins": 0.6137264370918274,
+      "rewards/rejected": -2.5579018592834473,
+      "step": 480
+    },
+    {
+      "epoch": 0.8737511353315168,
+      "grad_norm": 1.6375274658203125,
+      "learning_rate": 8.979565968870831e-06,
+      "logits/chosen": 0.07507088780403137,
+      "logits/rejected": 0.1289597451686859,
+      "logps/chosen": -74.78109741210938,
+      "logps/rejected": -84.55204772949219,
+      "loss": 2.4566,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.9041385650634766,
+      "rewards/margins": 0.42814433574676514,
+      "rewards/rejected": -2.332282781600952,
+      "step": 481
+    },
+    {
+      "epoch": 0.8755676657584015,
+      "grad_norm": 1.5601258277893066,
+      "learning_rate": 8.975487594424927e-06,
+      "logits/chosen": 0.012452262453734875,
+      "logits/rejected": 0.06930352002382278,
+      "logps/chosen": -83.01705169677734,
+      "logps/rejected": -92.38286590576172,
+      "loss": 2.2736,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.013011932373047,
+      "rewards/margins": 0.5085774064064026,
+      "rewards/rejected": -2.5215890407562256,
+      "step": 482
+    },
+    {
+      "epoch": 0.8773841961852861,
+      "grad_norm": 1.8685195446014404,
+      "learning_rate": 8.971402132854677e-06,
+      "logits/chosen": 0.09415426105260849,
+      "logits/rejected": 0.12719042599201202,
+      "logps/chosen": -79.85440063476562,
+      "logps/rejected": -82.80875396728516,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.114530086517334,
+      "rewards/margins": 0.46328115463256836,
+      "rewards/rejected": -2.5778112411499023,
+      "step": 483
+    },
+    {
+      "epoch": 0.8792007266121707,
+      "grad_norm": 1.8812834024429321,
+      "learning_rate": 8.967309592491052e-06,
+      "logits/chosen": 0.13547101616859436,
+      "logits/rejected": 0.11872326582670212,
+      "logps/chosen": -74.42122650146484,
+      "logps/rejected": -81.9700927734375,
+      "loss": 2.5016,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.1828627586364746,
+      "rewards/margins": 0.44510167837142944,
+      "rewards/rejected": -2.627964496612549,
+      "step": 484
+    },
+    {
+      "epoch": 0.8810172570390554,
+      "grad_norm": 1.510517954826355,
+      "learning_rate": 8.963209981679451e-06,
+      "logits/chosen": 0.03274242952466011,
+      "logits/rejected": 0.07188954204320908,
+      "logps/chosen": -82.55563354492188,
+      "logps/rejected": -100.33609771728516,
+      "loss": 1.9177,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9487042427062988,
+      "rewards/margins": 0.8189151287078857,
+      "rewards/rejected": -2.7676191329956055,
+      "step": 485
+    },
+    {
+      "epoch": 0.8828337874659401,
+      "grad_norm": 1.5973646640777588,
+      "learning_rate": 8.959103308779696e-06,
+      "logits/chosen": 0.0365552082657814,
+      "logits/rejected": 0.051129020750522614,
+      "logps/chosen": -72.4110107421875,
+      "logps/rejected": -83.10322570800781,
+      "loss": 2.1547,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9375314712524414,
+      "rewards/margins": 0.6991211175918579,
+      "rewards/rejected": -2.6366524696350098,
+      "step": 486
+    },
+    {
+      "epoch": 0.8846503178928247,
+      "grad_norm": 1.6707032918930054,
+      "learning_rate": 8.954989582166009e-06,
+      "logits/chosen": 0.0635693296790123,
+      "logits/rejected": 0.009854275733232498,
+      "logps/chosen": -87.12749481201172,
+      "logps/rejected": -88.9491195678711,
+      "loss": 2.2043,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9257593154907227,
+      "rewards/margins": 0.5429801344871521,
+      "rewards/rejected": -2.4687397480010986,
+      "step": 487
+    },
+    {
+      "epoch": 0.8864668483197093,
+      "grad_norm": 1.4422773122787476,
+      "learning_rate": 8.95086881022699e-06,
+      "logits/chosen": -0.024460218846797943,
+      "logits/rejected": 0.14137038588523865,
+      "logps/chosen": -74.60330200195312,
+      "logps/rejected": -95.01730346679688,
+      "loss": 2.0246,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.054253339767456,
+      "rewards/margins": 0.8260326981544495,
+      "rewards/rejected": -2.88028621673584,
+      "step": 488
+    },
+    {
+      "epoch": 0.888283378746594,
+      "grad_norm": 1.8121236562728882,
+      "learning_rate": 8.946741001365616e-06,
+      "logits/chosen": 0.05854415148496628,
+      "logits/rejected": 0.13758361339569092,
+      "logps/chosen": -74.46614074707031,
+      "logps/rejected": -83.96118927001953,
+      "loss": 2.5066,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.105980157852173,
+      "rewards/margins": 0.41478782892227173,
+      "rewards/rejected": -2.520768165588379,
+      "step": 489
+    },
+    {
+      "epoch": 0.8900999091734787,
+      "grad_norm": 2.023728370666504,
+      "learning_rate": 8.942606163999205e-06,
+      "logits/chosen": 0.04619833081960678,
+      "logits/rejected": 0.09330146014690399,
+      "logps/chosen": -75.35594940185547,
+      "logps/rejected": -89.18836975097656,
+      "loss": 2.2655,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9396051168441772,
+      "rewards/margins": 0.6726698875427246,
+      "rewards/rejected": -2.6122751235961914,
+      "step": 490
+    },
+    {
+      "epoch": 0.8919164396003633,
+      "grad_norm": 1.559735894203186,
+      "learning_rate": 8.938464306559412e-06,
+      "logits/chosen": 0.093504399061203,
+      "logits/rejected": 0.09722252935171127,
+      "logps/chosen": -80.77140808105469,
+      "logps/rejected": -87.86373901367188,
+      "loss": 2.1676,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.1073269844055176,
+      "rewards/margins": 0.5740070939064026,
+      "rewards/rejected": -2.6813340187072754,
+      "step": 491
+    },
+    {
+      "epoch": 0.8937329700272479,
+      "grad_norm": 1.6771574020385742,
+      "learning_rate": 8.934315437492203e-06,
+      "logits/chosen": 0.06368491798639297,
+      "logits/rejected": 0.04900998994708061,
+      "logps/chosen": -78.2313003540039,
+      "logps/rejected": -92.83306121826172,
+      "loss": 2.3435,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.110978364944458,
+      "rewards/margins": 0.6751725673675537,
+      "rewards/rejected": -2.786151170730591,
+      "step": 492
+    },
+    {
+      "epoch": 0.8955495004541326,
+      "grad_norm": 1.651183843612671,
+      "learning_rate": 8.930159565257846e-06,
+      "logits/chosen": 0.08393608778715134,
+      "logits/rejected": 0.0475153923034668,
+      "logps/chosen": -77.8431396484375,
+      "logps/rejected": -88.03681945800781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0274362564086914,
+      "rewards/margins": 0.8075671195983887,
+      "rewards/rejected": -2.835003137588501,
+      "step": 493
+    },
+    {
+      "epoch": 0.8973660308810173,
+      "grad_norm": 2.262662172317505,
+      "learning_rate": 8.925996698330887e-06,
+      "logits/chosen": 0.014226208440959454,
+      "logits/rejected": 0.09588178992271423,
+      "logps/chosen": -88.56185913085938,
+      "logps/rejected": -95.1131591796875,
+      "loss": 2.5853,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4549288749694824,
+      "rewards/margins": 0.5275144577026367,
+      "rewards/rejected": -2.982443332672119,
+      "step": 494
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 1.6070181131362915,
+      "learning_rate": 8.92182684520014e-06,
+      "logits/chosen": 0.11615607142448425,
+      "logits/rejected": 0.16645964980125427,
+      "logps/chosen": -78.97079467773438,
+      "logps/rejected": -90.49947357177734,
+      "loss": 2.1072,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.175706386566162,
+      "rewards/margins": 0.7435849905014038,
+      "rewards/rejected": -2.9192914962768555,
+      "step": 495
+    },
+    {
+      "epoch": 0.9009990917347865,
+      "grad_norm": 1.8343369960784912,
+      "learning_rate": 8.917650014368658e-06,
+      "logits/chosen": 0.12583515048027039,
+      "logits/rejected": 0.07308925688266754,
+      "logps/chosen": -83.38356018066406,
+      "logps/rejected": -90.73155975341797,
+      "loss": 2.3116,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3124871253967285,
+      "rewards/margins": 0.5869801640510559,
+      "rewards/rejected": -2.8994674682617188,
+      "step": 496
+    },
+    {
+      "epoch": 0.9028156221616712,
+      "grad_norm": 1.7175630331039429,
+      "learning_rate": 8.913466214353728e-06,
+      "logits/chosen": 0.14971572160720825,
+      "logits/rejected": 0.17372727394104004,
+      "logps/chosen": -86.66060638427734,
+      "logps/rejected": -94.42353820800781,
+      "loss": 2.3541,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.943577766418457,
+      "rewards/margins": 0.5955328345298767,
+      "rewards/rejected": -2.5391108989715576,
+      "step": 497
+    },
+    {
+      "epoch": 0.9046321525885559,
+      "grad_norm": 1.754708170890808,
+      "learning_rate": 8.909275453686845e-06,
+      "logits/chosen": 0.07779194414615631,
+      "logits/rejected": 0.09394712746143341,
+      "logps/chosen": -71.64151000976562,
+      "logps/rejected": -78.53131103515625,
+      "loss": 2.0436,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0102486610412598,
+      "rewards/margins": 0.6668623685836792,
+      "rewards/rejected": -2.6771109104156494,
+      "step": 498
+    },
+    {
+      "epoch": 0.9064486830154405,
+      "grad_norm": 1.9303216934204102,
+      "learning_rate": 8.905077740913701e-06,
+      "logits/chosen": 0.15853236615657806,
+      "logits/rejected": 0.11452697217464447,
+      "logps/chosen": -81.46524047851562,
+      "logps/rejected": -89.1716537475586,
+      "loss": 2.2477,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.264355182647705,
+      "rewards/margins": 0.6477693915367126,
+      "rewards/rejected": -2.9121243953704834,
+      "step": 499
+    },
+    {
+      "epoch": 0.9082652134423251,
+      "grad_norm": 2.1294379234313965,
+      "learning_rate": 8.900873084594164e-06,
+      "logits/chosen": 0.029480352997779846,
+      "logits/rejected": 0.09827219694852829,
+      "logps/chosen": -95.93231964111328,
+      "logps/rejected": -106.21436309814453,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.3455631732940674,
+      "rewards/margins": 0.5977038145065308,
+      "rewards/rejected": -2.9432668685913086,
+      "step": 500
+    },
+    {
+      "epoch": 0.9100817438692098,
+      "grad_norm": 1.7347708940505981,
+      "learning_rate": 8.896661493302258e-06,
+      "logits/chosen": 0.08531993627548218,
+      "logits/rejected": 0.14993277192115784,
+      "logps/chosen": -79.09967041015625,
+      "logps/rejected": -90.50935363769531,
+      "loss": 2.2969,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9908722639083862,
+      "rewards/margins": 0.5710910558700562,
+      "rewards/rejected": -2.5619633197784424,
+      "step": 501
+    },
+    {
+      "epoch": 0.9118982742960945,
+      "grad_norm": 2.065664529800415,
+      "learning_rate": 8.892442975626152e-06,
+      "logits/chosen": 0.18976512551307678,
+      "logits/rejected": 0.13936059176921844,
+      "logps/chosen": -82.80276489257812,
+      "logps/rejected": -84.57154083251953,
+      "loss": 2.5968,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4863569736480713,
+      "rewards/margins": 0.3651903569698334,
+      "rewards/rejected": -2.8515477180480957,
+      "step": 502
+    },
+    {
+      "epoch": 0.9137148047229791,
+      "grad_norm": 2.054955005645752,
+      "learning_rate": 8.888217540168139e-06,
+      "logits/chosen": 0.01604822278022766,
+      "logits/rejected": 0.1370609700679779,
+      "logps/chosen": -74.58015441894531,
+      "logps/rejected": -96.80826568603516,
+      "loss": 2.169,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.397087574005127,
+      "rewards/margins": 0.9368999004364014,
+      "rewards/rejected": -3.333987236022949,
+      "step": 503
+    },
+    {
+      "epoch": 0.9155313351498637,
+      "grad_norm": 2.179532527923584,
+      "learning_rate": 8.883985195544617e-06,
+      "logits/chosen": 0.03752445429563522,
+      "logits/rejected": 0.024428365752100945,
+      "logps/chosen": -83.81852722167969,
+      "logps/rejected": -88.1905288696289,
+      "loss": 2.5361,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.239609956741333,
+      "rewards/margins": 0.4188295900821686,
+      "rewards/rejected": -2.6584396362304688,
+      "step": 504
+    },
+    {
+      "epoch": 0.9173478655767484,
+      "grad_norm": 1.9171282052993774,
+      "learning_rate": 8.879745950386075e-06,
+      "logits/chosen": 0.17986616492271423,
+      "logits/rejected": 0.20335282385349274,
+      "logps/chosen": -72.679931640625,
+      "logps/rejected": -71.41309356689453,
+      "loss": 2.5452,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.261596918106079,
+      "rewards/margins": 0.3737923502922058,
+      "rewards/rejected": -2.6353893280029297,
+      "step": 505
+    },
+    {
+      "epoch": 0.9191643960036331,
+      "grad_norm": 2.3437929153442383,
+      "learning_rate": 8.87549981333707e-06,
+      "logits/chosen": 0.09010382741689682,
+      "logits/rejected": 0.040441811084747314,
+      "logps/chosen": -81.56139373779297,
+      "logps/rejected": -83.21439361572266,
+      "loss": 2.74,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2574033737182617,
+      "rewards/margins": 0.28719204664230347,
+      "rewards/rejected": -2.544595241546631,
+      "step": 506
+    },
+    {
+      "epoch": 0.9209809264305178,
+      "grad_norm": 2.4240329265594482,
+      "learning_rate": 8.871246793056215e-06,
+      "logits/chosen": 0.02888420596718788,
+      "logits/rejected": 0.11542786657810211,
+      "logps/chosen": -84.60144805908203,
+      "logps/rejected": -100.05682373046875,
+      "loss": 2.7052,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.863104820251465,
+      "rewards/margins": 0.41941094398498535,
+      "rewards/rejected": -3.28251576423645,
+      "step": 507
+    },
+    {
+      "epoch": 0.9227974568574023,
+      "grad_norm": 1.9247711896896362,
+      "learning_rate": 8.866986898216157e-06,
+      "logits/chosen": 0.06140238791704178,
+      "logits/rejected": 0.05623817816376686,
+      "logps/chosen": -77.40584564208984,
+      "logps/rejected": -86.71392822265625,
+      "loss": 2.2445,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4979231357574463,
+      "rewards/margins": 0.6161274313926697,
+      "rewards/rejected": -3.11405086517334,
+      "step": 508
+    },
+    {
+      "epoch": 0.924613987284287,
+      "grad_norm": 1.854956030845642,
+      "learning_rate": 8.862720137503568e-06,
+      "logits/chosen": 0.06755004823207855,
+      "logits/rejected": 0.03762562945485115,
+      "logps/chosen": -86.03059387207031,
+      "logps/rejected": -90.12970733642578,
+      "loss": 2.1459,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3017396926879883,
+      "rewards/margins": 0.6131142973899841,
+      "rewards/rejected": -2.914853572845459,
+      "step": 509
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.8044302463531494,
+      "learning_rate": 8.858446519619113e-06,
+      "logits/chosen": 0.08446178585290909,
+      "logits/rejected": 0.08818987011909485,
+      "logps/chosen": -77.58124542236328,
+      "logps/rejected": -85.02752685546875,
+      "loss": 2.1311,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2378087043762207,
+      "rewards/margins": 0.7067832350730896,
+      "rewards/rejected": -2.944591999053955,
+      "step": 510
+    },
+    {
+      "epoch": 0.9282470481380564,
+      "grad_norm": 1.4898866415023804,
+      "learning_rate": 8.854166053277443e-06,
+      "logits/chosen": 0.1039918065071106,
+      "logits/rejected": 0.1396111100912094,
+      "logps/chosen": -87.59841918945312,
+      "logps/rejected": -101.05659484863281,
+      "loss": 2.1092,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1167173385620117,
+      "rewards/margins": 0.7099149227142334,
+      "rewards/rejected": -2.826632499694824,
+      "step": 511
+    },
+    {
+      "epoch": 0.9300635785649409,
+      "grad_norm": 1.8009706735610962,
+      "learning_rate": 8.849878747207175e-06,
+      "logits/chosen": 0.11744043976068497,
+      "logits/rejected": 0.07621707767248154,
+      "logps/chosen": -78.37222290039062,
+      "logps/rejected": -74.32958221435547,
+      "loss": 2.6399,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.4002914428710938,
+      "rewards/margins": 0.21068021655082703,
+      "rewards/rejected": -2.610971450805664,
+      "step": 512
+    },
+    {
+      "epoch": 0.9318801089918256,
+      "grad_norm": 1.6368399858474731,
+      "learning_rate": 8.845584610150871e-06,
+      "logits/chosen": 0.09938757866621017,
+      "logits/rejected": 0.12675486505031586,
+      "logps/chosen": -82.24690246582031,
+      "logps/rejected": -92.60913848876953,
+      "loss": 1.9078,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.0825133323669434,
+      "rewards/margins": 0.8327144384384155,
+      "rewards/rejected": -2.9152278900146484,
+      "step": 513
+    },
+    {
+      "epoch": 0.9336966394187103,
+      "grad_norm": 1.8117483854293823,
+      "learning_rate": 8.841283650865027e-06,
+      "logits/chosen": 0.1355382651090622,
+      "logits/rejected": 0.14869986474514008,
+      "logps/chosen": -85.03866577148438,
+      "logps/rejected": -90.55638885498047,
+      "loss": 2.1243,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3315932750701904,
+      "rewards/margins": 0.6421911120414734,
+      "rewards/rejected": -2.9737846851348877,
+      "step": 514
+    },
+    {
+      "epoch": 0.935513169845595,
+      "grad_norm": 1.785927414894104,
+      "learning_rate": 8.836975878120046e-06,
+      "logits/chosen": 0.11557039618492126,
+      "logits/rejected": 0.11716655641794205,
+      "logps/chosen": -83.99951171875,
+      "logps/rejected": -92.36119079589844,
+      "loss": 2.4256,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.439833402633667,
+      "rewards/margins": 0.38893458247184753,
+      "rewards/rejected": -2.8287675380706787,
+      "step": 515
+    },
+    {
+      "epoch": 0.9373297002724795,
+      "grad_norm": 3.526102066040039,
+      "learning_rate": 8.832661300700228e-06,
+      "logits/chosen": 0.10778439044952393,
+      "logits/rejected": 0.18004637956619263,
+      "logps/chosen": -84.20158386230469,
+      "logps/rejected": -91.33497619628906,
+      "loss": 2.3568,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2032201290130615,
+      "rewards/margins": 0.5167797207832336,
+      "rewards/rejected": -2.7200000286102295,
+      "step": 516
+    },
+    {
+      "epoch": 0.9391462306993642,
+      "grad_norm": 1.7708752155303955,
+      "learning_rate": 8.828339927403745e-06,
+      "logits/chosen": 0.0862819030880928,
+      "logits/rejected": 0.16991934180259705,
+      "logps/chosen": -87.4437026977539,
+      "logps/rejected": -96.60844421386719,
+      "loss": 2.3041,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.230825185775757,
+      "rewards/margins": 0.5197271108627319,
+      "rewards/rejected": -2.7505524158477783,
+      "step": 517
+    },
+    {
+      "epoch": 0.9409627611262489,
+      "grad_norm": 2.1336302757263184,
+      "learning_rate": 8.824011767042631e-06,
+      "logits/chosen": 0.12070942670106888,
+      "logits/rejected": 0.19643370807170868,
+      "logps/chosen": -75.97718811035156,
+      "logps/rejected": -83.26056671142578,
+      "loss": 2.3468,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3768177032470703,
+      "rewards/margins": 0.5026894211769104,
+      "rewards/rejected": -2.879507064819336,
+      "step": 518
+    },
+    {
+      "epoch": 0.9427792915531336,
+      "grad_norm": 2.6361515522003174,
+      "learning_rate": 8.819676828442758e-06,
+      "logits/chosen": 0.04481849446892738,
+      "logits/rejected": -0.011357773095369339,
+      "logps/chosen": -84.09026336669922,
+      "logps/rejected": -87.71636199951172,
+      "loss": 2.8427,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.4249284267425537,
+      "rewards/margins": 0.41216185688972473,
+      "rewards/rejected": -2.837090015411377,
+      "step": 519
+    },
+    {
+      "epoch": 0.9445958219800181,
+      "grad_norm": 2.1068665981292725,
+      "learning_rate": 8.815335120443822e-06,
+      "logits/chosen": 0.1595858335494995,
+      "logits/rejected": 0.14017102122306824,
+      "logps/chosen": -94.00933837890625,
+      "logps/rejected": -96.91961669921875,
+      "loss": 2.6559,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.4731483459472656,
+      "rewards/margins": 0.3141450881958008,
+      "rewards/rejected": -2.7872931957244873,
+      "step": 520
+    },
+    {
+      "epoch": 0.9464123524069028,
+      "grad_norm": 1.7320570945739746,
+      "learning_rate": 8.810986651899322e-06,
+      "logits/chosen": 0.10371927917003632,
+      "logits/rejected": 0.17988254129886627,
+      "logps/chosen": -78.32308197021484,
+      "logps/rejected": -88.74859619140625,
+      "loss": 2.1866,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3973731994628906,
+      "rewards/margins": 0.7199669480323792,
+      "rewards/rejected": -3.117340087890625,
+      "step": 521
+    },
+    {
+      "epoch": 0.9482288828337875,
+      "grad_norm": 1.6995161771774292,
+      "learning_rate": 8.80663143167654e-06,
+      "logits/chosen": 0.14279219508171082,
+      "logits/rejected": 0.1058904379606247,
+      "logps/chosen": -87.79972839355469,
+      "logps/rejected": -94.35545349121094,
+      "loss": 2.1217,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.979785680770874,
+      "rewards/margins": 0.6210839152336121,
+      "rewards/rejected": -2.600869655609131,
+      "step": 522
+    },
+    {
+      "epoch": 0.9500454132606722,
+      "grad_norm": 2.1860764026641846,
+      "learning_rate": 8.80226946865653e-06,
+      "logits/chosen": 0.11601082980632782,
+      "logits/rejected": 0.15693408250808716,
+      "logps/chosen": -88.86433410644531,
+      "logps/rejected": -92.43590545654297,
+      "loss": 2.5629,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.2875983715057373,
+      "rewards/margins": 0.39402255415916443,
+      "rewards/rejected": -2.6816210746765137,
+      "step": 523
+    },
+    {
+      "epoch": 0.9518619436875567,
+      "grad_norm": 1.829032063484192,
+      "learning_rate": 8.797900771734094e-06,
+      "logits/chosen": 0.12328370660543442,
+      "logits/rejected": 0.042389824986457825,
+      "logps/chosen": -77.34864807128906,
+      "logps/rejected": -78.61405944824219,
+      "loss": 2.5327,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2514185905456543,
+      "rewards/margins": 0.3621898591518402,
+      "rewards/rejected": -2.6136088371276855,
+      "step": 524
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 2.235381841659546,
+      "learning_rate": 8.793525349817765e-06,
+      "logits/chosen": 0.12495981156826019,
+      "logits/rejected": 0.13291439414024353,
+      "logps/chosen": -76.7927474975586,
+      "logps/rejected": -86.1208267211914,
+      "loss": 2.5188,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3426284790039062,
+      "rewards/margins": 0.46077418327331543,
+      "rewards/rejected": -2.8034026622772217,
+      "step": 525
+    },
+    {
+      "epoch": 0.9554950045413261,
+      "grad_norm": 1.8071073293685913,
+      "learning_rate": 8.78914321182979e-06,
+      "logits/chosen": 0.13614074885845184,
+      "logits/rejected": 0.1179000735282898,
+      "logps/chosen": -74.44164276123047,
+      "logps/rejected": -77.69221496582031,
+      "loss": 2.6082,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0528411865234375,
+      "rewards/margins": 0.29493850469589233,
+      "rewards/rejected": -2.3477797508239746,
+      "step": 526
+    },
+    {
+      "epoch": 0.9573115349682108,
+      "grad_norm": 1.8623837232589722,
+      "learning_rate": 8.784754366706115e-06,
+      "logits/chosen": 0.02768833190202713,
+      "logits/rejected": 0.09584817290306091,
+      "logps/chosen": -77.13914489746094,
+      "logps/rejected": -86.53968048095703,
+      "loss": 2.2244,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.199755907058716,
+      "rewards/margins": 0.6271941661834717,
+      "rewards/rejected": -2.8269503116607666,
+      "step": 527
+    },
+    {
+      "epoch": 0.9591280653950953,
+      "grad_norm": 1.5299558639526367,
+      "learning_rate": 8.780358823396352e-06,
+      "logits/chosen": 0.11699292808771133,
+      "logits/rejected": 0.10016686469316483,
+      "logps/chosen": -86.1636962890625,
+      "logps/rejected": -87.16045379638672,
+      "loss": 2.0645,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.163857936859131,
+      "rewards/margins": 0.697452187538147,
+      "rewards/rejected": -2.8613100051879883,
+      "step": 528
+    },
+    {
+      "epoch": 0.96094459582198,
+      "grad_norm": 1.6236025094985962,
+      "learning_rate": 8.775956590863785e-06,
+      "logits/chosen": 0.11622033268213272,
+      "logits/rejected": 0.06629584729671478,
+      "logps/chosen": -76.42916870117188,
+      "logps/rejected": -80.24000549316406,
+      "loss": 1.8904,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.102374315261841,
+      "rewards/margins": 0.7230857014656067,
+      "rewards/rejected": -2.825460195541382,
+      "step": 529
+    },
+    {
+      "epoch": 0.9627611262488647,
+      "grad_norm": 1.6119427680969238,
+      "learning_rate": 8.771547678085332e-06,
+      "logits/chosen": 0.07009143382310867,
+      "logits/rejected": 0.16355563700199127,
+      "logps/chosen": -72.83528137207031,
+      "logps/rejected": -87.72441101074219,
+      "loss": 2.0136,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1697566509246826,
+      "rewards/margins": 0.8312336206436157,
+      "rewards/rejected": -3.000990390777588,
+      "step": 530
+    },
+    {
+      "epoch": 0.9645776566757494,
+      "grad_norm": 2.0324013233184814,
+      "learning_rate": 8.767132094051534e-06,
+      "logits/chosen": 0.14837129414081573,
+      "logits/rejected": 0.0700330063700676,
+      "logps/chosen": -84.65350341796875,
+      "logps/rejected": -91.84485626220703,
+      "loss": 2.2551,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5343966484069824,
+      "rewards/margins": 0.6457280516624451,
+      "rewards/rejected": -3.180124521255493,
+      "step": 531
+    },
+    {
+      "epoch": 0.9663941871026339,
+      "grad_norm": 1.4526044130325317,
+      "learning_rate": 8.762709847766532e-06,
+      "logits/chosen": 0.013620391488075256,
+      "logits/rejected": 0.10376289486885071,
+      "logps/chosen": -70.80522155761719,
+      "logps/rejected": -84.26114654541016,
+      "loss": 1.8607,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.249685287475586,
+      "rewards/margins": 0.8104004859924316,
+      "rewards/rejected": -3.0600855350494385,
+      "step": 532
+    },
+    {
+      "epoch": 0.9682107175295186,
+      "grad_norm": 2.0312576293945312,
+      "learning_rate": 8.758280948248059e-06,
+      "logits/chosen": 0.0965358167886734,
+      "logits/rejected": 0.10093791037797928,
+      "logps/chosen": -81.83265686035156,
+      "logps/rejected": -88.52181243896484,
+      "loss": 2.6128,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4901068210601807,
+      "rewards/margins": 0.30889907479286194,
+      "rewards/rejected": -2.7990057468414307,
+      "step": 533
+    },
+    {
+      "epoch": 0.9700272479564033,
+      "grad_norm": 2.547189712524414,
+      "learning_rate": 8.753845404527413e-06,
+      "logits/chosen": 0.020555848255753517,
+      "logits/rejected": 0.0945032387971878,
+      "logps/chosen": -72.16207122802734,
+      "logps/rejected": -86.27986907958984,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.148913860321045,
+      "rewards/margins": 0.7298768162727356,
+      "rewards/rejected": -2.8787906169891357,
+      "step": 534
+    },
+    {
+      "epoch": 0.971843778383288,
+      "grad_norm": 1.996645450592041,
+      "learning_rate": 8.74940322564944e-06,
+      "logits/chosen": 0.0345986932516098,
+      "logits/rejected": 0.08187264204025269,
+      "logps/chosen": -78.76763916015625,
+      "logps/rejected": -91.45555114746094,
+      "loss": 2.4587,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.4772286415100098,
+      "rewards/margins": 0.4609605073928833,
+      "rewards/rejected": -2.9381890296936035,
+      "step": 535
+    },
+    {
+      "epoch": 0.9736603088101726,
+      "grad_norm": 1.8145204782485962,
+      "learning_rate": 8.744954420672514e-06,
+      "logits/chosen": 0.03380711376667023,
+      "logits/rejected": 0.09039584547281265,
+      "logps/chosen": -79.83094787597656,
+      "logps/rejected": -90.87137603759766,
+      "loss": 2.3589,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.1433308124542236,
+      "rewards/margins": 0.5286959409713745,
+      "rewards/rejected": -2.6720268726348877,
+      "step": 536
+    },
+    {
+      "epoch": 0.9754768392370572,
+      "grad_norm": 1.9745460748672485,
+      "learning_rate": 8.740498998668523e-06,
+      "logits/chosen": 0.07233883440494537,
+      "logits/rejected": 0.17097605764865875,
+      "logps/chosen": -72.3458251953125,
+      "logps/rejected": -83.47319793701172,
+      "loss": 2.3469,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199948310852051,
+      "rewards/margins": 0.47777751088142395,
+      "rewards/rejected": -2.6777257919311523,
+      "step": 537
+    },
+    {
+      "epoch": 0.9772933696639419,
+      "grad_norm": 1.665255069732666,
+      "learning_rate": 8.736036968722851e-06,
+      "logits/chosen": 0.09596951305866241,
+      "logits/rejected": 0.06625551730394363,
+      "logps/chosen": -80.11689758300781,
+      "logps/rejected": -88.4573974609375,
+      "loss": 2.1889,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1831865310668945,
+      "rewards/margins": 0.6106584668159485,
+      "rewards/rejected": -2.7938451766967773,
+      "step": 538
+    },
+    {
+      "epoch": 0.9791099000908265,
+      "grad_norm": 1.5587108135223389,
+      "learning_rate": 8.73156833993435e-06,
+      "logits/chosen": 0.07950340211391449,
+      "logits/rejected": 0.0903201699256897,
+      "logps/chosen": -78.9182357788086,
+      "logps/rejected": -92.290283203125,
+      "loss": 2.048,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.12032413482666,
+      "rewards/margins": 0.7009568214416504,
+      "rewards/rejected": -2.8212809562683105,
+      "step": 539
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 1.7177670001983643,
+      "learning_rate": 8.727093121415338e-06,
+      "logits/chosen": 0.06052964925765991,
+      "logits/rejected": 0.14043518900871277,
+      "logps/chosen": -79.04943084716797,
+      "logps/rejected": -88.881591796875,
+      "loss": 2.2089,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2680561542510986,
+      "rewards/margins": 0.5380589962005615,
+      "rewards/rejected": -2.806114912033081,
+      "step": 540
+    },
+    {
+      "epoch": 0.9827429609445958,
+      "grad_norm": 1.7630436420440674,
+      "learning_rate": 8.722611322291558e-06,
+      "logits/chosen": 0.02032918483018875,
+      "logits/rejected": 0.11728623509407043,
+      "logps/chosen": -78.1409912109375,
+      "logps/rejected": -89.3685073852539,
+      "loss": 2.3303,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.356957197189331,
+      "rewards/margins": 0.5139226317405701,
+      "rewards/rejected": -2.870880126953125,
+      "step": 541
+    },
+    {
+      "epoch": 0.9845594913714805,
+      "grad_norm": 1.852418065071106,
+      "learning_rate": 8.718122951702183e-06,
+      "logits/chosen": 0.0439969077706337,
+      "logits/rejected": 0.16671347618103027,
+      "logps/chosen": -69.96346282958984,
+      "logps/rejected": -85.24410247802734,
+      "loss": 2.1393,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3295838832855225,
+      "rewards/margins": 0.6544574499130249,
+      "rewards/rejected": -2.984041213989258,
+      "step": 542
+    },
+    {
+      "epoch": 0.9863760217983651,
+      "grad_norm": 1.9028044939041138,
+      "learning_rate": 8.713628018799782e-06,
+      "logits/chosen": 0.04875154793262482,
+      "logits/rejected": -0.009283583611249924,
+      "logps/chosen": -75.80145263671875,
+      "logps/rejected": -78.19249725341797,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.241586208343506,
+      "rewards/margins": 0.4672737419605255,
+      "rewards/rejected": -2.708860158920288,
+      "step": 543
+    },
+    {
+      "epoch": 0.9881925522252498,
+      "grad_norm": 1.5648458003997803,
+      "learning_rate": 8.709126532750304e-06,
+      "logits/chosen": 0.02982347458600998,
+      "logits/rejected": 0.07835812866687775,
+      "logps/chosen": -75.35140991210938,
+      "logps/rejected": -91.11531829833984,
+      "loss": 2.0945,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.062386989593506,
+      "rewards/margins": 0.6832989454269409,
+      "rewards/rejected": -2.7456860542297363,
+      "step": 544
+    },
+    {
+      "epoch": 0.9900090826521344,
+      "grad_norm": 1.6214131116867065,
+      "learning_rate": 8.704618502733064e-06,
+      "logits/chosen": 0.09551770985126495,
+      "logits/rejected": 0.11310344189405441,
+      "logps/chosen": -80.15961456298828,
+      "logps/rejected": -93.38126373291016,
+      "loss": 2.1642,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.255985736846924,
+      "rewards/margins": 0.7712126970291138,
+      "rewards/rejected": -3.027198553085327,
+      "step": 545
+    },
+    {
+      "epoch": 0.9918256130790191,
+      "grad_norm": 1.762953519821167,
+      "learning_rate": 8.70010393794072e-06,
+      "logits/chosen": 0.0971008688211441,
+      "logits/rejected": 0.16590046882629395,
+      "logps/chosen": -86.24911499023438,
+      "logps/rejected": -102.59831237792969,
+      "loss": 2.0951,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3542394638061523,
+      "rewards/margins": 0.7857018709182739,
+      "rewards/rejected": -3.139941453933716,
+      "step": 546
+    },
+    {
+      "epoch": 0.9936421435059037,
+      "grad_norm": 1.6033780574798584,
+      "learning_rate": 8.695582847579254e-06,
+      "logits/chosen": 0.15147003531455994,
+      "logits/rejected": 0.13297875225543976,
+      "logps/chosen": -78.5172348022461,
+      "logps/rejected": -86.49285888671875,
+      "loss": 2.1101,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2025821208953857,
+      "rewards/margins": 0.7471145391464233,
+      "rewards/rejected": -2.9496963024139404,
+      "step": 547
+    },
+    {
+      "epoch": 0.9954586739327884,
+      "grad_norm": 1.4905331134796143,
+      "learning_rate": 8.69105524086796e-06,
+      "logits/chosen": 0.07559307664632797,
+      "logits/rejected": 0.09286531805992126,
+      "logps/chosen": -74.92215728759766,
+      "logps/rejected": -78.23614501953125,
+      "loss": 2.2745,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.24694561958313,
+      "rewards/margins": 0.5010396838188171,
+      "rewards/rejected": -2.747985363006592,
+      "step": 548
+    },
+    {
+      "epoch": 0.997275204359673,
+      "grad_norm": 1.823449730873108,
+      "learning_rate": 8.686521127039414e-06,
+      "logits/chosen": 0.03763032704591751,
+      "logits/rejected": 0.06642220914363861,
+      "logps/chosen": -79.37518310546875,
+      "logps/rejected": -87.17401885986328,
+      "loss": 2.5954,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158803701400757,
+      "rewards/margins": 0.4884320795536041,
+      "rewards/rejected": -2.647235870361328,
+      "step": 549
+    },
+    {
+      "epoch": 0.9990917347865577,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.681980515339464e-06,
+      "logits/chosen": 0.08186414837837219,
+      "logits/rejected": 0.08158639073371887,
+      "logps/chosen": -84.89871215820312,
+      "logps/rejected": -85.87324523925781,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.1884236335754395,
+      "rewards/margins": 0.2923518121242523,
+      "rewards/rejected": -2.4807753562927246,
+      "step": 550
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.677433415027209e-06,
+      "logits/chosen": 0.18737269937992096,
+      "logits/rejected": 0.11028257012367249,
+      "logps/chosen": -76.3554916381836,
+      "logps/rejected": -86.39812469482422,
+      "loss": 1.1517,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1043050289154053,
+      "rewards/margins": 0.7949569225311279,
+      "rewards/rejected": -2.8992621898651123,
+      "step": 551
+    },
+    {
+      "epoch": 1.0018165304268847,
+      "grad_norm": 1.6554900407791138,
+      "learning_rate": 8.672879835374976e-06,
+      "logits/chosen": 0.11259499937295914,
+      "logits/rejected": 0.07095417380332947,
+      "logps/chosen": -84.55628967285156,
+      "logps/rejected": -88.55287170410156,
+      "loss": 2.1275,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.2375316619873047,
+      "rewards/margins": 0.5453992486000061,
+      "rewards/rejected": -2.782930850982666,
+      "step": 552
+    },
+    {
+      "epoch": 1.0036330608537694,
+      "grad_norm": 1.6807721853256226,
+      "learning_rate": 8.668319785668307e-06,
+      "logits/chosen": 0.14847862720489502,
+      "logits/rejected": 0.1863957643508911,
+      "logps/chosen": -87.27869415283203,
+      "logps/rejected": -89.04036712646484,
+      "loss": 2.3209,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.087535858154297,
+      "rewards/margins": 0.5243443846702576,
+      "rewards/rejected": -2.61188006401062,
+      "step": 553
+    },
+    {
+      "epoch": 1.005449591280654,
+      "grad_norm": 1.8480761051177979,
+      "learning_rate": 8.66375327520594e-06,
+      "logits/chosen": 0.07236558198928833,
+      "logits/rejected": 0.13397395610809326,
+      "logps/chosen": -85.43444061279297,
+      "logps/rejected": -101.47248840332031,
+      "loss": 1.7106,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.157266616821289,
+      "rewards/margins": 0.9501473903656006,
+      "rewards/rejected": -3.1074140071868896,
+      "step": 554
+    },
+    {
+      "epoch": 1.0072661217075387,
+      "grad_norm": 1.586917757987976,
+      "learning_rate": 8.659180313299783e-06,
+      "logits/chosen": 0.05113527923822403,
+      "logits/rejected": 0.14182163774967194,
+      "logps/chosen": -78.31108856201172,
+      "logps/rejected": -90.46333312988281,
+      "loss": 1.8965,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.945315957069397,
+      "rewards/margins": 0.8104385137557983,
+      "rewards/rejected": -2.755754232406616,
+      "step": 555
+    },
+    {
+      "epoch": 1.0090826521344232,
+      "grad_norm": 1.406339406967163,
+      "learning_rate": 8.6546009092749e-06,
+      "logits/chosen": -0.01422208547592163,
+      "logits/rejected": 0.10859756916761398,
+      "logps/chosen": -75.02066040039062,
+      "logps/rejected": -87.69715118408203,
+      "loss": 2.027,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9996552467346191,
+      "rewards/margins": 0.7365851402282715,
+      "rewards/rejected": -2.7362406253814697,
+      "step": 556
+    },
+    {
+      "epoch": 1.0108991825613078,
+      "grad_norm": 1.5511010885238647,
+      "learning_rate": 8.650015072469496e-06,
+      "logits/chosen": 0.07154582440853119,
+      "logits/rejected": 0.11479673534631729,
+      "logps/chosen": -88.33872985839844,
+      "logps/rejected": -96.75527954101562,
+      "loss": 1.966,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2190096378326416,
+      "rewards/margins": 0.861441969871521,
+      "rewards/rejected": -3.080451250076294,
+      "step": 557
+    },
+    {
+      "epoch": 1.0127157129881925,
+      "grad_norm": 1.5951570272445679,
+      "learning_rate": 8.645422812234888e-06,
+      "logits/chosen": 0.17342886328697205,
+      "logits/rejected": 0.11909954994916916,
+      "logps/chosen": -79.63985443115234,
+      "logps/rejected": -83.41281127929688,
+      "loss": 2.1115,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3469128608703613,
+      "rewards/margins": 0.7317838668823242,
+      "rewards/rejected": -3.0786967277526855,
+      "step": 558
+    },
+    {
+      "epoch": 1.0145322434150772,
+      "grad_norm": 1.891517162322998,
+      "learning_rate": 8.640824137935498e-06,
+      "logits/chosen": 0.14266839623451233,
+      "logits/rejected": 0.14209085702896118,
+      "logps/chosen": -84.77491760253906,
+      "logps/rejected": -87.46759796142578,
+      "loss": 2.2691,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.022730827331543,
+      "rewards/margins": 0.42655885219573975,
+      "rewards/rejected": -2.449289560317993,
+      "step": 559
+    },
+    {
+      "epoch": 1.0163487738419619,
+      "grad_norm": 1.7919303178787231,
+      "learning_rate": 8.636219058948823e-06,
+      "logits/chosen": 0.01973957009613514,
+      "logits/rejected": 0.12187746167182922,
+      "logps/chosen": -78.26383209228516,
+      "logps/rejected": -92.714599609375,
+      "loss": 1.974,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2774088382720947,
+      "rewards/margins": 0.8670985698699951,
+      "rewards/rejected": -3.1445071697235107,
+      "step": 560
+    },
+    {
+      "epoch": 1.0181653042688465,
+      "grad_norm": 1.5417025089263916,
+      "learning_rate": 8.631607584665415e-06,
+      "logits/chosen": 0.07740732282400131,
+      "logits/rejected": 0.17052005231380463,
+      "logps/chosen": -83.54193115234375,
+      "logps/rejected": -95.1084976196289,
+      "loss": 1.9963,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1598939895629883,
+      "rewards/margins": 0.802270233631134,
+      "rewards/rejected": -2.9621646404266357,
+      "step": 561
+    },
+    {
+      "epoch": 1.0199818346957312,
+      "grad_norm": 1.9338072538375854,
+      "learning_rate": 8.62698972448888e-06,
+      "logits/chosen": 0.09191437065601349,
+      "logits/rejected": 0.1776769608259201,
+      "logps/chosen": -82.68082427978516,
+      "logps/rejected": -96.89566802978516,
+      "loss": 2.0399,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3505523204803467,
+      "rewards/margins": 0.7893310785293579,
+      "rewards/rejected": -3.1398837566375732,
+      "step": 562
+    },
+    {
+      "epoch": 1.021798365122616,
+      "grad_norm": 1.609601378440857,
+      "learning_rate": 8.622365487835836e-06,
+      "logits/chosen": 0.13105043768882751,
+      "logits/rejected": 0.138211190700531,
+      "logps/chosen": -77.19727325439453,
+      "logps/rejected": -84.1854476928711,
+      "loss": 2.3184,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.13655948638916,
+      "rewards/margins": 0.5744473934173584,
+      "rewards/rejected": -2.7110068798065186,
+      "step": 563
+    },
+    {
+      "epoch": 1.0236148955495004,
+      "grad_norm": 2.1222517490386963,
+      "learning_rate": 8.617734884135906e-06,
+      "logits/chosen": 0.14744582772254944,
+      "logits/rejected": 0.053739145398139954,
+      "logps/chosen": -81.11271667480469,
+      "logps/rejected": -81.40655517578125,
+      "loss": 2.2,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.419705390930176,
+      "rewards/margins": 0.7198007702827454,
+      "rewards/rejected": -3.1395058631896973,
+      "step": 564
+    },
+    {
+      "epoch": 1.025431425976385,
+      "grad_norm": 1.8075231313705444,
+      "learning_rate": 8.613097922831699e-06,
+      "logits/chosen": 0.07210052758455276,
+      "logits/rejected": 0.07022904604673386,
+      "logps/chosen": -88.72612762451172,
+      "logps/rejected": -98.2815933227539,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1057255268096924,
+      "rewards/margins": 0.8747395873069763,
+      "rewards/rejected": -2.9804654121398926,
+      "step": 565
+    },
+    {
+      "epoch": 1.0272479564032697,
+      "grad_norm": 1.3381479978561401,
+      "learning_rate": 8.608454613378783e-06,
+      "logits/chosen": 0.08187520503997803,
+      "logits/rejected": 0.1098201796412468,
+      "logps/chosen": -77.1299819946289,
+      "logps/rejected": -89.75099182128906,
+      "loss": 1.8687,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3585519790649414,
+      "rewards/margins": 0.8574986457824707,
+      "rewards/rejected": -3.216050386428833,
+      "step": 566
+    },
+    {
+      "epoch": 1.0290644868301544,
+      "grad_norm": 1.6325740814208984,
+      "learning_rate": 8.603804965245678e-06,
+      "logits/chosen": 0.12077341973781586,
+      "logits/rejected": 0.09103713929653168,
+      "logps/chosen": -74.03280639648438,
+      "logps/rejected": -79.61197662353516,
+      "loss": 2.0884,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.430152654647827,
+      "rewards/margins": 0.6819649934768677,
+      "rewards/rejected": -3.1121177673339844,
+      "step": 567
+    },
+    {
+      "epoch": 1.030881017257039,
+      "grad_norm": 1.782224178314209,
+      "learning_rate": 8.599148987913829e-06,
+      "logits/chosen": -0.008054995909333229,
+      "logits/rejected": 0.08848709613084793,
+      "logps/chosen": -88.39682006835938,
+      "logps/rejected": -98.20311737060547,
+      "loss": 2.3203,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.21012282371521,
+      "rewards/margins": 0.5451631546020508,
+      "rewards/rejected": -2.75528621673584,
+      "step": 568
+    },
+    {
+      "epoch": 1.0326975476839237,
+      "grad_norm": 1.7691792249679565,
+      "learning_rate": 8.594486690877577e-06,
+      "logits/chosen": 0.05038870871067047,
+      "logits/rejected": -0.0020756814628839493,
+      "logps/chosen": -88.35281372070312,
+      "logps/rejected": -87.42422485351562,
+      "loss": 2.1789,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6035892963409424,
+      "rewards/margins": 0.559741735458374,
+      "rewards/rejected": -3.1633315086364746,
+      "step": 569
+    },
+    {
+      "epoch": 1.0345140781108084,
+      "grad_norm": 1.7002201080322266,
+      "learning_rate": 8.589818083644161e-06,
+      "logits/chosen": 0.030099213123321533,
+      "logits/rejected": 0.10112117975950241,
+      "logps/chosen": -74.28582763671875,
+      "logps/rejected": -85.93130493164062,
+      "loss": 2.2068,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.324796676635742,
+      "rewards/margins": 0.5959895849227905,
+      "rewards/rejected": -2.9207863807678223,
+      "step": 570
+    },
+    {
+      "epoch": 1.036330608537693,
+      "grad_norm": 1.5611207485198975,
+      "learning_rate": 8.585143175733686e-06,
+      "logits/chosen": 0.09465142339468002,
+      "logits/rejected": 0.13904502987861633,
+      "logps/chosen": -86.5047378540039,
+      "logps/rejected": -95.09537506103516,
+      "loss": 1.9628,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5564661026000977,
+      "rewards/margins": 0.8365517258644104,
+      "rewards/rejected": -3.3930182456970215,
+      "step": 571
+    },
+    {
+      "epoch": 1.0381471389645776,
+      "grad_norm": 1.7271509170532227,
+      "learning_rate": 8.5804619766791e-06,
+      "logits/chosen": 0.08665300905704498,
+      "logits/rejected": 0.10898350924253464,
+      "logps/chosen": -85.17437744140625,
+      "logps/rejected": -94.50398254394531,
+      "loss": 2.2717,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3690195083618164,
+      "rewards/margins": 0.6290315389633179,
+      "rewards/rejected": -2.9980509281158447,
+      "step": 572
+    },
+    {
+      "epoch": 1.0399636693914622,
+      "grad_norm": 1.7832714319229126,
+      "learning_rate": 8.575774496026184e-06,
+      "logits/chosen": 0.033456090837717056,
+      "logits/rejected": 0.09503711760044098,
+      "logps/chosen": -87.27224731445312,
+      "logps/rejected": -94.98229217529297,
+      "loss": 2.1781,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.509322166442871,
+      "rewards/margins": 0.5943110585212708,
+      "rewards/rejected": -3.103632926940918,
+      "step": 573
+    },
+    {
+      "epoch": 1.041780199818347,
+      "grad_norm": 1.8239610195159912,
+      "learning_rate": 8.571080743333528e-06,
+      "logits/chosen": 0.11436322331428528,
+      "logits/rejected": 0.06905262917280197,
+      "logps/chosen": -81.12911987304688,
+      "logps/rejected": -88.73529052734375,
+      "loss": 2.1358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2652292251586914,
+      "rewards/margins": 0.6765917539596558,
+      "rewards/rejected": -2.9418206214904785,
+      "step": 574
+    },
+    {
+      "epoch": 1.0435967302452316,
+      "grad_norm": 1.9338902235031128,
+      "learning_rate": 8.566380728172512e-06,
+      "logits/chosen": 0.11198662221431732,
+      "logits/rejected": 0.11845988035202026,
+      "logps/chosen": -82.55985260009766,
+      "logps/rejected": -94.05873107910156,
+      "loss": 1.9609,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3074758052825928,
+      "rewards/margins": 0.8309231400489807,
+      "rewards/rejected": -3.1383988857269287,
+      "step": 575
+    },
+    {
+      "epoch": 1.0454132606721163,
+      "grad_norm": 1.6932177543640137,
+      "learning_rate": 8.561674460127281e-06,
+      "logits/chosen": 0.0954650342464447,
+      "logits/rejected": 0.139107346534729,
+      "logps/chosen": -83.95409393310547,
+      "logps/rejected": -98.31000518798828,
+      "loss": 2.311,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.476616859436035,
+      "rewards/margins": 0.7193690538406372,
+      "rewards/rejected": -3.195985794067383,
+      "step": 576
+    },
+    {
+      "epoch": 1.047229791099001,
+      "grad_norm": 1.6585469245910645,
+      "learning_rate": 8.556961948794738e-06,
+      "logits/chosen": 0.050676412880420685,
+      "logits/rejected": 0.13941077888011932,
+      "logps/chosen": -91.46603393554688,
+      "logps/rejected": -111.38130187988281,
+      "loss": 1.9574,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.630396604537964,
+      "rewards/margins": 0.8525002002716064,
+      "rewards/rejected": -3.4828968048095703,
+      "step": 577
+    },
+    {
+      "epoch": 1.0490463215258856,
+      "grad_norm": 1.8747247457504272,
+      "learning_rate": 8.552243203784514e-06,
+      "logits/chosen": 0.06109774485230446,
+      "logits/rejected": 0.03656052052974701,
+      "logps/chosen": -84.25686645507812,
+      "logps/rejected": -99.20081329345703,
+      "loss": 2.0244,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4541397094726562,
+      "rewards/margins": 1.0175232887268066,
+      "rewards/rejected": -3.471662759780884,
+      "step": 578
+    },
+    {
+      "epoch": 1.0508628519527703,
+      "grad_norm": 1.714175820350647,
+      "learning_rate": 8.547518234718954e-06,
+      "logits/chosen": 0.07959345728158951,
+      "logits/rejected": 0.17609870433807373,
+      "logps/chosen": -79.58773040771484,
+      "logps/rejected": -98.85990142822266,
+      "loss": 1.7738,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.431654930114746,
+      "rewards/margins": 1.0072871446609497,
+      "rewards/rejected": -3.4389421939849854,
+      "step": 579
+    },
+    {
+      "epoch": 1.052679382379655,
+      "grad_norm": 1.6469632387161255,
+      "learning_rate": 8.542787051233088e-06,
+      "logits/chosen": 0.06128916144371033,
+      "logits/rejected": 0.11901555210351944,
+      "logps/chosen": -90.99638366699219,
+      "logps/rejected": -104.79571533203125,
+      "loss": 1.8789,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3995859622955322,
+      "rewards/margins": 0.9405049085617065,
+      "rewards/rejected": -3.340090751647949,
+      "step": 580
+    },
+    {
+      "epoch": 1.0544959128065394,
+      "grad_norm": 1.8093998432159424,
+      "learning_rate": 8.538049662974625e-06,
+      "logits/chosen": 0.136922687292099,
+      "logits/rejected": 0.06489068269729614,
+      "logps/chosen": -84.12126159667969,
+      "logps/rejected": -85.51757049560547,
+      "loss": 2.2372,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7452549934387207,
+      "rewards/margins": 0.6496641635894775,
+      "rewards/rejected": -3.394918918609619,
+      "step": 581
+    },
+    {
+      "epoch": 1.056312443233424,
+      "grad_norm": 1.761516809463501,
+      "learning_rate": 8.533306079603928e-06,
+      "logits/chosen": 0.02220803126692772,
+      "logits/rejected": 0.03564752638339996,
+      "logps/chosen": -91.6943359375,
+      "logps/rejected": -105.33677673339844,
+      "loss": 1.831,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.1495776176452637,
+      "rewards/margins": 0.8695197701454163,
+      "rewards/rejected": -3.019097328186035,
+      "step": 582
+    },
+    {
+      "epoch": 1.0581289736603088,
+      "grad_norm": 1.8054577112197876,
+      "learning_rate": 8.52855631079398e-06,
+      "logits/chosen": 0.044406406581401825,
+      "logits/rejected": 0.13649103045463562,
+      "logps/chosen": -83.07669830322266,
+      "logps/rejected": -99.42033386230469,
+      "loss": 2.1295,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.674705743789673,
+      "rewards/margins": 0.8526960015296936,
+      "rewards/rejected": -3.52740216255188,
+      "step": 583
+    },
+    {
+      "epoch": 1.0599455040871935,
+      "grad_norm": 2.092067241668701,
+      "learning_rate": 8.523800366230397e-06,
+      "logits/chosen": 0.11083705723285675,
+      "logits/rejected": 0.13336583971977234,
+      "logps/chosen": -82.55166625976562,
+      "logps/rejected": -88.78050994873047,
+      "loss": 2.783,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.7093071937561035,
+      "rewards/margins": 0.34429922699928284,
+      "rewards/rejected": -3.0536062717437744,
+      "step": 584
+    },
+    {
+      "epoch": 1.0617620345140781,
+      "grad_norm": 2.0747568607330322,
+      "learning_rate": 8.519038255611372e-06,
+      "logits/chosen": 0.06060079485177994,
+      "logits/rejected": 0.06341060996055603,
+      "logps/chosen": -82.93423461914062,
+      "logps/rejected": -85.8175048828125,
+      "loss": 2.1046,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6574935913085938,
+      "rewards/margins": 0.7763374447822571,
+      "rewards/rejected": -3.433830738067627,
+      "step": 585
+    },
+    {
+      "epoch": 1.0635785649409628,
+      "grad_norm": 2.1435482501983643,
+      "learning_rate": 8.51426998864768e-06,
+      "logits/chosen": 0.044051673263311386,
+      "logits/rejected": 0.08015372604131699,
+      "logps/chosen": -86.49735260009766,
+      "logps/rejected": -97.29399108886719,
+      "loss": 2.1365,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.620311737060547,
+      "rewards/margins": 0.6845528483390808,
+      "rewards/rejected": -3.3048648834228516,
+      "step": 586
+    },
+    {
+      "epoch": 1.0653950953678475,
+      "grad_norm": 1.65945303440094,
+      "learning_rate": 8.509495575062647e-06,
+      "logits/chosen": 0.15405744314193726,
+      "logits/rejected": 0.12232419848442078,
+      "logps/chosen": -73.09071350097656,
+      "logps/rejected": -74.24645233154297,
+      "loss": 2.1912,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3465700149536133,
+      "rewards/margins": 0.6234737634658813,
+      "rewards/rejected": -2.970043420791626,
+      "step": 587
+    },
+    {
+      "epoch": 1.067211625794732,
+      "grad_norm": 2.166276216506958,
+      "learning_rate": 8.504715024592132e-06,
+      "logits/chosen": -0.05482687056064606,
+      "logits/rejected": -0.003489813767373562,
+      "logps/chosen": -84.82199096679688,
+      "logps/rejected": -97.53748321533203,
+      "loss": 2.1753,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5944690704345703,
+      "rewards/margins": 0.6753374934196472,
+      "rewards/rejected": -3.269806385040283,
+      "step": 588
+    },
+    {
+      "epoch": 1.0690281562216166,
+      "grad_norm": 1.8869906663894653,
+      "learning_rate": 8.499928346984512e-06,
+      "logits/chosen": 0.13489654660224915,
+      "logits/rejected": 0.052815720438957214,
+      "logps/chosen": -90.98455047607422,
+      "logps/rejected": -91.64305877685547,
+      "loss": 2.3005,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.8705921173095703,
+      "rewards/margins": 0.555056095123291,
+      "rewards/rejected": -3.4256482124328613,
+      "step": 589
+    },
+    {
+      "epoch": 1.0708446866485013,
+      "grad_norm": 2.286853790283203,
+      "learning_rate": 8.495135552000657e-06,
+      "logits/chosen": -0.0365738607943058,
+      "logits/rejected": 0.09564587473869324,
+      "logps/chosen": -93.99796295166016,
+      "logps/rejected": -111.09819793701172,
+      "loss": 2.1574,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.484596014022827,
+      "rewards/margins": 0.840925395488739,
+      "rewards/rejected": -3.325521469116211,
+      "step": 590
+    },
+    {
+      "epoch": 1.072661217075386,
+      "grad_norm": 1.7640435695648193,
+      "learning_rate": 8.490336649413909e-06,
+      "logits/chosen": 0.06625102460384369,
+      "logits/rejected": 0.14394444227218628,
+      "logps/chosen": -74.65644836425781,
+      "logps/rejected": -92.41899871826172,
+      "loss": 2.0387,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.453733205795288,
+      "rewards/margins": 0.7862347364425659,
+      "rewards/rejected": -3.2399678230285645,
+      "step": 591
+    },
+    {
+      "epoch": 1.0744777475022707,
+      "grad_norm": 1.5737273693084717,
+      "learning_rate": 8.485531649010063e-06,
+      "logits/chosen": -0.0012233639135956764,
+      "logits/rejected": 0.021886199712753296,
+      "logps/chosen": -79.36224365234375,
+      "logps/rejected": -94.19145965576172,
+      "loss": 2.3886,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.6413750648498535,
+      "rewards/margins": 0.7800988554954529,
+      "rewards/rejected": -3.421473979949951,
+      "step": 592
+    },
+    {
+      "epoch": 1.0762942779291553,
+      "grad_norm": 2.4172728061676025,
+      "learning_rate": 8.480720560587356e-06,
+      "logits/chosen": 0.04091513529419899,
+      "logits/rejected": 0.16047289967536926,
+      "logps/chosen": -76.38783264160156,
+      "logps/rejected": -94.16845703125,
+      "loss": 2.2219,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.719104290008545,
+      "rewards/margins": 0.9459499716758728,
+      "rewards/rejected": -3.6650545597076416,
+      "step": 593
+    },
+    {
+      "epoch": 1.07811080835604,
+      "grad_norm": 1.6836284399032593,
+      "learning_rate": 8.475903393956434e-06,
+      "logits/chosen": 0.07094185054302216,
+      "logits/rejected": 0.09420361369848251,
+      "logps/chosen": -82.47905731201172,
+      "logps/rejected": -90.06558990478516,
+      "loss": 2.1268,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3819637298583984,
+      "rewards/margins": 0.5947781801223755,
+      "rewards/rejected": -2.9767420291900635,
+      "step": 594
+    },
+    {
+      "epoch": 1.0799273387829247,
+      "grad_norm": 1.808720588684082,
+      "learning_rate": 8.471080158940336e-06,
+      "logits/chosen": 0.010593242943286896,
+      "logits/rejected": 0.03635484725236893,
+      "logps/chosen": -82.94917297363281,
+      "logps/rejected": -94.52174377441406,
+      "loss": 1.9352,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.120471239089966,
+      "rewards/margins": 0.8991338014602661,
+      "rewards/rejected": -3.0196051597595215,
+      "step": 595
+    },
+    {
+      "epoch": 1.0817438692098094,
+      "grad_norm": 1.809515118598938,
+      "learning_rate": 8.46625086537448e-06,
+      "logits/chosen": 0.0202183797955513,
+      "logits/rejected": 0.1138802096247673,
+      "logps/chosen": -84.1740951538086,
+      "logps/rejected": -99.66446685791016,
+      "loss": 2.1384,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7415525913238525,
+      "rewards/margins": 0.8209500908851624,
+      "rewards/rejected": -3.562502861022949,
+      "step": 596
+    },
+    {
+      "epoch": 1.0835603996366938,
+      "grad_norm": 2.023786783218384,
+      "learning_rate": 8.46141552310664e-06,
+      "logits/chosen": 0.11318185180425644,
+      "logits/rejected": 0.1275079846382141,
+      "logps/chosen": -78.02458953857422,
+      "logps/rejected": -89.36206817626953,
+      "loss": 1.9695,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5616073608398438,
+      "rewards/margins": 0.7586690783500671,
+      "rewards/rejected": -3.3202764987945557,
+      "step": 597
+    },
+    {
+      "epoch": 1.0853769300635785,
+      "grad_norm": 1.7670155763626099,
+      "learning_rate": 8.456574141996913e-06,
+      "logits/chosen": 0.09054378420114517,
+      "logits/rejected": 0.17597083747386932,
+      "logps/chosen": -79.7802505493164,
+      "logps/rejected": -95.21873474121094,
+      "loss": 1.9499,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.336801767349243,
+      "rewards/margins": 0.7675365209579468,
+      "rewards/rejected": -3.1043381690979004,
+      "step": 598
+    },
+    {
+      "epoch": 1.0871934604904632,
+      "grad_norm": 2.033724308013916,
+      "learning_rate": 8.451726731917722e-06,
+      "logits/chosen": 0.08810828626155853,
+      "logits/rejected": 0.1388731151819229,
+      "logps/chosen": -75.62480163574219,
+      "logps/rejected": -90.9974365234375,
+      "loss": 1.993,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2493839263916016,
+      "rewards/margins": 0.9519067406654358,
+      "rewards/rejected": -3.2012906074523926,
+      "step": 599
+    },
+    {
+      "epoch": 1.0890099909173478,
+      "grad_norm": 2.0463502407073975,
+      "learning_rate": 8.446873302753783e-06,
+      "logits/chosen": -0.017319753766059875,
+      "logits/rejected": 0.026935823261737823,
+      "logps/chosen": -91.20410919189453,
+      "logps/rejected": -97.56501770019531,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5798678398132324,
+      "rewards/margins": 0.46915650367736816,
+      "rewards/rejected": -3.0490241050720215,
+      "step": 600
+    },
+    {
+      "epoch": 1.0908265213442325,
+      "grad_norm": 1.944875717163086,
+      "learning_rate": 8.44201386440208e-06,
+      "logits/chosen": 0.1267091929912567,
+      "logits/rejected": 0.12764671444892883,
+      "logps/chosen": -82.59441375732422,
+      "logps/rejected": -91.89214324951172,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.495068311691284,
+      "rewards/margins": 0.6413819789886475,
+      "rewards/rejected": -3.1364500522613525,
+      "step": 601
+    },
+    {
+      "epoch": 1.0926430517711172,
+      "grad_norm": 1.8729236125946045,
+      "learning_rate": 8.437148426771852e-06,
+      "logits/chosen": 0.15159347653388977,
+      "logits/rejected": 0.16930025815963745,
+      "logps/chosen": -80.54000854492188,
+      "logps/rejected": -93.16973114013672,
+      "loss": 1.8605,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4500646591186523,
+      "rewards/margins": 0.9991447925567627,
+      "rewards/rejected": -3.449209690093994,
+      "step": 602
+    },
+    {
+      "epoch": 1.0944595821980019,
+      "grad_norm": 1.9284056425094604,
+      "learning_rate": 8.432276999784578e-06,
+      "logits/chosen": 0.16441625356674194,
+      "logits/rejected": 0.17125487327575684,
+      "logps/chosen": -83.31804656982422,
+      "logps/rejected": -91.39764404296875,
+      "loss": 1.9416,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.577606201171875,
+      "rewards/margins": 0.8178077936172485,
+      "rewards/rejected": -3.395413875579834,
+      "step": 603
+    },
+    {
+      "epoch": 1.0962761126248866,
+      "grad_norm": 2.0095536708831787,
+      "learning_rate": 8.427399593373941e-06,
+      "logits/chosen": 0.03275620937347412,
+      "logits/rejected": 0.033026132732629776,
+      "logps/chosen": -83.01813507080078,
+      "logps/rejected": -93.54053497314453,
+      "loss": 2.0358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.488004684448242,
+      "rewards/margins": 0.6699528694152832,
+      "rewards/rejected": -3.1579573154449463,
+      "step": 604
+    },
+    {
+      "epoch": 1.0980926430517712,
+      "grad_norm": 1.8138859272003174,
+      "learning_rate": 8.422516217485826e-06,
+      "logits/chosen": 0.1295854151248932,
+      "logits/rejected": 0.15053214132785797,
+      "logps/chosen": -87.74354553222656,
+      "logps/rejected": -94.38423919677734,
+      "loss": 2.6379,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.538604259490967,
+      "rewards/margins": 0.44896796345710754,
+      "rewards/rejected": -2.987572193145752,
+      "step": 605
+    },
+    {
+      "epoch": 1.0999091734786557,
+      "grad_norm": 2.262779951095581,
+      "learning_rate": 8.417626882078287e-06,
+      "logits/chosen": 0.07311846315860748,
+      "logits/rejected": 0.09329289197921753,
+      "logps/chosen": -88.92317199707031,
+      "logps/rejected": -105.15555572509766,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.468625783920288,
+      "rewards/margins": 1.0499695539474487,
+      "rewards/rejected": -3.5185956954956055,
+      "step": 606
+    },
+    {
+      "epoch": 1.1017257039055404,
+      "grad_norm": 1.7983455657958984,
+      "learning_rate": 8.412731597121527e-06,
+      "logits/chosen": 0.06970179080963135,
+      "logits/rejected": 0.170780211687088,
+      "logps/chosen": -84.99759674072266,
+      "logps/rejected": -91.03089904785156,
+      "loss": 2.3611,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.31594181060791,
+      "rewards/margins": 0.46637701988220215,
+      "rewards/rejected": -2.7823188304901123,
+      "step": 607
+    },
+    {
+      "epoch": 1.103542234332425,
+      "grad_norm": 1.8828606605529785,
+      "learning_rate": 8.407830372597884e-06,
+      "logits/chosen": 0.10343387722969055,
+      "logits/rejected": 0.15009143948554993,
+      "logps/chosen": -74.23157501220703,
+      "logps/rejected": -85.4384994506836,
+      "loss": 2.0535,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3184473514556885,
+      "rewards/margins": 0.7085882425308228,
+      "rewards/rejected": -3.027035713195801,
+      "step": 608
+    },
+    {
+      "epoch": 1.1053587647593097,
+      "grad_norm": 1.8570231199264526,
+      "learning_rate": 8.402923218501813e-06,
+      "logits/chosen": 0.0035794638097286224,
+      "logits/rejected": 0.044682763516902924,
+      "logps/chosen": -82.44414520263672,
+      "logps/rejected": -93.22355651855469,
+      "loss": 2.1751,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.584066867828369,
+      "rewards/margins": 0.7579058408737183,
+      "rewards/rejected": -3.3419723510742188,
+      "step": 609
+    },
+    {
+      "epoch": 1.1071752951861944,
+      "grad_norm": 1.9866671562194824,
+      "learning_rate": 8.39801014483985e-06,
+      "logits/chosen": -0.015922199934720993,
+      "logits/rejected": 0.10281078517436981,
+      "logps/chosen": -99.91399383544922,
+      "logps/rejected": -113.79450988769531,
+      "loss": 2.2522,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.595137119293213,
+      "rewards/margins": 0.7955703139305115,
+      "rewards/rejected": -3.390707492828369,
+      "step": 610
+    },
+    {
+      "epoch": 1.108991825613079,
+      "grad_norm": 2.148136854171753,
+      "learning_rate": 8.393091161630612e-06,
+      "logits/chosen": 0.04068015143275261,
+      "logits/rejected": 0.09121442586183548,
+      "logps/chosen": -85.49508666992188,
+      "logps/rejected": -97.27752685546875,
+      "loss": 1.9467,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.673704147338867,
+      "rewards/margins": 0.8281559348106384,
+      "rewards/rejected": -3.5018603801727295,
+      "step": 611
+    },
+    {
+      "epoch": 1.1108083560399638,
+      "grad_norm": 2.0157570838928223,
+      "learning_rate": 8.388166278904759e-06,
+      "logits/chosen": 0.03451048582792282,
+      "logits/rejected": 0.10373912751674652,
+      "logps/chosen": -79.24334716796875,
+      "logps/rejected": -91.64152526855469,
+      "loss": 2.041,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4238996505737305,
+      "rewards/margins": 0.7874577045440674,
+      "rewards/rejected": -3.211357593536377,
+      "step": 612
+    },
+    {
+      "epoch": 1.1126248864668482,
+      "grad_norm": 1.7337363958358765,
+      "learning_rate": 8.383235506704986e-06,
+      "logits/chosen": 0.06933741271495819,
+      "logits/rejected": 0.10624522715806961,
+      "logps/chosen": -77.62388610839844,
+      "logps/rejected": -91.96408081054688,
+      "loss": 2.0593,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.304389476776123,
+      "rewards/margins": 0.6717057228088379,
+      "rewards/rejected": -2.97609543800354,
+      "step": 613
+    },
+    {
+      "epoch": 1.1144414168937329,
+      "grad_norm": 3.6357574462890625,
+      "learning_rate": 8.378298855085996e-06,
+      "logits/chosen": 0.08138547092676163,
+      "logits/rejected": 0.0961306244134903,
+      "logps/chosen": -87.44918823242188,
+      "logps/rejected": -94.0655746459961,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5593717098236084,
+      "rewards/margins": 0.6174351572990417,
+      "rewards/rejected": -3.176806688308716,
+      "step": 614
+    },
+    {
+      "epoch": 1.1162579473206176,
+      "grad_norm": 2.273991107940674,
+      "learning_rate": 8.373356334114484e-06,
+      "logits/chosen": 0.029560726135969162,
+      "logits/rejected": 0.11155828088521957,
+      "logps/chosen": -91.85681915283203,
+      "logps/rejected": -103.11241912841797,
+      "loss": 2.2401,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5434679985046387,
+      "rewards/margins": 0.7232552170753479,
+      "rewards/rejected": -3.266723394393921,
+      "step": 615
+    },
+    {
+      "epoch": 1.1180744777475022,
+      "grad_norm": 2.1391029357910156,
+      "learning_rate": 8.368407953869105e-06,
+      "logits/chosen": 0.05583259090781212,
+      "logits/rejected": 0.04287164285778999,
+      "logps/chosen": -92.72290802001953,
+      "logps/rejected": -99.79463195800781,
+      "loss": 1.8293,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.395390272140503,
+      "rewards/margins": 0.9053142666816711,
+      "rewards/rejected": -3.3007044792175293,
+      "step": 616
+    },
+    {
+      "epoch": 1.119891008174387,
+      "grad_norm": 1.6499502658843994,
+      "learning_rate": 8.363453724440471e-06,
+      "logits/chosen": 0.06658023595809937,
+      "logits/rejected": 0.04786865413188934,
+      "logps/chosen": -88.62269592285156,
+      "logps/rejected": -95.76766967773438,
+      "loss": 1.9608,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.522400379180908,
+      "rewards/margins": 0.8233134746551514,
+      "rewards/rejected": -3.3457136154174805,
+      "step": 617
+    },
+    {
+      "epoch": 1.1217075386012716,
+      "grad_norm": 2.018575668334961,
+      "learning_rate": 8.358493655931119e-06,
+      "logits/chosen": 0.06770970672369003,
+      "logits/rejected": 0.08037856221199036,
+      "logps/chosen": -82.23643493652344,
+      "logps/rejected": -97.71478271484375,
+      "loss": 1.8448,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.201833724975586,
+      "rewards/margins": 1.0529935359954834,
+      "rewards/rejected": -3.2548277378082275,
+      "step": 618
+    },
+    {
+      "epoch": 1.1235240690281563,
+      "grad_norm": 2.0033349990844727,
+      "learning_rate": 8.353527758455491e-06,
+      "logits/chosen": 0.11088447272777557,
+      "logits/rejected": 0.09635978937149048,
+      "logps/chosen": -81.90861511230469,
+      "logps/rejected": -92.48001098632812,
+      "loss": 2.2057,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.494230270385742,
+      "rewards/margins": 0.7774206399917603,
+      "rewards/rejected": -3.271650791168213,
+      "step": 619
+    },
+    {
+      "epoch": 1.125340599455041,
+      "grad_norm": 1.700210452079773,
+      "learning_rate": 8.348556042139918e-06,
+      "logits/chosen": 0.08965969830751419,
+      "logits/rejected": 0.06873422861099243,
+      "logps/chosen": -80.28114318847656,
+      "logps/rejected": -84.04791259765625,
+      "loss": 2.4366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.773008108139038,
+      "rewards/margins": 0.5039690732955933,
+      "rewards/rejected": -3.276977300643921,
+      "step": 620
+    },
+    {
+      "epoch": 1.1271571298819256,
+      "grad_norm": 2.623145580291748,
+      "learning_rate": 8.343578517122594e-06,
+      "logits/chosen": 0.08407651633024216,
+      "logits/rejected": 0.09678006917238235,
+      "logps/chosen": -81.90699005126953,
+      "logps/rejected": -94.30622863769531,
+      "loss": 2.126,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4066429138183594,
+      "rewards/margins": 0.8733029365539551,
+      "rewards/rejected": -3.2799456119537354,
+      "step": 621
+    },
+    {
+      "epoch": 1.12897366030881,
+      "grad_norm": 1.8616830110549927,
+      "learning_rate": 8.338595193553559e-06,
+      "logits/chosen": 0.002603452652692795,
+      "logits/rejected": 0.13452833890914917,
+      "logps/chosen": -73.35641479492188,
+      "logps/rejected": -84.70191192626953,
+      "loss": 2.084,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3735039234161377,
+      "rewards/margins": 0.7769300937652588,
+      "rewards/rejected": -3.1504342555999756,
+      "step": 622
+    },
+    {
+      "epoch": 1.1307901907356948,
+      "grad_norm": 1.8286879062652588,
+      "learning_rate": 8.333606081594678e-06,
+      "logits/chosen": 0.06104264035820961,
+      "logits/rejected": 0.10199623554944992,
+      "logps/chosen": -76.39376831054688,
+      "logps/rejected": -88.08850860595703,
+      "loss": 1.7885,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.130338191986084,
+      "rewards/margins": 0.848843514919281,
+      "rewards/rejected": -2.9791817665100098,
+      "step": 623
+    },
+    {
+      "epoch": 1.1326067211625794,
+      "grad_norm": 1.4803669452667236,
+      "learning_rate": 8.32861119141962e-06,
+      "logits/chosen": 0.07387635856866837,
+      "logits/rejected": 0.11718127131462097,
+      "logps/chosen": -82.92483520507812,
+      "logps/rejected": -91.46949005126953,
+      "loss": 1.7624,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.291823148727417,
+      "rewards/margins": 0.9700915217399597,
+      "rewards/rejected": -3.2619142532348633,
+      "step": 624
+    },
+    {
+      "epoch": 1.134423251589464,
+      "grad_norm": 1.6750752925872803,
+      "learning_rate": 8.323610533213835e-06,
+      "logits/chosen": 0.10784655809402466,
+      "logits/rejected": 0.17897658050060272,
+      "logps/chosen": -75.01900482177734,
+      "logps/rejected": -88.1203384399414,
+      "loss": 2.1956,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4190964698791504,
+      "rewards/margins": 0.688554584980011,
+      "rewards/rejected": -3.1076512336730957,
+      "step": 625
+    },
+    {
+      "epoch": 1.1362397820163488,
+      "grad_norm": 1.8641157150268555,
+      "learning_rate": 8.318604117174533e-06,
+      "logits/chosen": 0.05499793961644173,
+      "logits/rejected": 0.0563310906291008,
+      "logps/chosen": -83.42897033691406,
+      "logps/rejected": -94.2099380493164,
+      "loss": 1.675,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2912588119506836,
+      "rewards/margins": 0.9781790375709534,
+      "rewards/rejected": -3.2694382667541504,
+      "step": 626
+    },
+    {
+      "epoch": 1.1380563124432335,
+      "grad_norm": 1.7625656127929688,
+      "learning_rate": 8.313591953510674e-06,
+      "logits/chosen": 0.0831976979970932,
+      "logits/rejected": 0.12448858469724655,
+      "logps/chosen": -80.08613586425781,
+      "logps/rejected": -85.6810302734375,
+      "loss": 2.3784,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.453763008117676,
+      "rewards/margins": 0.46461760997772217,
+      "rewards/rejected": -2.9183804988861084,
+      "step": 627
+    },
+    {
+      "epoch": 1.1398728428701181,
+      "grad_norm": 2.10945463180542,
+      "learning_rate": 8.308574052442928e-06,
+      "logits/chosen": 0.09317971020936966,
+      "logits/rejected": 0.12335637211799622,
+      "logps/chosen": -82.36027526855469,
+      "logps/rejected": -94.93681335449219,
+      "loss": 2.1623,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5503153800964355,
+      "rewards/margins": 0.6077253222465515,
+      "rewards/rejected": -3.158040761947632,
+      "step": 628
+    },
+    {
+      "epoch": 1.1416893732970028,
+      "grad_norm": 1.8368538618087769,
+      "learning_rate": 8.30355042420367e-06,
+      "logits/chosen": 0.08310627192258835,
+      "logits/rejected": 0.11056395620107651,
+      "logps/chosen": -81.20982360839844,
+      "logps/rejected": -92.69551086425781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2482149600982666,
+      "rewards/margins": 0.9621560573577881,
+      "rewards/rejected": -3.2103710174560547,
+      "step": 629
+    },
+    {
+      "epoch": 1.1435059037238875,
+      "grad_norm": 1.7961188554763794,
+      "learning_rate": 8.29852107903695e-06,
+      "logits/chosen": 0.13976231217384338,
+      "logits/rejected": 0.11740799248218536,
+      "logps/chosen": -70.01700592041016,
+      "logps/rejected": -76.1086654663086,
+      "loss": 2.3401,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.239321231842041,
+      "rewards/margins": 0.5643521547317505,
+      "rewards/rejected": -2.80367374420166,
+      "step": 630
+    },
+    {
+      "epoch": 1.145322434150772,
+      "grad_norm": 2.041840076446533,
+      "learning_rate": 8.293486027198483e-06,
+      "logits/chosen": 0.06489598006010056,
+      "logits/rejected": 0.11150355637073517,
+      "logps/chosen": -95.31941223144531,
+      "logps/rejected": -96.98028564453125,
+      "loss": 2.2151,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5194008350372314,
+      "rewards/margins": 0.5945936441421509,
+      "rewards/rejected": -3.113994836807251,
+      "step": 631
+    },
+    {
+      "epoch": 1.1471389645776566,
+      "grad_norm": 2.7394556999206543,
+      "learning_rate": 8.288445278955615e-06,
+      "logits/chosen": 0.10089049488306046,
+      "logits/rejected": 0.08729755133390427,
+      "logps/chosen": -95.64276123046875,
+      "logps/rejected": -99.41658020019531,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.5987305641174316,
+      "rewards/margins": 0.4612637162208557,
+      "rewards/rejected": -3.0599937438964844,
+      "step": 632
+    },
+    {
+      "epoch": 1.1489554950045413,
+      "grad_norm": 2.024982213973999,
+      "learning_rate": 8.283398844587308e-06,
+      "logits/chosen": 0.023156292736530304,
+      "logits/rejected": 0.06789979338645935,
+      "logps/chosen": -89.90803527832031,
+      "logps/rejected": -96.06836700439453,
+      "loss": 1.9289,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.1415646076202393,
+      "rewards/margins": 0.7636609673500061,
+      "rewards/rejected": -2.9052255153656006,
+      "step": 633
+    },
+    {
+      "epoch": 1.150772025431426,
+      "grad_norm": 1.8750615119934082,
+      "learning_rate": 8.278346734384122e-06,
+      "logits/chosen": 0.02243964746594429,
+      "logits/rejected": 0.12660253047943115,
+      "logps/chosen": -71.34333038330078,
+      "logps/rejected": -86.67922973632812,
+      "loss": 1.8554,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.2015068531036377,
+      "rewards/margins": 0.9946616291999817,
+      "rewards/rejected": -3.1961684226989746,
+      "step": 634
+    },
+    {
+      "epoch": 1.1525885558583107,
+      "grad_norm": 1.9947502613067627,
+      "learning_rate": 8.27328895864819e-06,
+      "logits/chosen": 0.05387189984321594,
+      "logits/rejected": 0.08354412019252777,
+      "logps/chosen": -85.38788604736328,
+      "logps/rejected": -97.08065795898438,
+      "loss": 2.3917,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.595083236694336,
+      "rewards/margins": 0.5549375414848328,
+      "rewards/rejected": -3.1500213146209717,
+      "step": 635
+    },
+    {
+      "epoch": 1.1544050862851953,
+      "grad_norm": 2.145918607711792,
+      "learning_rate": 8.268225527693193e-06,
+      "logits/chosen": 0.10176774114370346,
+      "logits/rejected": 0.14395561814308167,
+      "logps/chosen": -76.02487182617188,
+      "logps/rejected": -86.05180358886719,
+      "loss": 1.8909,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.0001683235168457,
+      "rewards/margins": 0.8724240064620972,
+      "rewards/rejected": -2.8725922107696533,
+      "step": 636
+    },
+    {
+      "epoch": 1.15622161671208,
+      "grad_norm": 1.6830824613571167,
+      "learning_rate": 8.263156451844353e-06,
+      "logits/chosen": 0.10138621181249619,
+      "logits/rejected": 0.11966821551322937,
+      "logps/chosen": -76.56681060791016,
+      "logps/rejected": -88.1928939819336,
+      "loss": 1.9508,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.247589588165283,
+      "rewards/margins": 0.8286643028259277,
+      "rewards/rejected": -3.076253890991211,
+      "step": 637
+    },
+    {
+      "epoch": 1.1580381471389645,
+      "grad_norm": 2.0992536544799805,
+      "learning_rate": 8.258081741438396e-06,
+      "logits/chosen": 0.07541397213935852,
+      "logits/rejected": 0.09907414764165878,
+      "logps/chosen": -90.69495391845703,
+      "logps/rejected": -101.71548461914062,
+      "loss": 2.1933,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4336788654327393,
+      "rewards/margins": 0.7834519147872925,
+      "rewards/rejected": -3.217130422592163,
+      "step": 638
+    },
+    {
+      "epoch": 1.1598546775658491,
+      "grad_norm": 1.8066256046295166,
+      "learning_rate": 8.25300140682354e-06,
+      "logits/chosen": 0.14558832347393036,
+      "logits/rejected": 0.14358943700790405,
+      "logps/chosen": -82.73606872558594,
+      "logps/rejected": -86.27526092529297,
+      "loss": 2.0025,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.393693208694458,
+      "rewards/margins": 0.7131556272506714,
+      "rewards/rejected": -3.106848955154419,
+      "step": 639
+    },
+    {
+      "epoch": 1.1616712079927338,
+      "grad_norm": 2.3521695137023926,
+      "learning_rate": 8.247915458359473e-06,
+      "logits/chosen": 0.05236164107918739,
+      "logits/rejected": 0.1095740869641304,
+      "logps/chosen": -83.70420837402344,
+      "logps/rejected": -100.09117889404297,
+      "loss": 2.3348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5044684410095215,
+      "rewards/margins": 0.7525762319564819,
+      "rewards/rejected": -3.257044553756714,
+      "step": 640
+    },
+    {
+      "epoch": 1.1634877384196185,
+      "grad_norm": 2.343412160873413,
+      "learning_rate": 8.242823906417329e-06,
+      "logits/chosen": -0.00503757456317544,
+      "logits/rejected": 0.06355556100606918,
+      "logps/chosen": -80.41883087158203,
+      "logps/rejected": -95.32416534423828,
+      "loss": 2.2004,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3209891319274902,
+      "rewards/margins": 0.745349109172821,
+      "rewards/rejected": -3.066338539123535,
+      "step": 641
+    },
+    {
+      "epoch": 1.1653042688465032,
+      "grad_norm": 1.9915282726287842,
+      "learning_rate": 8.237726761379668e-06,
+      "logits/chosen": 0.02087419107556343,
+      "logits/rejected": 0.08556030690670013,
+      "logps/chosen": -84.41456604003906,
+      "logps/rejected": -97.61527252197266,
+      "loss": 2.0052,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.385272741317749,
+      "rewards/margins": 0.7285323739051819,
+      "rewards/rejected": -3.113805055618286,
+      "step": 642
+    },
+    {
+      "epoch": 1.1671207992733879,
+      "grad_norm": 2.4018874168395996,
+      "learning_rate": 8.232624033640458e-06,
+      "logits/chosen": 0.0840807855129242,
+      "logits/rejected": 0.03736239671707153,
+      "logps/chosen": -83.6908950805664,
+      "logps/rejected": -88.65196228027344,
+      "loss": 2.5072,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.494255542755127,
+      "rewards/margins": 0.5595706701278687,
+      "rewards/rejected": -3.053826332092285,
+      "step": 643
+    },
+    {
+      "epoch": 1.1689373297002725,
+      "grad_norm": 2.3055484294891357,
+      "learning_rate": 8.227515733605049e-06,
+      "logits/chosen": 0.05399080738425255,
+      "logits/rejected": 0.13454601168632507,
+      "logps/chosen": -77.62064361572266,
+      "logps/rejected": -86.54924011230469,
+      "loss": 2.0759,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3616065979003906,
+      "rewards/margins": 0.7132005095481873,
+      "rewards/rejected": -3.0748066902160645,
+      "step": 644
+    },
+    {
+      "epoch": 1.1707538601271572,
+      "grad_norm": 1.989559292793274,
+      "learning_rate": 8.222401871690153e-06,
+      "logits/chosen": 0.03339620307087898,
+      "logits/rejected": 0.06584692001342773,
+      "logps/chosen": -74.14878845214844,
+      "logps/rejected": -89.0107421875,
+      "loss": 1.9992,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3560426235198975,
+      "rewards/margins": 0.7924355864524841,
+      "rewards/rejected": -3.1484780311584473,
+      "step": 645
+    },
+    {
+      "epoch": 1.1725703905540419,
+      "grad_norm": 1.7572104930877686,
+      "learning_rate": 8.217282458323825e-06,
+      "logits/chosen": 0.09702566266059875,
+      "logits/rejected": 0.06184859201312065,
+      "logps/chosen": -82.05592346191406,
+      "logps/rejected": -87.59920501708984,
+      "loss": 2.0353,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.121622085571289,
+      "rewards/margins": 0.707017183303833,
+      "rewards/rejected": -2.828639268875122,
+      "step": 646
+    },
+    {
+      "epoch": 1.1743869209809263,
+      "grad_norm": 1.8265388011932373,
+      "learning_rate": 8.21215750394544e-06,
+      "logits/chosen": 0.12195339798927307,
+      "logits/rejected": 0.09588825702667236,
+      "logps/chosen": -76.14807891845703,
+      "logps/rejected": -85.36613464355469,
+      "loss": 2.0684,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2251272201538086,
+      "rewards/margins": 0.8967663049697876,
+      "rewards/rejected": -3.1218934059143066,
+      "step": 647
+    },
+    {
+      "epoch": 1.176203451407811,
+      "grad_norm": 1.9061174392700195,
+      "learning_rate": 8.207027019005675e-06,
+      "logits/chosen": 0.09651574492454529,
+      "logits/rejected": 0.10208628326654434,
+      "logps/chosen": -86.28733825683594,
+      "logps/rejected": -97.5046157836914,
+      "loss": 2.1319,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3506147861480713,
+      "rewards/margins": 0.7415947318077087,
+      "rewards/rejected": -3.0922091007232666,
+      "step": 648
+    },
+    {
+      "epoch": 1.1780199818346957,
+      "grad_norm": 1.999624252319336,
+      "learning_rate": 8.20189101396648e-06,
+      "logits/chosen": 0.08931821584701538,
+      "logits/rejected": 0.09364073723554611,
+      "logps/chosen": -84.32461547851562,
+      "logps/rejected": -96.70581817626953,
+      "loss": 1.7919,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4468250274658203,
+      "rewards/margins": 0.8684096932411194,
+      "rewards/rejected": -3.315234661102295,
+      "step": 649
+    },
+    {
+      "epoch": 1.1798365122615804,
+      "grad_norm": 1.5666590929031372,
+      "learning_rate": 8.196749499301062e-06,
+      "logits/chosen": 0.09945614635944366,
+      "logits/rejected": 0.12043865770101547,
+      "logps/chosen": -81.99935150146484,
+      "logps/rejected": -96.36174011230469,
+      "loss": 1.8932,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4319562911987305,
+      "rewards/margins": 0.8513790369033813,
+      "rewards/rejected": -3.2833354473114014,
+      "step": 650
+    },
+    {
+      "epoch": 1.181653042688465,
+      "grad_norm": 1.9403899908065796,
+      "learning_rate": 8.191602485493868e-06,
+      "logits/chosen": 0.12361271679401398,
+      "logits/rejected": 0.14710690081119537,
+      "logps/chosen": -82.11424255371094,
+      "logps/rejected": -89.37992858886719,
+      "loss": 2.0615,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.161297082901001,
+      "rewards/margins": 0.7220393419265747,
+      "rewards/rejected": -2.8833365440368652,
+      "step": 651
+    },
+    {
+      "epoch": 1.1834695731153497,
+      "grad_norm": 1.8558425903320312,
+      "learning_rate": 8.186449983040552e-06,
+      "logits/chosen": 0.07215605676174164,
+      "logits/rejected": 0.06390087306499481,
+      "logps/chosen": -81.64322662353516,
+      "logps/rejected": -85.39015197753906,
+      "loss": 2.0064,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.320831775665283,
+      "rewards/margins": 0.7350802421569824,
+      "rewards/rejected": -3.0559120178222656,
+      "step": 652
+    },
+    {
+      "epoch": 1.1852861035422344,
+      "grad_norm": 1.7956428527832031,
+      "learning_rate": 8.181292002447966e-06,
+      "logits/chosen": 0.11276388168334961,
+      "logits/rejected": 0.06517648696899414,
+      "logps/chosen": -76.14613342285156,
+      "logps/rejected": -82.24524688720703,
+      "loss": 2.0473,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2219183444976807,
+      "rewards/margins": 0.6231520175933838,
+      "rewards/rejected": -2.8450703620910645,
+      "step": 653
+    },
+    {
+      "epoch": 1.1871026339691189,
+      "grad_norm": 2.1648740768432617,
+      "learning_rate": 8.17612855423413e-06,
+      "logits/chosen": 0.0334198959171772,
+      "logits/rejected": 0.013921715319156647,
+      "logps/chosen": -83.02323150634766,
+      "logps/rejected": -92.95353698730469,
+      "loss": 1.9884,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.0769619941711426,
+      "rewards/margins": 0.851192057132721,
+      "rewards/rejected": -2.9281537532806396,
+      "step": 654
+    },
+    {
+      "epoch": 1.1889191643960035,
+      "grad_norm": 1.921157717704773,
+      "learning_rate": 8.170959648928214e-06,
+      "logits/chosen": 0.073361836373806,
+      "logits/rejected": 0.13112977147102356,
+      "logps/chosen": -83.41218566894531,
+      "logps/rejected": -92.03260803222656,
+      "loss": 2.0203,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7512502670288086,
+      "rewards/margins": 0.8279457092285156,
+      "rewards/rejected": -3.579195737838745,
+      "step": 655
+    },
+    {
+      "epoch": 1.1907356948228882,
+      "grad_norm": 2.4273247718811035,
+      "learning_rate": 8.165785297070516e-06,
+      "logits/chosen": 0.027622584253549576,
+      "logits/rejected": 0.08521658927202225,
+      "logps/chosen": -84.34210205078125,
+      "logps/rejected": -101.76324462890625,
+      "loss": 1.9023,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6731176376342773,
+      "rewards/margins": 1.004132866859436,
+      "rewards/rejected": -3.677250385284424,
+      "step": 656
+    },
+    {
+      "epoch": 1.192552225249773,
+      "grad_norm": 1.8262163400650024,
+      "learning_rate": 8.16060550921244e-06,
+      "logits/chosen": 0.03431132063269615,
+      "logits/rejected": 0.06454990059137344,
+      "logps/chosen": -76.59049987792969,
+      "logps/rejected": -97.41972351074219,
+      "loss": 2.0105,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3002851009368896,
+      "rewards/margins": 0.7388467788696289,
+      "rewards/rejected": -3.0391316413879395,
+      "step": 657
+    },
+    {
+      "epoch": 1.1943687556766576,
+      "grad_norm": 1.9906315803527832,
+      "learning_rate": 8.155420295916474e-06,
+      "logits/chosen": 0.08427983522415161,
+      "logits/rejected": 0.025464089587330818,
+      "logps/chosen": -84.34020233154297,
+      "logps/rejected": -92.31369018554688,
+      "loss": 1.9895,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3555400371551514,
+      "rewards/margins": 0.8297086954116821,
+      "rewards/rejected": -3.185249090194702,
+      "step": 658
+    },
+    {
+      "epoch": 1.1961852861035422,
+      "grad_norm": 1.9980812072753906,
+      "learning_rate": 8.150229667756172e-06,
+      "logits/chosen": 0.07556813955307007,
+      "logits/rejected": 0.07593435049057007,
+      "logps/chosen": -85.46858215332031,
+      "logps/rejected": -90.94758605957031,
+      "loss": 2.2687,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.679569959640503,
+      "rewards/margins": 0.5908665657043457,
+      "rewards/rejected": -3.2704365253448486,
+      "step": 659
+    },
+    {
+      "epoch": 1.198001816530427,
+      "grad_norm": 2.0758330821990967,
+      "learning_rate": 8.14503363531613e-06,
+      "logits/chosen": 0.11312856525182724,
+      "logits/rejected": 0.08777043223381042,
+      "logps/chosen": -80.32341003417969,
+      "logps/rejected": -84.89898681640625,
+      "loss": 2.1371,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5170207023620605,
+      "rewards/margins": 0.632743239402771,
+      "rewards/rejected": -3.149764060974121,
+      "step": 660
+    },
+    {
+      "epoch": 1.1998183469573116,
+      "grad_norm": 2.765526533126831,
+      "learning_rate": 8.139832209191961e-06,
+      "logits/chosen": 0.17475244402885437,
+      "logits/rejected": 0.13103297352790833,
+      "logps/chosen": -92.03308868408203,
+      "logps/rejected": -94.15422821044922,
+      "loss": 2.4151,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.3920867443084717,
+      "rewards/margins": 0.518372118473053,
+      "rewards/rejected": -2.910458564758301,
+      "step": 661
+    },
+    {
+      "epoch": 1.2016348773841963,
+      "grad_norm": 2.0190646648406982,
+      "learning_rate": 8.13462539999028e-06,
+      "logits/chosen": 0.040579523891210556,
+      "logits/rejected": 0.060216568410396576,
+      "logps/chosen": -96.32977294921875,
+      "logps/rejected": -105.37261962890625,
+      "loss": 2.1216,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5936977863311768,
+      "rewards/margins": 1.040100336074829,
+      "rewards/rejected": -3.6337978839874268,
+      "step": 662
+    },
+    {
+      "epoch": 1.2034514078110807,
+      "grad_norm": 2.53027606010437,
+      "learning_rate": 8.129413218328674e-06,
+      "logits/chosen": 0.07498883455991745,
+      "logits/rejected": 0.042916588485240936,
+      "logps/chosen": -83.17558288574219,
+      "logps/rejected": -89.55461883544922,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6350390911102295,
+      "rewards/margins": 0.7082449793815613,
+      "rewards/rejected": -3.3432838916778564,
+      "step": 663
+    },
+    {
+      "epoch": 1.2052679382379654,
+      "grad_norm": 1.6480835676193237,
+      "learning_rate": 8.124195674835695e-06,
+      "logits/chosen": 0.1062261089682579,
+      "logits/rejected": 0.09598012268543243,
+      "logps/chosen": -84.5193099975586,
+      "logps/rejected": -95.05423736572266,
+      "loss": 1.7544,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.1563069820404053,
+      "rewards/margins": 0.8859087228775024,
+      "rewards/rejected": -3.042215585708618,
+      "step": 664
+    },
+    {
+      "epoch": 1.20708446866485,
+      "grad_norm": 2.126664161682129,
+      "learning_rate": 8.118972780150817e-06,
+      "logits/chosen": -0.01870904304087162,
+      "logits/rejected": 0.1347496211528778,
+      "logps/chosen": -79.35530853271484,
+      "logps/rejected": -105.41546630859375,
+      "loss": 2.0524,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6006765365600586,
+      "rewards/margins": 1.0105364322662354,
+      "rewards/rejected": -3.611213207244873,
+      "step": 665
+    },
+    {
+      "epoch": 1.2089009990917348,
+      "grad_norm": 2.117645263671875,
+      "learning_rate": 8.113744544924434e-06,
+      "logits/chosen": 0.01287349034100771,
+      "logits/rejected": 0.019422955811023712,
+      "logps/chosen": -105.86250305175781,
+      "logps/rejected": -117.42082977294922,
+      "loss": 1.8973,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.636453151702881,
+      "rewards/margins": 0.8021495342254639,
+      "rewards/rejected": -3.4386026859283447,
+      "step": 666
+    },
+    {
+      "epoch": 1.2107175295186194,
+      "grad_norm": 2.956285238265991,
+      "learning_rate": 8.108510979817828e-06,
+      "logits/chosen": 0.047061942517757416,
+      "logits/rejected": 0.11715231090784073,
+      "logps/chosen": -89.56261444091797,
+      "logps/rejected": -103.57279968261719,
+      "loss": 2.0627,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5317227840423584,
+      "rewards/margins": 0.8301056623458862,
+      "rewards/rejected": -3.361828327178955,
+      "step": 667
+    },
+    {
+      "epoch": 1.2125340599455041,
+      "grad_norm": 1.897623062133789,
+      "learning_rate": 8.10327209550315e-06,
+      "logits/chosen": 0.017857536673545837,
+      "logits/rejected": 0.12088865786790848,
+      "logps/chosen": -95.75634002685547,
+      "logps/rejected": -113.0296630859375,
+      "loss": 1.9091,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6507534980773926,
+      "rewards/margins": 0.963871955871582,
+      "rewards/rejected": -3.6146254539489746,
+      "step": 668
+    },
+    {
+      "epoch": 1.2143505903723888,
+      "grad_norm": 2.176003932952881,
+      "learning_rate": 8.098027902663396e-06,
+      "logits/chosen": -0.0014993082731962204,
+      "logits/rejected": 0.11425516754388809,
+      "logps/chosen": -79.22991943359375,
+      "logps/rejected": -96.39210510253906,
+      "loss": 1.8977,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.582876443862915,
+      "rewards/margins": 0.9855522513389587,
+      "rewards/rejected": -3.5684289932250977,
+      "step": 669
+    },
+    {
+      "epoch": 1.2161671207992735,
+      "grad_norm": 2.140666961669922,
+      "learning_rate": 8.092778411992388e-06,
+      "logits/chosen": 0.04144468903541565,
+      "logits/rejected": 0.0556509830057621,
+      "logps/chosen": -80.35415649414062,
+      "logps/rejected": -97.16475677490234,
+      "loss": 1.9182,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.66176176071167,
+      "rewards/margins": 0.9624839425086975,
+      "rewards/rejected": -3.6242458820343018,
+      "step": 670
+    },
+    {
+      "epoch": 1.2179836512261581,
+      "grad_norm": 1.937705636024475,
+      "learning_rate": 8.087523634194755e-06,
+      "logits/chosen": 0.03470912575721741,
+      "logits/rejected": 0.12844812870025635,
+      "logps/chosen": -81.73896026611328,
+      "logps/rejected": -98.39558410644531,
+      "loss": 2.032,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.406296730041504,
+      "rewards/margins": 0.9326379895210266,
+      "rewards/rejected": -3.3389346599578857,
+      "step": 671
+    },
+    {
+      "epoch": 1.2198001816530426,
+      "grad_norm": 1.8903594017028809,
+      "learning_rate": 8.082263579985898e-06,
+      "logits/chosen": 0.03509046137332916,
+      "logits/rejected": 0.0663415864109993,
+      "logps/chosen": -81.79936981201172,
+      "logps/rejected": -96.15756225585938,
+      "loss": 1.7671,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2738354206085205,
+      "rewards/margins": 1.1403576135635376,
+      "rewards/rejected": -3.4141931533813477,
+      "step": 672
+    },
+    {
+      "epoch": 1.2216167120799273,
+      "grad_norm": 2.206265687942505,
+      "learning_rate": 8.076998260091989e-06,
+      "logits/chosen": 0.016251683235168457,
+      "logits/rejected": 0.07814528048038483,
+      "logps/chosen": -80.23258209228516,
+      "logps/rejected": -85.3895034790039,
+      "loss": 2.2589,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6450579166412354,
+      "rewards/margins": 0.5741185545921326,
+      "rewards/rejected": -3.2191762924194336,
+      "step": 673
+    },
+    {
+      "epoch": 1.223433242506812,
+      "grad_norm": 2.3217952251434326,
+      "learning_rate": 8.071727685249929e-06,
+      "logits/chosen": 0.15181973576545715,
+      "logits/rejected": 0.11617043614387512,
+      "logps/chosen": -85.15460205078125,
+      "logps/rejected": -83.82669830322266,
+      "loss": 2.2534,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.530156135559082,
+      "rewards/margins": 0.6635551452636719,
+      "rewards/rejected": -3.193711519241333,
+      "step": 674
+    },
+    {
+      "epoch": 1.2252497729336966,
+      "grad_norm": 2.011244297027588,
+      "learning_rate": 8.06645186620734e-06,
+      "logits/chosen": 0.02763158269226551,
+      "logits/rejected": 0.04441550746560097,
+      "logps/chosen": -81.14067840576172,
+      "logps/rejected": -95.907958984375,
+      "loss": 1.8979,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.693267583847046,
+      "rewards/margins": 0.9938480854034424,
+      "rewards/rejected": -3.6871156692504883,
+      "step": 675
+    },
+    {
+      "epoch": 1.2270663033605813,
+      "grad_norm": 2.3554768562316895,
+      "learning_rate": 8.061170813722533e-06,
+      "logits/chosen": 0.06379646062850952,
+      "logits/rejected": 0.052463918924331665,
+      "logps/chosen": -92.42412567138672,
+      "logps/rejected": -92.37240600585938,
+      "loss": 2.4093,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.7538669109344482,
+      "rewards/margins": 0.5145044326782227,
+      "rewards/rejected": -3.268371105194092,
+      "step": 676
+    },
+    {
+      "epoch": 1.228882833787466,
+      "grad_norm": 2.2725136280059814,
+      "learning_rate": 8.055884538564493e-06,
+      "logits/chosen": 0.028828933835029602,
+      "logits/rejected": 0.08082776516675949,
+      "logps/chosen": -76.20146179199219,
+      "logps/rejected": -94.82791900634766,
+      "loss": 1.7849,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.348304510116577,
+      "rewards/margins": 1.1067698001861572,
+      "rewards/rejected": -3.4550740718841553,
+      "step": 677
+    },
+    {
+      "epoch": 1.2306993642143507,
+      "grad_norm": 2.3565657138824463,
+      "learning_rate": 8.050593051512859e-06,
+      "logits/chosen": 0.09367989003658295,
+      "logits/rejected": 0.06132106855511665,
+      "logps/chosen": -84.85393524169922,
+      "logps/rejected": -91.87137603759766,
+      "loss": 2.162,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5945935249328613,
+      "rewards/margins": 0.7115421295166016,
+      "rewards/rejected": -3.306135416030884,
+      "step": 678
+    },
+    {
+      "epoch": 1.2325158946412351,
+      "grad_norm": 2.2214298248291016,
+      "learning_rate": 8.045296363357891e-06,
+      "logits/chosen": 0.015750454738736153,
+      "logits/rejected": -0.001254781149327755,
+      "logps/chosen": -94.99285125732422,
+      "logps/rejected": -107.42745971679688,
+      "loss": 2.0311,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.769968032836914,
+      "rewards/margins": 0.9211719632148743,
+      "rewards/rejected": -3.6911399364471436,
+      "step": 679
+    },
+    {
+      "epoch": 1.2343324250681198,
+      "grad_norm": 1.9377881288528442,
+      "learning_rate": 8.039994484900463e-06,
+      "logits/chosen": -0.011005287989974022,
+      "logits/rejected": 0.05877537280321121,
+      "logps/chosen": -84.62346649169922,
+      "logps/rejected": -99.59872436523438,
+      "loss": 1.8343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.554427146911621,
+      "rewards/margins": 1.086029291152954,
+      "rewards/rejected": -3.640456199645996,
+      "step": 680
+    },
+    {
+      "epoch": 1.2361489554950045,
+      "grad_norm": 2.665781259536743,
+      "learning_rate": 8.034687426952023e-06,
+      "logits/chosen": 0.030874190852046013,
+      "logits/rejected": 0.09339022636413574,
+      "logps/chosen": -82.25083923339844,
+      "logps/rejected": -99.68885803222656,
+      "loss": 2.0966,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.847683906555176,
+      "rewards/margins": 0.8753347396850586,
+      "rewards/rejected": -3.7230184078216553,
+      "step": 681
+    },
+    {
+      "epoch": 1.2379654859218892,
+      "grad_norm": 1.8999613523483276,
+      "learning_rate": 8.029375200334588e-06,
+      "logits/chosen": 0.07361680269241333,
+      "logits/rejected": 0.08493717014789581,
+      "logps/chosen": -88.70954895019531,
+      "logps/rejected": -102.10923767089844,
+      "loss": 1.7236,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.518230676651001,
+      "rewards/margins": 1.0492362976074219,
+      "rewards/rejected": -3.567467212677002,
+      "step": 682
+    },
+    {
+      "epoch": 1.2397820163487738,
+      "grad_norm": 2.179049491882324,
+      "learning_rate": 8.024057815880716e-06,
+      "logits/chosen": 0.12260061502456665,
+      "logits/rejected": 0.09896722435951233,
+      "logps/chosen": -82.61366271972656,
+      "logps/rejected": -89.37825775146484,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3436989784240723,
+      "rewards/margins": 0.7696498036384583,
+      "rewards/rejected": -3.1133487224578857,
+      "step": 683
+    },
+    {
+      "epoch": 1.2415985467756585,
+      "grad_norm": 1.8527759313583374,
+      "learning_rate": 8.018735284433475e-06,
+      "logits/chosen": 0.007166730239987373,
+      "logits/rejected": 0.04468049108982086,
+      "logps/chosen": -88.59947204589844,
+      "logps/rejected": -97.66645050048828,
+      "loss": 1.9353,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.669034957885742,
+      "rewards/margins": 0.8667396306991577,
+      "rewards/rejected": -3.5357747077941895,
+      "step": 684
+    },
+    {
+      "epoch": 1.2434150772025432,
+      "grad_norm": 2.013792037963867,
+      "learning_rate": 8.013407616846436e-06,
+      "logits/chosen": 0.05774620547890663,
+      "logits/rejected": 0.06977065652608871,
+      "logps/chosen": -83.88308715820312,
+      "logps/rejected": -93.36985778808594,
+      "loss": 1.9716,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5362911224365234,
+      "rewards/margins": 0.9699288010597229,
+      "rewards/rejected": -3.5062198638916016,
+      "step": 685
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.224138049042514e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-685/training_args.bin b/checkpoint-685/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-685/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-685/zero_to_fp32.py b/checkpoint-685/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-685/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-822/README.md b/checkpoint-822/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-822/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-822/adapter_config.json b/checkpoint-822/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-822/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-822/adapter_model.safetensors b/checkpoint-822/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..8a013fb00a3c4de332df33d6810ff89252b84315
--- /dev/null
+++ b/checkpoint-822/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8ce15419ddab7694c79babfaf2695078b3171ebd22bfd9af571f100e4400051
+size 207244392
diff --git a/checkpoint-822/latest b/checkpoint-822/latest
new file mode 100644
index 0000000000000000000000000000000000000000..3159aab1f7bb3903604150491f83c05295b87c00
--- /dev/null
+++ b/checkpoint-822/latest
@@ -0,0 +1 @@
+global_step821
\ No newline at end of file
diff --git a/checkpoint-822/rng_state_0.pth b/checkpoint-822/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b346349ce12dd5a17d4b91ed2a5722bb52550950
--- /dev/null
+++ b/checkpoint-822/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
+size 15984
diff --git a/checkpoint-822/rng_state_1.pth b/checkpoint-822/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..68f3c6994456cb8d0592a5375d99503c8924b1c4
--- /dev/null
+++ b/checkpoint-822/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73
+size 15984
diff --git a/checkpoint-822/rng_state_2.pth b/checkpoint-822/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..be044f6ceeed587d30e80c2f72d5aa19fdc9947b
--- /dev/null
+++ b/checkpoint-822/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b
+size 15984
diff --git a/checkpoint-822/rng_state_3.pth b/checkpoint-822/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..fc825249656a9b858782542bd3f4386250f1dfe0
--- /dev/null
+++ b/checkpoint-822/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc
+size 15984
diff --git a/checkpoint-822/rng_state_4.pth b/checkpoint-822/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d30f52a44be563c152ae09db6ae934da6da0d3ed
--- /dev/null
+++ b/checkpoint-822/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972
+size 15984
diff --git a/checkpoint-822/rng_state_5.pth b/checkpoint-822/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..c8715d27ab23ae545d58039cf949cc44ecc1da5e
--- /dev/null
+++ b/checkpoint-822/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991
+size 15984
diff --git a/checkpoint-822/rng_state_6.pth b/checkpoint-822/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1ed791b6ef76eadf0b0c55a5733411771e2ae027
--- /dev/null
+++ b/checkpoint-822/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa
+size 15984
diff --git a/checkpoint-822/rng_state_7.pth b/checkpoint-822/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..800c3bbbc5edf7db01a8316069d439c5fb8d8c30
--- /dev/null
+++ b/checkpoint-822/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773
+size 15984
diff --git a/checkpoint-822/special_tokens_map.json b/checkpoint-822/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-822/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-822/tokenizer.json b/checkpoint-822/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-822/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-822/tokenizer_config.json b/checkpoint-822/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-822/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-822/trainer_state.json b/checkpoint-822/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dee02ce3e60ea0b47f13a0689f9c3a50e28e7b29
--- /dev/null
+++ b/checkpoint-822/trainer_state.json
@@ -0,0 +1,12363 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.4922797456857402,
+  "eval_steps": 500,
+  "global_step": 822,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    },
+    {
+      "epoch": 0.49954586739327883,
+      "grad_norm": 2.1769776344299316,
+      "learning_rate": 9.65745789630079e-06,
+      "logits/chosen": 0.11112834513187408,
+      "logits/rejected": 0.10816515237092972,
+      "logps/chosen": -83.5405044555664,
+      "logps/rejected": -83.06329345703125,
+      "loss": 3.0939,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.142024040222168,
+      "rewards/margins": 0.1820124089717865,
+      "rewards/rejected": -2.3240363597869873,
+      "step": 275
+    },
+    {
+      "epoch": 0.5013623978201635,
+      "grad_norm": 1.7329221963882446,
+      "learning_rate": 9.654994540260396e-06,
+      "logits/chosen": 0.0653618574142456,
+      "logits/rejected": 0.08004368096590042,
+      "logps/chosen": -80.83209228515625,
+      "logps/rejected": -82.97142028808594,
+      "loss": 2.7161,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.851841688156128,
+      "rewards/margins": 0.2948168218135834,
+      "rewards/rejected": -2.146658420562744,
+      "step": 276
+    },
+    {
+      "epoch": 0.5031789282470481,
+      "grad_norm": 1.592657208442688,
+      "learning_rate": 9.65252271146268e-06,
+      "logits/chosen": 0.09880789369344711,
+      "logits/rejected": 0.14229761064052582,
+      "logps/chosen": -67.1727294921875,
+      "logps/rejected": -75.03417205810547,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8714643716812134,
+      "rewards/margins": 0.37937116622924805,
+      "rewards/rejected": -2.250835657119751,
+      "step": 277
+    },
+    {
+      "epoch": 0.5049954586739328,
+      "grad_norm": 1.6487712860107422,
+      "learning_rate": 9.650042414948133e-06,
+      "logits/chosen": 0.13465353846549988,
+      "logits/rejected": 0.12865689396858215,
+      "logps/chosen": -76.4417724609375,
+      "logps/rejected": -78.6947021484375,
+      "loss": 2.6085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.875291347503662,
+      "rewards/margins": 0.35175251960754395,
+      "rewards/rejected": -2.227043867111206,
+      "step": 278
+    },
+    {
+      "epoch": 0.5068119891008175,
+      "grad_norm": 1.6523009538650513,
+      "learning_rate": 9.64755365577451e-06,
+      "logits/chosen": 0.04238567873835564,
+      "logits/rejected": 0.07994347810745239,
+      "logps/chosen": -76.92097473144531,
+      "logps/rejected": -83.20886993408203,
+      "loss": 2.3156,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9759610891342163,
+      "rewards/margins": 0.5398927330970764,
+      "rewards/rejected": -2.5158536434173584,
+      "step": 279
+    },
+    {
+      "epoch": 0.508628519527702,
+      "grad_norm": 1.7999261617660522,
+      "learning_rate": 9.645056439016827e-06,
+      "logits/chosen": 0.07349395751953125,
+      "logits/rejected": 0.07518415153026581,
+      "logps/chosen": -77.56079864501953,
+      "logps/rejected": -84.97645568847656,
+      "loss": 2.8085,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9408115148544312,
+      "rewards/margins": 0.3470652997493744,
+      "rewards/rejected": -2.287877082824707,
+      "step": 280
+    },
+    {
+      "epoch": 0.5104450499545867,
+      "grad_norm": 1.5684200525283813,
+      "learning_rate": 9.642550769767342e-06,
+      "logits/chosen": 0.16188879311084747,
+      "logits/rejected": 0.12772323191165924,
+      "logps/chosen": -89.29315185546875,
+      "logps/rejected": -94.35065460205078,
+      "loss": 2.2314,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8958841562271118,
+      "rewards/margins": 0.5102108716964722,
+      "rewards/rejected": -2.406095027923584,
+      "step": 281
+    },
+    {
+      "epoch": 0.5122615803814714,
+      "grad_norm": 1.7878178358078003,
+      "learning_rate": 9.640036653135548e-06,
+      "logits/chosen": 0.060573749244213104,
+      "logits/rejected": 0.13457715511322021,
+      "logps/chosen": -68.9404525756836,
+      "logps/rejected": -74.77693176269531,
+      "loss": 2.4359,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.889530897140503,
+      "rewards/margins": 0.45911547541618347,
+      "rewards/rejected": -2.348646402359009,
+      "step": 282
+    },
+    {
+      "epoch": 0.5140781108083561,
+      "grad_norm": 1.4741288423538208,
+      "learning_rate": 9.637514094248172e-06,
+      "logits/chosen": 0.10433132946491241,
+      "logits/rejected": 0.1439315229654312,
+      "logps/chosen": -74.89447021484375,
+      "logps/rejected": -86.58551788330078,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9090807437896729,
+      "rewards/margins": 0.6733560562133789,
+      "rewards/rejected": -2.5824368000030518,
+      "step": 283
+    },
+    {
+      "epoch": 0.5158946412352406,
+      "grad_norm": 1.2864018678665161,
+      "learning_rate": 9.634983098249146e-06,
+      "logits/chosen": 0.10891089588403702,
+      "logits/rejected": 0.11755162477493286,
+      "logps/chosen": -66.90985107421875,
+      "logps/rejected": -71.8125228881836,
+      "loss": 2.1538,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.7779250144958496,
+      "rewards/margins": 0.5625147223472595,
+      "rewards/rejected": -2.340439796447754,
+      "step": 284
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 1.465747594833374,
+      "learning_rate": 9.632443670299616e-06,
+      "logits/chosen": 0.08224496245384216,
+      "logits/rejected": 0.12130744755268097,
+      "logps/chosen": -75.4281997680664,
+      "logps/rejected": -85.0781021118164,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7217226028442383,
+      "rewards/margins": 0.4750482439994812,
+      "rewards/rejected": -2.196770668029785,
+      "step": 285
+    },
+    {
+      "epoch": 0.51952770208901,
+      "grad_norm": 1.492859959602356,
+      "learning_rate": 9.629895815577915e-06,
+      "logits/chosen": 0.06619664281606674,
+      "logits/rejected": 0.13152630627155304,
+      "logps/chosen": -96.65383911132812,
+      "logps/rejected": -111.93521881103516,
+      "loss": 2.2831,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7611618041992188,
+      "rewards/margins": 0.5886337757110596,
+      "rewards/rejected": -2.349795341491699,
+      "step": 286
+    },
+    {
+      "epoch": 0.5213442325158947,
+      "grad_norm": 1.5534065961837769,
+      "learning_rate": 9.627339539279564e-06,
+      "logits/chosen": 0.06637927144765854,
+      "logits/rejected": 0.09107412397861481,
+      "logps/chosen": -71.92534637451172,
+      "logps/rejected": -82.98391723632812,
+      "loss": 2.5101,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7867075204849243,
+      "rewards/margins": 0.39493298530578613,
+      "rewards/rejected": -2.181640625,
+      "step": 287
+    },
+    {
+      "epoch": 0.5231607629427792,
+      "grad_norm": 1.6778221130371094,
+      "learning_rate": 9.624774846617254e-06,
+      "logits/chosen": 0.14700329303741455,
+      "logits/rejected": 0.12778782844543457,
+      "logps/chosen": -65.2364730834961,
+      "logps/rejected": -74.39017486572266,
+      "loss": 2.5236,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6942007541656494,
+      "rewards/margins": 0.35334131121635437,
+      "rewards/rejected": -2.047542095184326,
+      "step": 288
+    },
+    {
+      "epoch": 0.5249772933696639,
+      "grad_norm": 1.462215542793274,
+      "learning_rate": 9.622201742820839e-06,
+      "logits/chosen": 0.11467991769313812,
+      "logits/rejected": 0.1028795838356018,
+      "logps/chosen": -67.85939025878906,
+      "logps/rejected": -74.3462142944336,
+      "loss": 2.3891,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7194863557815552,
+      "rewards/margins": 0.3846975564956665,
+      "rewards/rejected": -2.1041836738586426,
+      "step": 289
+    },
+    {
+      "epoch": 0.5267938237965486,
+      "grad_norm": 1.7591735124588013,
+      "learning_rate": 9.619620233137328e-06,
+      "logits/chosen": 0.08407986164093018,
+      "logits/rejected": 0.12201236933469772,
+      "logps/chosen": -82.57340240478516,
+      "logps/rejected": -86.94192504882812,
+      "loss": 2.4569,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8536537885665894,
+      "rewards/margins": 0.38806283473968506,
+      "rewards/rejected": -2.2417166233062744,
+      "step": 290
+    },
+    {
+      "epoch": 0.5286103542234333,
+      "grad_norm": 1.5584774017333984,
+      "learning_rate": 9.617030322830868e-06,
+      "logits/chosen": 0.1126309484243393,
+      "logits/rejected": 0.08907752484083176,
+      "logps/chosen": -73.36901092529297,
+      "logps/rejected": -76.65506744384766,
+      "loss": 2.4021,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8091729879379272,
+      "rewards/margins": 0.3603411316871643,
+      "rewards/rejected": -2.1695141792297363,
+      "step": 291
+    },
+    {
+      "epoch": 0.5304268846503178,
+      "grad_norm": 1.487938642501831,
+      "learning_rate": 9.614432017182736e-06,
+      "logits/chosen": 0.09620118141174316,
+      "logits/rejected": 0.08514149487018585,
+      "logps/chosen": -88.66265106201172,
+      "logps/rejected": -92.26091766357422,
+      "loss": 2.3636,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.5980674028396606,
+      "rewards/margins": 0.4888036549091339,
+      "rewards/rejected": -2.0868711471557617,
+      "step": 292
+    },
+    {
+      "epoch": 0.5322434150772025,
+      "grad_norm": 1.445408582687378,
+      "learning_rate": 9.611825321491331e-06,
+      "logits/chosen": 0.10014252364635468,
+      "logits/rejected": 0.12857215106487274,
+      "logps/chosen": -82.46649932861328,
+      "logps/rejected": -92.84205627441406,
+      "loss": 2.343,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4620139598846436,
+      "rewards/margins": 0.5616316199302673,
+      "rewards/rejected": -2.0236456394195557,
+      "step": 293
+    },
+    {
+      "epoch": 0.5340599455040872,
+      "grad_norm": 1.262193202972412,
+      "learning_rate": 9.609210241072158e-06,
+      "logits/chosen": 0.09868282079696655,
+      "logits/rejected": 0.16074486076831818,
+      "logps/chosen": -67.0407485961914,
+      "logps/rejected": -82.36201477050781,
+      "loss": 2.1163,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5362765789031982,
+      "rewards/margins": 0.7091963887214661,
+      "rewards/rejected": -2.2454731464385986,
+      "step": 294
+    },
+    {
+      "epoch": 0.5358764759309719,
+      "grad_norm": 1.601891279220581,
+      "learning_rate": 9.606586781257822e-06,
+      "logits/chosen": 0.1236480325460434,
+      "logits/rejected": 0.12187935411930084,
+      "logps/chosen": -78.2726058959961,
+      "logps/rejected": -84.49217987060547,
+      "loss": 2.5075,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.779215931892395,
+      "rewards/margins": 0.38987410068511963,
+      "rewards/rejected": -2.1690900325775146,
+      "step": 295
+    },
+    {
+      "epoch": 0.5376930063578564,
+      "grad_norm": 1.4211548566818237,
+      "learning_rate": 9.603954947398016e-06,
+      "logits/chosen": 0.18375667929649353,
+      "logits/rejected": 0.21945703029632568,
+      "logps/chosen": -70.29829406738281,
+      "logps/rejected": -78.42284393310547,
+      "loss": 2.1959,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.7448753118515015,
+      "rewards/margins": 0.5106840133666992,
+      "rewards/rejected": -2.2555594444274902,
+      "step": 296
+    },
+    {
+      "epoch": 0.5395095367847411,
+      "grad_norm": 1.458617925643921,
+      "learning_rate": 9.601314744859504e-06,
+      "logits/chosen": 0.07360847294330597,
+      "logits/rejected": 0.1596718430519104,
+      "logps/chosen": -72.28546142578125,
+      "logps/rejected": -87.13935089111328,
+      "loss": 2.3538,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8007893562316895,
+      "rewards/margins": 0.5433144569396973,
+      "rewards/rejected": -2.3441038131713867,
+      "step": 297
+    },
+    {
+      "epoch": 0.5413260672116258,
+      "grad_norm": 1.6526339054107666,
+      "learning_rate": 9.598666179026123e-06,
+      "logits/chosen": 0.12013350427150726,
+      "logits/rejected": 0.10237178206443787,
+      "logps/chosen": -83.06564331054688,
+      "logps/rejected": -85.60771942138672,
+      "loss": 2.4531,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.911988615989685,
+      "rewards/margins": 0.35885563492774963,
+      "rewards/rejected": -2.2708444595336914,
+      "step": 298
+    },
+    {
+      "epoch": 0.5431425976385105,
+      "grad_norm": 1.9808340072631836,
+      "learning_rate": 9.596009255298755e-06,
+      "logits/chosen": 0.062342576682567596,
+      "logits/rejected": 0.09053834527730942,
+      "logps/chosen": -96.33489990234375,
+      "logps/rejected": -93.43024444580078,
+      "loss": 2.8549,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.845227599143982,
+      "rewards/margins": 0.22291362285614014,
+      "rewards/rejected": -2.068141222000122,
+      "step": 299
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 1.6276236772537231,
+      "learning_rate": 9.593343979095334e-06,
+      "logits/chosen": 0.2073422521352768,
+      "logits/rejected": 0.13107003271579742,
+      "logps/chosen": -72.45758056640625,
+      "logps/rejected": -77.42770385742188,
+      "loss": 2.4644,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9407453536987305,
+      "rewards/margins": 0.3134301006793976,
+      "rewards/rejected": -2.2541751861572266,
+      "step": 300
+    },
+    {
+      "epoch": 0.5467756584922797,
+      "grad_norm": 1.5001753568649292,
+      "learning_rate": 9.590670355850819e-06,
+      "logits/chosen": 0.15097060799598694,
+      "logits/rejected": 0.16254279017448425,
+      "logps/chosen": -70.44524383544922,
+      "logps/rejected": -70.64158630371094,
+      "loss": 2.4159,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.6921097040176392,
+      "rewards/margins": 0.3336741626262665,
+      "rewards/rejected": -2.0257837772369385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5485921889191644,
+      "grad_norm": 1.8360233306884766,
+      "learning_rate": 9.587988391017198e-06,
+      "logits/chosen": 0.14593760669231415,
+      "logits/rejected": 0.12311654537916183,
+      "logps/chosen": -78.32576751708984,
+      "logps/rejected": -88.25840759277344,
+      "loss": 2.5826,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.6591644287109375,
+      "rewards/margins": 0.5071319341659546,
+      "rewards/rejected": -2.1662964820861816,
+      "step": 302
+    },
+    {
+      "epoch": 0.5504087193460491,
+      "grad_norm": 1.4822838306427002,
+      "learning_rate": 9.585298090063459e-06,
+      "logits/chosen": 0.20818498730659485,
+      "logits/rejected": 0.16436657309532166,
+      "logps/chosen": -67.59429931640625,
+      "logps/rejected": -77.14763641357422,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6798239946365356,
+      "rewards/margins": 0.5954271554946899,
+      "rewards/rejected": -2.2752511501312256,
+      "step": 303
+    },
+    {
+      "epoch": 0.5522252497729337,
+      "grad_norm": 1.6118305921554565,
+      "learning_rate": 9.582599458475598e-06,
+      "logits/chosen": 0.0827561467885971,
+      "logits/rejected": 0.09151773154735565,
+      "logps/chosen": -73.28964233398438,
+      "logps/rejected": -76.90730285644531,
+      "loss": 2.3792,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.8056440353393555,
+      "rewards/margins": 0.39551618695259094,
+      "rewards/rejected": -2.201160192489624,
+      "step": 304
+    },
+    {
+      "epoch": 0.5540417801998183,
+      "grad_norm": 1.4609856605529785,
+      "learning_rate": 9.579892501756593e-06,
+      "logits/chosen": 0.05332394689321518,
+      "logits/rejected": 0.1590057611465454,
+      "logps/chosen": -77.51653289794922,
+      "logps/rejected": -95.51261138916016,
+      "loss": 2.2105,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.740310549736023,
+      "rewards/margins": 0.6771562695503235,
+      "rewards/rejected": -2.4174671173095703,
+      "step": 305
+    },
+    {
+      "epoch": 0.555858310626703,
+      "grad_norm": 1.7181960344314575,
+      "learning_rate": 9.5771772254264e-06,
+      "logits/chosen": 0.05865276977419853,
+      "logits/rejected": 0.03500773757696152,
+      "logps/chosen": -83.72650146484375,
+      "logps/rejected": -93.69804382324219,
+      "loss": 2.1687,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6025928258895874,
+      "rewards/margins": 0.7339029312133789,
+      "rewards/rejected": -2.3364956378936768,
+      "step": 306
+    },
+    {
+      "epoch": 0.5576748410535877,
+      "grad_norm": 1.497698187828064,
+      "learning_rate": 9.57445363502194e-06,
+      "logits/chosen": 0.1225174218416214,
+      "logits/rejected": 0.09580346196889877,
+      "logps/chosen": -76.27323150634766,
+      "logps/rejected": -76.90716552734375,
+      "loss": 2.2116,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.715336799621582,
+      "rewards/margins": 0.4648086428642273,
+      "rewards/rejected": -2.180145263671875,
+      "step": 307
+    },
+    {
+      "epoch": 0.5594913714804723,
+      "grad_norm": 1.3957064151763916,
+      "learning_rate": 9.571721736097089e-06,
+      "logits/chosen": 0.07596514374017715,
+      "logits/rejected": 0.14832191169261932,
+      "logps/chosen": -84.65274810791016,
+      "logps/rejected": -105.76422119140625,
+      "loss": 1.9942,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.6643587350845337,
+      "rewards/margins": 0.7884883284568787,
+      "rewards/rejected": -2.4528470039367676,
+      "step": 308
+    },
+    {
+      "epoch": 0.5613079019073569,
+      "grad_norm": 1.6824349164962769,
+      "learning_rate": 9.568981534222664e-06,
+      "logits/chosen": 0.049865882843732834,
+      "logits/rejected": 0.023348212242126465,
+      "logps/chosen": -80.3697509765625,
+      "logps/rejected": -86.07074737548828,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.009049654006958,
+      "rewards/margins": 0.5470253825187683,
+      "rewards/rejected": -2.556075096130371,
+      "step": 309
+    },
+    {
+      "epoch": 0.5631244323342416,
+      "grad_norm": 1.8752086162567139,
+      "learning_rate": 9.566233034986413e-06,
+      "logits/chosen": 0.07250591367483139,
+      "logits/rejected": 0.12809628248214722,
+      "logps/chosen": -74.84794616699219,
+      "logps/rejected": -81.79740905761719,
+      "loss": 2.6286,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.75887131690979,
+      "rewards/margins": 0.33979111909866333,
+      "rewards/rejected": -2.0986623764038086,
+      "step": 310
+    },
+    {
+      "epoch": 0.5649409627611263,
+      "grad_norm": 1.7387233972549438,
+      "learning_rate": 9.563476243993008e-06,
+      "logits/chosen": 0.130618155002594,
+      "logits/rejected": 0.12952059507369995,
+      "logps/chosen": -80.75495910644531,
+      "logps/rejected": -90.4281234741211,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.836004614830017,
+      "rewards/margins": 0.4418962597846985,
+      "rewards/rejected": -2.2779006958007812,
+      "step": 311
+    },
+    {
+      "epoch": 0.5667574931880109,
+      "grad_norm": 1.5147444009780884,
+      "learning_rate": 9.56071116686402e-06,
+      "logits/chosen": 0.10193713754415512,
+      "logits/rejected": 0.22481802105903625,
+      "logps/chosen": -73.21926879882812,
+      "logps/rejected": -81.27864837646484,
+      "loss": 2.578,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6358015537261963,
+      "rewards/margins": 0.38279739022254944,
+      "rewards/rejected": -2.018598794937134,
+      "step": 312
+    },
+    {
+      "epoch": 0.5685740236148955,
+      "grad_norm": 1.6253665685653687,
+      "learning_rate": 9.557937809237927e-06,
+      "logits/chosen": 0.09468917548656464,
+      "logits/rejected": 0.09415112435817719,
+      "logps/chosen": -80.62995147705078,
+      "logps/rejected": -86.65946197509766,
+      "loss": 2.3854,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8635404109954834,
+      "rewards/margins": 0.4438764154911041,
+      "rewards/rejected": -2.3074169158935547,
+      "step": 313
+    },
+    {
+      "epoch": 0.5703905540417802,
+      "grad_norm": 1.7893344163894653,
+      "learning_rate": 9.555156176770087e-06,
+      "logits/chosen": 0.15863659977912903,
+      "logits/rejected": 0.09485571831464767,
+      "logps/chosen": -76.10442352294922,
+      "logps/rejected": -72.73162078857422,
+      "loss": 2.6795,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.935206413269043,
+      "rewards/margins": 0.22100940346717834,
+      "rewards/rejected": -2.1562156677246094,
+      "step": 314
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 1.696327805519104,
+      "learning_rate": 9.552366275132733e-06,
+      "logits/chosen": 0.07012113183736801,
+      "logits/rejected": 0.13891686499118805,
+      "logps/chosen": -78.08012390136719,
+      "logps/rejected": -83.05044555664062,
+      "loss": 2.6181,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7869051694869995,
+      "rewards/margins": 0.23424415290355682,
+      "rewards/rejected": -2.0211493968963623,
+      "step": 315
+    },
+    {
+      "epoch": 0.5740236148955495,
+      "grad_norm": 1.6044729948043823,
+      "learning_rate": 9.54956811001496e-06,
+      "logits/chosen": 0.11298641562461853,
+      "logits/rejected": 0.2595204710960388,
+      "logps/chosen": -75.5998306274414,
+      "logps/rejected": -85.31849670410156,
+      "loss": 2.5849,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.5757699012756348,
+      "rewards/margins": 0.35041162371635437,
+      "rewards/rejected": -1.9261815547943115,
+      "step": 316
+    },
+    {
+      "epoch": 0.5758401453224341,
+      "grad_norm": 1.7399548292160034,
+      "learning_rate": 9.546761687122715e-06,
+      "logits/chosen": 0.14933931827545166,
+      "logits/rejected": 0.1868455857038498,
+      "logps/chosen": -68.05965423583984,
+      "logps/rejected": -82.37442779541016,
+      "loss": 2.3208,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8474600315093994,
+      "rewards/margins": 0.5766161680221558,
+      "rewards/rejected": -2.4240763187408447,
+      "step": 317
+    },
+    {
+      "epoch": 0.5776566757493188,
+      "grad_norm": 1.597495675086975,
+      "learning_rate": 9.54394701217878e-06,
+      "logits/chosen": 0.04729313403367996,
+      "logits/rejected": 0.05815067142248154,
+      "logps/chosen": -97.36541748046875,
+      "logps/rejected": -104.02650451660156,
+      "loss": 2.2106,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8631618022918701,
+      "rewards/margins": 0.6027428507804871,
+      "rewards/rejected": -2.465904474258423,
+      "step": 318
+    },
+    {
+      "epoch": 0.5794732061762035,
+      "grad_norm": 1.8646149635314941,
+      "learning_rate": 9.541124090922771e-06,
+      "logits/chosen": 0.1399674415588379,
+      "logits/rejected": 0.11107950657606125,
+      "logps/chosen": -82.39740753173828,
+      "logps/rejected": -76.80902099609375,
+      "loss": 2.6435,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.7670300006866455,
+      "rewards/margins": 0.15345275402069092,
+      "rewards/rejected": -1.920482873916626,
+      "step": 319
+    },
+    {
+      "epoch": 0.5812897366030881,
+      "grad_norm": 1.502447247505188,
+      "learning_rate": 9.538292929111114e-06,
+      "logits/chosen": 0.10672347247600555,
+      "logits/rejected": 0.08290571719408035,
+      "logps/chosen": -74.72683715820312,
+      "logps/rejected": -86.37158203125,
+      "loss": 2.2137,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8041073083877563,
+      "rewards/margins": 0.596272885799408,
+      "rewards/rejected": -2.4003803730010986,
+      "step": 320
+    },
+    {
+      "epoch": 0.5831062670299727,
+      "grad_norm": 1.7287745475769043,
+      "learning_rate": 9.535453532517039e-06,
+      "logits/chosen": 0.13742896914482117,
+      "logits/rejected": 0.07607734203338623,
+      "logps/chosen": -86.82475280761719,
+      "logps/rejected": -77.8934555053711,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7918696403503418,
+      "rewards/margins": 0.2591831088066101,
+      "rewards/rejected": -2.0510525703430176,
+      "step": 321
+    },
+    {
+      "epoch": 0.5849227974568574,
+      "grad_norm": 1.5914596319198608,
+      "learning_rate": 9.532605906930575e-06,
+      "logits/chosen": 0.11301672458648682,
+      "logits/rejected": 0.1944851577281952,
+      "logps/chosen": -68.9898452758789,
+      "logps/rejected": -73.17451477050781,
+      "loss": 2.4565,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6999026536941528,
+      "rewards/margins": 0.30462026596069336,
+      "rewards/rejected": -2.0045228004455566,
+      "step": 322
+    },
+    {
+      "epoch": 0.5867393278837421,
+      "grad_norm": 1.6072031259536743,
+      "learning_rate": 9.529750058158522e-06,
+      "logits/chosen": 0.07092760503292084,
+      "logits/rejected": 0.08035591244697571,
+      "logps/chosen": -81.5555419921875,
+      "logps/rejected": -81.93560028076172,
+      "loss": 2.2729,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6834638118743896,
+      "rewards/margins": 0.48882579803466797,
+      "rewards/rejected": -2.1722893714904785,
+      "step": 323
+    },
+    {
+      "epoch": 0.5885558583106267,
+      "grad_norm": 1.6433221101760864,
+      "learning_rate": 9.526885992024453e-06,
+      "logits/chosen": 0.13823899626731873,
+      "logits/rejected": 0.13610433042049408,
+      "logps/chosen": -78.9625244140625,
+      "logps/rejected": -83.2666015625,
+      "loss": 2.2075,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6939194202423096,
+      "rewards/margins": 0.5755473971366882,
+      "rewards/rejected": -2.2694668769836426,
+      "step": 324
+    },
+    {
+      "epoch": 0.5903723887375113,
+      "grad_norm": 1.7674918174743652,
+      "learning_rate": 9.524013714368702e-06,
+      "logits/chosen": 0.1704932153224945,
+      "logits/rejected": 0.1073535829782486,
+      "logps/chosen": -72.47895050048828,
+      "logps/rejected": -73.0809326171875,
+      "loss": 2.5473,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8390110731124878,
+      "rewards/margins": 0.34467947483062744,
+      "rewards/rejected": -2.1836905479431152,
+      "step": 325
+    },
+    {
+      "epoch": 0.592188919164396,
+      "grad_norm": 1.8219250440597534,
+      "learning_rate": 9.521133231048338e-06,
+      "logits/chosen": 0.07941028475761414,
+      "logits/rejected": 0.13045310974121094,
+      "logps/chosen": -80.73494720458984,
+      "logps/rejected": -91.58990478515625,
+      "loss": 2.4866,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8594518899917603,
+      "rewards/margins": 0.4554288685321808,
+      "rewards/rejected": -2.314880847930908,
+      "step": 326
+    },
+    {
+      "epoch": 0.5940054495912807,
+      "grad_norm": 1.3753328323364258,
+      "learning_rate": 9.51824454793717e-06,
+      "logits/chosen": 0.08879546821117401,
+      "logits/rejected": 0.04693777486681938,
+      "logps/chosen": -82.9569091796875,
+      "logps/rejected": -91.44571685791016,
+      "loss": 2.2602,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.549119472503662,
+      "rewards/margins": 0.5145683288574219,
+      "rewards/rejected": -2.063688039779663,
+      "step": 327
+    },
+    {
+      "epoch": 0.5958219800181653,
+      "grad_norm": 1.4296562671661377,
+      "learning_rate": 9.515347670925728e-06,
+      "logits/chosen": 0.15614314377307892,
+      "logits/rejected": 0.1598319411277771,
+      "logps/chosen": -75.60345458984375,
+      "logps/rejected": -80.81770324707031,
+      "loss": 2.3946,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.556883454322815,
+      "rewards/margins": 0.4525718092918396,
+      "rewards/rejected": -2.0094552040100098,
+      "step": 328
+    },
+    {
+      "epoch": 0.59763851044505,
+      "grad_norm": 1.7932195663452148,
+      "learning_rate": 9.512442605921245e-06,
+      "logits/chosen": 0.05863601714372635,
+      "logits/rejected": 0.0788806602358818,
+      "logps/chosen": -75.95397186279297,
+      "logps/rejected": -84.50177001953125,
+      "loss": 2.5085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9159862995147705,
+      "rewards/margins": 0.4600725769996643,
+      "rewards/rejected": -2.376059055328369,
+      "step": 329
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 1.6558383703231812,
+      "learning_rate": 9.509529358847655e-06,
+      "logits/chosen": 0.08205496519804001,
+      "logits/rejected": 0.13091425597667694,
+      "logps/chosen": -82.73474884033203,
+      "logps/rejected": -94.02273559570312,
+      "loss": 2.4419,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.083285093307495,
+      "rewards/margins": 0.5300815105438232,
+      "rewards/rejected": -2.6133666038513184,
+      "step": 330
+    },
+    {
+      "epoch": 0.6012715712988193,
+      "grad_norm": 1.7888520956039429,
+      "learning_rate": 9.506607935645579e-06,
+      "logits/chosen": 0.11793217062950134,
+      "logits/rejected": 0.19568441808223724,
+      "logps/chosen": -79.86200714111328,
+      "logps/rejected": -89.25471496582031,
+      "loss": 2.4273,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9273384809494019,
+      "rewards/margins": 0.5115458965301514,
+      "rewards/rejected": -2.4388844966888428,
+      "step": 331
+    },
+    {
+      "epoch": 0.6030881017257039,
+      "grad_norm": 2.0436160564422607,
+      "learning_rate": 9.503678342272306e-06,
+      "logits/chosen": 0.043473344296216965,
+      "logits/rejected": 0.15554016828536987,
+      "logps/chosen": -75.2901611328125,
+      "logps/rejected": -83.20555877685547,
+      "loss": 2.6941,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.725769281387329,
+      "rewards/margins": 0.2789250314235687,
+      "rewards/rejected": -2.0046942234039307,
+      "step": 332
+    },
+    {
+      "epoch": 0.6049046321525886,
+      "grad_norm": 1.665578842163086,
+      "learning_rate": 9.500740584701785e-06,
+      "logits/chosen": 0.17119848728179932,
+      "logits/rejected": 0.14128939807415009,
+      "logps/chosen": -83.1322021484375,
+      "logps/rejected": -93.38603210449219,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.8975954055786133,
+      "rewards/margins": 0.7810046672821045,
+      "rewards/rejected": -2.6786000728607178,
+      "step": 333
+    },
+    {
+      "epoch": 0.6067211625794732,
+      "grad_norm": 1.9258419275283813,
+      "learning_rate": 9.497794668924617e-06,
+      "logits/chosen": 0.028591612353920937,
+      "logits/rejected": 0.08894480764865875,
+      "logps/chosen": -78.12230682373047,
+      "logps/rejected": -90.00184631347656,
+      "loss": 2.4006,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8113291263580322,
+      "rewards/margins": 0.5277000665664673,
+      "rewards/rejected": -2.339029550552368,
+      "step": 334
+    },
+    {
+      "epoch": 0.6085376930063578,
+      "grad_norm": 1.4878523349761963,
+      "learning_rate": 9.494840600948038e-06,
+      "logits/chosen": 0.03548199310898781,
+      "logits/rejected": 0.09170147776603699,
+      "logps/chosen": -73.15607452392578,
+      "logps/rejected": -83.77317810058594,
+      "loss": 2.3301,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6781296730041504,
+      "rewards/margins": 0.4735565185546875,
+      "rewards/rejected": -2.151685953140259,
+      "step": 335
+    },
+    {
+      "epoch": 0.6103542234332425,
+      "grad_norm": 1.8042774200439453,
+      "learning_rate": 9.491878386795906e-06,
+      "logits/chosen": 0.046132348477840424,
+      "logits/rejected": 0.0721711814403534,
+      "logps/chosen": -81.01045227050781,
+      "logps/rejected": -89.28679656982422,
+      "loss": 2.5593,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.012852191925049,
+      "rewards/margins": 0.44549623131752014,
+      "rewards/rejected": -2.458348512649536,
+      "step": 336
+    },
+    {
+      "epoch": 0.6121707538601272,
+      "grad_norm": 2.048952102661133,
+      "learning_rate": 9.488908032508691e-06,
+      "logits/chosen": 0.10774732381105423,
+      "logits/rejected": 0.11935572326183319,
+      "logps/chosen": -91.25210571289062,
+      "logps/rejected": -90.42224884033203,
+      "loss": 2.8282,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0127718448638916,
+      "rewards/margins": 0.16889013350009918,
+      "rewards/rejected": -2.18166184425354,
+      "step": 337
+    },
+    {
+      "epoch": 0.6139872842870118,
+      "grad_norm": 1.6447219848632812,
+      "learning_rate": 9.485929544143462e-06,
+      "logits/chosen": 0.12260966747999191,
+      "logits/rejected": 0.1517799198627472,
+      "logps/chosen": -72.49166870117188,
+      "logps/rejected": -82.3083724975586,
+      "loss": 2.4029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.624878168106079,
+      "rewards/margins": 0.5487539768218994,
+      "rewards/rejected": -2.1736321449279785,
+      "step": 338
+    },
+    {
+      "epoch": 0.6158038147138964,
+      "grad_norm": 1.599530577659607,
+      "learning_rate": 9.482942927773876e-06,
+      "logits/chosen": 0.09051798284053802,
+      "logits/rejected": 0.11202570050954819,
+      "logps/chosen": -84.42019653320312,
+      "logps/rejected": -92.61677551269531,
+      "loss": 2.2269,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6424546241760254,
+      "rewards/margins": 0.5715836882591248,
+      "rewards/rejected": -2.214038133621216,
+      "step": 339
+    },
+    {
+      "epoch": 0.6176203451407811,
+      "grad_norm": 1.5411081314086914,
+      "learning_rate": 9.479948189490164e-06,
+      "logits/chosen": 0.06563656777143478,
+      "logits/rejected": 0.17349205911159515,
+      "logps/chosen": -70.98219299316406,
+      "logps/rejected": -78.88846588134766,
+      "loss": 2.3328,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.824373722076416,
+      "rewards/margins": 0.46550822257995605,
+      "rewards/rejected": -2.289881706237793,
+      "step": 340
+    },
+    {
+      "epoch": 0.6194368755676658,
+      "grad_norm": 1.8855030536651611,
+      "learning_rate": 9.476945335399122e-06,
+      "logits/chosen": 0.11399642378091812,
+      "logits/rejected": 0.12690946459770203,
+      "logps/chosen": -85.86933898925781,
+      "logps/rejected": -89.75601196289062,
+      "loss": 2.4269,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9560539722442627,
+      "rewards/margins": 0.38427016139030457,
+      "rewards/rejected": -2.3403239250183105,
+      "step": 341
+    },
+    {
+      "epoch": 0.6212534059945504,
+      "grad_norm": 1.9973480701446533,
+      "learning_rate": 9.473934371624087e-06,
+      "logits/chosen": 0.029105912894010544,
+      "logits/rejected": 0.12763622403144836,
+      "logps/chosen": -80.68119049072266,
+      "logps/rejected": -93.90747833251953,
+      "loss": 2.4709,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.906477928161621,
+      "rewards/margins": 0.6422742009162903,
+      "rewards/rejected": -2.5487518310546875,
+      "step": 342
+    },
+    {
+      "epoch": 0.623069936421435,
+      "grad_norm": 1.7752224206924438,
+      "learning_rate": 9.47091530430494e-06,
+      "logits/chosen": 0.14096824824810028,
+      "logits/rejected": 0.1774113029241562,
+      "logps/chosen": -71.80216217041016,
+      "logps/rejected": -79.65301513671875,
+      "loss": 2.2678,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.5735952854156494,
+      "rewards/margins": 0.5823659300804138,
+      "rewards/rejected": -2.155961036682129,
+      "step": 343
+    },
+    {
+      "epoch": 0.6248864668483197,
+      "grad_norm": 1.855377197265625,
+      "learning_rate": 9.467888139598086e-06,
+      "logits/chosen": 0.10637074708938599,
+      "logits/rejected": 0.0798158049583435,
+      "logps/chosen": -79.01347351074219,
+      "logps/rejected": -80.77239990234375,
+      "loss": 2.7867,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9598251581192017,
+      "rewards/margins": 0.3184160590171814,
+      "rewards/rejected": -2.2782411575317383,
+      "step": 344
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 1.423040747642517,
+      "learning_rate": 9.464852883676441e-06,
+      "logits/chosen": 0.1304859220981598,
+      "logits/rejected": 0.1759231686592102,
+      "logps/chosen": -74.09733581542969,
+      "logps/rejected": -89.27589416503906,
+      "loss": 2.2777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7697277069091797,
+      "rewards/margins": 0.635611355304718,
+      "rewards/rejected": -2.405339002609253,
+      "step": 345
+    },
+    {
+      "epoch": 0.628519527702089,
+      "grad_norm": 1.6246038675308228,
+      "learning_rate": 9.461809542729421e-06,
+      "logits/chosen": 0.03750050812959671,
+      "logits/rejected": 0.10337980091571808,
+      "logps/chosen": -81.89110565185547,
+      "logps/rejected": -95.36811828613281,
+      "loss": 2.0438,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7299752235412598,
+      "rewards/margins": 0.7404756546020508,
+      "rewards/rejected": -2.4704508781433105,
+      "step": 346
+    },
+    {
+      "epoch": 0.6303360581289736,
+      "grad_norm": 1.6618752479553223,
+      "learning_rate": 9.458758122962926e-06,
+      "logits/chosen": 0.05359608680009842,
+      "logits/rejected": 0.10455545783042908,
+      "logps/chosen": -85.34078216552734,
+      "logps/rejected": -90.46200561523438,
+      "loss": 2.3299,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9331045150756836,
+      "rewards/margins": 0.4755082130432129,
+      "rewards/rejected": -2.4086129665374756,
+      "step": 347
+    },
+    {
+      "epoch": 0.6321525885558583,
+      "grad_norm": 1.5805696249008179,
+      "learning_rate": 9.455698630599332e-06,
+      "logits/chosen": 0.10048776119947433,
+      "logits/rejected": 0.1140337884426117,
+      "logps/chosen": -80.25875854492188,
+      "logps/rejected": -94.44998168945312,
+      "loss": 2.3439,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8421945571899414,
+      "rewards/margins": 0.5977468490600586,
+      "rewards/rejected": -2.43994140625,
+      "step": 348
+    },
+    {
+      "epoch": 0.633969118982743,
+      "grad_norm": 1.37093985080719,
+      "learning_rate": 9.452631071877478e-06,
+      "logits/chosen": 0.11764326691627502,
+      "logits/rejected": 0.10735289752483368,
+      "logps/chosen": -72.02367401123047,
+      "logps/rejected": -77.9522705078125,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.6527085304260254,
+      "rewards/margins": 0.6065118908882141,
+      "rewards/rejected": -2.259220600128174,
+      "step": 349
+    },
+    {
+      "epoch": 0.6357856494096276,
+      "grad_norm": 2.0214192867279053,
+      "learning_rate": 9.449555453052652e-06,
+      "logits/chosen": 0.13177426159381866,
+      "logits/rejected": 0.12408209592103958,
+      "logps/chosen": -76.62931823730469,
+      "logps/rejected": -81.16517639160156,
+      "loss": 2.9175,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.9575048685073853,
+      "rewards/margins": 0.21441945433616638,
+      "rewards/rejected": -2.171924352645874,
+      "step": 350
+    },
+    {
+      "epoch": 0.6376021798365122,
+      "grad_norm": 1.6041687726974487,
+      "learning_rate": 9.446471780396573e-06,
+      "logits/chosen": 0.18574532866477966,
+      "logits/rejected": 0.22683225572109222,
+      "logps/chosen": -73.95652770996094,
+      "logps/rejected": -80.29524993896484,
+      "loss": 2.3833,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8684748411178589,
+      "rewards/margins": 0.37268272042274475,
+      "rewards/rejected": -2.2411575317382812,
+      "step": 351
+    },
+    {
+      "epoch": 0.6394187102633969,
+      "grad_norm": 1.7410838603973389,
+      "learning_rate": 9.443380060197387e-06,
+      "logits/chosen": 0.07876043766736984,
+      "logits/rejected": 0.14103996753692627,
+      "logps/chosen": -74.64432525634766,
+      "logps/rejected": -83.6999740600586,
+      "loss": 2.5558,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.029916763305664,
+      "rewards/margins": 0.4088842272758484,
+      "rewards/rejected": -2.438800811767578,
+      "step": 352
+    },
+    {
+      "epoch": 0.6412352406902816,
+      "grad_norm": 1.87971031665802,
+      "learning_rate": 9.440280298759653e-06,
+      "logits/chosen": 0.13997013866901398,
+      "logits/rejected": 0.1322249174118042,
+      "logps/chosen": -88.78776550292969,
+      "logps/rejected": -88.22732543945312,
+      "loss": 2.6547,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9983258247375488,
+      "rewards/margins": 0.2458382397890091,
+      "rewards/rejected": -2.244164228439331,
+      "step": 353
+    },
+    {
+      "epoch": 0.6430517711171662,
+      "grad_norm": 1.7056363821029663,
+      "learning_rate": 9.437172502404318e-06,
+      "logits/chosen": 0.07248476892709732,
+      "logits/rejected": 0.13019773364067078,
+      "logps/chosen": -78.4591064453125,
+      "logps/rejected": -80.49126434326172,
+      "loss": 2.3809,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9740041494369507,
+      "rewards/margins": 0.4853326082229614,
+      "rewards/rejected": -2.459336757659912,
+      "step": 354
+    },
+    {
+      "epoch": 0.6448683015440508,
+      "grad_norm": 1.435718059539795,
+      "learning_rate": 9.434056677468726e-06,
+      "logits/chosen": 0.09164869040250778,
+      "logits/rejected": 0.14243285357952118,
+      "logps/chosen": -77.83367156982422,
+      "logps/rejected": -86.18836212158203,
+      "loss": 2.006,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.7355122566223145,
+      "rewards/margins": 0.6869419813156128,
+      "rewards/rejected": -2.422454357147217,
+      "step": 355
+    },
+    {
+      "epoch": 0.6466848319709355,
+      "grad_norm": 1.6553188562393188,
+      "learning_rate": 9.430932830306587e-06,
+      "logits/chosen": 0.04967673122882843,
+      "logits/rejected": 0.1456151008605957,
+      "logps/chosen": -73.50204467773438,
+      "logps/rejected": -85.31363677978516,
+      "loss": 2.4388,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.978384256362915,
+      "rewards/margins": 0.48918718099594116,
+      "rewards/rejected": -2.467571496963501,
+      "step": 356
+    },
+    {
+      "epoch": 0.6485013623978202,
+      "grad_norm": 1.8443480730056763,
+      "learning_rate": 9.427800967287963e-06,
+      "logits/chosen": 0.06455090641975403,
+      "logits/rejected": 0.12608012557029724,
+      "logps/chosen": -75.49232482910156,
+      "logps/rejected": -77.49136352539062,
+      "loss": 2.706,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.8730812072753906,
+      "rewards/margins": 0.2989741861820221,
+      "rewards/rejected": -2.17205548286438,
+      "step": 357
+    },
+    {
+      "epoch": 0.6503178928247049,
+      "grad_norm": 1.6912306547164917,
+      "learning_rate": 9.424661094799273e-06,
+      "logits/chosen": 0.11878645420074463,
+      "logits/rejected": 0.13628609478473663,
+      "logps/chosen": -73.09883880615234,
+      "logps/rejected": -84.19624328613281,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.928421139717102,
+      "rewards/margins": 0.5017052292823792,
+      "rewards/rejected": -2.430126190185547,
+      "step": 358
+    },
+    {
+      "epoch": 0.6521344232515894,
+      "grad_norm": 1.3643461465835571,
+      "learning_rate": 9.421513219243262e-06,
+      "logits/chosen": 0.07683826237916946,
+      "logits/rejected": 0.15765298902988434,
+      "logps/chosen": -76.0871353149414,
+      "logps/rejected": -97.47781372070312,
+      "loss": 1.7844,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.8300602436065674,
+      "rewards/margins": 0.911116361618042,
+      "rewards/rejected": -2.7411766052246094,
+      "step": 359
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 1.6327749490737915,
+      "learning_rate": 9.418357347038999e-06,
+      "logits/chosen": 0.1078951433300972,
+      "logits/rejected": 0.12233921140432358,
+      "logps/chosen": -75.89913177490234,
+      "logps/rejected": -78.4587173461914,
+      "loss": 2.4976,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.7992844581604004,
+      "rewards/margins": 0.311392605304718,
+      "rewards/rejected": -2.1106772422790527,
+      "step": 360
+    },
+    {
+      "epoch": 0.6557674841053588,
+      "grad_norm": 1.463025689125061,
+      "learning_rate": 9.415193484621852e-06,
+      "logits/chosen": 0.057331383228302,
+      "logits/rejected": 0.1563551127910614,
+      "logps/chosen": -80.7889633178711,
+      "logps/rejected": -93.21602630615234,
+      "loss": 2.1398,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.744189739227295,
+      "rewards/margins": 0.576555073261261,
+      "rewards/rejected": -2.320744752883911,
+      "step": 361
+    },
+    {
+      "epoch": 0.6575840145322435,
+      "grad_norm": 1.743695616722107,
+      "learning_rate": 9.412021638443491e-06,
+      "logits/chosen": 0.03781123086810112,
+      "logits/rejected": 0.154897540807724,
+      "logps/chosen": -77.72047424316406,
+      "logps/rejected": -88.80599975585938,
+      "loss": 2.2361,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9849369525909424,
+      "rewards/margins": 0.6562294363975525,
+      "rewards/rejected": -2.6411664485931396,
+      "step": 362
+    },
+    {
+      "epoch": 0.659400544959128,
+      "grad_norm": 1.5648279190063477,
+      "learning_rate": 9.408841814971862e-06,
+      "logits/chosen": 0.10401102900505066,
+      "logits/rejected": 0.12699122726917267,
+      "logps/chosen": -83.10671997070312,
+      "logps/rejected": -86.8663101196289,
+      "loss": 2.4009,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7048468589782715,
+      "rewards/margins": 0.4090174734592438,
+      "rewards/rejected": -2.1138644218444824,
+      "step": 363
+    },
+    {
+      "epoch": 0.6612170753860127,
+      "grad_norm": 1.3385239839553833,
+      "learning_rate": 9.405654020691178e-06,
+      "logits/chosen": 0.06100422143936157,
+      "logits/rejected": 0.10872650146484375,
+      "logps/chosen": -76.72837829589844,
+      "logps/rejected": -87.63470458984375,
+      "loss": 2.141,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.747226595878601,
+      "rewards/margins": 0.615075945854187,
+      "rewards/rejected": -2.362302780151367,
+      "step": 364
+    },
+    {
+      "epoch": 0.6630336058128974,
+      "grad_norm": 1.4060734510421753,
+      "learning_rate": 9.402458262101906e-06,
+      "logits/chosen": 0.10638861358165741,
+      "logits/rejected": 0.16951681673526764,
+      "logps/chosen": -77.35757446289062,
+      "logps/rejected": -92.03012084960938,
+      "loss": 2.0511,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8359463214874268,
+      "rewards/margins": 0.7444422841072083,
+      "rewards/rejected": -2.5803885459899902,
+      "step": 365
+    },
+    {
+      "epoch": 0.6648501362397821,
+      "grad_norm": 1.4371014833450317,
+      "learning_rate": 9.399254545720757e-06,
+      "logits/chosen": 0.0383220911026001,
+      "logits/rejected": 0.05856206640601158,
+      "logps/chosen": -82.40064239501953,
+      "logps/rejected": -92.1117935180664,
+      "loss": 2.1756,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8580785989761353,
+      "rewards/margins": 0.6475991010665894,
+      "rewards/rejected": -2.5056777000427246,
+      "step": 366
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.6673945188522339,
+      "learning_rate": 9.396042878080661e-06,
+      "logits/chosen": 0.14657820761203766,
+      "logits/rejected": 0.19021111726760864,
+      "logps/chosen": -73.56724548339844,
+      "logps/rejected": -78.85279846191406,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9079951047897339,
+      "rewards/margins": 0.3709837794303894,
+      "rewards/rejected": -2.2789790630340576,
+      "step": 367
+    },
+    {
+      "epoch": 0.6684831970935513,
+      "grad_norm": 1.7402448654174805,
+      "learning_rate": 9.392823265730775e-06,
+      "logits/chosen": 0.16515754163265228,
+      "logits/rejected": 0.12460774928331375,
+      "logps/chosen": -69.89266204833984,
+      "logps/rejected": -74.53837585449219,
+      "loss": 2.4262,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.0627622604370117,
+      "rewards/margins": 0.5317320823669434,
+      "rewards/rejected": -2.594494342803955,
+      "step": 368
+    },
+    {
+      "epoch": 0.670299727520436,
+      "grad_norm": 1.5290364027023315,
+      "learning_rate": 9.389595715236446e-06,
+      "logits/chosen": 0.0954434722661972,
+      "logits/rejected": 0.16226956248283386,
+      "logps/chosen": -77.17019653320312,
+      "logps/rejected": -91.07398986816406,
+      "loss": 2.2511,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.241170644760132,
+      "rewards/margins": 0.6319360136985779,
+      "rewards/rejected": -2.8731067180633545,
+      "step": 369
+    },
+    {
+      "epoch": 0.6721162579473207,
+      "grad_norm": 1.4633573293685913,
+      "learning_rate": 9.386360233179206e-06,
+      "logits/chosen": 0.06517557799816132,
+      "logits/rejected": 0.07250035554170609,
+      "logps/chosen": -79.23770141601562,
+      "logps/rejected": -88.50403594970703,
+      "loss": 2.3858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.751698613166809,
+      "rewards/margins": 0.49529415369033813,
+      "rewards/rejected": -2.246993064880371,
+      "step": 370
+    },
+    {
+      "epoch": 0.6739327883742052,
+      "grad_norm": 2.007215976715088,
+      "learning_rate": 9.383116826156775e-06,
+      "logits/chosen": 0.13584929704666138,
+      "logits/rejected": 0.10950647294521332,
+      "logps/chosen": -80.75408935546875,
+      "logps/rejected": -79.3903579711914,
+      "loss": 2.6967,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.2421724796295166,
+      "rewards/margins": 0.34328368306159973,
+      "rewards/rejected": -2.585456132888794,
+      "step": 371
+    },
+    {
+      "epoch": 0.6757493188010899,
+      "grad_norm": 1.7323477268218994,
+      "learning_rate": 9.37986550078302e-06,
+      "logits/chosen": 0.011626070365309715,
+      "logits/rejected": 0.06207559257745743,
+      "logps/chosen": -75.51494598388672,
+      "logps/rejected": -85.83218383789062,
+      "loss": 2.4608,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9569413661956787,
+      "rewards/margins": 0.5160585641860962,
+      "rewards/rejected": -2.4729998111724854,
+      "step": 372
+    },
+    {
+      "epoch": 0.6775658492279746,
+      "grad_norm": 1.6365752220153809,
+      "learning_rate": 9.376606263687959e-06,
+      "logits/chosen": 0.10213632136583328,
+      "logits/rejected": 0.08355780690908432,
+      "logps/chosen": -79.01370239257812,
+      "logps/rejected": -81.9417953491211,
+      "loss": 2.4701,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.064467668533325,
+      "rewards/margins": 0.4157037138938904,
+      "rewards/rejected": -2.4801712036132812,
+      "step": 373
+    },
+    {
+      "epoch": 0.6793823796548593,
+      "grad_norm": 1.5918922424316406,
+      "learning_rate": 9.373339121517748e-06,
+      "logits/chosen": 0.09486684203147888,
+      "logits/rejected": 0.07311725616455078,
+      "logps/chosen": -82.15351867675781,
+      "logps/rejected": -91.7690658569336,
+      "loss": 2.1297,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0153114795684814,
+      "rewards/margins": 0.669613242149353,
+      "rewards/rejected": -2.684924840927124,
+      "step": 374
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 1.6511566638946533,
+      "learning_rate": 9.370064080934654e-06,
+      "logits/chosen": 0.1406637728214264,
+      "logits/rejected": 0.20172299444675446,
+      "logps/chosen": -69.41023254394531,
+      "logps/rejected": -74.65138244628906,
+      "loss": 2.4704,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.061182737350464,
+      "rewards/margins": 0.3090137243270874,
+      "rewards/rejected": -2.3701963424682617,
+      "step": 375
+    },
+    {
+      "epoch": 0.6830154405086285,
+      "grad_norm": 1.770624041557312,
+      "learning_rate": 9.366781148617056e-06,
+      "logits/chosen": 0.10267248749732971,
+      "logits/rejected": 0.07900385558605194,
+      "logps/chosen": -77.27003479003906,
+      "logps/rejected": -88.09040832519531,
+      "loss": 2.2813,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.132927656173706,
+      "rewards/margins": 0.6629016399383545,
+      "rewards/rejected": -2.7958290576934814,
+      "step": 376
+    },
+    {
+      "epoch": 0.6848319709355132,
+      "grad_norm": 1.5172028541564941,
+      "learning_rate": 9.363490331259426e-06,
+      "logits/chosen": 0.05240853130817413,
+      "logits/rejected": 0.10445387661457062,
+      "logps/chosen": -75.5218276977539,
+      "logps/rejected": -85.49366760253906,
+      "loss": 2.1442,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9302880764007568,
+      "rewards/margins": 0.5776917338371277,
+      "rewards/rejected": -2.5079798698425293,
+      "step": 377
+    },
+    {
+      "epoch": 0.6866485013623979,
+      "grad_norm": 1.6540950536727905,
+      "learning_rate": 9.360191635572313e-06,
+      "logits/chosen": 0.14195458590984344,
+      "logits/rejected": 0.0978107899427414,
+      "logps/chosen": -85.00608825683594,
+      "logps/rejected": -85.1456298828125,
+      "loss": 2.423,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8765841722488403,
+      "rewards/margins": 0.49441370368003845,
+      "rewards/rejected": -2.370997905731201,
+      "step": 378
+    },
+    {
+      "epoch": 0.6884650317892824,
+      "grad_norm": 1.6463801860809326,
+      "learning_rate": 9.356885068282334e-06,
+      "logits/chosen": 0.13197994232177734,
+      "logits/rejected": 0.09924699366092682,
+      "logps/chosen": -86.94219970703125,
+      "logps/rejected": -86.85704803466797,
+      "loss": 2.4801,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9674016237258911,
+      "rewards/margins": 0.4530283510684967,
+      "rewards/rejected": -2.4204299449920654,
+      "step": 379
+    },
+    {
+      "epoch": 0.6902815622161671,
+      "grad_norm": 1.5428895950317383,
+      "learning_rate": 9.353570636132151e-06,
+      "logits/chosen": 0.10434838384389877,
+      "logits/rejected": 0.12619757652282715,
+      "logps/chosen": -78.56607818603516,
+      "logps/rejected": -86.33320617675781,
+      "loss": 2.2711,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9667280912399292,
+      "rewards/margins": 0.6350463628768921,
+      "rewards/rejected": -2.601774215698242,
+      "step": 380
+    },
+    {
+      "epoch": 0.6920980926430518,
+      "grad_norm": 1.6211251020431519,
+      "learning_rate": 9.350248345880471e-06,
+      "logits/chosen": 0.14081251621246338,
+      "logits/rejected": 0.1453506350517273,
+      "logps/chosen": -74.38184356689453,
+      "logps/rejected": -80.61524200439453,
+      "loss": 2.3863,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.0189671516418457,
+      "rewards/margins": 0.4528267979621887,
+      "rewards/rejected": -2.4717938899993896,
+      "step": 381
+    },
+    {
+      "epoch": 0.6939146230699365,
+      "grad_norm": 2.6346240043640137,
+      "learning_rate": 9.346918204302022e-06,
+      "logits/chosen": 0.0519830696284771,
+      "logits/rejected": 0.04284125566482544,
+      "logps/chosen": -82.38153839111328,
+      "logps/rejected": -86.35488891601562,
+      "loss": 2.5434,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9824391603469849,
+      "rewards/margins": 0.4156208634376526,
+      "rewards/rejected": -2.3980600833892822,
+      "step": 382
+    },
+    {
+      "epoch": 0.695731153496821,
+      "grad_norm": 1.4762026071548462,
+      "learning_rate": 9.343580218187544e-06,
+      "logits/chosen": 0.02196469157934189,
+      "logits/rejected": 0.04618150740861893,
+      "logps/chosen": -72.66182708740234,
+      "logps/rejected": -88.68983459472656,
+      "loss": 2.1699,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9227774143218994,
+      "rewards/margins": 0.9200201630592346,
+      "rewards/rejected": -2.8427975177764893,
+      "step": 383
+    },
+    {
+      "epoch": 0.6975476839237057,
+      "grad_norm": 1.9675711393356323,
+      "learning_rate": 9.340234394343768e-06,
+      "logits/chosen": 0.12956203520298004,
+      "logits/rejected": 0.16079677641391754,
+      "logps/chosen": -87.38492584228516,
+      "logps/rejected": -87.80662536621094,
+      "loss": 2.3138,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.144991397857666,
+      "rewards/margins": 0.4893474578857422,
+      "rewards/rejected": -2.634338855743408,
+      "step": 384
+    },
+    {
+      "epoch": 0.6993642143505904,
+      "grad_norm": 1.9116860628128052,
+      "learning_rate": 9.336880739593415e-06,
+      "logits/chosen": 0.06013559550046921,
+      "logits/rejected": 0.09889352321624756,
+      "logps/chosen": -86.55718231201172,
+      "logps/rejected": -91.01994323730469,
+      "loss": 2.4864,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9528815746307373,
+      "rewards/margins": 0.44100096821784973,
+      "rewards/rejected": -2.3938825130462646,
+      "step": 385
+    },
+    {
+      "epoch": 0.701180744777475,
+      "grad_norm": 2.0582289695739746,
+      "learning_rate": 9.33351926077517e-06,
+      "logits/chosen": 0.11538423597812653,
+      "logits/rejected": 0.17568480968475342,
+      "logps/chosen": -76.24241638183594,
+      "logps/rejected": -87.1357650756836,
+      "loss": 2.6878,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.102504253387451,
+      "rewards/margins": 0.3671773076057434,
+      "rewards/rejected": -2.469681739807129,
+      "step": 386
+    },
+    {
+      "epoch": 0.7029972752043597,
+      "grad_norm": 1.6962640285491943,
+      "learning_rate": 9.330149964743674e-06,
+      "logits/chosen": 0.12425235658884048,
+      "logits/rejected": 0.1852879822254181,
+      "logps/chosen": -82.67887115478516,
+      "logps/rejected": -98.66133117675781,
+      "loss": 2.343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.182525634765625,
+      "rewards/margins": 0.6602460145950317,
+      "rewards/rejected": -2.842771530151367,
+      "step": 387
+    },
+    {
+      "epoch": 0.7048138056312443,
+      "grad_norm": 1.6582266092300415,
+      "learning_rate": 9.326772858369506e-06,
+      "logits/chosen": 0.14438432455062866,
+      "logits/rejected": 0.0922938883304596,
+      "logps/chosen": -78.87794494628906,
+      "logps/rejected": -84.929443359375,
+      "loss": 2.3036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.988234281539917,
+      "rewards/margins": 0.4737466275691986,
+      "rewards/rejected": -2.4619810581207275,
+      "step": 388
+    },
+    {
+      "epoch": 0.706630336058129,
+      "grad_norm": 1.6689919233322144,
+      "learning_rate": 9.323387948539176e-06,
+      "logits/chosen": 0.1282673329114914,
+      "logits/rejected": 0.13633723556995392,
+      "logps/chosen": -70.22606658935547,
+      "logps/rejected": -79.26126861572266,
+      "loss": 2.241,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9162399768829346,
+      "rewards/margins": 0.5345122814178467,
+      "rewards/rejected": -2.4507524967193604,
+      "step": 389
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 1.5861207246780396,
+      "learning_rate": 9.319995242155102e-06,
+      "logits/chosen": 0.13024169206619263,
+      "logits/rejected": 0.14390775561332703,
+      "logps/chosen": -87.00923919677734,
+      "logps/rejected": -91.24594116210938,
+      "loss": 2.284,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9127002954483032,
+      "rewards/margins": 0.5265774130821228,
+      "rewards/rejected": -2.4392776489257812,
+      "step": 390
+    },
+    {
+      "epoch": 0.7102633969118983,
+      "grad_norm": 2.0118792057037354,
+      "learning_rate": 9.316594746135608e-06,
+      "logits/chosen": 0.14648675918579102,
+      "logits/rejected": 0.11786539107561111,
+      "logps/chosen": -82.68962097167969,
+      "logps/rejected": -88.26126861572266,
+      "loss": 2.7281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9463748931884766,
+      "rewards/margins": 0.20282456278800964,
+      "rewards/rejected": -2.1491994857788086,
+      "step": 391
+    },
+    {
+      "epoch": 0.7120799273387829,
+      "grad_norm": 1.8176332712173462,
+      "learning_rate": 9.313186467414892e-06,
+      "logits/chosen": 0.12084021419286728,
+      "logits/rejected": 0.1433698982000351,
+      "logps/chosen": -73.46707153320312,
+      "logps/rejected": -82.46024322509766,
+      "loss": 2.5247,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0802624225616455,
+      "rewards/margins": 0.41601985692977905,
+      "rewards/rejected": -2.4962823390960693,
+      "step": 392
+    },
+    {
+      "epoch": 0.7138964577656676,
+      "grad_norm": 1.7207351922988892,
+      "learning_rate": 9.30977041294303e-06,
+      "logits/chosen": 0.11983273923397064,
+      "logits/rejected": 0.16106371581554413,
+      "logps/chosen": -76.09751892089844,
+      "logps/rejected": -85.20323944091797,
+      "loss": 2.5284,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9018669128417969,
+      "rewards/margins": 0.4894491136074066,
+      "rewards/rejected": -2.3913159370422363,
+      "step": 393
+    },
+    {
+      "epoch": 0.7157129881925522,
+      "grad_norm": 1.6489367485046387,
+      "learning_rate": 9.306346589685956e-06,
+      "logits/chosen": 0.14845696091651917,
+      "logits/rejected": 0.10356368124485016,
+      "logps/chosen": -83.94526672363281,
+      "logps/rejected": -86.53378295898438,
+      "loss": 2.3394,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9089899063110352,
+      "rewards/margins": 0.4619474411010742,
+      "rewards/rejected": -2.3709375858306885,
+      "step": 394
+    },
+    {
+      "epoch": 0.7175295186194369,
+      "grad_norm": 1.792077898979187,
+      "learning_rate": 9.302915004625435e-06,
+      "logits/chosen": 0.06936248391866684,
+      "logits/rejected": 0.13349927961826324,
+      "logps/chosen": -91.82075500488281,
+      "logps/rejected": -98.7391357421875,
+      "loss": 2.3875,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9371806383132935,
+      "rewards/margins": 0.6781354546546936,
+      "rewards/rejected": -2.6153156757354736,
+      "step": 395
+    },
+    {
+      "epoch": 0.7193460490463215,
+      "grad_norm": 2.049694538116455,
+      "learning_rate": 9.29947566475907e-06,
+      "logits/chosen": 0.08474650233983994,
+      "logits/rejected": 0.13380834460258484,
+      "logps/chosen": -84.59528350830078,
+      "logps/rejected": -96.24267578125,
+      "loss": 2.7001,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9375882148742676,
+      "rewards/margins": 0.4607384204864502,
+      "rewards/rejected": -2.3983266353607178,
+      "step": 396
+    },
+    {
+      "epoch": 0.7211625794732062,
+      "grad_norm": 1.9235875606536865,
+      "learning_rate": 9.296028577100271e-06,
+      "logits/chosen": 0.06255945563316345,
+      "logits/rejected": 0.17074811458587646,
+      "logps/chosen": -68.99498748779297,
+      "logps/rejected": -84.45899963378906,
+      "loss": 2.347,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9344629049301147,
+      "rewards/margins": 0.688963770866394,
+      "rewards/rejected": -2.6234264373779297,
+      "step": 397
+    },
+    {
+      "epoch": 0.7229791099000908,
+      "grad_norm": 1.6617276668548584,
+      "learning_rate": 9.292573748678254e-06,
+      "logits/chosen": 0.11962493509054184,
+      "logits/rejected": 0.11647717654705048,
+      "logps/chosen": -85.3626708984375,
+      "logps/rejected": -95.70562744140625,
+      "loss": 2.2229,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.8968594074249268,
+      "rewards/margins": 0.6918852925300598,
+      "rewards/rejected": -2.588744640350342,
+      "step": 398
+    },
+    {
+      "epoch": 0.7247956403269755,
+      "grad_norm": 1.4650444984436035,
+      "learning_rate": 9.289111186538013e-06,
+      "logits/chosen": 0.06390775740146637,
+      "logits/rejected": 0.11076060682535172,
+      "logps/chosen": -69.9189453125,
+      "logps/rejected": -81.5618667602539,
+      "loss": 2.2271,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.857581377029419,
+      "rewards/margins": 0.5540103316307068,
+      "rewards/rejected": -2.4115917682647705,
+      "step": 399
+    },
+    {
+      "epoch": 0.7266121707538601,
+      "grad_norm": 1.5728726387023926,
+      "learning_rate": 9.285640897740316e-06,
+      "logits/chosen": 0.08816932141780853,
+      "logits/rejected": 0.1330798715353012,
+      "logps/chosen": -78.04735565185547,
+      "logps/rejected": -87.89193725585938,
+      "loss": 2.3595,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.823161244392395,
+      "rewards/margins": 0.4568220376968384,
+      "rewards/rejected": -2.2799830436706543,
+      "step": 400
+    },
+    {
+      "epoch": 0.7284287011807448,
+      "grad_norm": 1.6291779279708862,
+      "learning_rate": 9.282162889361686e-06,
+      "logits/chosen": 0.17718347907066345,
+      "logits/rejected": 0.20603135228157043,
+      "logps/chosen": -67.99884033203125,
+      "logps/rejected": -75.93153381347656,
+      "loss": 2.3002,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.671633243560791,
+      "rewards/margins": 0.45543500781059265,
+      "rewards/rejected": -2.127068281173706,
+      "step": 401
+    },
+    {
+      "epoch": 0.7302452316076294,
+      "grad_norm": 1.5066276788711548,
+      "learning_rate": 9.278677168494388e-06,
+      "logits/chosen": 0.16840124130249023,
+      "logits/rejected": 0.17102479934692383,
+      "logps/chosen": -78.1556625366211,
+      "logps/rejected": -84.91495513916016,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9279460906982422,
+      "rewards/margins": 0.6659258008003235,
+      "rewards/rejected": -2.593871831893921,
+      "step": 402
+    },
+    {
+      "epoch": 0.7320617620345141,
+      "grad_norm": 1.71995210647583,
+      "learning_rate": 9.275183742246412e-06,
+      "logits/chosen": -0.0005522281862795353,
+      "logits/rejected": 0.1412215530872345,
+      "logps/chosen": -73.37832641601562,
+      "logps/rejected": -92.71226501464844,
+      "loss": 2.3315,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6880759000778198,
+      "rewards/margins": 0.5844776630401611,
+      "rewards/rejected": -2.2725534439086914,
+      "step": 403
+    },
+    {
+      "epoch": 0.7338782924613987,
+      "grad_norm": 1.676321029663086,
+      "learning_rate": 9.271682617741466e-06,
+      "logits/chosen": 0.15319044888019562,
+      "logits/rejected": 0.14503052830696106,
+      "logps/chosen": -75.85136413574219,
+      "logps/rejected": -82.61698150634766,
+      "loss": 2.3528,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.819403052330017,
+      "rewards/margins": 0.3919827342033386,
+      "rewards/rejected": -2.211385726928711,
+      "step": 404
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 1.4811294078826904,
+      "learning_rate": 9.268173802118949e-06,
+      "logits/chosen": 0.07122528553009033,
+      "logits/rejected": 0.12146291136741638,
+      "logps/chosen": -76.78916931152344,
+      "logps/rejected": -85.33805084228516,
+      "loss": 2.1505,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9886473417282104,
+      "rewards/margins": 0.5911746025085449,
+      "rewards/rejected": -2.579822301864624,
+      "step": 405
+    },
+    {
+      "epoch": 0.737511353315168,
+      "grad_norm": 1.629601001739502,
+      "learning_rate": 9.264657302533947e-06,
+      "logits/chosen": 0.11274963617324829,
+      "logits/rejected": 0.0651293620467186,
+      "logps/chosen": -82.86985778808594,
+      "logps/rejected": -88.68803405761719,
+      "loss": 2.1873,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8538885116577148,
+      "rewards/margins": 0.6726438403129578,
+      "rewards/rejected": -2.5265324115753174,
+      "step": 406
+    },
+    {
+      "epoch": 0.7393278837420527,
+      "grad_norm": 1.5687483549118042,
+      "learning_rate": 9.261133126157218e-06,
+      "logits/chosen": 0.06054290384054184,
+      "logits/rejected": 0.1543959081172943,
+      "logps/chosen": -68.5821304321289,
+      "logps/rejected": -80.544921875,
+      "loss": 2.3587,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7550634145736694,
+      "rewards/margins": 0.46927812695503235,
+      "rewards/rejected": -2.224341630935669,
+      "step": 407
+    },
+    {
+      "epoch": 0.7411444141689373,
+      "grad_norm": 1.849908471107483,
+      "learning_rate": 9.257601280175167e-06,
+      "logits/chosen": 0.15177126228809357,
+      "logits/rejected": 0.1485670655965805,
+      "logps/chosen": -72.81680297851562,
+      "logps/rejected": -80.58684539794922,
+      "loss": 2.444,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9739688634872437,
+      "rewards/margins": 0.6166737079620361,
+      "rewards/rejected": -2.5906424522399902,
+      "step": 408
+    },
+    {
+      "epoch": 0.742960944595822,
+      "grad_norm": 1.779342770576477,
+      "learning_rate": 9.254061771789847e-06,
+      "logits/chosen": 0.05093669891357422,
+      "logits/rejected": 0.1048060953617096,
+      "logps/chosen": -84.48141479492188,
+      "logps/rejected": -91.49285125732422,
+      "loss": 2.5324,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9952929019927979,
+      "rewards/margins": 0.3900759220123291,
+      "rewards/rejected": -2.385368585586548,
+      "step": 409
+    },
+    {
+      "epoch": 0.7447774750227066,
+      "grad_norm": 1.71133291721344,
+      "learning_rate": 9.25051460821893e-06,
+      "logits/chosen": 0.1308833658695221,
+      "logits/rejected": 0.08994461596012115,
+      "logps/chosen": -74.09046936035156,
+      "logps/rejected": -80.80438995361328,
+      "loss": 2.1895,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8457674980163574,
+      "rewards/margins": 0.5731581449508667,
+      "rewards/rejected": -2.4189257621765137,
+      "step": 410
+    },
+    {
+      "epoch": 0.7465940054495913,
+      "grad_norm": 1.6980112791061401,
+      "learning_rate": 9.2469597966957e-06,
+      "logits/chosen": 0.05856658145785332,
+      "logits/rejected": 0.08371179550886154,
+      "logps/chosen": -78.58245849609375,
+      "logps/rejected": -88.30054473876953,
+      "loss": 2.46,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9803493022918701,
+      "rewards/margins": 0.6115507483482361,
+      "rewards/rejected": -2.591899871826172,
+      "step": 411
+    },
+    {
+      "epoch": 0.7484105358764759,
+      "grad_norm": 1.8350476026535034,
+      "learning_rate": 9.243397344469037e-06,
+      "logits/chosen": 0.05285171419382095,
+      "logits/rejected": 0.13240401446819305,
+      "logps/chosen": -73.79881286621094,
+      "logps/rejected": -86.56880187988281,
+      "loss": 2.2929,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.060075283050537,
+      "rewards/margins": 0.54813152551651,
+      "rewards/rejected": -2.6082065105438232,
+      "step": 412
+    },
+    {
+      "epoch": 0.7502270663033606,
+      "grad_norm": 1.4403142929077148,
+      "learning_rate": 9.239827258803402e-06,
+      "logits/chosen": 0.07327421009540558,
+      "logits/rejected": 0.18387822806835175,
+      "logps/chosen": -68.3246078491211,
+      "logps/rejected": -81.33067321777344,
+      "loss": 1.8794,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8658640384674072,
+      "rewards/margins": 0.7600168585777283,
+      "rewards/rejected": -2.625880718231201,
+      "step": 413
+    },
+    {
+      "epoch": 0.7520435967302452,
+      "grad_norm": 1.883155345916748,
+      "learning_rate": 9.23624954697882e-06,
+      "logits/chosen": 0.10411994159221649,
+      "logits/rejected": 0.06234448403120041,
+      "logps/chosen": -78.31743621826172,
+      "logps/rejected": -80.0100326538086,
+      "loss": 2.5646,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.934099555015564,
+      "rewards/margins": 0.28737103939056396,
+      "rewards/rejected": -2.221470832824707,
+      "step": 414
+    },
+    {
+      "epoch": 0.7538601271571299,
+      "grad_norm": 1.5926934480667114,
+      "learning_rate": 9.232664216290868e-06,
+      "logits/chosen": 0.09589770436286926,
+      "logits/rejected": 0.041689179837703705,
+      "logps/chosen": -77.89552307128906,
+      "logps/rejected": -84.02517700195312,
+      "loss": 2.2284,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.787453532218933,
+      "rewards/margins": 0.49444466829299927,
+      "rewards/rejected": -2.281898260116577,
+      "step": 415
+    },
+    {
+      "epoch": 0.7556766575840145,
+      "grad_norm": 1.511832356452942,
+      "learning_rate": 9.229071274050663e-06,
+      "logits/chosen": 0.07688678801059723,
+      "logits/rejected": 0.14106512069702148,
+      "logps/chosen": -78.53334045410156,
+      "logps/rejected": -86.92496490478516,
+      "loss": 2.2211,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0686798095703125,
+      "rewards/margins": 0.5378819704055786,
+      "rewards/rejected": -2.6065618991851807,
+      "step": 416
+    },
+    {
+      "epoch": 0.7574931880108992,
+      "grad_norm": 1.7858117818832397,
+      "learning_rate": 9.225470727584835e-06,
+      "logits/chosen": 0.14854730665683746,
+      "logits/rejected": 0.06465649604797363,
+      "logps/chosen": -87.31832122802734,
+      "logps/rejected": -79.3746566772461,
+      "loss": 2.4446,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9429931640625,
+      "rewards/margins": 0.38861486315727234,
+      "rewards/rejected": -2.3316078186035156,
+      "step": 417
+    },
+    {
+      "epoch": 0.7593097184377838,
+      "grad_norm": 1.9230687618255615,
+      "learning_rate": 9.221862584235527e-06,
+      "logits/chosen": 0.07233145087957382,
+      "logits/rejected": 0.13825634121894836,
+      "logps/chosen": -81.53103637695312,
+      "logps/rejected": -90.54696655273438,
+      "loss": 2.7171,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.065859794616699,
+      "rewards/margins": 0.30683329701423645,
+      "rewards/rejected": -2.3726933002471924,
+      "step": 418
+    },
+    {
+      "epoch": 0.7611262488646685,
+      "grad_norm": 2.063098192214966,
+      "learning_rate": 9.218246851360374e-06,
+      "logits/chosen": 0.10360075533390045,
+      "logits/rejected": 0.14677830040454865,
+      "logps/chosen": -85.08120727539062,
+      "logps/rejected": -89.35393524169922,
+      "loss": 2.4292,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.300565719604492,
+      "rewards/margins": 0.4277467727661133,
+      "rewards/rejected": -2.7283124923706055,
+      "step": 419
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 2.4728786945343018,
+      "learning_rate": 9.214623536332483e-06,
+      "logits/chosen": 0.08115474879741669,
+      "logits/rejected": 0.07649822533130646,
+      "logps/chosen": -83.79824829101562,
+      "logps/rejected": -87.47264099121094,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.0838184356689453,
+      "rewards/margins": 0.41770505905151367,
+      "rewards/rejected": -2.501523494720459,
+      "step": 420
+    },
+    {
+      "epoch": 0.7647593097184378,
+      "grad_norm": 1.7078697681427002,
+      "learning_rate": 9.210992646540425e-06,
+      "logits/chosen": 0.08155008405447006,
+      "logits/rejected": 0.08723931759595871,
+      "logps/chosen": -74.4200210571289,
+      "logps/rejected": -87.49038696289062,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9076974391937256,
+      "rewards/margins": 0.6807244420051575,
+      "rewards/rejected": -2.5884220600128174,
+      "step": 421
+    },
+    {
+      "epoch": 0.7665758401453224,
+      "grad_norm": 1.5644993782043457,
+      "learning_rate": 9.207354189388214e-06,
+      "logits/chosen": 0.16107802093029022,
+      "logits/rejected": 0.15951679646968842,
+      "logps/chosen": -74.85030364990234,
+      "logps/rejected": -81.25472259521484,
+      "loss": 2.2389,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0082590579986572,
+      "rewards/margins": 0.5728211402893066,
+      "rewards/rejected": -2.581080198287964,
+      "step": 422
+    },
+    {
+      "epoch": 0.7683923705722071,
+      "grad_norm": 1.6519557237625122,
+      "learning_rate": 9.203708172295299e-06,
+      "logits/chosen": 0.08725707978010178,
+      "logits/rejected": 0.11393023282289505,
+      "logps/chosen": -79.56071472167969,
+      "logps/rejected": -90.27279663085938,
+      "loss": 2.2853,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9312469959259033,
+      "rewards/margins": 0.6229196786880493,
+      "rewards/rejected": -2.554166793823242,
+      "step": 423
+    },
+    {
+      "epoch": 0.7702089009990918,
+      "grad_norm": 1.668648600578308,
+      "learning_rate": 9.200054602696544e-06,
+      "logits/chosen": 0.1369432657957077,
+      "logits/rejected": 0.08549812436103821,
+      "logps/chosen": -80.69041442871094,
+      "logps/rejected": -85.47969055175781,
+      "loss": 2.3022,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7405030727386475,
+      "rewards/margins": 0.4421282112598419,
+      "rewards/rejected": -2.182631492614746,
+      "step": 424
+    },
+    {
+      "epoch": 0.7720254314259763,
+      "grad_norm": 1.7834622859954834,
+      "learning_rate": 9.196393488042213e-06,
+      "logits/chosen": 0.06559979915618896,
+      "logits/rejected": 0.05396304652094841,
+      "logps/chosen": -76.51896667480469,
+      "logps/rejected": -88.19717407226562,
+      "loss": 2.5216,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.050562620162964,
+      "rewards/margins": 0.5811472535133362,
+      "rewards/rejected": -2.631709575653076,
+      "step": 425
+    },
+    {
+      "epoch": 0.773841961852861,
+      "grad_norm": 1.5034058094024658,
+      "learning_rate": 9.192724835797956e-06,
+      "logits/chosen": 0.10820607095956802,
+      "logits/rejected": 0.14427171647548676,
+      "logps/chosen": -81.0079116821289,
+      "logps/rejected": -95.6849136352539,
+      "loss": 2.0472,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.19939923286438,
+      "rewards/margins": 0.8476728796958923,
+      "rewards/rejected": -3.047071933746338,
+      "step": 426
+    },
+    {
+      "epoch": 0.7756584922797457,
+      "grad_norm": 1.7205973863601685,
+      "learning_rate": 9.189048653444796e-06,
+      "logits/chosen": 0.08522580564022064,
+      "logits/rejected": 0.09135682880878448,
+      "logps/chosen": -77.4021224975586,
+      "logps/rejected": -85.47643280029297,
+      "loss": 2.2618,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.002034902572632,
+      "rewards/margins": 0.5705874562263489,
+      "rewards/rejected": -2.572622776031494,
+      "step": 427
+    },
+    {
+      "epoch": 0.7774750227066304,
+      "grad_norm": 2.131197214126587,
+      "learning_rate": 9.185364948479109e-06,
+      "logits/chosen": 0.05514610558748245,
+      "logits/rejected": 0.0597330704331398,
+      "logps/chosen": -86.83938598632812,
+      "logps/rejected": -94.71990966796875,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.088331699371338,
+      "rewards/margins": 0.47913864254951477,
+      "rewards/rejected": -2.567470073699951,
+      "step": 428
+    },
+    {
+      "epoch": 0.779291553133515,
+      "grad_norm": 3.0228030681610107,
+      "learning_rate": 9.181673728412605e-06,
+      "logits/chosen": 0.09502460807561874,
+      "logits/rejected": 0.10438862442970276,
+      "logps/chosen": -77.20478820800781,
+      "logps/rejected": -80.07164764404297,
+      "loss": 2.5924,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.107954978942871,
+      "rewards/margins": 0.4795181155204773,
+      "rewards/rejected": -2.5874733924865723,
+      "step": 429
+    },
+    {
+      "epoch": 0.7811080835603996,
+      "grad_norm": 2.630490779876709,
+      "learning_rate": 9.17797500077233e-06,
+      "logits/chosen": 0.053769052028656006,
+      "logits/rejected": 0.1357191652059555,
+      "logps/chosen": -74.01277160644531,
+      "logps/rejected": -86.05644989013672,
+      "loss": 2.6208,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0179696083068848,
+      "rewards/margins": 0.3262583613395691,
+      "rewards/rejected": -2.3442280292510986,
+      "step": 430
+    },
+    {
+      "epoch": 0.7829246139872843,
+      "grad_norm": 1.7841664552688599,
+      "learning_rate": 9.174268773100636e-06,
+      "logits/chosen": 0.09468546509742737,
+      "logits/rejected": 0.16804733872413635,
+      "logps/chosen": -80.07379913330078,
+      "logps/rejected": -87.17518615722656,
+      "loss": 2.4676,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.27384877204895,
+      "rewards/margins": 0.4068894386291504,
+      "rewards/rejected": -2.6807379722595215,
+      "step": 431
+    },
+    {
+      "epoch": 0.784741144414169,
+      "grad_norm": 1.7030746936798096,
+      "learning_rate": 9.170555052955158e-06,
+      "logits/chosen": 0.10446357727050781,
+      "logits/rejected": 0.1235620528459549,
+      "logps/chosen": -89.81584930419922,
+      "logps/rejected": -94.43669128417969,
+      "loss": 2.3366,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.98858642578125,
+      "rewards/margins": 0.5402320623397827,
+      "rewards/rejected": -2.5288188457489014,
+      "step": 432
+    },
+    {
+      "epoch": 0.7865576748410535,
+      "grad_norm": 1.4258465766906738,
+      "learning_rate": 9.166833847908825e-06,
+      "logits/chosen": 0.10827025771141052,
+      "logits/rejected": 0.10875076055526733,
+      "logps/chosen": -69.29400634765625,
+      "logps/rejected": -80.72003936767578,
+      "loss": 2.2501,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9478650093078613,
+      "rewards/margins": 0.5940475463867188,
+      "rewards/rejected": -2.54191255569458,
+      "step": 433
+    },
+    {
+      "epoch": 0.7883742052679382,
+      "grad_norm": 1.8069814443588257,
+      "learning_rate": 9.163105165549819e-06,
+      "logits/chosen": 0.08856840431690216,
+      "logits/rejected": 0.07843751460313797,
+      "logps/chosen": -79.93275451660156,
+      "logps/rejected": -83.75475311279297,
+      "loss": 2.5214,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0315518379211426,
+      "rewards/margins": 0.45457565784454346,
+      "rewards/rejected": -2.4861276149749756,
+      "step": 434
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 1.7881503105163574,
+      "learning_rate": 9.159369013481574e-06,
+      "logits/chosen": 0.08166562020778656,
+      "logits/rejected": 0.08240097761154175,
+      "logps/chosen": -92.98298645019531,
+      "logps/rejected": -96.15763092041016,
+      "loss": 2.053,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4435133934020996,
+      "rewards/margins": 0.7140956521034241,
+      "rewards/rejected": -3.157609462738037,
+      "step": 435
+    },
+    {
+      "epoch": 0.7920072661217076,
+      "grad_norm": 1.7574496269226074,
+      "learning_rate": 9.155625399322754e-06,
+      "logits/chosen": 0.08114133030176163,
+      "logits/rejected": 0.11279730498790741,
+      "logps/chosen": -74.07096862792969,
+      "logps/rejected": -80.94386291503906,
+      "loss": 2.4289,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2497410774230957,
+      "rewards/margins": 0.4424628019332886,
+      "rewards/rejected": -2.6922037601470947,
+      "step": 436
+    },
+    {
+      "epoch": 0.7938237965485921,
+      "grad_norm": 1.4814083576202393,
+      "learning_rate": 9.15187433070724e-06,
+      "logits/chosen": 0.026073571294546127,
+      "logits/rejected": 0.1382063925266266,
+      "logps/chosen": -72.39479064941406,
+      "logps/rejected": -86.76889038085938,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9130635261535645,
+      "rewards/margins": 0.6362709999084473,
+      "rewards/rejected": -2.549334764480591,
+      "step": 437
+    },
+    {
+      "epoch": 0.7956403269754768,
+      "grad_norm": 1.6441129446029663,
+      "learning_rate": 9.148115815284113e-06,
+      "logits/chosen": 0.07432619482278824,
+      "logits/rejected": 0.10215617716312408,
+      "logps/chosen": -77.05310821533203,
+      "logps/rejected": -80.81717681884766,
+      "loss": 2.338,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9012683629989624,
+      "rewards/margins": 0.43776583671569824,
+      "rewards/rejected": -2.339034080505371,
+      "step": 438
+    },
+    {
+      "epoch": 0.7974568574023615,
+      "grad_norm": 1.8839560747146606,
+      "learning_rate": 9.144349860717643e-06,
+      "logits/chosen": 0.07576426863670349,
+      "logits/rejected": 0.08868670463562012,
+      "logps/chosen": -74.13190460205078,
+      "logps/rejected": -80.07232666015625,
+      "loss": 2.3485,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9554617404937744,
+      "rewards/margins": 0.531810998916626,
+      "rewards/rejected": -2.4872727394104004,
+      "step": 439
+    },
+    {
+      "epoch": 0.7992733878292462,
+      "grad_norm": 1.826811671257019,
+      "learning_rate": 9.140576474687263e-06,
+      "logits/chosen": 0.09120994061231613,
+      "logits/rejected": 0.08987519890069962,
+      "logps/chosen": -75.1444320678711,
+      "logps/rejected": -78.58987426757812,
+      "loss": 2.6513,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158261299133301,
+      "rewards/margins": 0.28016677498817444,
+      "rewards/rejected": -2.4384284019470215,
+      "step": 440
+    },
+    {
+      "epoch": 0.8010899182561307,
+      "grad_norm": 1.8647096157073975,
+      "learning_rate": 9.13679566488757e-06,
+      "logits/chosen": 0.12438184767961502,
+      "logits/rejected": 0.13951222598552704,
+      "logps/chosen": -83.79389953613281,
+      "logps/rejected": -90.33230590820312,
+      "loss": 2.3956,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0185930728912354,
+      "rewards/margins": 0.43685024976730347,
+      "rewards/rejected": -2.4554433822631836,
+      "step": 441
+    },
+    {
+      "epoch": 0.8029064486830154,
+      "grad_norm": 1.7610801458358765,
+      "learning_rate": 9.133007439028288e-06,
+      "logits/chosen": 0.08218151330947876,
+      "logits/rejected": 0.072816863656044,
+      "logps/chosen": -86.35633087158203,
+      "logps/rejected": -84.08189392089844,
+      "loss": 2.243,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8836033344268799,
+      "rewards/margins": 0.516740083694458,
+      "rewards/rejected": -2.400343418121338,
+      "step": 442
+    },
+    {
+      "epoch": 0.8047229791099001,
+      "grad_norm": 1.6049703359603882,
+      "learning_rate": 9.129211804834271e-06,
+      "logits/chosen": 0.059631846845149994,
+      "logits/rejected": 0.09744230657815933,
+      "logps/chosen": -76.90062713623047,
+      "logps/rejected": -82.29356384277344,
+      "loss": 2.2574,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9436204433441162,
+      "rewards/margins": 0.47444695234298706,
+      "rewards/rejected": -2.418067216873169,
+      "step": 443
+    },
+    {
+      "epoch": 0.8065395095367848,
+      "grad_norm": 1.9819791316986084,
+      "learning_rate": 9.12540877004548e-06,
+      "logits/chosen": 0.07371871173381805,
+      "logits/rejected": 0.058843065053224564,
+      "logps/chosen": -78.42516326904297,
+      "logps/rejected": -82.23822784423828,
+      "loss": 2.6959,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.273909330368042,
+      "rewards/margins": 0.3514346182346344,
+      "rewards/rejected": -2.6253440380096436,
+      "step": 444
+    },
+    {
+      "epoch": 0.8083560399636693,
+      "grad_norm": 1.9540653228759766,
+      "learning_rate": 9.12159834241696e-06,
+      "logits/chosen": 0.13155074417591095,
+      "logits/rejected": 0.11775672435760498,
+      "logps/chosen": -69.17151641845703,
+      "logps/rejected": -77.9190902709961,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.073258638381958,
+      "rewards/margins": 0.5955159664154053,
+      "rewards/rejected": -2.6687746047973633,
+      "step": 445
+    },
+    {
+      "epoch": 0.810172570390554,
+      "grad_norm": 1.9634339809417725,
+      "learning_rate": 9.117780529718843e-06,
+      "logits/chosen": 0.05265370383858681,
+      "logits/rejected": 0.09651105850934982,
+      "logps/chosen": -78.77818298339844,
+      "logps/rejected": -84.63584899902344,
+      "loss": 2.2157,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.084688186645508,
+      "rewards/margins": 0.541061282157898,
+      "rewards/rejected": -2.625749349594116,
+      "step": 446
+    },
+    {
+      "epoch": 0.8119891008174387,
+      "grad_norm": 1.9086997509002686,
+      "learning_rate": 9.113955339736309e-06,
+      "logits/chosen": 0.08473093807697296,
+      "logits/rejected": 0.047919195145368576,
+      "logps/chosen": -83.44226837158203,
+      "logps/rejected": -87.24287414550781,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9542481899261475,
+      "rewards/margins": 0.4577428698539734,
+      "rewards/rejected": -2.4119908809661865,
+      "step": 447
+    },
+    {
+      "epoch": 0.8138056312443234,
+      "grad_norm": 1.6967483758926392,
+      "learning_rate": 9.11012278026959e-06,
+      "logits/chosen": 0.062431350350379944,
+      "logits/rejected": 0.10980932414531708,
+      "logps/chosen": -90.98541259765625,
+      "logps/rejected": -102.51634979248047,
+      "loss": 2.3339,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1866049766540527,
+      "rewards/margins": 0.6549904346466064,
+      "rewards/rejected": -2.841595411300659,
+      "step": 448
+    },
+    {
+      "epoch": 0.815622161671208,
+      "grad_norm": 2.295504331588745,
+      "learning_rate": 9.106282859133936e-06,
+      "logits/chosen": 0.04481218010187149,
+      "logits/rejected": 0.002619542181491852,
+      "logps/chosen": -81.81380462646484,
+      "logps/rejected": -80.60113525390625,
+      "loss": 2.8324,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.1727848052978516,
+      "rewards/margins": 0.2777697443962097,
+      "rewards/rejected": -2.450554609298706,
+      "step": 449
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 1.7951968908309937,
+      "learning_rate": 9.102435584159623e-06,
+      "logits/chosen": 0.029065577313303947,
+      "logits/rejected": 0.11651361733675003,
+      "logps/chosen": -78.08231353759766,
+      "logps/rejected": -88.59742736816406,
+      "loss": 2.5562,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.2762022018432617,
+      "rewards/margins": 0.4410257041454315,
+      "rewards/rejected": -2.7172276973724365,
+      "step": 450
+    },
+    {
+      "epoch": 0.8192552225249773,
+      "grad_norm": 1.6301803588867188,
+      "learning_rate": 9.098580963191908e-06,
+      "logits/chosen": 0.06471782922744751,
+      "logits/rejected": 0.11030608415603638,
+      "logps/chosen": -78.3387451171875,
+      "logps/rejected": -91.82540130615234,
+      "loss": 2.1826,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.885233998298645,
+      "rewards/margins": 0.6971657872200012,
+      "rewards/rejected": -2.582399845123291,
+      "step": 451
+    },
+    {
+      "epoch": 0.821071752951862,
+      "grad_norm": 1.705962061882019,
+      "learning_rate": 9.094719004091039e-06,
+      "logits/chosen": 0.12591060996055603,
+      "logits/rejected": 0.1107010543346405,
+      "logps/chosen": -76.82783508300781,
+      "logps/rejected": -84.23365020751953,
+      "loss": 2.2843,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.04795241355896,
+      "rewards/margins": 0.5103356242179871,
+      "rewards/rejected": -2.558288335800171,
+      "step": 452
+    },
+    {
+      "epoch": 0.8228882833787466,
+      "grad_norm": 1.6535886526107788,
+      "learning_rate": 9.090849714732217e-06,
+      "logits/chosen": 0.11322569847106934,
+      "logits/rejected": 0.13092264533042908,
+      "logps/chosen": -80.16563415527344,
+      "logps/rejected": -82.96965026855469,
+      "loss": 2.3556,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9074076414108276,
+      "rewards/margins": 0.39431121945381165,
+      "rewards/rejected": -2.3017189502716064,
+      "step": 453
+    },
+    {
+      "epoch": 0.8247048138056312,
+      "grad_norm": 1.7194557189941406,
+      "learning_rate": 9.086973103005602e-06,
+      "logits/chosen": 0.1069367378950119,
+      "logits/rejected": 0.06237747147679329,
+      "logps/chosen": -72.59986877441406,
+      "logps/rejected": -77.62615203857422,
+      "loss": 2.2768,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.023085355758667,
+      "rewards/margins": 0.6052648425102234,
+      "rewards/rejected": -2.6283504962921143,
+      "step": 454
+    },
+    {
+      "epoch": 0.8265213442325159,
+      "grad_norm": 1.7921446561813354,
+      "learning_rate": 9.08308917681628e-06,
+      "logits/chosen": 0.12843580543994904,
+      "logits/rejected": 0.08002308011054993,
+      "logps/chosen": -80.6107177734375,
+      "logps/rejected": -80.92134857177734,
+      "loss": 2.3881,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9404933452606201,
+      "rewards/margins": 0.37478840351104736,
+      "rewards/rejected": -2.315281629562378,
+      "step": 455
+    },
+    {
+      "epoch": 0.8283378746594006,
+      "grad_norm": 2.3353819847106934,
+      "learning_rate": 9.07919794408425e-06,
+      "logits/chosen": 0.05452323704957962,
+      "logits/rejected": 0.12883196771144867,
+      "logps/chosen": -83.56001281738281,
+      "logps/rejected": -95.23291015625,
+      "loss": 2.8503,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.2597994804382324,
+      "rewards/margins": 0.4092620015144348,
+      "rewards/rejected": -2.6690614223480225,
+      "step": 456
+    },
+    {
+      "epoch": 0.8301544050862852,
+      "grad_norm": 1.6829804182052612,
+      "learning_rate": 9.075299412744417e-06,
+      "logits/chosen": 0.1223950982093811,
+      "logits/rejected": 0.10596577078104019,
+      "logps/chosen": -77.75115203857422,
+      "logps/rejected": -83.63490295410156,
+      "loss": 2.1639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.026258945465088,
+      "rewards/margins": 0.6107547879219055,
+      "rewards/rejected": -2.6370139122009277,
+      "step": 457
+    },
+    {
+      "epoch": 0.8319709355131698,
+      "grad_norm": 2.053755283355713,
+      "learning_rate": 9.07139359074656e-06,
+      "logits/chosen": 0.08021271228790283,
+      "logits/rejected": 0.11617676913738251,
+      "logps/chosen": -75.00261688232422,
+      "logps/rejected": -87.25222778320312,
+      "loss": 2.4048,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3207621574401855,
+      "rewards/margins": 0.5956254005432129,
+      "rewards/rejected": -2.9163873195648193,
+      "step": 458
+    },
+    {
+      "epoch": 0.8337874659400545,
+      "grad_norm": 2.1006064414978027,
+      "learning_rate": 9.067480486055333e-06,
+      "logits/chosen": 0.1115679144859314,
+      "logits/rejected": 0.08167213946580887,
+      "logps/chosen": -81.78447723388672,
+      "logps/rejected": -82.21318817138672,
+      "loss": 2.8725,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.165053367614746,
+      "rewards/margins": 0.2777022123336792,
+      "rewards/rejected": -2.442755699157715,
+      "step": 459
+    },
+    {
+      "epoch": 0.8356039963669392,
+      "grad_norm": 1.6339340209960938,
+      "learning_rate": 9.063560106650238e-06,
+      "logits/chosen": 0.0966312363743782,
+      "logits/rejected": 0.09943810105323792,
+      "logps/chosen": -71.0906982421875,
+      "logps/rejected": -76.22811126708984,
+      "loss": 2.2922,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.0007636547088623,
+      "rewards/margins": 0.47423097491264343,
+      "rewards/rejected": -2.474994659423828,
+      "step": 460
+    },
+    {
+      "epoch": 0.8374205267938238,
+      "grad_norm": 1.8504972457885742,
+      "learning_rate": 9.059632460525613e-06,
+      "logits/chosen": 0.09517084062099457,
+      "logits/rejected": 0.14127981662750244,
+      "logps/chosen": -79.06185913085938,
+      "logps/rejected": -89.11494445800781,
+      "loss": 2.3288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0016026496887207,
+      "rewards/margins": 0.5282158255577087,
+      "rewards/rejected": -2.529818534851074,
+      "step": 461
+    },
+    {
+      "epoch": 0.8392370572207084,
+      "grad_norm": 1.8151304721832275,
+      "learning_rate": 9.055697555690607e-06,
+      "logits/chosen": 0.08868349343538284,
+      "logits/rejected": 0.14090785384178162,
+      "logps/chosen": -77.26272583007812,
+      "logps/rejected": -85.14620971679688,
+      "loss": 2.3694,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.087087869644165,
+      "rewards/margins": 0.5032299757003784,
+      "rewards/rejected": -2.590317964553833,
+      "step": 462
+    },
+    {
+      "epoch": 0.8410535876475931,
+      "grad_norm": 1.8596280813217163,
+      "learning_rate": 9.051755400169182e-06,
+      "logits/chosen": 0.06272133439779282,
+      "logits/rejected": 0.12342572212219238,
+      "logps/chosen": -75.46456909179688,
+      "logps/rejected": -81.9759292602539,
+      "loss": 2.5167,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0463178157806396,
+      "rewards/margins": 0.3146142363548279,
+      "rewards/rejected": -2.3609323501586914,
+      "step": 463
+    },
+    {
+      "epoch": 0.8428701180744778,
+      "grad_norm": 1.7658417224884033,
+      "learning_rate": 9.047806002000075e-06,
+      "logits/chosen": -0.007259421981871128,
+      "logits/rejected": 0.09611339122056961,
+      "logps/chosen": -73.88439178466797,
+      "logps/rejected": -80.77063751220703,
+      "loss": 2.4938,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9998183250427246,
+      "rewards/margins": 0.4115346670150757,
+      "rewards/rejected": -2.4113528728485107,
+      "step": 464
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 1.6596916913986206,
+      "learning_rate": 9.043849369236799e-06,
+      "logits/chosen": 0.03223409131169319,
+      "logits/rejected": 0.05010119825601578,
+      "logps/chosen": -66.32562255859375,
+      "logps/rejected": -69.36509704589844,
+      "loss": 2.5881,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.909837245941162,
+      "rewards/margins": 0.2923263907432556,
+      "rewards/rejected": -2.2021636962890625,
+      "step": 465
+    },
+    {
+      "epoch": 0.846503178928247,
+      "grad_norm": 1.6832820177078247,
+      "learning_rate": 9.039885509947616e-06,
+      "logits/chosen": 0.06287454813718796,
+      "logits/rejected": 0.12746769189834595,
+      "logps/chosen": -76.6998519897461,
+      "logps/rejected": -87.36117553710938,
+      "loss": 2.3914,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.029526710510254,
+      "rewards/margins": 0.48344865441322327,
+      "rewards/rejected": -2.5129752159118652,
+      "step": 466
+    },
+    {
+      "epoch": 0.8483197093551317,
+      "grad_norm": 1.6972366571426392,
+      "learning_rate": 9.035914432215527e-06,
+      "logits/chosen": 0.1179896891117096,
+      "logits/rejected": 0.14000467956066132,
+      "logps/chosen": -75.3287353515625,
+      "logps/rejected": -78.60227966308594,
+      "loss": 2.4194,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.088515043258667,
+      "rewards/margins": 0.42544490098953247,
+      "rewards/rejected": -2.513960123062134,
+      "step": 467
+    },
+    {
+      "epoch": 0.8501362397820164,
+      "grad_norm": 2.5104973316192627,
+      "learning_rate": 9.031936144138247e-06,
+      "logits/chosen": 0.08423591405153275,
+      "logits/rejected": 0.1387752890586853,
+      "logps/chosen": -78.6176528930664,
+      "logps/rejected": -83.52703094482422,
+      "loss": 2.7262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1089184284210205,
+      "rewards/margins": 0.44402289390563965,
+      "rewards/rejected": -2.55294132232666,
+      "step": 468
+    },
+    {
+      "epoch": 0.851952770208901,
+      "grad_norm": 1.944931149482727,
+      "learning_rate": 9.027950653828202e-06,
+      "logits/chosen": 0.12252221256494522,
+      "logits/rejected": 0.14959384500980377,
+      "logps/chosen": -72.63184356689453,
+      "logps/rejected": -70.89137268066406,
+      "loss": 2.7323,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9612714052200317,
+      "rewards/margins": 0.24262896180152893,
+      "rewards/rejected": -2.2039003372192383,
+      "step": 469
+    },
+    {
+      "epoch": 0.8537693006357856,
+      "grad_norm": 1.6699979305267334,
+      "learning_rate": 9.0239579694125e-06,
+      "logits/chosen": 0.11827167868614197,
+      "logits/rejected": 0.11933022737503052,
+      "logps/chosen": -70.45188903808594,
+      "logps/rejected": -76.2336654663086,
+      "loss": 2.457,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.854758381843567,
+      "rewards/margins": 0.47020024061203003,
+      "rewards/rejected": -2.3249588012695312,
+      "step": 470
+    },
+    {
+      "epoch": 0.8555858310626703,
+      "grad_norm": 1.5790536403656006,
+      "learning_rate": 9.019958099032919e-06,
+      "logits/chosen": 0.04856352508068085,
+      "logits/rejected": 0.11418008804321289,
+      "logps/chosen": -75.7929458618164,
+      "logps/rejected": -86.55818176269531,
+      "loss": 2.2809,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.003899335861206,
+      "rewards/margins": 0.5960561037063599,
+      "rewards/rejected": -2.5999553203582764,
+      "step": 471
+    },
+    {
+      "epoch": 0.857402361489555,
+      "grad_norm": 1.3594144582748413,
+      "learning_rate": 9.015951050845891e-06,
+      "logits/chosen": 0.10341258347034454,
+      "logits/rejected": 0.12394269555807114,
+      "logps/chosen": -78.8187255859375,
+      "logps/rejected": -84.04349517822266,
+      "loss": 1.9536,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9372440576553345,
+      "rewards/margins": 0.6742948889732361,
+      "rewards/rejected": -2.611538887023926,
+      "step": 472
+    },
+    {
+      "epoch": 0.8592188919164396,
+      "grad_norm": 1.7592689990997314,
+      "learning_rate": 9.011936833022485e-06,
+      "logits/chosen": 0.10970023274421692,
+      "logits/rejected": 0.09040558338165283,
+      "logps/chosen": -80.59556579589844,
+      "logps/rejected": -84.2996826171875,
+      "loss": 2.5658,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.935511827468872,
+      "rewards/margins": 0.35643547773361206,
+      "rewards/rejected": -2.29194712638855,
+      "step": 473
+    },
+    {
+      "epoch": 0.8610354223433242,
+      "grad_norm": 1.6418087482452393,
+      "learning_rate": 9.00791545374839e-06,
+      "logits/chosen": 0.16755647957324982,
+      "logits/rejected": 0.1745520681142807,
+      "logps/chosen": -82.31010437011719,
+      "logps/rejected": -85.66128540039062,
+      "loss": 2.1854,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.9519617557525635,
+      "rewards/margins": 0.5059336423873901,
+      "rewards/rejected": -2.457895278930664,
+      "step": 474
+    },
+    {
+      "epoch": 0.8628519527702089,
+      "grad_norm": 1.4616092443466187,
+      "learning_rate": 9.0038869212239e-06,
+      "logits/chosen": 0.03978392109274864,
+      "logits/rejected": 0.1235240027308464,
+      "logps/chosen": -77.78251647949219,
+      "logps/rejected": -85.66397094726562,
+      "loss": 2.1995,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8721034526824951,
+      "rewards/margins": 0.45375847816467285,
+      "rewards/rejected": -2.325861930847168,
+      "step": 475
+    },
+    {
+      "epoch": 0.8646684831970936,
+      "grad_norm": 1.838180422782898,
+      "learning_rate": 8.99985124366389e-06,
+      "logits/chosen": 0.08277668058872223,
+      "logits/rejected": 0.035900432616472244,
+      "logps/chosen": -83.23967742919922,
+      "logps/rejected": -83.16535949707031,
+      "loss": 2.7347,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.2339279651641846,
+      "rewards/margins": 0.16711921989917755,
+      "rewards/rejected": -2.4010472297668457,
+      "step": 476
+    },
+    {
+      "epoch": 0.8664850136239782,
+      "grad_norm": 1.6746636629104614,
+      "learning_rate": 8.995808429297815e-06,
+      "logits/chosen": 0.14687396585941315,
+      "logits/rejected": 0.09570137411355972,
+      "logps/chosen": -78.79898071289062,
+      "logps/rejected": -77.9619369506836,
+      "loss": 2.4449,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8334993124008179,
+      "rewards/margins": 0.3442971706390381,
+      "rewards/rejected": -2.1777963638305664,
+      "step": 477
+    },
+    {
+      "epoch": 0.8683015440508629,
+      "grad_norm": 1.7042535543441772,
+      "learning_rate": 8.991758486369675e-06,
+      "logits/chosen": 0.047281138598918915,
+      "logits/rejected": -0.023666318506002426,
+      "logps/chosen": -83.798583984375,
+      "logps/rejected": -91.88795471191406,
+      "loss": 2.3072,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9728612899780273,
+      "rewards/margins": 0.5659723281860352,
+      "rewards/rejected": -2.5388338565826416,
+      "step": 478
+    },
+    {
+      "epoch": 0.8701180744777475,
+      "grad_norm": 1.5552817583084106,
+      "learning_rate": 8.987701423138007e-06,
+      "logits/chosen": 0.0772751122713089,
+      "logits/rejected": 0.08708472549915314,
+      "logps/chosen": -72.42391204833984,
+      "logps/rejected": -78.93070220947266,
+      "loss": 2.2815,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9803509712219238,
+      "rewards/margins": 0.5386512279510498,
+      "rewards/rejected": -2.5190021991729736,
+      "step": 479
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.6372044086456299,
+      "learning_rate": 8.983637247875872e-06,
+      "logits/chosen": -0.03352706879377365,
+      "logits/rejected": 0.01581352949142456,
+      "logps/chosen": -81.29138946533203,
+      "logps/rejected": -89.1744155883789,
+      "loss": 2.09,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9441754817962646,
+      "rewards/margins": 0.6137264370918274,
+      "rewards/rejected": -2.5579018592834473,
+      "step": 480
+    },
+    {
+      "epoch": 0.8737511353315168,
+      "grad_norm": 1.6375274658203125,
+      "learning_rate": 8.979565968870831e-06,
+      "logits/chosen": 0.07507088780403137,
+      "logits/rejected": 0.1289597451686859,
+      "logps/chosen": -74.78109741210938,
+      "logps/rejected": -84.55204772949219,
+      "loss": 2.4566,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.9041385650634766,
+      "rewards/margins": 0.42814433574676514,
+      "rewards/rejected": -2.332282781600952,
+      "step": 481
+    },
+    {
+      "epoch": 0.8755676657584015,
+      "grad_norm": 1.5601258277893066,
+      "learning_rate": 8.975487594424927e-06,
+      "logits/chosen": 0.012452262453734875,
+      "logits/rejected": 0.06930352002382278,
+      "logps/chosen": -83.01705169677734,
+      "logps/rejected": -92.38286590576172,
+      "loss": 2.2736,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.013011932373047,
+      "rewards/margins": 0.5085774064064026,
+      "rewards/rejected": -2.5215890407562256,
+      "step": 482
+    },
+    {
+      "epoch": 0.8773841961852861,
+      "grad_norm": 1.8685195446014404,
+      "learning_rate": 8.971402132854677e-06,
+      "logits/chosen": 0.09415426105260849,
+      "logits/rejected": 0.12719042599201202,
+      "logps/chosen": -79.85440063476562,
+      "logps/rejected": -82.80875396728516,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.114530086517334,
+      "rewards/margins": 0.46328115463256836,
+      "rewards/rejected": -2.5778112411499023,
+      "step": 483
+    },
+    {
+      "epoch": 0.8792007266121707,
+      "grad_norm": 1.8812834024429321,
+      "learning_rate": 8.967309592491052e-06,
+      "logits/chosen": 0.13547101616859436,
+      "logits/rejected": 0.11872326582670212,
+      "logps/chosen": -74.42122650146484,
+      "logps/rejected": -81.9700927734375,
+      "loss": 2.5016,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.1828627586364746,
+      "rewards/margins": 0.44510167837142944,
+      "rewards/rejected": -2.627964496612549,
+      "step": 484
+    },
+    {
+      "epoch": 0.8810172570390554,
+      "grad_norm": 1.510517954826355,
+      "learning_rate": 8.963209981679451e-06,
+      "logits/chosen": 0.03274242952466011,
+      "logits/rejected": 0.07188954204320908,
+      "logps/chosen": -82.55563354492188,
+      "logps/rejected": -100.33609771728516,
+      "loss": 1.9177,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9487042427062988,
+      "rewards/margins": 0.8189151287078857,
+      "rewards/rejected": -2.7676191329956055,
+      "step": 485
+    },
+    {
+      "epoch": 0.8828337874659401,
+      "grad_norm": 1.5973646640777588,
+      "learning_rate": 8.959103308779696e-06,
+      "logits/chosen": 0.0365552082657814,
+      "logits/rejected": 0.051129020750522614,
+      "logps/chosen": -72.4110107421875,
+      "logps/rejected": -83.10322570800781,
+      "loss": 2.1547,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9375314712524414,
+      "rewards/margins": 0.6991211175918579,
+      "rewards/rejected": -2.6366524696350098,
+      "step": 486
+    },
+    {
+      "epoch": 0.8846503178928247,
+      "grad_norm": 1.6707032918930054,
+      "learning_rate": 8.954989582166009e-06,
+      "logits/chosen": 0.0635693296790123,
+      "logits/rejected": 0.009854275733232498,
+      "logps/chosen": -87.12749481201172,
+      "logps/rejected": -88.9491195678711,
+      "loss": 2.2043,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9257593154907227,
+      "rewards/margins": 0.5429801344871521,
+      "rewards/rejected": -2.4687397480010986,
+      "step": 487
+    },
+    {
+      "epoch": 0.8864668483197093,
+      "grad_norm": 1.4422773122787476,
+      "learning_rate": 8.95086881022699e-06,
+      "logits/chosen": -0.024460218846797943,
+      "logits/rejected": 0.14137038588523865,
+      "logps/chosen": -74.60330200195312,
+      "logps/rejected": -95.01730346679688,
+      "loss": 2.0246,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.054253339767456,
+      "rewards/margins": 0.8260326981544495,
+      "rewards/rejected": -2.88028621673584,
+      "step": 488
+    },
+    {
+      "epoch": 0.888283378746594,
+      "grad_norm": 1.8121236562728882,
+      "learning_rate": 8.946741001365616e-06,
+      "logits/chosen": 0.05854415148496628,
+      "logits/rejected": 0.13758361339569092,
+      "logps/chosen": -74.46614074707031,
+      "logps/rejected": -83.96118927001953,
+      "loss": 2.5066,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.105980157852173,
+      "rewards/margins": 0.41478782892227173,
+      "rewards/rejected": -2.520768165588379,
+      "step": 489
+    },
+    {
+      "epoch": 0.8900999091734787,
+      "grad_norm": 2.023728370666504,
+      "learning_rate": 8.942606163999205e-06,
+      "logits/chosen": 0.04619833081960678,
+      "logits/rejected": 0.09330146014690399,
+      "logps/chosen": -75.35594940185547,
+      "logps/rejected": -89.18836975097656,
+      "loss": 2.2655,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9396051168441772,
+      "rewards/margins": 0.6726698875427246,
+      "rewards/rejected": -2.6122751235961914,
+      "step": 490
+    },
+    {
+      "epoch": 0.8919164396003633,
+      "grad_norm": 1.559735894203186,
+      "learning_rate": 8.938464306559412e-06,
+      "logits/chosen": 0.093504399061203,
+      "logits/rejected": 0.09722252935171127,
+      "logps/chosen": -80.77140808105469,
+      "logps/rejected": -87.86373901367188,
+      "loss": 2.1676,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.1073269844055176,
+      "rewards/margins": 0.5740070939064026,
+      "rewards/rejected": -2.6813340187072754,
+      "step": 491
+    },
+    {
+      "epoch": 0.8937329700272479,
+      "grad_norm": 1.6771574020385742,
+      "learning_rate": 8.934315437492203e-06,
+      "logits/chosen": 0.06368491798639297,
+      "logits/rejected": 0.04900998994708061,
+      "logps/chosen": -78.2313003540039,
+      "logps/rejected": -92.83306121826172,
+      "loss": 2.3435,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.110978364944458,
+      "rewards/margins": 0.6751725673675537,
+      "rewards/rejected": -2.786151170730591,
+      "step": 492
+    },
+    {
+      "epoch": 0.8955495004541326,
+      "grad_norm": 1.651183843612671,
+      "learning_rate": 8.930159565257846e-06,
+      "logits/chosen": 0.08393608778715134,
+      "logits/rejected": 0.0475153923034668,
+      "logps/chosen": -77.8431396484375,
+      "logps/rejected": -88.03681945800781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0274362564086914,
+      "rewards/margins": 0.8075671195983887,
+      "rewards/rejected": -2.835003137588501,
+      "step": 493
+    },
+    {
+      "epoch": 0.8973660308810173,
+      "grad_norm": 2.262662172317505,
+      "learning_rate": 8.925996698330887e-06,
+      "logits/chosen": 0.014226208440959454,
+      "logits/rejected": 0.09588178992271423,
+      "logps/chosen": -88.56185913085938,
+      "logps/rejected": -95.1131591796875,
+      "loss": 2.5853,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4549288749694824,
+      "rewards/margins": 0.5275144577026367,
+      "rewards/rejected": -2.982443332672119,
+      "step": 494
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 1.6070181131362915,
+      "learning_rate": 8.92182684520014e-06,
+      "logits/chosen": 0.11615607142448425,
+      "logits/rejected": 0.16645964980125427,
+      "logps/chosen": -78.97079467773438,
+      "logps/rejected": -90.49947357177734,
+      "loss": 2.1072,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.175706386566162,
+      "rewards/margins": 0.7435849905014038,
+      "rewards/rejected": -2.9192914962768555,
+      "step": 495
+    },
+    {
+      "epoch": 0.9009990917347865,
+      "grad_norm": 1.8343369960784912,
+      "learning_rate": 8.917650014368658e-06,
+      "logits/chosen": 0.12583515048027039,
+      "logits/rejected": 0.07308925688266754,
+      "logps/chosen": -83.38356018066406,
+      "logps/rejected": -90.73155975341797,
+      "loss": 2.3116,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3124871253967285,
+      "rewards/margins": 0.5869801640510559,
+      "rewards/rejected": -2.8994674682617188,
+      "step": 496
+    },
+    {
+      "epoch": 0.9028156221616712,
+      "grad_norm": 1.7175630331039429,
+      "learning_rate": 8.913466214353728e-06,
+      "logits/chosen": 0.14971572160720825,
+      "logits/rejected": 0.17372727394104004,
+      "logps/chosen": -86.66060638427734,
+      "logps/rejected": -94.42353820800781,
+      "loss": 2.3541,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.943577766418457,
+      "rewards/margins": 0.5955328345298767,
+      "rewards/rejected": -2.5391108989715576,
+      "step": 497
+    },
+    {
+      "epoch": 0.9046321525885559,
+      "grad_norm": 1.754708170890808,
+      "learning_rate": 8.909275453686845e-06,
+      "logits/chosen": 0.07779194414615631,
+      "logits/rejected": 0.09394712746143341,
+      "logps/chosen": -71.64151000976562,
+      "logps/rejected": -78.53131103515625,
+      "loss": 2.0436,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0102486610412598,
+      "rewards/margins": 0.6668623685836792,
+      "rewards/rejected": -2.6771109104156494,
+      "step": 498
+    },
+    {
+      "epoch": 0.9064486830154405,
+      "grad_norm": 1.9303216934204102,
+      "learning_rate": 8.905077740913701e-06,
+      "logits/chosen": 0.15853236615657806,
+      "logits/rejected": 0.11452697217464447,
+      "logps/chosen": -81.46524047851562,
+      "logps/rejected": -89.1716537475586,
+      "loss": 2.2477,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.264355182647705,
+      "rewards/margins": 0.6477693915367126,
+      "rewards/rejected": -2.9121243953704834,
+      "step": 499
+    },
+    {
+      "epoch": 0.9082652134423251,
+      "grad_norm": 2.1294379234313965,
+      "learning_rate": 8.900873084594164e-06,
+      "logits/chosen": 0.029480352997779846,
+      "logits/rejected": 0.09827219694852829,
+      "logps/chosen": -95.93231964111328,
+      "logps/rejected": -106.21436309814453,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.3455631732940674,
+      "rewards/margins": 0.5977038145065308,
+      "rewards/rejected": -2.9432668685913086,
+      "step": 500
+    },
+    {
+      "epoch": 0.9100817438692098,
+      "grad_norm": 1.7347708940505981,
+      "learning_rate": 8.896661493302258e-06,
+      "logits/chosen": 0.08531993627548218,
+      "logits/rejected": 0.14993277192115784,
+      "logps/chosen": -79.09967041015625,
+      "logps/rejected": -90.50935363769531,
+      "loss": 2.2969,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9908722639083862,
+      "rewards/margins": 0.5710910558700562,
+      "rewards/rejected": -2.5619633197784424,
+      "step": 501
+    },
+    {
+      "epoch": 0.9118982742960945,
+      "grad_norm": 2.065664529800415,
+      "learning_rate": 8.892442975626152e-06,
+      "logits/chosen": 0.18976512551307678,
+      "logits/rejected": 0.13936059176921844,
+      "logps/chosen": -82.80276489257812,
+      "logps/rejected": -84.57154083251953,
+      "loss": 2.5968,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4863569736480713,
+      "rewards/margins": 0.3651903569698334,
+      "rewards/rejected": -2.8515477180480957,
+      "step": 502
+    },
+    {
+      "epoch": 0.9137148047229791,
+      "grad_norm": 2.054955005645752,
+      "learning_rate": 8.888217540168139e-06,
+      "logits/chosen": 0.01604822278022766,
+      "logits/rejected": 0.1370609700679779,
+      "logps/chosen": -74.58015441894531,
+      "logps/rejected": -96.80826568603516,
+      "loss": 2.169,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.397087574005127,
+      "rewards/margins": 0.9368999004364014,
+      "rewards/rejected": -3.333987236022949,
+      "step": 503
+    },
+    {
+      "epoch": 0.9155313351498637,
+      "grad_norm": 2.179532527923584,
+      "learning_rate": 8.883985195544617e-06,
+      "logits/chosen": 0.03752445429563522,
+      "logits/rejected": 0.024428365752100945,
+      "logps/chosen": -83.81852722167969,
+      "logps/rejected": -88.1905288696289,
+      "loss": 2.5361,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.239609956741333,
+      "rewards/margins": 0.4188295900821686,
+      "rewards/rejected": -2.6584396362304688,
+      "step": 504
+    },
+    {
+      "epoch": 0.9173478655767484,
+      "grad_norm": 1.9171282052993774,
+      "learning_rate": 8.879745950386075e-06,
+      "logits/chosen": 0.17986616492271423,
+      "logits/rejected": 0.20335282385349274,
+      "logps/chosen": -72.679931640625,
+      "logps/rejected": -71.41309356689453,
+      "loss": 2.5452,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.261596918106079,
+      "rewards/margins": 0.3737923502922058,
+      "rewards/rejected": -2.6353893280029297,
+      "step": 505
+    },
+    {
+      "epoch": 0.9191643960036331,
+      "grad_norm": 2.3437929153442383,
+      "learning_rate": 8.87549981333707e-06,
+      "logits/chosen": 0.09010382741689682,
+      "logits/rejected": 0.040441811084747314,
+      "logps/chosen": -81.56139373779297,
+      "logps/rejected": -83.21439361572266,
+      "loss": 2.74,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2574033737182617,
+      "rewards/margins": 0.28719204664230347,
+      "rewards/rejected": -2.544595241546631,
+      "step": 506
+    },
+    {
+      "epoch": 0.9209809264305178,
+      "grad_norm": 2.4240329265594482,
+      "learning_rate": 8.871246793056215e-06,
+      "logits/chosen": 0.02888420596718788,
+      "logits/rejected": 0.11542786657810211,
+      "logps/chosen": -84.60144805908203,
+      "logps/rejected": -100.05682373046875,
+      "loss": 2.7052,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.863104820251465,
+      "rewards/margins": 0.41941094398498535,
+      "rewards/rejected": -3.28251576423645,
+      "step": 507
+    },
+    {
+      "epoch": 0.9227974568574023,
+      "grad_norm": 1.9247711896896362,
+      "learning_rate": 8.866986898216157e-06,
+      "logits/chosen": 0.06140238791704178,
+      "logits/rejected": 0.05623817816376686,
+      "logps/chosen": -77.40584564208984,
+      "logps/rejected": -86.71392822265625,
+      "loss": 2.2445,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4979231357574463,
+      "rewards/margins": 0.6161274313926697,
+      "rewards/rejected": -3.11405086517334,
+      "step": 508
+    },
+    {
+      "epoch": 0.924613987284287,
+      "grad_norm": 1.854956030845642,
+      "learning_rate": 8.862720137503568e-06,
+      "logits/chosen": 0.06755004823207855,
+      "logits/rejected": 0.03762562945485115,
+      "logps/chosen": -86.03059387207031,
+      "logps/rejected": -90.12970733642578,
+      "loss": 2.1459,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3017396926879883,
+      "rewards/margins": 0.6131142973899841,
+      "rewards/rejected": -2.914853572845459,
+      "step": 509
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.8044302463531494,
+      "learning_rate": 8.858446519619113e-06,
+      "logits/chosen": 0.08446178585290909,
+      "logits/rejected": 0.08818987011909485,
+      "logps/chosen": -77.58124542236328,
+      "logps/rejected": -85.02752685546875,
+      "loss": 2.1311,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2378087043762207,
+      "rewards/margins": 0.7067832350730896,
+      "rewards/rejected": -2.944591999053955,
+      "step": 510
+    },
+    {
+      "epoch": 0.9282470481380564,
+      "grad_norm": 1.4898866415023804,
+      "learning_rate": 8.854166053277443e-06,
+      "logits/chosen": 0.1039918065071106,
+      "logits/rejected": 0.1396111100912094,
+      "logps/chosen": -87.59841918945312,
+      "logps/rejected": -101.05659484863281,
+      "loss": 2.1092,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1167173385620117,
+      "rewards/margins": 0.7099149227142334,
+      "rewards/rejected": -2.826632499694824,
+      "step": 511
+    },
+    {
+      "epoch": 0.9300635785649409,
+      "grad_norm": 1.8009706735610962,
+      "learning_rate": 8.849878747207175e-06,
+      "logits/chosen": 0.11744043976068497,
+      "logits/rejected": 0.07621707767248154,
+      "logps/chosen": -78.37222290039062,
+      "logps/rejected": -74.32958221435547,
+      "loss": 2.6399,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.4002914428710938,
+      "rewards/margins": 0.21068021655082703,
+      "rewards/rejected": -2.610971450805664,
+      "step": 512
+    },
+    {
+      "epoch": 0.9318801089918256,
+      "grad_norm": 1.6368399858474731,
+      "learning_rate": 8.845584610150871e-06,
+      "logits/chosen": 0.09938757866621017,
+      "logits/rejected": 0.12675486505031586,
+      "logps/chosen": -82.24690246582031,
+      "logps/rejected": -92.60913848876953,
+      "loss": 1.9078,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.0825133323669434,
+      "rewards/margins": 0.8327144384384155,
+      "rewards/rejected": -2.9152278900146484,
+      "step": 513
+    },
+    {
+      "epoch": 0.9336966394187103,
+      "grad_norm": 1.8117483854293823,
+      "learning_rate": 8.841283650865027e-06,
+      "logits/chosen": 0.1355382651090622,
+      "logits/rejected": 0.14869986474514008,
+      "logps/chosen": -85.03866577148438,
+      "logps/rejected": -90.55638885498047,
+      "loss": 2.1243,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3315932750701904,
+      "rewards/margins": 0.6421911120414734,
+      "rewards/rejected": -2.9737846851348877,
+      "step": 514
+    },
+    {
+      "epoch": 0.935513169845595,
+      "grad_norm": 1.785927414894104,
+      "learning_rate": 8.836975878120046e-06,
+      "logits/chosen": 0.11557039618492126,
+      "logits/rejected": 0.11716655641794205,
+      "logps/chosen": -83.99951171875,
+      "logps/rejected": -92.36119079589844,
+      "loss": 2.4256,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.439833402633667,
+      "rewards/margins": 0.38893458247184753,
+      "rewards/rejected": -2.8287675380706787,
+      "step": 515
+    },
+    {
+      "epoch": 0.9373297002724795,
+      "grad_norm": 3.526102066040039,
+      "learning_rate": 8.832661300700228e-06,
+      "logits/chosen": 0.10778439044952393,
+      "logits/rejected": 0.18004637956619263,
+      "logps/chosen": -84.20158386230469,
+      "logps/rejected": -91.33497619628906,
+      "loss": 2.3568,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2032201290130615,
+      "rewards/margins": 0.5167797207832336,
+      "rewards/rejected": -2.7200000286102295,
+      "step": 516
+    },
+    {
+      "epoch": 0.9391462306993642,
+      "grad_norm": 1.7708752155303955,
+      "learning_rate": 8.828339927403745e-06,
+      "logits/chosen": 0.0862819030880928,
+      "logits/rejected": 0.16991934180259705,
+      "logps/chosen": -87.4437026977539,
+      "logps/rejected": -96.60844421386719,
+      "loss": 2.3041,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.230825185775757,
+      "rewards/margins": 0.5197271108627319,
+      "rewards/rejected": -2.7505524158477783,
+      "step": 517
+    },
+    {
+      "epoch": 0.9409627611262489,
+      "grad_norm": 2.1336302757263184,
+      "learning_rate": 8.824011767042631e-06,
+      "logits/chosen": 0.12070942670106888,
+      "logits/rejected": 0.19643370807170868,
+      "logps/chosen": -75.97718811035156,
+      "logps/rejected": -83.26056671142578,
+      "loss": 2.3468,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3768177032470703,
+      "rewards/margins": 0.5026894211769104,
+      "rewards/rejected": -2.879507064819336,
+      "step": 518
+    },
+    {
+      "epoch": 0.9427792915531336,
+      "grad_norm": 2.6361515522003174,
+      "learning_rate": 8.819676828442758e-06,
+      "logits/chosen": 0.04481849446892738,
+      "logits/rejected": -0.011357773095369339,
+      "logps/chosen": -84.09026336669922,
+      "logps/rejected": -87.71636199951172,
+      "loss": 2.8427,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.4249284267425537,
+      "rewards/margins": 0.41216185688972473,
+      "rewards/rejected": -2.837090015411377,
+      "step": 519
+    },
+    {
+      "epoch": 0.9445958219800181,
+      "grad_norm": 2.1068665981292725,
+      "learning_rate": 8.815335120443822e-06,
+      "logits/chosen": 0.1595858335494995,
+      "logits/rejected": 0.14017102122306824,
+      "logps/chosen": -94.00933837890625,
+      "logps/rejected": -96.91961669921875,
+      "loss": 2.6559,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.4731483459472656,
+      "rewards/margins": 0.3141450881958008,
+      "rewards/rejected": -2.7872931957244873,
+      "step": 520
+    },
+    {
+      "epoch": 0.9464123524069028,
+      "grad_norm": 1.7320570945739746,
+      "learning_rate": 8.810986651899322e-06,
+      "logits/chosen": 0.10371927917003632,
+      "logits/rejected": 0.17988254129886627,
+      "logps/chosen": -78.32308197021484,
+      "logps/rejected": -88.74859619140625,
+      "loss": 2.1866,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3973731994628906,
+      "rewards/margins": 0.7199669480323792,
+      "rewards/rejected": -3.117340087890625,
+      "step": 521
+    },
+    {
+      "epoch": 0.9482288828337875,
+      "grad_norm": 1.6995161771774292,
+      "learning_rate": 8.80663143167654e-06,
+      "logits/chosen": 0.14279219508171082,
+      "logits/rejected": 0.1058904379606247,
+      "logps/chosen": -87.79972839355469,
+      "logps/rejected": -94.35545349121094,
+      "loss": 2.1217,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.979785680770874,
+      "rewards/margins": 0.6210839152336121,
+      "rewards/rejected": -2.600869655609131,
+      "step": 522
+    },
+    {
+      "epoch": 0.9500454132606722,
+      "grad_norm": 2.1860764026641846,
+      "learning_rate": 8.80226946865653e-06,
+      "logits/chosen": 0.11601082980632782,
+      "logits/rejected": 0.15693408250808716,
+      "logps/chosen": -88.86433410644531,
+      "logps/rejected": -92.43590545654297,
+      "loss": 2.5629,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.2875983715057373,
+      "rewards/margins": 0.39402255415916443,
+      "rewards/rejected": -2.6816210746765137,
+      "step": 523
+    },
+    {
+      "epoch": 0.9518619436875567,
+      "grad_norm": 1.829032063484192,
+      "learning_rate": 8.797900771734094e-06,
+      "logits/chosen": 0.12328370660543442,
+      "logits/rejected": 0.042389824986457825,
+      "logps/chosen": -77.34864807128906,
+      "logps/rejected": -78.61405944824219,
+      "loss": 2.5327,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2514185905456543,
+      "rewards/margins": 0.3621898591518402,
+      "rewards/rejected": -2.6136088371276855,
+      "step": 524
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 2.235381841659546,
+      "learning_rate": 8.793525349817765e-06,
+      "logits/chosen": 0.12495981156826019,
+      "logits/rejected": 0.13291439414024353,
+      "logps/chosen": -76.7927474975586,
+      "logps/rejected": -86.1208267211914,
+      "loss": 2.5188,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3426284790039062,
+      "rewards/margins": 0.46077418327331543,
+      "rewards/rejected": -2.8034026622772217,
+      "step": 525
+    },
+    {
+      "epoch": 0.9554950045413261,
+      "grad_norm": 1.8071073293685913,
+      "learning_rate": 8.78914321182979e-06,
+      "logits/chosen": 0.13614074885845184,
+      "logits/rejected": 0.1179000735282898,
+      "logps/chosen": -74.44164276123047,
+      "logps/rejected": -77.69221496582031,
+      "loss": 2.6082,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0528411865234375,
+      "rewards/margins": 0.29493850469589233,
+      "rewards/rejected": -2.3477797508239746,
+      "step": 526
+    },
+    {
+      "epoch": 0.9573115349682108,
+      "grad_norm": 1.8623837232589722,
+      "learning_rate": 8.784754366706115e-06,
+      "logits/chosen": 0.02768833190202713,
+      "logits/rejected": 0.09584817290306091,
+      "logps/chosen": -77.13914489746094,
+      "logps/rejected": -86.53968048095703,
+      "loss": 2.2244,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.199755907058716,
+      "rewards/margins": 0.6271941661834717,
+      "rewards/rejected": -2.8269503116607666,
+      "step": 527
+    },
+    {
+      "epoch": 0.9591280653950953,
+      "grad_norm": 1.5299558639526367,
+      "learning_rate": 8.780358823396352e-06,
+      "logits/chosen": 0.11699292808771133,
+      "logits/rejected": 0.10016686469316483,
+      "logps/chosen": -86.1636962890625,
+      "logps/rejected": -87.16045379638672,
+      "loss": 2.0645,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.163857936859131,
+      "rewards/margins": 0.697452187538147,
+      "rewards/rejected": -2.8613100051879883,
+      "step": 528
+    },
+    {
+      "epoch": 0.96094459582198,
+      "grad_norm": 1.6236025094985962,
+      "learning_rate": 8.775956590863785e-06,
+      "logits/chosen": 0.11622033268213272,
+      "logits/rejected": 0.06629584729671478,
+      "logps/chosen": -76.42916870117188,
+      "logps/rejected": -80.24000549316406,
+      "loss": 1.8904,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.102374315261841,
+      "rewards/margins": 0.7230857014656067,
+      "rewards/rejected": -2.825460195541382,
+      "step": 529
+    },
+    {
+      "epoch": 0.9627611262488647,
+      "grad_norm": 1.6119427680969238,
+      "learning_rate": 8.771547678085332e-06,
+      "logits/chosen": 0.07009143382310867,
+      "logits/rejected": 0.16355563700199127,
+      "logps/chosen": -72.83528137207031,
+      "logps/rejected": -87.72441101074219,
+      "loss": 2.0136,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1697566509246826,
+      "rewards/margins": 0.8312336206436157,
+      "rewards/rejected": -3.000990390777588,
+      "step": 530
+    },
+    {
+      "epoch": 0.9645776566757494,
+      "grad_norm": 2.0324013233184814,
+      "learning_rate": 8.767132094051534e-06,
+      "logits/chosen": 0.14837129414081573,
+      "logits/rejected": 0.0700330063700676,
+      "logps/chosen": -84.65350341796875,
+      "logps/rejected": -91.84485626220703,
+      "loss": 2.2551,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5343966484069824,
+      "rewards/margins": 0.6457280516624451,
+      "rewards/rejected": -3.180124521255493,
+      "step": 531
+    },
+    {
+      "epoch": 0.9663941871026339,
+      "grad_norm": 1.4526044130325317,
+      "learning_rate": 8.762709847766532e-06,
+      "logits/chosen": 0.013620391488075256,
+      "logits/rejected": 0.10376289486885071,
+      "logps/chosen": -70.80522155761719,
+      "logps/rejected": -84.26114654541016,
+      "loss": 1.8607,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.249685287475586,
+      "rewards/margins": 0.8104004859924316,
+      "rewards/rejected": -3.0600855350494385,
+      "step": 532
+    },
+    {
+      "epoch": 0.9682107175295186,
+      "grad_norm": 2.0312576293945312,
+      "learning_rate": 8.758280948248059e-06,
+      "logits/chosen": 0.0965358167886734,
+      "logits/rejected": 0.10093791037797928,
+      "logps/chosen": -81.83265686035156,
+      "logps/rejected": -88.52181243896484,
+      "loss": 2.6128,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4901068210601807,
+      "rewards/margins": 0.30889907479286194,
+      "rewards/rejected": -2.7990057468414307,
+      "step": 533
+    },
+    {
+      "epoch": 0.9700272479564033,
+      "grad_norm": 2.547189712524414,
+      "learning_rate": 8.753845404527413e-06,
+      "logits/chosen": 0.020555848255753517,
+      "logits/rejected": 0.0945032387971878,
+      "logps/chosen": -72.16207122802734,
+      "logps/rejected": -86.27986907958984,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.148913860321045,
+      "rewards/margins": 0.7298768162727356,
+      "rewards/rejected": -2.8787906169891357,
+      "step": 534
+    },
+    {
+      "epoch": 0.971843778383288,
+      "grad_norm": 1.996645450592041,
+      "learning_rate": 8.74940322564944e-06,
+      "logits/chosen": 0.0345986932516098,
+      "logits/rejected": 0.08187264204025269,
+      "logps/chosen": -78.76763916015625,
+      "logps/rejected": -91.45555114746094,
+      "loss": 2.4587,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.4772286415100098,
+      "rewards/margins": 0.4609605073928833,
+      "rewards/rejected": -2.9381890296936035,
+      "step": 535
+    },
+    {
+      "epoch": 0.9736603088101726,
+      "grad_norm": 1.8145204782485962,
+      "learning_rate": 8.744954420672514e-06,
+      "logits/chosen": 0.03380711376667023,
+      "logits/rejected": 0.09039584547281265,
+      "logps/chosen": -79.83094787597656,
+      "logps/rejected": -90.87137603759766,
+      "loss": 2.3589,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.1433308124542236,
+      "rewards/margins": 0.5286959409713745,
+      "rewards/rejected": -2.6720268726348877,
+      "step": 536
+    },
+    {
+      "epoch": 0.9754768392370572,
+      "grad_norm": 1.9745460748672485,
+      "learning_rate": 8.740498998668523e-06,
+      "logits/chosen": 0.07233883440494537,
+      "logits/rejected": 0.17097605764865875,
+      "logps/chosen": -72.3458251953125,
+      "logps/rejected": -83.47319793701172,
+      "loss": 2.3469,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199948310852051,
+      "rewards/margins": 0.47777751088142395,
+      "rewards/rejected": -2.6777257919311523,
+      "step": 537
+    },
+    {
+      "epoch": 0.9772933696639419,
+      "grad_norm": 1.665255069732666,
+      "learning_rate": 8.736036968722851e-06,
+      "logits/chosen": 0.09596951305866241,
+      "logits/rejected": 0.06625551730394363,
+      "logps/chosen": -80.11689758300781,
+      "logps/rejected": -88.4573974609375,
+      "loss": 2.1889,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1831865310668945,
+      "rewards/margins": 0.6106584668159485,
+      "rewards/rejected": -2.7938451766967773,
+      "step": 538
+    },
+    {
+      "epoch": 0.9791099000908265,
+      "grad_norm": 1.5587108135223389,
+      "learning_rate": 8.73156833993435e-06,
+      "logits/chosen": 0.07950340211391449,
+      "logits/rejected": 0.0903201699256897,
+      "logps/chosen": -78.9182357788086,
+      "logps/rejected": -92.290283203125,
+      "loss": 2.048,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.12032413482666,
+      "rewards/margins": 0.7009568214416504,
+      "rewards/rejected": -2.8212809562683105,
+      "step": 539
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 1.7177670001983643,
+      "learning_rate": 8.727093121415338e-06,
+      "logits/chosen": 0.06052964925765991,
+      "logits/rejected": 0.14043518900871277,
+      "logps/chosen": -79.04943084716797,
+      "logps/rejected": -88.881591796875,
+      "loss": 2.2089,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2680561542510986,
+      "rewards/margins": 0.5380589962005615,
+      "rewards/rejected": -2.806114912033081,
+      "step": 540
+    },
+    {
+      "epoch": 0.9827429609445958,
+      "grad_norm": 1.7630436420440674,
+      "learning_rate": 8.722611322291558e-06,
+      "logits/chosen": 0.02032918483018875,
+      "logits/rejected": 0.11728623509407043,
+      "logps/chosen": -78.1409912109375,
+      "logps/rejected": -89.3685073852539,
+      "loss": 2.3303,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.356957197189331,
+      "rewards/margins": 0.5139226317405701,
+      "rewards/rejected": -2.870880126953125,
+      "step": 541
+    },
+    {
+      "epoch": 0.9845594913714805,
+      "grad_norm": 1.852418065071106,
+      "learning_rate": 8.718122951702183e-06,
+      "logits/chosen": 0.0439969077706337,
+      "logits/rejected": 0.16671347618103027,
+      "logps/chosen": -69.96346282958984,
+      "logps/rejected": -85.24410247802734,
+      "loss": 2.1393,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3295838832855225,
+      "rewards/margins": 0.6544574499130249,
+      "rewards/rejected": -2.984041213989258,
+      "step": 542
+    },
+    {
+      "epoch": 0.9863760217983651,
+      "grad_norm": 1.9028044939041138,
+      "learning_rate": 8.713628018799782e-06,
+      "logits/chosen": 0.04875154793262482,
+      "logits/rejected": -0.009283583611249924,
+      "logps/chosen": -75.80145263671875,
+      "logps/rejected": -78.19249725341797,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.241586208343506,
+      "rewards/margins": 0.4672737419605255,
+      "rewards/rejected": -2.708860158920288,
+      "step": 543
+    },
+    {
+      "epoch": 0.9881925522252498,
+      "grad_norm": 1.5648458003997803,
+      "learning_rate": 8.709126532750304e-06,
+      "logits/chosen": 0.02982347458600998,
+      "logits/rejected": 0.07835812866687775,
+      "logps/chosen": -75.35140991210938,
+      "logps/rejected": -91.11531829833984,
+      "loss": 2.0945,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.062386989593506,
+      "rewards/margins": 0.6832989454269409,
+      "rewards/rejected": -2.7456860542297363,
+      "step": 544
+    },
+    {
+      "epoch": 0.9900090826521344,
+      "grad_norm": 1.6214131116867065,
+      "learning_rate": 8.704618502733064e-06,
+      "logits/chosen": 0.09551770985126495,
+      "logits/rejected": 0.11310344189405441,
+      "logps/chosen": -80.15961456298828,
+      "logps/rejected": -93.38126373291016,
+      "loss": 2.1642,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.255985736846924,
+      "rewards/margins": 0.7712126970291138,
+      "rewards/rejected": -3.027198553085327,
+      "step": 545
+    },
+    {
+      "epoch": 0.9918256130790191,
+      "grad_norm": 1.762953519821167,
+      "learning_rate": 8.70010393794072e-06,
+      "logits/chosen": 0.0971008688211441,
+      "logits/rejected": 0.16590046882629395,
+      "logps/chosen": -86.24911499023438,
+      "logps/rejected": -102.59831237792969,
+      "loss": 2.0951,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3542394638061523,
+      "rewards/margins": 0.7857018709182739,
+      "rewards/rejected": -3.139941453933716,
+      "step": 546
+    },
+    {
+      "epoch": 0.9936421435059037,
+      "grad_norm": 1.6033780574798584,
+      "learning_rate": 8.695582847579254e-06,
+      "logits/chosen": 0.15147003531455994,
+      "logits/rejected": 0.13297875225543976,
+      "logps/chosen": -78.5172348022461,
+      "logps/rejected": -86.49285888671875,
+      "loss": 2.1101,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2025821208953857,
+      "rewards/margins": 0.7471145391464233,
+      "rewards/rejected": -2.9496963024139404,
+      "step": 547
+    },
+    {
+      "epoch": 0.9954586739327884,
+      "grad_norm": 1.4905331134796143,
+      "learning_rate": 8.69105524086796e-06,
+      "logits/chosen": 0.07559307664632797,
+      "logits/rejected": 0.09286531805992126,
+      "logps/chosen": -74.92215728759766,
+      "logps/rejected": -78.23614501953125,
+      "loss": 2.2745,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.24694561958313,
+      "rewards/margins": 0.5010396838188171,
+      "rewards/rejected": -2.747985363006592,
+      "step": 548
+    },
+    {
+      "epoch": 0.997275204359673,
+      "grad_norm": 1.823449730873108,
+      "learning_rate": 8.686521127039414e-06,
+      "logits/chosen": 0.03763032704591751,
+      "logits/rejected": 0.06642220914363861,
+      "logps/chosen": -79.37518310546875,
+      "logps/rejected": -87.17401885986328,
+      "loss": 2.5954,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158803701400757,
+      "rewards/margins": 0.4884320795536041,
+      "rewards/rejected": -2.647235870361328,
+      "step": 549
+    },
+    {
+      "epoch": 0.9990917347865577,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.681980515339464e-06,
+      "logits/chosen": 0.08186414837837219,
+      "logits/rejected": 0.08158639073371887,
+      "logps/chosen": -84.89871215820312,
+      "logps/rejected": -85.87324523925781,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.1884236335754395,
+      "rewards/margins": 0.2923518121242523,
+      "rewards/rejected": -2.4807753562927246,
+      "step": 550
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.677433415027209e-06,
+      "logits/chosen": 0.18737269937992096,
+      "logits/rejected": 0.11028257012367249,
+      "logps/chosen": -76.3554916381836,
+      "logps/rejected": -86.39812469482422,
+      "loss": 1.1517,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1043050289154053,
+      "rewards/margins": 0.7949569225311279,
+      "rewards/rejected": -2.8992621898651123,
+      "step": 551
+    },
+    {
+      "epoch": 1.0018165304268847,
+      "grad_norm": 1.6554900407791138,
+      "learning_rate": 8.672879835374976e-06,
+      "logits/chosen": 0.11259499937295914,
+      "logits/rejected": 0.07095417380332947,
+      "logps/chosen": -84.55628967285156,
+      "logps/rejected": -88.55287170410156,
+      "loss": 2.1275,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.2375316619873047,
+      "rewards/margins": 0.5453992486000061,
+      "rewards/rejected": -2.782930850982666,
+      "step": 552
+    },
+    {
+      "epoch": 1.0036330608537694,
+      "grad_norm": 1.6807721853256226,
+      "learning_rate": 8.668319785668307e-06,
+      "logits/chosen": 0.14847862720489502,
+      "logits/rejected": 0.1863957643508911,
+      "logps/chosen": -87.27869415283203,
+      "logps/rejected": -89.04036712646484,
+      "loss": 2.3209,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.087535858154297,
+      "rewards/margins": 0.5243443846702576,
+      "rewards/rejected": -2.61188006401062,
+      "step": 553
+    },
+    {
+      "epoch": 1.005449591280654,
+      "grad_norm": 1.8480761051177979,
+      "learning_rate": 8.66375327520594e-06,
+      "logits/chosen": 0.07236558198928833,
+      "logits/rejected": 0.13397395610809326,
+      "logps/chosen": -85.43444061279297,
+      "logps/rejected": -101.47248840332031,
+      "loss": 1.7106,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.157266616821289,
+      "rewards/margins": 0.9501473903656006,
+      "rewards/rejected": -3.1074140071868896,
+      "step": 554
+    },
+    {
+      "epoch": 1.0072661217075387,
+      "grad_norm": 1.586917757987976,
+      "learning_rate": 8.659180313299783e-06,
+      "logits/chosen": 0.05113527923822403,
+      "logits/rejected": 0.14182163774967194,
+      "logps/chosen": -78.31108856201172,
+      "logps/rejected": -90.46333312988281,
+      "loss": 1.8965,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.945315957069397,
+      "rewards/margins": 0.8104385137557983,
+      "rewards/rejected": -2.755754232406616,
+      "step": 555
+    },
+    {
+      "epoch": 1.0090826521344232,
+      "grad_norm": 1.406339406967163,
+      "learning_rate": 8.6546009092749e-06,
+      "logits/chosen": -0.01422208547592163,
+      "logits/rejected": 0.10859756916761398,
+      "logps/chosen": -75.02066040039062,
+      "logps/rejected": -87.69715118408203,
+      "loss": 2.027,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9996552467346191,
+      "rewards/margins": 0.7365851402282715,
+      "rewards/rejected": -2.7362406253814697,
+      "step": 556
+    },
+    {
+      "epoch": 1.0108991825613078,
+      "grad_norm": 1.5511010885238647,
+      "learning_rate": 8.650015072469496e-06,
+      "logits/chosen": 0.07154582440853119,
+      "logits/rejected": 0.11479673534631729,
+      "logps/chosen": -88.33872985839844,
+      "logps/rejected": -96.75527954101562,
+      "loss": 1.966,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2190096378326416,
+      "rewards/margins": 0.861441969871521,
+      "rewards/rejected": -3.080451250076294,
+      "step": 557
+    },
+    {
+      "epoch": 1.0127157129881925,
+      "grad_norm": 1.5951570272445679,
+      "learning_rate": 8.645422812234888e-06,
+      "logits/chosen": 0.17342886328697205,
+      "logits/rejected": 0.11909954994916916,
+      "logps/chosen": -79.63985443115234,
+      "logps/rejected": -83.41281127929688,
+      "loss": 2.1115,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3469128608703613,
+      "rewards/margins": 0.7317838668823242,
+      "rewards/rejected": -3.0786967277526855,
+      "step": 558
+    },
+    {
+      "epoch": 1.0145322434150772,
+      "grad_norm": 1.891517162322998,
+      "learning_rate": 8.640824137935498e-06,
+      "logits/chosen": 0.14266839623451233,
+      "logits/rejected": 0.14209085702896118,
+      "logps/chosen": -84.77491760253906,
+      "logps/rejected": -87.46759796142578,
+      "loss": 2.2691,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.022730827331543,
+      "rewards/margins": 0.42655885219573975,
+      "rewards/rejected": -2.449289560317993,
+      "step": 559
+    },
+    {
+      "epoch": 1.0163487738419619,
+      "grad_norm": 1.7919303178787231,
+      "learning_rate": 8.636219058948823e-06,
+      "logits/chosen": 0.01973957009613514,
+      "logits/rejected": 0.12187746167182922,
+      "logps/chosen": -78.26383209228516,
+      "logps/rejected": -92.714599609375,
+      "loss": 1.974,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2774088382720947,
+      "rewards/margins": 0.8670985698699951,
+      "rewards/rejected": -3.1445071697235107,
+      "step": 560
+    },
+    {
+      "epoch": 1.0181653042688465,
+      "grad_norm": 1.5417025089263916,
+      "learning_rate": 8.631607584665415e-06,
+      "logits/chosen": 0.07740732282400131,
+      "logits/rejected": 0.17052005231380463,
+      "logps/chosen": -83.54193115234375,
+      "logps/rejected": -95.1084976196289,
+      "loss": 1.9963,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1598939895629883,
+      "rewards/margins": 0.802270233631134,
+      "rewards/rejected": -2.9621646404266357,
+      "step": 561
+    },
+    {
+      "epoch": 1.0199818346957312,
+      "grad_norm": 1.9338072538375854,
+      "learning_rate": 8.62698972448888e-06,
+      "logits/chosen": 0.09191437065601349,
+      "logits/rejected": 0.1776769608259201,
+      "logps/chosen": -82.68082427978516,
+      "logps/rejected": -96.89566802978516,
+      "loss": 2.0399,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3505523204803467,
+      "rewards/margins": 0.7893310785293579,
+      "rewards/rejected": -3.1398837566375732,
+      "step": 562
+    },
+    {
+      "epoch": 1.021798365122616,
+      "grad_norm": 1.609601378440857,
+      "learning_rate": 8.622365487835836e-06,
+      "logits/chosen": 0.13105043768882751,
+      "logits/rejected": 0.138211190700531,
+      "logps/chosen": -77.19727325439453,
+      "logps/rejected": -84.1854476928711,
+      "loss": 2.3184,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.13655948638916,
+      "rewards/margins": 0.5744473934173584,
+      "rewards/rejected": -2.7110068798065186,
+      "step": 563
+    },
+    {
+      "epoch": 1.0236148955495004,
+      "grad_norm": 2.1222517490386963,
+      "learning_rate": 8.617734884135906e-06,
+      "logits/chosen": 0.14744582772254944,
+      "logits/rejected": 0.053739145398139954,
+      "logps/chosen": -81.11271667480469,
+      "logps/rejected": -81.40655517578125,
+      "loss": 2.2,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.419705390930176,
+      "rewards/margins": 0.7198007702827454,
+      "rewards/rejected": -3.1395058631896973,
+      "step": 564
+    },
+    {
+      "epoch": 1.025431425976385,
+      "grad_norm": 1.8075231313705444,
+      "learning_rate": 8.613097922831699e-06,
+      "logits/chosen": 0.07210052758455276,
+      "logits/rejected": 0.07022904604673386,
+      "logps/chosen": -88.72612762451172,
+      "logps/rejected": -98.2815933227539,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1057255268096924,
+      "rewards/margins": 0.8747395873069763,
+      "rewards/rejected": -2.9804654121398926,
+      "step": 565
+    },
+    {
+      "epoch": 1.0272479564032697,
+      "grad_norm": 1.3381479978561401,
+      "learning_rate": 8.608454613378783e-06,
+      "logits/chosen": 0.08187520503997803,
+      "logits/rejected": 0.1098201796412468,
+      "logps/chosen": -77.1299819946289,
+      "logps/rejected": -89.75099182128906,
+      "loss": 1.8687,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3585519790649414,
+      "rewards/margins": 0.8574986457824707,
+      "rewards/rejected": -3.216050386428833,
+      "step": 566
+    },
+    {
+      "epoch": 1.0290644868301544,
+      "grad_norm": 1.6325740814208984,
+      "learning_rate": 8.603804965245678e-06,
+      "logits/chosen": 0.12077341973781586,
+      "logits/rejected": 0.09103713929653168,
+      "logps/chosen": -74.03280639648438,
+      "logps/rejected": -79.61197662353516,
+      "loss": 2.0884,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.430152654647827,
+      "rewards/margins": 0.6819649934768677,
+      "rewards/rejected": -3.1121177673339844,
+      "step": 567
+    },
+    {
+      "epoch": 1.030881017257039,
+      "grad_norm": 1.782224178314209,
+      "learning_rate": 8.599148987913829e-06,
+      "logits/chosen": -0.008054995909333229,
+      "logits/rejected": 0.08848709613084793,
+      "logps/chosen": -88.39682006835938,
+      "logps/rejected": -98.20311737060547,
+      "loss": 2.3203,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.21012282371521,
+      "rewards/margins": 0.5451631546020508,
+      "rewards/rejected": -2.75528621673584,
+      "step": 568
+    },
+    {
+      "epoch": 1.0326975476839237,
+      "grad_norm": 1.7691792249679565,
+      "learning_rate": 8.594486690877577e-06,
+      "logits/chosen": 0.05038870871067047,
+      "logits/rejected": -0.0020756814628839493,
+      "logps/chosen": -88.35281372070312,
+      "logps/rejected": -87.42422485351562,
+      "loss": 2.1789,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6035892963409424,
+      "rewards/margins": 0.559741735458374,
+      "rewards/rejected": -3.1633315086364746,
+      "step": 569
+    },
+    {
+      "epoch": 1.0345140781108084,
+      "grad_norm": 1.7002201080322266,
+      "learning_rate": 8.589818083644161e-06,
+      "logits/chosen": 0.030099213123321533,
+      "logits/rejected": 0.10112117975950241,
+      "logps/chosen": -74.28582763671875,
+      "logps/rejected": -85.93130493164062,
+      "loss": 2.2068,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.324796676635742,
+      "rewards/margins": 0.5959895849227905,
+      "rewards/rejected": -2.9207863807678223,
+      "step": 570
+    },
+    {
+      "epoch": 1.036330608537693,
+      "grad_norm": 1.5611207485198975,
+      "learning_rate": 8.585143175733686e-06,
+      "logits/chosen": 0.09465142339468002,
+      "logits/rejected": 0.13904502987861633,
+      "logps/chosen": -86.5047378540039,
+      "logps/rejected": -95.09537506103516,
+      "loss": 1.9628,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5564661026000977,
+      "rewards/margins": 0.8365517258644104,
+      "rewards/rejected": -3.3930182456970215,
+      "step": 571
+    },
+    {
+      "epoch": 1.0381471389645776,
+      "grad_norm": 1.7271509170532227,
+      "learning_rate": 8.5804619766791e-06,
+      "logits/chosen": 0.08665300905704498,
+      "logits/rejected": 0.10898350924253464,
+      "logps/chosen": -85.17437744140625,
+      "logps/rejected": -94.50398254394531,
+      "loss": 2.2717,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3690195083618164,
+      "rewards/margins": 0.6290315389633179,
+      "rewards/rejected": -2.9980509281158447,
+      "step": 572
+    },
+    {
+      "epoch": 1.0399636693914622,
+      "grad_norm": 1.7832714319229126,
+      "learning_rate": 8.575774496026184e-06,
+      "logits/chosen": 0.033456090837717056,
+      "logits/rejected": 0.09503711760044098,
+      "logps/chosen": -87.27224731445312,
+      "logps/rejected": -94.98229217529297,
+      "loss": 2.1781,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.509322166442871,
+      "rewards/margins": 0.5943110585212708,
+      "rewards/rejected": -3.103632926940918,
+      "step": 573
+    },
+    {
+      "epoch": 1.041780199818347,
+      "grad_norm": 1.8239610195159912,
+      "learning_rate": 8.571080743333528e-06,
+      "logits/chosen": 0.11436322331428528,
+      "logits/rejected": 0.06905262917280197,
+      "logps/chosen": -81.12911987304688,
+      "logps/rejected": -88.73529052734375,
+      "loss": 2.1358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2652292251586914,
+      "rewards/margins": 0.6765917539596558,
+      "rewards/rejected": -2.9418206214904785,
+      "step": 574
+    },
+    {
+      "epoch": 1.0435967302452316,
+      "grad_norm": 1.9338902235031128,
+      "learning_rate": 8.566380728172512e-06,
+      "logits/chosen": 0.11198662221431732,
+      "logits/rejected": 0.11845988035202026,
+      "logps/chosen": -82.55985260009766,
+      "logps/rejected": -94.05873107910156,
+      "loss": 1.9609,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3074758052825928,
+      "rewards/margins": 0.8309231400489807,
+      "rewards/rejected": -3.1383988857269287,
+      "step": 575
+    },
+    {
+      "epoch": 1.0454132606721163,
+      "grad_norm": 1.6932177543640137,
+      "learning_rate": 8.561674460127281e-06,
+      "logits/chosen": 0.0954650342464447,
+      "logits/rejected": 0.139107346534729,
+      "logps/chosen": -83.95409393310547,
+      "logps/rejected": -98.31000518798828,
+      "loss": 2.311,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.476616859436035,
+      "rewards/margins": 0.7193690538406372,
+      "rewards/rejected": -3.195985794067383,
+      "step": 576
+    },
+    {
+      "epoch": 1.047229791099001,
+      "grad_norm": 1.6585469245910645,
+      "learning_rate": 8.556961948794738e-06,
+      "logits/chosen": 0.050676412880420685,
+      "logits/rejected": 0.13941077888011932,
+      "logps/chosen": -91.46603393554688,
+      "logps/rejected": -111.38130187988281,
+      "loss": 1.9574,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.630396604537964,
+      "rewards/margins": 0.8525002002716064,
+      "rewards/rejected": -3.4828968048095703,
+      "step": 577
+    },
+    {
+      "epoch": 1.0490463215258856,
+      "grad_norm": 1.8747247457504272,
+      "learning_rate": 8.552243203784514e-06,
+      "logits/chosen": 0.06109774485230446,
+      "logits/rejected": 0.03656052052974701,
+      "logps/chosen": -84.25686645507812,
+      "logps/rejected": -99.20081329345703,
+      "loss": 2.0244,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4541397094726562,
+      "rewards/margins": 1.0175232887268066,
+      "rewards/rejected": -3.471662759780884,
+      "step": 578
+    },
+    {
+      "epoch": 1.0508628519527703,
+      "grad_norm": 1.714175820350647,
+      "learning_rate": 8.547518234718954e-06,
+      "logits/chosen": 0.07959345728158951,
+      "logits/rejected": 0.17609870433807373,
+      "logps/chosen": -79.58773040771484,
+      "logps/rejected": -98.85990142822266,
+      "loss": 1.7738,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.431654930114746,
+      "rewards/margins": 1.0072871446609497,
+      "rewards/rejected": -3.4389421939849854,
+      "step": 579
+    },
+    {
+      "epoch": 1.052679382379655,
+      "grad_norm": 1.6469632387161255,
+      "learning_rate": 8.542787051233088e-06,
+      "logits/chosen": 0.06128916144371033,
+      "logits/rejected": 0.11901555210351944,
+      "logps/chosen": -90.99638366699219,
+      "logps/rejected": -104.79571533203125,
+      "loss": 1.8789,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3995859622955322,
+      "rewards/margins": 0.9405049085617065,
+      "rewards/rejected": -3.340090751647949,
+      "step": 580
+    },
+    {
+      "epoch": 1.0544959128065394,
+      "grad_norm": 1.8093998432159424,
+      "learning_rate": 8.538049662974625e-06,
+      "logits/chosen": 0.136922687292099,
+      "logits/rejected": 0.06489068269729614,
+      "logps/chosen": -84.12126159667969,
+      "logps/rejected": -85.51757049560547,
+      "loss": 2.2372,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7452549934387207,
+      "rewards/margins": 0.6496641635894775,
+      "rewards/rejected": -3.394918918609619,
+      "step": 581
+    },
+    {
+      "epoch": 1.056312443233424,
+      "grad_norm": 1.761516809463501,
+      "learning_rate": 8.533306079603928e-06,
+      "logits/chosen": 0.02220803126692772,
+      "logits/rejected": 0.03564752638339996,
+      "logps/chosen": -91.6943359375,
+      "logps/rejected": -105.33677673339844,
+      "loss": 1.831,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.1495776176452637,
+      "rewards/margins": 0.8695197701454163,
+      "rewards/rejected": -3.019097328186035,
+      "step": 582
+    },
+    {
+      "epoch": 1.0581289736603088,
+      "grad_norm": 1.8054577112197876,
+      "learning_rate": 8.52855631079398e-06,
+      "logits/chosen": 0.044406406581401825,
+      "logits/rejected": 0.13649103045463562,
+      "logps/chosen": -83.07669830322266,
+      "logps/rejected": -99.42033386230469,
+      "loss": 2.1295,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.674705743789673,
+      "rewards/margins": 0.8526960015296936,
+      "rewards/rejected": -3.52740216255188,
+      "step": 583
+    },
+    {
+      "epoch": 1.0599455040871935,
+      "grad_norm": 2.092067241668701,
+      "learning_rate": 8.523800366230397e-06,
+      "logits/chosen": 0.11083705723285675,
+      "logits/rejected": 0.13336583971977234,
+      "logps/chosen": -82.55166625976562,
+      "logps/rejected": -88.78050994873047,
+      "loss": 2.783,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.7093071937561035,
+      "rewards/margins": 0.34429922699928284,
+      "rewards/rejected": -3.0536062717437744,
+      "step": 584
+    },
+    {
+      "epoch": 1.0617620345140781,
+      "grad_norm": 2.0747568607330322,
+      "learning_rate": 8.519038255611372e-06,
+      "logits/chosen": 0.06060079485177994,
+      "logits/rejected": 0.06341060996055603,
+      "logps/chosen": -82.93423461914062,
+      "logps/rejected": -85.8175048828125,
+      "loss": 2.1046,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6574935913085938,
+      "rewards/margins": 0.7763374447822571,
+      "rewards/rejected": -3.433830738067627,
+      "step": 585
+    },
+    {
+      "epoch": 1.0635785649409628,
+      "grad_norm": 2.1435482501983643,
+      "learning_rate": 8.51426998864768e-06,
+      "logits/chosen": 0.044051673263311386,
+      "logits/rejected": 0.08015372604131699,
+      "logps/chosen": -86.49735260009766,
+      "logps/rejected": -97.29399108886719,
+      "loss": 2.1365,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.620311737060547,
+      "rewards/margins": 0.6845528483390808,
+      "rewards/rejected": -3.3048648834228516,
+      "step": 586
+    },
+    {
+      "epoch": 1.0653950953678475,
+      "grad_norm": 1.65945303440094,
+      "learning_rate": 8.509495575062647e-06,
+      "logits/chosen": 0.15405744314193726,
+      "logits/rejected": 0.12232419848442078,
+      "logps/chosen": -73.09071350097656,
+      "logps/rejected": -74.24645233154297,
+      "loss": 2.1912,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3465700149536133,
+      "rewards/margins": 0.6234737634658813,
+      "rewards/rejected": -2.970043420791626,
+      "step": 587
+    },
+    {
+      "epoch": 1.067211625794732,
+      "grad_norm": 2.166276216506958,
+      "learning_rate": 8.504715024592132e-06,
+      "logits/chosen": -0.05482687056064606,
+      "logits/rejected": -0.003489813767373562,
+      "logps/chosen": -84.82199096679688,
+      "logps/rejected": -97.53748321533203,
+      "loss": 2.1753,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5944690704345703,
+      "rewards/margins": 0.6753374934196472,
+      "rewards/rejected": -3.269806385040283,
+      "step": 588
+    },
+    {
+      "epoch": 1.0690281562216166,
+      "grad_norm": 1.8869906663894653,
+      "learning_rate": 8.499928346984512e-06,
+      "logits/chosen": 0.13489654660224915,
+      "logits/rejected": 0.052815720438957214,
+      "logps/chosen": -90.98455047607422,
+      "logps/rejected": -91.64305877685547,
+      "loss": 2.3005,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.8705921173095703,
+      "rewards/margins": 0.555056095123291,
+      "rewards/rejected": -3.4256482124328613,
+      "step": 589
+    },
+    {
+      "epoch": 1.0708446866485013,
+      "grad_norm": 2.286853790283203,
+      "learning_rate": 8.495135552000657e-06,
+      "logits/chosen": -0.0365738607943058,
+      "logits/rejected": 0.09564587473869324,
+      "logps/chosen": -93.99796295166016,
+      "logps/rejected": -111.09819793701172,
+      "loss": 2.1574,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.484596014022827,
+      "rewards/margins": 0.840925395488739,
+      "rewards/rejected": -3.325521469116211,
+      "step": 590
+    },
+    {
+      "epoch": 1.072661217075386,
+      "grad_norm": 1.7640435695648193,
+      "learning_rate": 8.490336649413909e-06,
+      "logits/chosen": 0.06625102460384369,
+      "logits/rejected": 0.14394444227218628,
+      "logps/chosen": -74.65644836425781,
+      "logps/rejected": -92.41899871826172,
+      "loss": 2.0387,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.453733205795288,
+      "rewards/margins": 0.7862347364425659,
+      "rewards/rejected": -3.2399678230285645,
+      "step": 591
+    },
+    {
+      "epoch": 1.0744777475022707,
+      "grad_norm": 1.5737273693084717,
+      "learning_rate": 8.485531649010063e-06,
+      "logits/chosen": -0.0012233639135956764,
+      "logits/rejected": 0.021886199712753296,
+      "logps/chosen": -79.36224365234375,
+      "logps/rejected": -94.19145965576172,
+      "loss": 2.3886,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.6413750648498535,
+      "rewards/margins": 0.7800988554954529,
+      "rewards/rejected": -3.421473979949951,
+      "step": 592
+    },
+    {
+      "epoch": 1.0762942779291553,
+      "grad_norm": 2.4172728061676025,
+      "learning_rate": 8.480720560587356e-06,
+      "logits/chosen": 0.04091513529419899,
+      "logits/rejected": 0.16047289967536926,
+      "logps/chosen": -76.38783264160156,
+      "logps/rejected": -94.16845703125,
+      "loss": 2.2219,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.719104290008545,
+      "rewards/margins": 0.9459499716758728,
+      "rewards/rejected": -3.6650545597076416,
+      "step": 593
+    },
+    {
+      "epoch": 1.07811080835604,
+      "grad_norm": 1.6836284399032593,
+      "learning_rate": 8.475903393956434e-06,
+      "logits/chosen": 0.07094185054302216,
+      "logits/rejected": 0.09420361369848251,
+      "logps/chosen": -82.47905731201172,
+      "logps/rejected": -90.06558990478516,
+      "loss": 2.1268,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3819637298583984,
+      "rewards/margins": 0.5947781801223755,
+      "rewards/rejected": -2.9767420291900635,
+      "step": 594
+    },
+    {
+      "epoch": 1.0799273387829247,
+      "grad_norm": 1.808720588684082,
+      "learning_rate": 8.471080158940336e-06,
+      "logits/chosen": 0.010593242943286896,
+      "logits/rejected": 0.03635484725236893,
+      "logps/chosen": -82.94917297363281,
+      "logps/rejected": -94.52174377441406,
+      "loss": 1.9352,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.120471239089966,
+      "rewards/margins": 0.8991338014602661,
+      "rewards/rejected": -3.0196051597595215,
+      "step": 595
+    },
+    {
+      "epoch": 1.0817438692098094,
+      "grad_norm": 1.809515118598938,
+      "learning_rate": 8.46625086537448e-06,
+      "logits/chosen": 0.0202183797955513,
+      "logits/rejected": 0.1138802096247673,
+      "logps/chosen": -84.1740951538086,
+      "logps/rejected": -99.66446685791016,
+      "loss": 2.1384,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7415525913238525,
+      "rewards/margins": 0.8209500908851624,
+      "rewards/rejected": -3.562502861022949,
+      "step": 596
+    },
+    {
+      "epoch": 1.0835603996366938,
+      "grad_norm": 2.023786783218384,
+      "learning_rate": 8.46141552310664e-06,
+      "logits/chosen": 0.11318185180425644,
+      "logits/rejected": 0.1275079846382141,
+      "logps/chosen": -78.02458953857422,
+      "logps/rejected": -89.36206817626953,
+      "loss": 1.9695,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5616073608398438,
+      "rewards/margins": 0.7586690783500671,
+      "rewards/rejected": -3.3202764987945557,
+      "step": 597
+    },
+    {
+      "epoch": 1.0853769300635785,
+      "grad_norm": 1.7670155763626099,
+      "learning_rate": 8.456574141996913e-06,
+      "logits/chosen": 0.09054378420114517,
+      "logits/rejected": 0.17597083747386932,
+      "logps/chosen": -79.7802505493164,
+      "logps/rejected": -95.21873474121094,
+      "loss": 1.9499,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.336801767349243,
+      "rewards/margins": 0.7675365209579468,
+      "rewards/rejected": -3.1043381690979004,
+      "step": 598
+    },
+    {
+      "epoch": 1.0871934604904632,
+      "grad_norm": 2.033724308013916,
+      "learning_rate": 8.451726731917722e-06,
+      "logits/chosen": 0.08810828626155853,
+      "logits/rejected": 0.1388731151819229,
+      "logps/chosen": -75.62480163574219,
+      "logps/rejected": -90.9974365234375,
+      "loss": 1.993,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2493839263916016,
+      "rewards/margins": 0.9519067406654358,
+      "rewards/rejected": -3.2012906074523926,
+      "step": 599
+    },
+    {
+      "epoch": 1.0890099909173478,
+      "grad_norm": 2.0463502407073975,
+      "learning_rate": 8.446873302753783e-06,
+      "logits/chosen": -0.017319753766059875,
+      "logits/rejected": 0.026935823261737823,
+      "logps/chosen": -91.20410919189453,
+      "logps/rejected": -97.56501770019531,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5798678398132324,
+      "rewards/margins": 0.46915650367736816,
+      "rewards/rejected": -3.0490241050720215,
+      "step": 600
+    },
+    {
+      "epoch": 1.0908265213442325,
+      "grad_norm": 1.944875717163086,
+      "learning_rate": 8.44201386440208e-06,
+      "logits/chosen": 0.1267091929912567,
+      "logits/rejected": 0.12764671444892883,
+      "logps/chosen": -82.59441375732422,
+      "logps/rejected": -91.89214324951172,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.495068311691284,
+      "rewards/margins": 0.6413819789886475,
+      "rewards/rejected": -3.1364500522613525,
+      "step": 601
+    },
+    {
+      "epoch": 1.0926430517711172,
+      "grad_norm": 1.8729236125946045,
+      "learning_rate": 8.437148426771852e-06,
+      "logits/chosen": 0.15159347653388977,
+      "logits/rejected": 0.16930025815963745,
+      "logps/chosen": -80.54000854492188,
+      "logps/rejected": -93.16973114013672,
+      "loss": 1.8605,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4500646591186523,
+      "rewards/margins": 0.9991447925567627,
+      "rewards/rejected": -3.449209690093994,
+      "step": 602
+    },
+    {
+      "epoch": 1.0944595821980019,
+      "grad_norm": 1.9284056425094604,
+      "learning_rate": 8.432276999784578e-06,
+      "logits/chosen": 0.16441625356674194,
+      "logits/rejected": 0.17125487327575684,
+      "logps/chosen": -83.31804656982422,
+      "logps/rejected": -91.39764404296875,
+      "loss": 1.9416,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.577606201171875,
+      "rewards/margins": 0.8178077936172485,
+      "rewards/rejected": -3.395413875579834,
+      "step": 603
+    },
+    {
+      "epoch": 1.0962761126248866,
+      "grad_norm": 2.0095536708831787,
+      "learning_rate": 8.427399593373941e-06,
+      "logits/chosen": 0.03275620937347412,
+      "logits/rejected": 0.033026132732629776,
+      "logps/chosen": -83.01813507080078,
+      "logps/rejected": -93.54053497314453,
+      "loss": 2.0358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.488004684448242,
+      "rewards/margins": 0.6699528694152832,
+      "rewards/rejected": -3.1579573154449463,
+      "step": 604
+    },
+    {
+      "epoch": 1.0980926430517712,
+      "grad_norm": 1.8138859272003174,
+      "learning_rate": 8.422516217485826e-06,
+      "logits/chosen": 0.1295854151248932,
+      "logits/rejected": 0.15053214132785797,
+      "logps/chosen": -87.74354553222656,
+      "logps/rejected": -94.38423919677734,
+      "loss": 2.6379,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.538604259490967,
+      "rewards/margins": 0.44896796345710754,
+      "rewards/rejected": -2.987572193145752,
+      "step": 605
+    },
+    {
+      "epoch": 1.0999091734786557,
+      "grad_norm": 2.262779951095581,
+      "learning_rate": 8.417626882078287e-06,
+      "logits/chosen": 0.07311846315860748,
+      "logits/rejected": 0.09329289197921753,
+      "logps/chosen": -88.92317199707031,
+      "logps/rejected": -105.15555572509766,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.468625783920288,
+      "rewards/margins": 1.0499695539474487,
+      "rewards/rejected": -3.5185956954956055,
+      "step": 606
+    },
+    {
+      "epoch": 1.1017257039055404,
+      "grad_norm": 1.7983455657958984,
+      "learning_rate": 8.412731597121527e-06,
+      "logits/chosen": 0.06970179080963135,
+      "logits/rejected": 0.170780211687088,
+      "logps/chosen": -84.99759674072266,
+      "logps/rejected": -91.03089904785156,
+      "loss": 2.3611,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.31594181060791,
+      "rewards/margins": 0.46637701988220215,
+      "rewards/rejected": -2.7823188304901123,
+      "step": 607
+    },
+    {
+      "epoch": 1.103542234332425,
+      "grad_norm": 1.8828606605529785,
+      "learning_rate": 8.407830372597884e-06,
+      "logits/chosen": 0.10343387722969055,
+      "logits/rejected": 0.15009143948554993,
+      "logps/chosen": -74.23157501220703,
+      "logps/rejected": -85.4384994506836,
+      "loss": 2.0535,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3184473514556885,
+      "rewards/margins": 0.7085882425308228,
+      "rewards/rejected": -3.027035713195801,
+      "step": 608
+    },
+    {
+      "epoch": 1.1053587647593097,
+      "grad_norm": 1.8570231199264526,
+      "learning_rate": 8.402923218501813e-06,
+      "logits/chosen": 0.0035794638097286224,
+      "logits/rejected": 0.044682763516902924,
+      "logps/chosen": -82.44414520263672,
+      "logps/rejected": -93.22355651855469,
+      "loss": 2.1751,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.584066867828369,
+      "rewards/margins": 0.7579058408737183,
+      "rewards/rejected": -3.3419723510742188,
+      "step": 609
+    },
+    {
+      "epoch": 1.1071752951861944,
+      "grad_norm": 1.9866671562194824,
+      "learning_rate": 8.39801014483985e-06,
+      "logits/chosen": -0.015922199934720993,
+      "logits/rejected": 0.10281078517436981,
+      "logps/chosen": -99.91399383544922,
+      "logps/rejected": -113.79450988769531,
+      "loss": 2.2522,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.595137119293213,
+      "rewards/margins": 0.7955703139305115,
+      "rewards/rejected": -3.390707492828369,
+      "step": 610
+    },
+    {
+      "epoch": 1.108991825613079,
+      "grad_norm": 2.148136854171753,
+      "learning_rate": 8.393091161630612e-06,
+      "logits/chosen": 0.04068015143275261,
+      "logits/rejected": 0.09121442586183548,
+      "logps/chosen": -85.49508666992188,
+      "logps/rejected": -97.27752685546875,
+      "loss": 1.9467,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.673704147338867,
+      "rewards/margins": 0.8281559348106384,
+      "rewards/rejected": -3.5018603801727295,
+      "step": 611
+    },
+    {
+      "epoch": 1.1108083560399638,
+      "grad_norm": 2.0157570838928223,
+      "learning_rate": 8.388166278904759e-06,
+      "logits/chosen": 0.03451048582792282,
+      "logits/rejected": 0.10373912751674652,
+      "logps/chosen": -79.24334716796875,
+      "logps/rejected": -91.64152526855469,
+      "loss": 2.041,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4238996505737305,
+      "rewards/margins": 0.7874577045440674,
+      "rewards/rejected": -3.211357593536377,
+      "step": 612
+    },
+    {
+      "epoch": 1.1126248864668482,
+      "grad_norm": 1.7337363958358765,
+      "learning_rate": 8.383235506704986e-06,
+      "logits/chosen": 0.06933741271495819,
+      "logits/rejected": 0.10624522715806961,
+      "logps/chosen": -77.62388610839844,
+      "logps/rejected": -91.96408081054688,
+      "loss": 2.0593,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.304389476776123,
+      "rewards/margins": 0.6717057228088379,
+      "rewards/rejected": -2.97609543800354,
+      "step": 613
+    },
+    {
+      "epoch": 1.1144414168937329,
+      "grad_norm": 3.6357574462890625,
+      "learning_rate": 8.378298855085996e-06,
+      "logits/chosen": 0.08138547092676163,
+      "logits/rejected": 0.0961306244134903,
+      "logps/chosen": -87.44918823242188,
+      "logps/rejected": -94.0655746459961,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5593717098236084,
+      "rewards/margins": 0.6174351572990417,
+      "rewards/rejected": -3.176806688308716,
+      "step": 614
+    },
+    {
+      "epoch": 1.1162579473206176,
+      "grad_norm": 2.273991107940674,
+      "learning_rate": 8.373356334114484e-06,
+      "logits/chosen": 0.029560726135969162,
+      "logits/rejected": 0.11155828088521957,
+      "logps/chosen": -91.85681915283203,
+      "logps/rejected": -103.11241912841797,
+      "loss": 2.2401,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5434679985046387,
+      "rewards/margins": 0.7232552170753479,
+      "rewards/rejected": -3.266723394393921,
+      "step": 615
+    },
+    {
+      "epoch": 1.1180744777475022,
+      "grad_norm": 2.1391029357910156,
+      "learning_rate": 8.368407953869105e-06,
+      "logits/chosen": 0.05583259090781212,
+      "logits/rejected": 0.04287164285778999,
+      "logps/chosen": -92.72290802001953,
+      "logps/rejected": -99.79463195800781,
+      "loss": 1.8293,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.395390272140503,
+      "rewards/margins": 0.9053142666816711,
+      "rewards/rejected": -3.3007044792175293,
+      "step": 616
+    },
+    {
+      "epoch": 1.119891008174387,
+      "grad_norm": 1.6499502658843994,
+      "learning_rate": 8.363453724440471e-06,
+      "logits/chosen": 0.06658023595809937,
+      "logits/rejected": 0.04786865413188934,
+      "logps/chosen": -88.62269592285156,
+      "logps/rejected": -95.76766967773438,
+      "loss": 1.9608,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.522400379180908,
+      "rewards/margins": 0.8233134746551514,
+      "rewards/rejected": -3.3457136154174805,
+      "step": 617
+    },
+    {
+      "epoch": 1.1217075386012716,
+      "grad_norm": 2.018575668334961,
+      "learning_rate": 8.358493655931119e-06,
+      "logits/chosen": 0.06770970672369003,
+      "logits/rejected": 0.08037856221199036,
+      "logps/chosen": -82.23643493652344,
+      "logps/rejected": -97.71478271484375,
+      "loss": 1.8448,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.201833724975586,
+      "rewards/margins": 1.0529935359954834,
+      "rewards/rejected": -3.2548277378082275,
+      "step": 618
+    },
+    {
+      "epoch": 1.1235240690281563,
+      "grad_norm": 2.0033349990844727,
+      "learning_rate": 8.353527758455491e-06,
+      "logits/chosen": 0.11088447272777557,
+      "logits/rejected": 0.09635978937149048,
+      "logps/chosen": -81.90861511230469,
+      "logps/rejected": -92.48001098632812,
+      "loss": 2.2057,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.494230270385742,
+      "rewards/margins": 0.7774206399917603,
+      "rewards/rejected": -3.271650791168213,
+      "step": 619
+    },
+    {
+      "epoch": 1.125340599455041,
+      "grad_norm": 1.700210452079773,
+      "learning_rate": 8.348556042139918e-06,
+      "logits/chosen": 0.08965969830751419,
+      "logits/rejected": 0.06873422861099243,
+      "logps/chosen": -80.28114318847656,
+      "logps/rejected": -84.04791259765625,
+      "loss": 2.4366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.773008108139038,
+      "rewards/margins": 0.5039690732955933,
+      "rewards/rejected": -3.276977300643921,
+      "step": 620
+    },
+    {
+      "epoch": 1.1271571298819256,
+      "grad_norm": 2.623145580291748,
+      "learning_rate": 8.343578517122594e-06,
+      "logits/chosen": 0.08407651633024216,
+      "logits/rejected": 0.09678006917238235,
+      "logps/chosen": -81.90699005126953,
+      "logps/rejected": -94.30622863769531,
+      "loss": 2.126,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4066429138183594,
+      "rewards/margins": 0.8733029365539551,
+      "rewards/rejected": -3.2799456119537354,
+      "step": 621
+    },
+    {
+      "epoch": 1.12897366030881,
+      "grad_norm": 1.8616830110549927,
+      "learning_rate": 8.338595193553559e-06,
+      "logits/chosen": 0.002603452652692795,
+      "logits/rejected": 0.13452833890914917,
+      "logps/chosen": -73.35641479492188,
+      "logps/rejected": -84.70191192626953,
+      "loss": 2.084,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3735039234161377,
+      "rewards/margins": 0.7769300937652588,
+      "rewards/rejected": -3.1504342555999756,
+      "step": 622
+    },
+    {
+      "epoch": 1.1307901907356948,
+      "grad_norm": 1.8286879062652588,
+      "learning_rate": 8.333606081594678e-06,
+      "logits/chosen": 0.06104264035820961,
+      "logits/rejected": 0.10199623554944992,
+      "logps/chosen": -76.39376831054688,
+      "logps/rejected": -88.08850860595703,
+      "loss": 1.7885,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.130338191986084,
+      "rewards/margins": 0.848843514919281,
+      "rewards/rejected": -2.9791817665100098,
+      "step": 623
+    },
+    {
+      "epoch": 1.1326067211625794,
+      "grad_norm": 1.4803669452667236,
+      "learning_rate": 8.32861119141962e-06,
+      "logits/chosen": 0.07387635856866837,
+      "logits/rejected": 0.11718127131462097,
+      "logps/chosen": -82.92483520507812,
+      "logps/rejected": -91.46949005126953,
+      "loss": 1.7624,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.291823148727417,
+      "rewards/margins": 0.9700915217399597,
+      "rewards/rejected": -3.2619142532348633,
+      "step": 624
+    },
+    {
+      "epoch": 1.134423251589464,
+      "grad_norm": 1.6750752925872803,
+      "learning_rate": 8.323610533213835e-06,
+      "logits/chosen": 0.10784655809402466,
+      "logits/rejected": 0.17897658050060272,
+      "logps/chosen": -75.01900482177734,
+      "logps/rejected": -88.1203384399414,
+      "loss": 2.1956,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4190964698791504,
+      "rewards/margins": 0.688554584980011,
+      "rewards/rejected": -3.1076512336730957,
+      "step": 625
+    },
+    {
+      "epoch": 1.1362397820163488,
+      "grad_norm": 1.8641157150268555,
+      "learning_rate": 8.318604117174533e-06,
+      "logits/chosen": 0.05499793961644173,
+      "logits/rejected": 0.0563310906291008,
+      "logps/chosen": -83.42897033691406,
+      "logps/rejected": -94.2099380493164,
+      "loss": 1.675,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2912588119506836,
+      "rewards/margins": 0.9781790375709534,
+      "rewards/rejected": -3.2694382667541504,
+      "step": 626
+    },
+    {
+      "epoch": 1.1380563124432335,
+      "grad_norm": 1.7625656127929688,
+      "learning_rate": 8.313591953510674e-06,
+      "logits/chosen": 0.0831976979970932,
+      "logits/rejected": 0.12448858469724655,
+      "logps/chosen": -80.08613586425781,
+      "logps/rejected": -85.6810302734375,
+      "loss": 2.3784,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.453763008117676,
+      "rewards/margins": 0.46461760997772217,
+      "rewards/rejected": -2.9183804988861084,
+      "step": 627
+    },
+    {
+      "epoch": 1.1398728428701181,
+      "grad_norm": 2.10945463180542,
+      "learning_rate": 8.308574052442928e-06,
+      "logits/chosen": 0.09317971020936966,
+      "logits/rejected": 0.12335637211799622,
+      "logps/chosen": -82.36027526855469,
+      "logps/rejected": -94.93681335449219,
+      "loss": 2.1623,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5503153800964355,
+      "rewards/margins": 0.6077253222465515,
+      "rewards/rejected": -3.158040761947632,
+      "step": 628
+    },
+    {
+      "epoch": 1.1416893732970028,
+      "grad_norm": 1.8368538618087769,
+      "learning_rate": 8.30355042420367e-06,
+      "logits/chosen": 0.08310627192258835,
+      "logits/rejected": 0.11056395620107651,
+      "logps/chosen": -81.20982360839844,
+      "logps/rejected": -92.69551086425781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2482149600982666,
+      "rewards/margins": 0.9621560573577881,
+      "rewards/rejected": -3.2103710174560547,
+      "step": 629
+    },
+    {
+      "epoch": 1.1435059037238875,
+      "grad_norm": 1.7961188554763794,
+      "learning_rate": 8.29852107903695e-06,
+      "logits/chosen": 0.13976231217384338,
+      "logits/rejected": 0.11740799248218536,
+      "logps/chosen": -70.01700592041016,
+      "logps/rejected": -76.1086654663086,
+      "loss": 2.3401,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.239321231842041,
+      "rewards/margins": 0.5643521547317505,
+      "rewards/rejected": -2.80367374420166,
+      "step": 630
+    },
+    {
+      "epoch": 1.145322434150772,
+      "grad_norm": 2.041840076446533,
+      "learning_rate": 8.293486027198483e-06,
+      "logits/chosen": 0.06489598006010056,
+      "logits/rejected": 0.11150355637073517,
+      "logps/chosen": -95.31941223144531,
+      "logps/rejected": -96.98028564453125,
+      "loss": 2.2151,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5194008350372314,
+      "rewards/margins": 0.5945936441421509,
+      "rewards/rejected": -3.113994836807251,
+      "step": 631
+    },
+    {
+      "epoch": 1.1471389645776566,
+      "grad_norm": 2.7394556999206543,
+      "learning_rate": 8.288445278955615e-06,
+      "logits/chosen": 0.10089049488306046,
+      "logits/rejected": 0.08729755133390427,
+      "logps/chosen": -95.64276123046875,
+      "logps/rejected": -99.41658020019531,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.5987305641174316,
+      "rewards/margins": 0.4612637162208557,
+      "rewards/rejected": -3.0599937438964844,
+      "step": 632
+    },
+    {
+      "epoch": 1.1489554950045413,
+      "grad_norm": 2.024982213973999,
+      "learning_rate": 8.283398844587308e-06,
+      "logits/chosen": 0.023156292736530304,
+      "logits/rejected": 0.06789979338645935,
+      "logps/chosen": -89.90803527832031,
+      "logps/rejected": -96.06836700439453,
+      "loss": 1.9289,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.1415646076202393,
+      "rewards/margins": 0.7636609673500061,
+      "rewards/rejected": -2.9052255153656006,
+      "step": 633
+    },
+    {
+      "epoch": 1.150772025431426,
+      "grad_norm": 1.8750615119934082,
+      "learning_rate": 8.278346734384122e-06,
+      "logits/chosen": 0.02243964746594429,
+      "logits/rejected": 0.12660253047943115,
+      "logps/chosen": -71.34333038330078,
+      "logps/rejected": -86.67922973632812,
+      "loss": 1.8554,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.2015068531036377,
+      "rewards/margins": 0.9946616291999817,
+      "rewards/rejected": -3.1961684226989746,
+      "step": 634
+    },
+    {
+      "epoch": 1.1525885558583107,
+      "grad_norm": 1.9947502613067627,
+      "learning_rate": 8.27328895864819e-06,
+      "logits/chosen": 0.05387189984321594,
+      "logits/rejected": 0.08354412019252777,
+      "logps/chosen": -85.38788604736328,
+      "logps/rejected": -97.08065795898438,
+      "loss": 2.3917,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.595083236694336,
+      "rewards/margins": 0.5549375414848328,
+      "rewards/rejected": -3.1500213146209717,
+      "step": 635
+    },
+    {
+      "epoch": 1.1544050862851953,
+      "grad_norm": 2.145918607711792,
+      "learning_rate": 8.268225527693193e-06,
+      "logits/chosen": 0.10176774114370346,
+      "logits/rejected": 0.14395561814308167,
+      "logps/chosen": -76.02487182617188,
+      "logps/rejected": -86.05180358886719,
+      "loss": 1.8909,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.0001683235168457,
+      "rewards/margins": 0.8724240064620972,
+      "rewards/rejected": -2.8725922107696533,
+      "step": 636
+    },
+    {
+      "epoch": 1.15622161671208,
+      "grad_norm": 1.6830824613571167,
+      "learning_rate": 8.263156451844353e-06,
+      "logits/chosen": 0.10138621181249619,
+      "logits/rejected": 0.11966821551322937,
+      "logps/chosen": -76.56681060791016,
+      "logps/rejected": -88.1928939819336,
+      "loss": 1.9508,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.247589588165283,
+      "rewards/margins": 0.8286643028259277,
+      "rewards/rejected": -3.076253890991211,
+      "step": 637
+    },
+    {
+      "epoch": 1.1580381471389645,
+      "grad_norm": 2.0992536544799805,
+      "learning_rate": 8.258081741438396e-06,
+      "logits/chosen": 0.07541397213935852,
+      "logits/rejected": 0.09907414764165878,
+      "logps/chosen": -90.69495391845703,
+      "logps/rejected": -101.71548461914062,
+      "loss": 2.1933,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4336788654327393,
+      "rewards/margins": 0.7834519147872925,
+      "rewards/rejected": -3.217130422592163,
+      "step": 638
+    },
+    {
+      "epoch": 1.1598546775658491,
+      "grad_norm": 1.8066256046295166,
+      "learning_rate": 8.25300140682354e-06,
+      "logits/chosen": 0.14558832347393036,
+      "logits/rejected": 0.14358943700790405,
+      "logps/chosen": -82.73606872558594,
+      "logps/rejected": -86.27526092529297,
+      "loss": 2.0025,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.393693208694458,
+      "rewards/margins": 0.7131556272506714,
+      "rewards/rejected": -3.106848955154419,
+      "step": 639
+    },
+    {
+      "epoch": 1.1616712079927338,
+      "grad_norm": 2.3521695137023926,
+      "learning_rate": 8.247915458359473e-06,
+      "logits/chosen": 0.05236164107918739,
+      "logits/rejected": 0.1095740869641304,
+      "logps/chosen": -83.70420837402344,
+      "logps/rejected": -100.09117889404297,
+      "loss": 2.3348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5044684410095215,
+      "rewards/margins": 0.7525762319564819,
+      "rewards/rejected": -3.257044553756714,
+      "step": 640
+    },
+    {
+      "epoch": 1.1634877384196185,
+      "grad_norm": 2.343412160873413,
+      "learning_rate": 8.242823906417329e-06,
+      "logits/chosen": -0.00503757456317544,
+      "logits/rejected": 0.06355556100606918,
+      "logps/chosen": -80.41883087158203,
+      "logps/rejected": -95.32416534423828,
+      "loss": 2.2004,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3209891319274902,
+      "rewards/margins": 0.745349109172821,
+      "rewards/rejected": -3.066338539123535,
+      "step": 641
+    },
+    {
+      "epoch": 1.1653042688465032,
+      "grad_norm": 1.9915282726287842,
+      "learning_rate": 8.237726761379668e-06,
+      "logits/chosen": 0.02087419107556343,
+      "logits/rejected": 0.08556030690670013,
+      "logps/chosen": -84.41456604003906,
+      "logps/rejected": -97.61527252197266,
+      "loss": 2.0052,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.385272741317749,
+      "rewards/margins": 0.7285323739051819,
+      "rewards/rejected": -3.113805055618286,
+      "step": 642
+    },
+    {
+      "epoch": 1.1671207992733879,
+      "grad_norm": 2.4018874168395996,
+      "learning_rate": 8.232624033640458e-06,
+      "logits/chosen": 0.0840807855129242,
+      "logits/rejected": 0.03736239671707153,
+      "logps/chosen": -83.6908950805664,
+      "logps/rejected": -88.65196228027344,
+      "loss": 2.5072,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.494255542755127,
+      "rewards/margins": 0.5595706701278687,
+      "rewards/rejected": -3.053826332092285,
+      "step": 643
+    },
+    {
+      "epoch": 1.1689373297002725,
+      "grad_norm": 2.3055484294891357,
+      "learning_rate": 8.227515733605049e-06,
+      "logits/chosen": 0.05399080738425255,
+      "logits/rejected": 0.13454601168632507,
+      "logps/chosen": -77.62064361572266,
+      "logps/rejected": -86.54924011230469,
+      "loss": 2.0759,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3616065979003906,
+      "rewards/margins": 0.7132005095481873,
+      "rewards/rejected": -3.0748066902160645,
+      "step": 644
+    },
+    {
+      "epoch": 1.1707538601271572,
+      "grad_norm": 1.989559292793274,
+      "learning_rate": 8.222401871690153e-06,
+      "logits/chosen": 0.03339620307087898,
+      "logits/rejected": 0.06584692001342773,
+      "logps/chosen": -74.14878845214844,
+      "logps/rejected": -89.0107421875,
+      "loss": 1.9992,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3560426235198975,
+      "rewards/margins": 0.7924355864524841,
+      "rewards/rejected": -3.1484780311584473,
+      "step": 645
+    },
+    {
+      "epoch": 1.1725703905540419,
+      "grad_norm": 1.7572104930877686,
+      "learning_rate": 8.217282458323825e-06,
+      "logits/chosen": 0.09702566266059875,
+      "logits/rejected": 0.06184859201312065,
+      "logps/chosen": -82.05592346191406,
+      "logps/rejected": -87.59920501708984,
+      "loss": 2.0353,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.121622085571289,
+      "rewards/margins": 0.707017183303833,
+      "rewards/rejected": -2.828639268875122,
+      "step": 646
+    },
+    {
+      "epoch": 1.1743869209809263,
+      "grad_norm": 1.8265388011932373,
+      "learning_rate": 8.21215750394544e-06,
+      "logits/chosen": 0.12195339798927307,
+      "logits/rejected": 0.09588825702667236,
+      "logps/chosen": -76.14807891845703,
+      "logps/rejected": -85.36613464355469,
+      "loss": 2.0684,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2251272201538086,
+      "rewards/margins": 0.8967663049697876,
+      "rewards/rejected": -3.1218934059143066,
+      "step": 647
+    },
+    {
+      "epoch": 1.176203451407811,
+      "grad_norm": 1.9061174392700195,
+      "learning_rate": 8.207027019005675e-06,
+      "logits/chosen": 0.09651574492454529,
+      "logits/rejected": 0.10208628326654434,
+      "logps/chosen": -86.28733825683594,
+      "logps/rejected": -97.5046157836914,
+      "loss": 2.1319,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3506147861480713,
+      "rewards/margins": 0.7415947318077087,
+      "rewards/rejected": -3.0922091007232666,
+      "step": 648
+    },
+    {
+      "epoch": 1.1780199818346957,
+      "grad_norm": 1.999624252319336,
+      "learning_rate": 8.20189101396648e-06,
+      "logits/chosen": 0.08931821584701538,
+      "logits/rejected": 0.09364073723554611,
+      "logps/chosen": -84.32461547851562,
+      "logps/rejected": -96.70581817626953,
+      "loss": 1.7919,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4468250274658203,
+      "rewards/margins": 0.8684096932411194,
+      "rewards/rejected": -3.315234661102295,
+      "step": 649
+    },
+    {
+      "epoch": 1.1798365122615804,
+      "grad_norm": 1.5666590929031372,
+      "learning_rate": 8.196749499301062e-06,
+      "logits/chosen": 0.09945614635944366,
+      "logits/rejected": 0.12043865770101547,
+      "logps/chosen": -81.99935150146484,
+      "logps/rejected": -96.36174011230469,
+      "loss": 1.8932,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4319562911987305,
+      "rewards/margins": 0.8513790369033813,
+      "rewards/rejected": -3.2833354473114014,
+      "step": 650
+    },
+    {
+      "epoch": 1.181653042688465,
+      "grad_norm": 1.9403899908065796,
+      "learning_rate": 8.191602485493868e-06,
+      "logits/chosen": 0.12361271679401398,
+      "logits/rejected": 0.14710690081119537,
+      "logps/chosen": -82.11424255371094,
+      "logps/rejected": -89.37992858886719,
+      "loss": 2.0615,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.161297082901001,
+      "rewards/margins": 0.7220393419265747,
+      "rewards/rejected": -2.8833365440368652,
+      "step": 651
+    },
+    {
+      "epoch": 1.1834695731153497,
+      "grad_norm": 1.8558425903320312,
+      "learning_rate": 8.186449983040552e-06,
+      "logits/chosen": 0.07215605676174164,
+      "logits/rejected": 0.06390087306499481,
+      "logps/chosen": -81.64322662353516,
+      "logps/rejected": -85.39015197753906,
+      "loss": 2.0064,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.320831775665283,
+      "rewards/margins": 0.7350802421569824,
+      "rewards/rejected": -3.0559120178222656,
+      "step": 652
+    },
+    {
+      "epoch": 1.1852861035422344,
+      "grad_norm": 1.7956428527832031,
+      "learning_rate": 8.181292002447966e-06,
+      "logits/chosen": 0.11276388168334961,
+      "logits/rejected": 0.06517648696899414,
+      "logps/chosen": -76.14613342285156,
+      "logps/rejected": -82.24524688720703,
+      "loss": 2.0473,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2219183444976807,
+      "rewards/margins": 0.6231520175933838,
+      "rewards/rejected": -2.8450703620910645,
+      "step": 653
+    },
+    {
+      "epoch": 1.1871026339691189,
+      "grad_norm": 2.1648740768432617,
+      "learning_rate": 8.17612855423413e-06,
+      "logits/chosen": 0.0334198959171772,
+      "logits/rejected": 0.013921715319156647,
+      "logps/chosen": -83.02323150634766,
+      "logps/rejected": -92.95353698730469,
+      "loss": 1.9884,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.0769619941711426,
+      "rewards/margins": 0.851192057132721,
+      "rewards/rejected": -2.9281537532806396,
+      "step": 654
+    },
+    {
+      "epoch": 1.1889191643960035,
+      "grad_norm": 1.921157717704773,
+      "learning_rate": 8.170959648928214e-06,
+      "logits/chosen": 0.073361836373806,
+      "logits/rejected": 0.13112977147102356,
+      "logps/chosen": -83.41218566894531,
+      "logps/rejected": -92.03260803222656,
+      "loss": 2.0203,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7512502670288086,
+      "rewards/margins": 0.8279457092285156,
+      "rewards/rejected": -3.579195737838745,
+      "step": 655
+    },
+    {
+      "epoch": 1.1907356948228882,
+      "grad_norm": 2.4273247718811035,
+      "learning_rate": 8.165785297070516e-06,
+      "logits/chosen": 0.027622584253549576,
+      "logits/rejected": 0.08521658927202225,
+      "logps/chosen": -84.34210205078125,
+      "logps/rejected": -101.76324462890625,
+      "loss": 1.9023,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6731176376342773,
+      "rewards/margins": 1.004132866859436,
+      "rewards/rejected": -3.677250385284424,
+      "step": 656
+    },
+    {
+      "epoch": 1.192552225249773,
+      "grad_norm": 1.8262163400650024,
+      "learning_rate": 8.16060550921244e-06,
+      "logits/chosen": 0.03431132063269615,
+      "logits/rejected": 0.06454990059137344,
+      "logps/chosen": -76.59049987792969,
+      "logps/rejected": -97.41972351074219,
+      "loss": 2.0105,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3002851009368896,
+      "rewards/margins": 0.7388467788696289,
+      "rewards/rejected": -3.0391316413879395,
+      "step": 657
+    },
+    {
+      "epoch": 1.1943687556766576,
+      "grad_norm": 1.9906315803527832,
+      "learning_rate": 8.155420295916474e-06,
+      "logits/chosen": 0.08427983522415161,
+      "logits/rejected": 0.025464089587330818,
+      "logps/chosen": -84.34020233154297,
+      "logps/rejected": -92.31369018554688,
+      "loss": 1.9895,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3555400371551514,
+      "rewards/margins": 0.8297086954116821,
+      "rewards/rejected": -3.185249090194702,
+      "step": 658
+    },
+    {
+      "epoch": 1.1961852861035422,
+      "grad_norm": 1.9980812072753906,
+      "learning_rate": 8.150229667756172e-06,
+      "logits/chosen": 0.07556813955307007,
+      "logits/rejected": 0.07593435049057007,
+      "logps/chosen": -85.46858215332031,
+      "logps/rejected": -90.94758605957031,
+      "loss": 2.2687,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.679569959640503,
+      "rewards/margins": 0.5908665657043457,
+      "rewards/rejected": -3.2704365253448486,
+      "step": 659
+    },
+    {
+      "epoch": 1.198001816530427,
+      "grad_norm": 2.0758330821990967,
+      "learning_rate": 8.14503363531613e-06,
+      "logits/chosen": 0.11312856525182724,
+      "logits/rejected": 0.08777043223381042,
+      "logps/chosen": -80.32341003417969,
+      "logps/rejected": -84.89898681640625,
+      "loss": 2.1371,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5170207023620605,
+      "rewards/margins": 0.632743239402771,
+      "rewards/rejected": -3.149764060974121,
+      "step": 660
+    },
+    {
+      "epoch": 1.1998183469573116,
+      "grad_norm": 2.765526533126831,
+      "learning_rate": 8.139832209191961e-06,
+      "logits/chosen": 0.17475244402885437,
+      "logits/rejected": 0.13103297352790833,
+      "logps/chosen": -92.03308868408203,
+      "logps/rejected": -94.15422821044922,
+      "loss": 2.4151,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.3920867443084717,
+      "rewards/margins": 0.518372118473053,
+      "rewards/rejected": -2.910458564758301,
+      "step": 661
+    },
+    {
+      "epoch": 1.2016348773841963,
+      "grad_norm": 2.0190646648406982,
+      "learning_rate": 8.13462539999028e-06,
+      "logits/chosen": 0.040579523891210556,
+      "logits/rejected": 0.060216568410396576,
+      "logps/chosen": -96.32977294921875,
+      "logps/rejected": -105.37261962890625,
+      "loss": 2.1216,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5936977863311768,
+      "rewards/margins": 1.040100336074829,
+      "rewards/rejected": -3.6337978839874268,
+      "step": 662
+    },
+    {
+      "epoch": 1.2034514078110807,
+      "grad_norm": 2.53027606010437,
+      "learning_rate": 8.129413218328674e-06,
+      "logits/chosen": 0.07498883455991745,
+      "logits/rejected": 0.042916588485240936,
+      "logps/chosen": -83.17558288574219,
+      "logps/rejected": -89.55461883544922,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6350390911102295,
+      "rewards/margins": 0.7082449793815613,
+      "rewards/rejected": -3.3432838916778564,
+      "step": 663
+    },
+    {
+      "epoch": 1.2052679382379654,
+      "grad_norm": 1.6480835676193237,
+      "learning_rate": 8.124195674835695e-06,
+      "logits/chosen": 0.1062261089682579,
+      "logits/rejected": 0.09598012268543243,
+      "logps/chosen": -84.5193099975586,
+      "logps/rejected": -95.05423736572266,
+      "loss": 1.7544,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.1563069820404053,
+      "rewards/margins": 0.8859087228775024,
+      "rewards/rejected": -3.042215585708618,
+      "step": 664
+    },
+    {
+      "epoch": 1.20708446866485,
+      "grad_norm": 2.126664161682129,
+      "learning_rate": 8.118972780150817e-06,
+      "logits/chosen": -0.01870904304087162,
+      "logits/rejected": 0.1347496211528778,
+      "logps/chosen": -79.35530853271484,
+      "logps/rejected": -105.41546630859375,
+      "loss": 2.0524,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6006765365600586,
+      "rewards/margins": 1.0105364322662354,
+      "rewards/rejected": -3.611213207244873,
+      "step": 665
+    },
+    {
+      "epoch": 1.2089009990917348,
+      "grad_norm": 2.117645263671875,
+      "learning_rate": 8.113744544924434e-06,
+      "logits/chosen": 0.01287349034100771,
+      "logits/rejected": 0.019422955811023712,
+      "logps/chosen": -105.86250305175781,
+      "logps/rejected": -117.42082977294922,
+      "loss": 1.8973,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.636453151702881,
+      "rewards/margins": 0.8021495342254639,
+      "rewards/rejected": -3.4386026859283447,
+      "step": 666
+    },
+    {
+      "epoch": 1.2107175295186194,
+      "grad_norm": 2.956285238265991,
+      "learning_rate": 8.108510979817828e-06,
+      "logits/chosen": 0.047061942517757416,
+      "logits/rejected": 0.11715231090784073,
+      "logps/chosen": -89.56261444091797,
+      "logps/rejected": -103.57279968261719,
+      "loss": 2.0627,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5317227840423584,
+      "rewards/margins": 0.8301056623458862,
+      "rewards/rejected": -3.361828327178955,
+      "step": 667
+    },
+    {
+      "epoch": 1.2125340599455041,
+      "grad_norm": 1.897623062133789,
+      "learning_rate": 8.10327209550315e-06,
+      "logits/chosen": 0.017857536673545837,
+      "logits/rejected": 0.12088865786790848,
+      "logps/chosen": -95.75634002685547,
+      "logps/rejected": -113.0296630859375,
+      "loss": 1.9091,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6507534980773926,
+      "rewards/margins": 0.963871955871582,
+      "rewards/rejected": -3.6146254539489746,
+      "step": 668
+    },
+    {
+      "epoch": 1.2143505903723888,
+      "grad_norm": 2.176003932952881,
+      "learning_rate": 8.098027902663396e-06,
+      "logits/chosen": -0.0014993082731962204,
+      "logits/rejected": 0.11425516754388809,
+      "logps/chosen": -79.22991943359375,
+      "logps/rejected": -96.39210510253906,
+      "loss": 1.8977,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.582876443862915,
+      "rewards/margins": 0.9855522513389587,
+      "rewards/rejected": -3.5684289932250977,
+      "step": 669
+    },
+    {
+      "epoch": 1.2161671207992735,
+      "grad_norm": 2.140666961669922,
+      "learning_rate": 8.092778411992388e-06,
+      "logits/chosen": 0.04144468903541565,
+      "logits/rejected": 0.0556509830057621,
+      "logps/chosen": -80.35415649414062,
+      "logps/rejected": -97.16475677490234,
+      "loss": 1.9182,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.66176176071167,
+      "rewards/margins": 0.9624839425086975,
+      "rewards/rejected": -3.6242458820343018,
+      "step": 670
+    },
+    {
+      "epoch": 1.2179836512261581,
+      "grad_norm": 1.937705636024475,
+      "learning_rate": 8.087523634194755e-06,
+      "logits/chosen": 0.03470912575721741,
+      "logits/rejected": 0.12844812870025635,
+      "logps/chosen": -81.73896026611328,
+      "logps/rejected": -98.39558410644531,
+      "loss": 2.032,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.406296730041504,
+      "rewards/margins": 0.9326379895210266,
+      "rewards/rejected": -3.3389346599578857,
+      "step": 671
+    },
+    {
+      "epoch": 1.2198001816530426,
+      "grad_norm": 1.8903594017028809,
+      "learning_rate": 8.082263579985898e-06,
+      "logits/chosen": 0.03509046137332916,
+      "logits/rejected": 0.0663415864109993,
+      "logps/chosen": -81.79936981201172,
+      "logps/rejected": -96.15756225585938,
+      "loss": 1.7671,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2738354206085205,
+      "rewards/margins": 1.1403576135635376,
+      "rewards/rejected": -3.4141931533813477,
+      "step": 672
+    },
+    {
+      "epoch": 1.2216167120799273,
+      "grad_norm": 2.206265687942505,
+      "learning_rate": 8.076998260091989e-06,
+      "logits/chosen": 0.016251683235168457,
+      "logits/rejected": 0.07814528048038483,
+      "logps/chosen": -80.23258209228516,
+      "logps/rejected": -85.3895034790039,
+      "loss": 2.2589,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6450579166412354,
+      "rewards/margins": 0.5741185545921326,
+      "rewards/rejected": -3.2191762924194336,
+      "step": 673
+    },
+    {
+      "epoch": 1.223433242506812,
+      "grad_norm": 2.3217952251434326,
+      "learning_rate": 8.071727685249929e-06,
+      "logits/chosen": 0.15181973576545715,
+      "logits/rejected": 0.11617043614387512,
+      "logps/chosen": -85.15460205078125,
+      "logps/rejected": -83.82669830322266,
+      "loss": 2.2534,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.530156135559082,
+      "rewards/margins": 0.6635551452636719,
+      "rewards/rejected": -3.193711519241333,
+      "step": 674
+    },
+    {
+      "epoch": 1.2252497729336966,
+      "grad_norm": 2.011244297027588,
+      "learning_rate": 8.06645186620734e-06,
+      "logits/chosen": 0.02763158269226551,
+      "logits/rejected": 0.04441550746560097,
+      "logps/chosen": -81.14067840576172,
+      "logps/rejected": -95.907958984375,
+      "loss": 1.8979,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.693267583847046,
+      "rewards/margins": 0.9938480854034424,
+      "rewards/rejected": -3.6871156692504883,
+      "step": 675
+    },
+    {
+      "epoch": 1.2270663033605813,
+      "grad_norm": 2.3554768562316895,
+      "learning_rate": 8.061170813722533e-06,
+      "logits/chosen": 0.06379646062850952,
+      "logits/rejected": 0.052463918924331665,
+      "logps/chosen": -92.42412567138672,
+      "logps/rejected": -92.37240600585938,
+      "loss": 2.4093,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.7538669109344482,
+      "rewards/margins": 0.5145044326782227,
+      "rewards/rejected": -3.268371105194092,
+      "step": 676
+    },
+    {
+      "epoch": 1.228882833787466,
+      "grad_norm": 2.2725136280059814,
+      "learning_rate": 8.055884538564493e-06,
+      "logits/chosen": 0.028828933835029602,
+      "logits/rejected": 0.08082776516675949,
+      "logps/chosen": -76.20146179199219,
+      "logps/rejected": -94.82791900634766,
+      "loss": 1.7849,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.348304510116577,
+      "rewards/margins": 1.1067698001861572,
+      "rewards/rejected": -3.4550740718841553,
+      "step": 677
+    },
+    {
+      "epoch": 1.2306993642143507,
+      "grad_norm": 2.3565657138824463,
+      "learning_rate": 8.050593051512859e-06,
+      "logits/chosen": 0.09367989003658295,
+      "logits/rejected": 0.06132106855511665,
+      "logps/chosen": -84.85393524169922,
+      "logps/rejected": -91.87137603759766,
+      "loss": 2.162,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5945935249328613,
+      "rewards/margins": 0.7115421295166016,
+      "rewards/rejected": -3.306135416030884,
+      "step": 678
+    },
+    {
+      "epoch": 1.2325158946412351,
+      "grad_norm": 2.2214298248291016,
+      "learning_rate": 8.045296363357891e-06,
+      "logits/chosen": 0.015750454738736153,
+      "logits/rejected": -0.001254781149327755,
+      "logps/chosen": -94.99285125732422,
+      "logps/rejected": -107.42745971679688,
+      "loss": 2.0311,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.769968032836914,
+      "rewards/margins": 0.9211719632148743,
+      "rewards/rejected": -3.6911399364471436,
+      "step": 679
+    },
+    {
+      "epoch": 1.2343324250681198,
+      "grad_norm": 1.9377881288528442,
+      "learning_rate": 8.039994484900463e-06,
+      "logits/chosen": -0.011005287989974022,
+      "logits/rejected": 0.05877537280321121,
+      "logps/chosen": -84.62346649169922,
+      "logps/rejected": -99.59872436523438,
+      "loss": 1.8343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.554427146911621,
+      "rewards/margins": 1.086029291152954,
+      "rewards/rejected": -3.640456199645996,
+      "step": 680
+    },
+    {
+      "epoch": 1.2361489554950045,
+      "grad_norm": 2.665781259536743,
+      "learning_rate": 8.034687426952023e-06,
+      "logits/chosen": 0.030874190852046013,
+      "logits/rejected": 0.09339022636413574,
+      "logps/chosen": -82.25083923339844,
+      "logps/rejected": -99.68885803222656,
+      "loss": 2.0966,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.847683906555176,
+      "rewards/margins": 0.8753347396850586,
+      "rewards/rejected": -3.7230184078216553,
+      "step": 681
+    },
+    {
+      "epoch": 1.2379654859218892,
+      "grad_norm": 1.8999613523483276,
+      "learning_rate": 8.029375200334588e-06,
+      "logits/chosen": 0.07361680269241333,
+      "logits/rejected": 0.08493717014789581,
+      "logps/chosen": -88.70954895019531,
+      "logps/rejected": -102.10923767089844,
+      "loss": 1.7236,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.518230676651001,
+      "rewards/margins": 1.0492362976074219,
+      "rewards/rejected": -3.567467212677002,
+      "step": 682
+    },
+    {
+      "epoch": 1.2397820163487738,
+      "grad_norm": 2.179049491882324,
+      "learning_rate": 8.024057815880716e-06,
+      "logits/chosen": 0.12260061502456665,
+      "logits/rejected": 0.09896722435951233,
+      "logps/chosen": -82.61366271972656,
+      "logps/rejected": -89.37825775146484,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3436989784240723,
+      "rewards/margins": 0.7696498036384583,
+      "rewards/rejected": -3.1133487224578857,
+      "step": 683
+    },
+    {
+      "epoch": 1.2415985467756585,
+      "grad_norm": 1.8527759313583374,
+      "learning_rate": 8.018735284433475e-06,
+      "logits/chosen": 0.007166730239987373,
+      "logits/rejected": 0.04468049108982086,
+      "logps/chosen": -88.59947204589844,
+      "logps/rejected": -97.66645050048828,
+      "loss": 1.9353,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.669034957885742,
+      "rewards/margins": 0.8667396306991577,
+      "rewards/rejected": -3.5357747077941895,
+      "step": 684
+    },
+    {
+      "epoch": 1.2434150772025432,
+      "grad_norm": 2.013792037963867,
+      "learning_rate": 8.013407616846436e-06,
+      "logits/chosen": 0.05774620547890663,
+      "logits/rejected": 0.06977065652608871,
+      "logps/chosen": -83.88308715820312,
+      "logps/rejected": -93.36985778808594,
+      "loss": 1.9716,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5362911224365234,
+      "rewards/margins": 0.9699288010597229,
+      "rewards/rejected": -3.5062198638916016,
+      "step": 685
+    },
+    {
+      "epoch": 1.2452316076294279,
+      "grad_norm": 1.9147852659225464,
+      "learning_rate": 8.008074823983642e-06,
+      "logits/chosen": 0.10071337223052979,
+      "logits/rejected": 0.09696929156780243,
+      "logps/chosen": -87.06709289550781,
+      "logps/rejected": -98.84783172607422,
+      "loss": 1.7989,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7123823165893555,
+      "rewards/margins": 0.9586774110794067,
+      "rewards/rejected": -3.6710598468780518,
+      "step": 686
+    },
+    {
+      "epoch": 1.2470481380563125,
+      "grad_norm": 3.8477859497070312,
+      "learning_rate": 8.002736916719582e-06,
+      "logits/chosen": 0.09164638817310333,
+      "logits/rejected": 0.14311164617538452,
+      "logps/chosen": -94.75381469726562,
+      "logps/rejected": -101.14093017578125,
+      "loss": 2.3677,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7352569103240967,
+      "rewards/margins": 0.5952449440956116,
+      "rewards/rejected": -3.3305020332336426,
+      "step": 687
+    },
+    {
+      "epoch": 1.248864668483197,
+      "grad_norm": 2.1601197719573975,
+      "learning_rate": 7.997393905939183e-06,
+      "logits/chosen": 0.14653781056404114,
+      "logits/rejected": 0.217861607670784,
+      "logps/chosen": -78.05258178710938,
+      "logps/rejected": -87.37406921386719,
+      "loss": 2.2395,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4850776195526123,
+      "rewards/margins": 0.5597268342971802,
+      "rewards/rejected": -3.044804573059082,
+      "step": 688
+    },
+    {
+      "epoch": 1.2506811989100817,
+      "grad_norm": 2.1657280921936035,
+      "learning_rate": 7.992045802537772e-06,
+      "logits/chosen": 0.008183173835277557,
+      "logits/rejected": 0.07762319594621658,
+      "logps/chosen": -76.81200408935547,
+      "logps/rejected": -89.77823638916016,
+      "loss": 2.0968,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.395460844039917,
+      "rewards/margins": 0.7663928270339966,
+      "rewards/rejected": -3.161853790283203,
+      "step": 689
+    },
+    {
+      "epoch": 1.2524977293369663,
+      "grad_norm": 2.0229945182800293,
+      "learning_rate": 7.986692617421063e-06,
+      "logits/chosen": 0.049828238785266876,
+      "logits/rejected": 0.12920916080474854,
+      "logps/chosen": -89.49681854248047,
+      "logps/rejected": -94.05538177490234,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5402889251708984,
+      "rewards/margins": 0.6041077375411987,
+      "rewards/rejected": -3.1443967819213867,
+      "step": 690
+    },
+    {
+      "epoch": 1.254314259763851,
+      "grad_norm": 2.1632978916168213,
+      "learning_rate": 7.981334361505131e-06,
+      "logits/chosen": 0.01283620297908783,
+      "logits/rejected": 0.06363464891910553,
+      "logps/chosen": -90.30138397216797,
+      "logps/rejected": -109.25923156738281,
+      "loss": 1.9464,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5346744060516357,
+      "rewards/margins": 1.0385441780090332,
+      "rewards/rejected": -3.573218584060669,
+      "step": 691
+    },
+    {
+      "epoch": 1.2561307901907357,
+      "grad_norm": 2.1562981605529785,
+      "learning_rate": 7.975971045716398e-06,
+      "logits/chosen": 0.12652680277824402,
+      "logits/rejected": 0.15553821623325348,
+      "logps/chosen": -74.45803833007812,
+      "logps/rejected": -90.87694549560547,
+      "loss": 1.8599,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.415904998779297,
+      "rewards/margins": 0.9220768809318542,
+      "rewards/rejected": -3.337982177734375,
+      "step": 692
+    },
+    {
+      "epoch": 1.2579473206176204,
+      "grad_norm": 2.007781982421875,
+      "learning_rate": 7.970602680991594e-06,
+      "logits/chosen": 0.05228351801633835,
+      "logits/rejected": 0.12716761231422424,
+      "logps/chosen": -96.6414794921875,
+      "logps/rejected": -115.6935043334961,
+      "loss": 2.0777,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5941665172576904,
+      "rewards/margins": 1.015256643295288,
+      "rewards/rejected": -3.6094233989715576,
+      "step": 693
+    },
+    {
+      "epoch": 1.259763851044505,
+      "grad_norm": 2.2393176555633545,
+      "learning_rate": 7.965229278277749e-06,
+      "logits/chosen": 0.11106812953948975,
+      "logits/rejected": 0.10329913347959518,
+      "logps/chosen": -91.51345825195312,
+      "logps/rejected": -95.78245544433594,
+      "loss": 2.1322,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6126160621643066,
+      "rewards/margins": 0.8137804865837097,
+      "rewards/rejected": -3.426396608352661,
+      "step": 694
+    },
+    {
+      "epoch": 1.2615803814713895,
+      "grad_norm": 1.9000614881515503,
+      "learning_rate": 7.959850848532172e-06,
+      "logits/chosen": 0.053980953991413116,
+      "logits/rejected": 0.1285300850868225,
+      "logps/chosen": -80.65451049804688,
+      "logps/rejected": -96.31999206542969,
+      "loss": 1.8157,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.449226140975952,
+      "rewards/margins": 0.9224156141281128,
+      "rewards/rejected": -3.3716418743133545,
+      "step": 695
+    },
+    {
+      "epoch": 1.2633969118982744,
+      "grad_norm": 2.1616997718811035,
+      "learning_rate": 7.95446740272241e-06,
+      "logits/chosen": 0.0943533331155777,
+      "logits/rejected": 0.09738673269748688,
+      "logps/chosen": -86.31460571289062,
+      "logps/rejected": -95.93961334228516,
+      "loss": 2.3068,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4675066471099854,
+      "rewards/margins": 0.5155179500579834,
+      "rewards/rejected": -2.9830245971679688,
+      "step": 696
+    },
+    {
+      "epoch": 1.2652134423251589,
+      "grad_norm": 2.040947675704956,
+      "learning_rate": 7.949078951826254e-06,
+      "logits/chosen": 0.05092187598347664,
+      "logits/rejected": 0.10424528270959854,
+      "logps/chosen": -78.41385650634766,
+      "logps/rejected": -96.27998352050781,
+      "loss": 1.9729,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.392333507537842,
+      "rewards/margins": 0.8541865348815918,
+      "rewards/rejected": -3.246520519256592,
+      "step": 697
+    },
+    {
+      "epoch": 1.2670299727520435,
+      "grad_norm": 2.123605251312256,
+      "learning_rate": 7.943685506831689e-06,
+      "logits/chosen": 0.1472083032131195,
+      "logits/rejected": 0.07182697206735611,
+      "logps/chosen": -87.48722839355469,
+      "logps/rejected": -89.31311798095703,
+      "loss": 2.1536,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.34977388381958,
+      "rewards/margins": 0.6212242245674133,
+      "rewards/rejected": -2.9709980487823486,
+      "step": 698
+    },
+    {
+      "epoch": 1.2688465031789282,
+      "grad_norm": 2.11053729057312,
+      "learning_rate": 7.938287078736889e-06,
+      "logits/chosen": 0.08341722190380096,
+      "logits/rejected": 0.12655527889728546,
+      "logps/chosen": -79.16002655029297,
+      "logps/rejected": -91.57137298583984,
+      "loss": 2.0396,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4659664630889893,
+      "rewards/margins": 0.7776908874511719,
+      "rewards/rejected": -3.2436575889587402,
+      "step": 699
+    },
+    {
+      "epoch": 1.270663033605813,
+      "grad_norm": 1.9555423259735107,
+      "learning_rate": 7.932883678550191e-06,
+      "logits/chosen": -0.06148137152194977,
+      "logits/rejected": 0.055927492678165436,
+      "logps/chosen": -81.51953125,
+      "logps/rejected": -95.884521484375,
+      "loss": 1.9223,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5550904273986816,
+      "rewards/margins": 0.8093082904815674,
+      "rewards/rejected": -3.364398956298828,
+      "step": 700
+    },
+    {
+      "epoch": 1.2724795640326976,
+      "grad_norm": 2.031662702560425,
+      "learning_rate": 7.927475317290065e-06,
+      "logits/chosen": 0.04235261678695679,
+      "logits/rejected": 0.1255102902650833,
+      "logps/chosen": -74.28742218017578,
+      "logps/rejected": -86.79241180419922,
+      "loss": 1.9669,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.138885498046875,
+      "rewards/margins": 0.8455573916435242,
+      "rewards/rejected": -2.984442949295044,
+      "step": 701
+    },
+    {
+      "epoch": 1.2742960944595823,
+      "grad_norm": 2.131220817565918,
+      "learning_rate": 7.922062005985105e-06,
+      "logits/chosen": 0.15667293965816498,
+      "logits/rejected": 0.12293773889541626,
+      "logps/chosen": -76.34903717041016,
+      "logps/rejected": -80.57623291015625,
+      "loss": 2.2422,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.408144474029541,
+      "rewards/margins": 0.5897243618965149,
+      "rewards/rejected": -2.997868776321411,
+      "step": 702
+    },
+    {
+      "epoch": 1.276112624886467,
+      "grad_norm": 1.9162285327911377,
+      "learning_rate": 7.916643755673994e-06,
+      "logits/chosen": -0.00397343747317791,
+      "logits/rejected": -0.01795889064669609,
+      "logps/chosen": -72.12644958496094,
+      "logps/rejected": -80.48601531982422,
+      "loss": 2.286,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1793017387390137,
+      "rewards/margins": 0.6844039559364319,
+      "rewards/rejected": -2.863705635070801,
+      "step": 703
+    },
+    {
+      "epoch": 1.2779291553133514,
+      "grad_norm": 2.877826690673828,
+      "learning_rate": 7.911220577405485e-06,
+      "logits/chosen": 0.031911637634038925,
+      "logits/rejected": -0.02112415060400963,
+      "logps/chosen": -92.17039489746094,
+      "logps/rejected": -100.05145263671875,
+      "loss": 2.2664,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5374324321746826,
+      "rewards/margins": 0.6277021765708923,
+      "rewards/rejected": -3.1651346683502197,
+      "step": 704
+    },
+    {
+      "epoch": 1.2797456857402363,
+      "grad_norm": 2.437596082687378,
+      "learning_rate": 7.905792482238387e-06,
+      "logits/chosen": 0.05161967873573303,
+      "logits/rejected": 0.07132750749588013,
+      "logps/chosen": -83.6578140258789,
+      "logps/rejected": -91.72476959228516,
+      "loss": 2.2835,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.62272310256958,
+      "rewards/margins": 0.7025635242462158,
+      "rewards/rejected": -3.325286865234375,
+      "step": 705
+    },
+    {
+      "epoch": 1.2815622161671207,
+      "grad_norm": 1.9355360269546509,
+      "learning_rate": 7.900359481241531e-06,
+      "logits/chosen": 0.11531795561313629,
+      "logits/rejected": 0.1957155019044876,
+      "logps/chosen": -70.70565795898438,
+      "logps/rejected": -83.15331268310547,
+      "loss": 1.9752,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.568023681640625,
+      "rewards/margins": 0.748910129070282,
+      "rewards/rejected": -3.3169333934783936,
+      "step": 706
+    },
+    {
+      "epoch": 1.2833787465940054,
+      "grad_norm": 1.9978153705596924,
+      "learning_rate": 7.89492158549375e-06,
+      "logits/chosen": 0.04371386021375656,
+      "logits/rejected": 0.07748796790838242,
+      "logps/chosen": -80.5776138305664,
+      "logps/rejected": -90.11798095703125,
+      "loss": 2.0422,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1579325199127197,
+      "rewards/margins": 0.726677656173706,
+      "rewards/rejected": -2.8846099376678467,
+      "step": 707
+    },
+    {
+      "epoch": 1.28519527702089,
+      "grad_norm": 1.982696771621704,
+      "learning_rate": 7.889478806083859e-06,
+      "logits/chosen": 0.11542633175849915,
+      "logits/rejected": 0.09008399397134781,
+      "logps/chosen": -80.96235656738281,
+      "logps/rejected": -84.60002899169922,
+      "loss": 2.1817,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.0073299407958984,
+      "rewards/margins": 0.626833438873291,
+      "rewards/rejected": -2.6341633796691895,
+      "step": 708
+    },
+    {
+      "epoch": 1.2870118074477748,
+      "grad_norm": 2.2399492263793945,
+      "learning_rate": 7.884031154110636e-06,
+      "logits/chosen": 0.13016283512115479,
+      "logits/rejected": 0.10277393460273743,
+      "logps/chosen": -76.04215240478516,
+      "logps/rejected": -82.029296875,
+      "loss": 2.1021,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.4932732582092285,
+      "rewards/margins": 0.7273231744766235,
+      "rewards/rejected": -3.2205960750579834,
+      "step": 709
+    },
+    {
+      "epoch": 1.2888283378746594,
+      "grad_norm": 1.7780169248580933,
+      "learning_rate": 7.87857864068279e-06,
+      "logits/chosen": 0.015022790059447289,
+      "logits/rejected": 0.08707302063703537,
+      "logps/chosen": -85.66944122314453,
+      "logps/rejected": -101.31831359863281,
+      "loss": 1.8779,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4976251125335693,
+      "rewards/margins": 0.9009003639221191,
+      "rewards/rejected": -3.3985257148742676,
+      "step": 710
+    },
+    {
+      "epoch": 1.2906448683015441,
+      "grad_norm": 2.2210092544555664,
+      "learning_rate": 7.873121276918947e-06,
+      "logits/chosen": 0.02679639309644699,
+      "logits/rejected": 0.00411232328042388,
+      "logps/chosen": -87.88114929199219,
+      "logps/rejected": -98.84779357910156,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.56654953956604,
+      "rewards/margins": 0.7424182295799255,
+      "rewards/rejected": -3.3089675903320312,
+      "step": 711
+    },
+    {
+      "epoch": 1.2924613987284288,
+      "grad_norm": 2.3803563117980957,
+      "learning_rate": 7.867659073947618e-06,
+      "logits/chosen": 0.12869741022586823,
+      "logits/rejected": 0.17313969135284424,
+      "logps/chosen": -78.2127685546875,
+      "logps/rejected": -89.54530334472656,
+      "loss": 2.2313,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2719531059265137,
+      "rewards/margins": 0.6112346649169922,
+      "rewards/rejected": -2.883187770843506,
+      "step": 712
+    },
+    {
+      "epoch": 1.2942779291553133,
+      "grad_norm": 1.8979357481002808,
+      "learning_rate": 7.862192042907191e-06,
+      "logits/chosen": -0.002289058640599251,
+      "logits/rejected": 0.01941063068807125,
+      "logps/chosen": -79.63894653320312,
+      "logps/rejected": -99.8336410522461,
+      "loss": 1.8824,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.396660566329956,
+      "rewards/margins": 1.1458237171173096,
+      "rewards/rejected": -3.5424838066101074,
+      "step": 713
+    },
+    {
+      "epoch": 1.296094459582198,
+      "grad_norm": 2.069608449935913,
+      "learning_rate": 7.856720194945892e-06,
+      "logits/chosen": 0.061066634953022,
+      "logits/rejected": 0.0689641609787941,
+      "logps/chosen": -78.00001525878906,
+      "logps/rejected": -87.9764404296875,
+      "loss": 2.0765,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3631246089935303,
+      "rewards/margins": 0.8202179670333862,
+      "rewards/rejected": -3.183342695236206,
+      "step": 714
+    },
+    {
+      "epoch": 1.2979109900090826,
+      "grad_norm": 2.080536365509033,
+      "learning_rate": 7.85124354122177e-06,
+      "logits/chosen": 0.055848486721515656,
+      "logits/rejected": 0.09090499579906464,
+      "logps/chosen": -86.07003784179688,
+      "logps/rejected": -95.0962905883789,
+      "loss": 2.0101,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5897908210754395,
+      "rewards/margins": 0.6455370187759399,
+      "rewards/rejected": -3.23532772064209,
+      "step": 715
+    },
+    {
+      "epoch": 1.2997275204359673,
+      "grad_norm": 1.9409488439559937,
+      "learning_rate": 7.845762092902678e-06,
+      "logits/chosen": 0.17071956396102905,
+      "logits/rejected": 0.11957748234272003,
+      "logps/chosen": -85.08267211914062,
+      "logps/rejected": -100.11673736572266,
+      "loss": 1.9544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2637064456939697,
+      "rewards/margins": 0.9579347968101501,
+      "rewards/rejected": -3.2216408252716064,
+      "step": 716
+    },
+    {
+      "epoch": 1.301544050862852,
+      "grad_norm": 2.3988044261932373,
+      "learning_rate": 7.840275861166243e-06,
+      "logits/chosen": 0.07587475329637527,
+      "logits/rejected": 0.018805254250764847,
+      "logps/chosen": -92.8540267944336,
+      "logps/rejected": -97.47180938720703,
+      "loss": 2.3578,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.429079294204712,
+      "rewards/margins": 0.5412003993988037,
+      "rewards/rejected": -2.9702794551849365,
+      "step": 717
+    },
+    {
+      "epoch": 1.3033605812897366,
+      "grad_norm": 2.509941577911377,
+      "learning_rate": 7.834784857199846e-06,
+      "logits/chosen": 0.0847601592540741,
+      "logits/rejected": 0.039042674005031586,
+      "logps/chosen": -80.2916030883789,
+      "logps/rejected": -81.85697174072266,
+      "loss": 2.0909,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4719648361206055,
+      "rewards/margins": 0.6977176666259766,
+      "rewards/rejected": -3.169682264328003,
+      "step": 718
+    },
+    {
+      "epoch": 1.3051771117166213,
+      "grad_norm": 2.192031145095825,
+      "learning_rate": 7.8292890922006e-06,
+      "logits/chosen": 0.048360832035541534,
+      "logits/rejected": 0.10979942232370377,
+      "logps/chosen": -72.04978942871094,
+      "logps/rejected": -85.74111938476562,
+      "loss": 2.1544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.367837905883789,
+      "rewards/margins": 0.6129733920097351,
+      "rewards/rejected": -2.98081111907959,
+      "step": 719
+    },
+    {
+      "epoch": 1.3069936421435058,
+      "grad_norm": 2.33526873588562,
+      "learning_rate": 7.82378857737533e-06,
+      "logits/chosen": 0.09924662113189697,
+      "logits/rejected": 0.10467047989368439,
+      "logps/chosen": -85.21525573730469,
+      "logps/rejected": -94.86468505859375,
+      "loss": 2.0783,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.466061592102051,
+      "rewards/margins": 0.7927428483963013,
+      "rewards/rejected": -3.2588045597076416,
+      "step": 720
+    },
+    {
+      "epoch": 1.3088101725703907,
+      "grad_norm": 2.381256341934204,
+      "learning_rate": 7.818283323940539e-06,
+      "logits/chosen": 0.1228807270526886,
+      "logits/rejected": 0.1338506042957306,
+      "logps/chosen": -90.72122192382812,
+      "logps/rejected": -96.53125,
+      "loss": 2.0252,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4308130741119385,
+      "rewards/margins": 0.6035470366477966,
+      "rewards/rejected": -3.03436017036438,
+      "step": 721
+    },
+    {
+      "epoch": 1.3106267029972751,
+      "grad_norm": 1.942866325378418,
+      "learning_rate": 7.8127733431224e-06,
+      "logits/chosen": 0.05967065691947937,
+      "logits/rejected": 0.08946052193641663,
+      "logps/chosen": -78.91156768798828,
+      "logps/rejected": -90.6188735961914,
+      "loss": 1.9435,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.333373546600342,
+      "rewards/margins": 0.7353364825248718,
+      "rewards/rejected": -3.0687103271484375,
+      "step": 722
+    },
+    {
+      "epoch": 1.3124432334241598,
+      "grad_norm": 2.068608283996582,
+      "learning_rate": 7.80725864615672e-06,
+      "logits/chosen": 0.01783425733447075,
+      "logits/rejected": 0.06930528581142426,
+      "logps/chosen": -81.48194885253906,
+      "logps/rejected": -96.17157745361328,
+      "loss": 2.0554,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.423696756362915,
+      "rewards/margins": 0.8400081396102905,
+      "rewards/rejected": -3.263705253601074,
+      "step": 723
+    },
+    {
+      "epoch": 1.3142597638510445,
+      "grad_norm": 2.349059820175171,
+      "learning_rate": 7.801739244288929e-06,
+      "logits/chosen": 0.08266536891460419,
+      "logits/rejected": 0.09768007695674896,
+      "logps/chosen": -82.81443786621094,
+      "logps/rejected": -92.47273254394531,
+      "loss": 1.8288,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.347062587738037,
+      "rewards/margins": 0.839061975479126,
+      "rewards/rejected": -3.186124563217163,
+      "step": 724
+    },
+    {
+      "epoch": 1.3160762942779292,
+      "grad_norm": 2.1312410831451416,
+      "learning_rate": 7.79621514877405e-06,
+      "logits/chosen": 0.05873649939894676,
+      "logits/rejected": 0.10713300108909607,
+      "logps/chosen": -99.603759765625,
+      "logps/rejected": -110.6041259765625,
+      "loss": 1.8875,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4636247158050537,
+      "rewards/margins": 0.8619670271873474,
+      "rewards/rejected": -3.325591564178467,
+      "step": 725
+    },
+    {
+      "epoch": 1.3178928247048138,
+      "grad_norm": 2.137871742248535,
+      "learning_rate": 7.79068637087667e-06,
+      "logits/chosen": 0.07839182764291763,
+      "logits/rejected": 0.1225314810872078,
+      "logps/chosen": -82.73197937011719,
+      "logps/rejected": -101.63593292236328,
+      "loss": 1.7664,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.491330862045288,
+      "rewards/margins": 1.1549298763275146,
+      "rewards/rejected": -3.646261215209961,
+      "step": 726
+    },
+    {
+      "epoch": 1.3197093551316985,
+      "grad_norm": 2.030689001083374,
+      "learning_rate": 7.785152921870939e-06,
+      "logits/chosen": 0.034951481968164444,
+      "logits/rejected": 0.06388817727565765,
+      "logps/chosen": -85.28425598144531,
+      "logps/rejected": -101.43257904052734,
+      "loss": 2.3856,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6969833374023438,
+      "rewards/margins": 0.727889895439148,
+      "rewards/rejected": -3.4248733520507812,
+      "step": 727
+    },
+    {
+      "epoch": 1.3215258855585832,
+      "grad_norm": 2.65952205657959,
+      "learning_rate": 7.779614813040513e-06,
+      "logits/chosen": 0.1679934561252594,
+      "logits/rejected": 0.18826396763324738,
+      "logps/chosen": -80.64529418945312,
+      "logps/rejected": -87.35059356689453,
+      "loss": 1.9723,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.259064197540283,
+      "rewards/margins": 0.7135547995567322,
+      "rewards/rejected": -2.97261905670166,
+      "step": 728
+    },
+    {
+      "epoch": 1.3233424159854676,
+      "grad_norm": 2.1841068267822266,
+      "learning_rate": 7.774072055678565e-06,
+      "logits/chosen": 0.11853618919849396,
+      "logits/rejected": 0.1412927508354187,
+      "logps/chosen": -80.26649475097656,
+      "logps/rejected": -88.19939422607422,
+      "loss": 2.1344,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5820248126983643,
+      "rewards/margins": 0.8093788623809814,
+      "rewards/rejected": -3.3914034366607666,
+      "step": 729
+    },
+    {
+      "epoch": 1.3251589464123525,
+      "grad_norm": 2.463083505630493,
+      "learning_rate": 7.768524661087744e-06,
+      "logits/chosen": 0.05979528650641441,
+      "logits/rejected": 0.10566453635692596,
+      "logps/chosen": -88.6331787109375,
+      "logps/rejected": -97.57271575927734,
+      "loss": 2.1545,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4209208488464355,
+      "rewards/margins": 0.7680116295814514,
+      "rewards/rejected": -3.188931941986084,
+      "step": 730
+    },
+    {
+      "epoch": 1.326975476839237,
+      "grad_norm": 2.3693671226501465,
+      "learning_rate": 7.76297264058015e-06,
+      "logits/chosen": 0.05617160722613335,
+      "logits/rejected": 0.017447492107748985,
+      "logps/chosen": -91.15104675292969,
+      "logps/rejected": -91.32392883300781,
+      "loss": 2.0652,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.469024181365967,
+      "rewards/margins": 0.7454814910888672,
+      "rewards/rejected": -3.214505672454834,
+      "step": 731
+    },
+    {
+      "epoch": 1.3287920072661217,
+      "grad_norm": 2.0553715229034424,
+      "learning_rate": 7.757416005477323e-06,
+      "logits/chosen": 0.10786914825439453,
+      "logits/rejected": 0.1281609684228897,
+      "logps/chosen": -83.35604095458984,
+      "logps/rejected": -91.0313949584961,
+      "loss": 2.0932,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.75126314163208,
+      "rewards/margins": 0.7679738402366638,
+      "rewards/rejected": -3.5192370414733887,
+      "step": 732
+    },
+    {
+      "epoch": 1.3306085376930064,
+      "grad_norm": 2.6259214878082275,
+      "learning_rate": 7.751854767110205e-06,
+      "logits/chosen": 0.025147411972284317,
+      "logits/rejected": 0.11643064022064209,
+      "logps/chosen": -74.00305938720703,
+      "logps/rejected": -99.70087432861328,
+      "loss": 1.9757,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.617929220199585,
+      "rewards/margins": 1.131290078163147,
+      "rewards/rejected": -3.7492191791534424,
+      "step": 733
+    },
+    {
+      "epoch": 1.332425068119891,
+      "grad_norm": 2.21140193939209,
+      "learning_rate": 7.746288936819134e-06,
+      "logits/chosen": 0.036435484886169434,
+      "logits/rejected": 0.09736932814121246,
+      "logps/chosen": -79.6103515625,
+      "logps/rejected": -97.5123519897461,
+      "loss": 1.999,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.716804027557373,
+      "rewards/margins": 0.9280580282211304,
+      "rewards/rejected": -3.644861936569214,
+      "step": 734
+    },
+    {
+      "epoch": 1.3342415985467757,
+      "grad_norm": 1.8063753843307495,
+      "learning_rate": 7.740718525953804e-06,
+      "logits/chosen": 0.012957969680428505,
+      "logits/rejected": 0.06407644599676132,
+      "logps/chosen": -86.7850112915039,
+      "logps/rejected": -104.12468719482422,
+      "loss": 1.8674,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5402910709381104,
+      "rewards/margins": 0.9172405004501343,
+      "rewards/rejected": -3.457531690597534,
+      "step": 735
+    },
+    {
+      "epoch": 1.3360581289736604,
+      "grad_norm": 2.384291410446167,
+      "learning_rate": 7.735143545873252e-06,
+      "logits/chosen": 0.15138015151023865,
+      "logits/rejected": 0.1674022376537323,
+      "logps/chosen": -78.28397369384766,
+      "logps/rejected": -88.92375183105469,
+      "loss": 1.8622,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6077139377593994,
+      "rewards/margins": 0.9488463401794434,
+      "rewards/rejected": -3.5565602779388428,
+      "step": 736
+    },
+    {
+      "epoch": 1.337874659400545,
+      "grad_norm": 2.1717827320098877,
+      "learning_rate": 7.729564007945834e-06,
+      "logits/chosen": 0.07187292724847794,
+      "logits/rejected": 0.12309285253286362,
+      "logps/chosen": -87.5180892944336,
+      "logps/rejected": -93.48638153076172,
+      "loss": 2.2686,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4702401161193848,
+      "rewards/margins": 0.6287782192230225,
+      "rewards/rejected": -3.099018096923828,
+      "step": 737
+    },
+    {
+      "epoch": 1.3396911898274295,
+      "grad_norm": 2.6149754524230957,
+      "learning_rate": 7.723979923549201e-06,
+      "logits/chosen": 0.10500997304916382,
+      "logits/rejected": 0.13640286028385162,
+      "logps/chosen": -76.4052963256836,
+      "logps/rejected": -84.28958892822266,
+      "loss": 2.4316,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.9436187744140625,
+      "rewards/margins": 0.5192191004753113,
+      "rewards/rejected": -3.4628381729125977,
+      "step": 738
+    },
+    {
+      "epoch": 1.3415077202543142,
+      "grad_norm": 2.2604503631591797,
+      "learning_rate": 7.718391304070274e-06,
+      "logits/chosen": 0.019504152238368988,
+      "logits/rejected": 0.0677918866276741,
+      "logps/chosen": -96.57980346679688,
+      "logps/rejected": -111.43043518066406,
+      "loss": 2.1929,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.63236141204834,
+      "rewards/margins": 0.8397969007492065,
+      "rewards/rejected": -3.4721579551696777,
+      "step": 739
+    },
+    {
+      "epoch": 1.3433242506811989,
+      "grad_norm": 2.827580451965332,
+      "learning_rate": 7.712798160905218e-06,
+      "logits/chosen": 0.028927450999617577,
+      "logits/rejected": 0.060569021850824356,
+      "logps/chosen": -85.33325958251953,
+      "logps/rejected": -87.87812042236328,
+      "loss": 2.2098,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.669008255004883,
+      "rewards/margins": 0.6498023271560669,
+      "rewards/rejected": -3.31881046295166,
+      "step": 740
+    },
+    {
+      "epoch": 1.3451407811080835,
+      "grad_norm": 2.375645160675049,
+      "learning_rate": 7.707200505459427e-06,
+      "logits/chosen": 0.04526009410619736,
+      "logits/rejected": 0.0799727588891983,
+      "logps/chosen": -89.62974548339844,
+      "logps/rejected": -99.64259338378906,
+      "loss": 2.2352,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.709292411804199,
+      "rewards/margins": 0.6063066720962524,
+      "rewards/rejected": -3.3155994415283203,
+      "step": 741
+    },
+    {
+      "epoch": 1.3469573115349682,
+      "grad_norm": 2.4972751140594482,
+      "learning_rate": 7.701598349147496e-06,
+      "logits/chosen": -0.020723650231957436,
+      "logits/rejected": 0.04764425382018089,
+      "logps/chosen": -85.17987060546875,
+      "logps/rejected": -98.80146789550781,
+      "loss": 2.0034,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.656331777572632,
+      "rewards/margins": 0.9565580487251282,
+      "rewards/rejected": -3.6128900051116943,
+      "step": 742
+    },
+    {
+      "epoch": 1.348773841961853,
+      "grad_norm": 2.23154616355896,
+      "learning_rate": 7.695991703393196e-06,
+      "logits/chosen": 0.062117062509059906,
+      "logits/rejected": 0.018935926258563995,
+      "logps/chosen": -99.47203063964844,
+      "logps/rejected": -104.25543212890625,
+      "loss": 2.3257,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.538417339324951,
+      "rewards/margins": 0.5829829573631287,
+      "rewards/rejected": -3.1214003562927246,
+      "step": 743
+    },
+    {
+      "epoch": 1.3505903723887376,
+      "grad_norm": 2.1097710132598877,
+      "learning_rate": 7.690380579629454e-06,
+      "logits/chosen": 0.07178185135126114,
+      "logits/rejected": 0.13060711324214935,
+      "logps/chosen": -82.14419555664062,
+      "logps/rejected": -94.78469848632812,
+      "loss": 1.9158,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.5844056606292725,
+      "rewards/margins": 0.95278000831604,
+      "rewards/rejected": -3.5371859073638916,
+      "step": 744
+    },
+    {
+      "epoch": 1.352406902815622,
+      "grad_norm": 2.6826353073120117,
+      "learning_rate": 7.684764989298329e-06,
+      "logits/chosen": 0.10977941751480103,
+      "logits/rejected": 0.11871352046728134,
+      "logps/chosen": -87.78453826904297,
+      "logps/rejected": -88.22404479980469,
+      "loss": 2.405,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.647420644760132,
+      "rewards/margins": 0.5518589019775391,
+      "rewards/rejected": -3.199279546737671,
+      "step": 745
+    },
+    {
+      "epoch": 1.354223433242507,
+      "grad_norm": 2.235769510269165,
+      "learning_rate": 7.679144943850986e-06,
+      "logits/chosen": 0.05031013488769531,
+      "logits/rejected": 0.01593434251844883,
+      "logps/chosen": -94.69499206542969,
+      "logps/rejected": -99.25582885742188,
+      "loss": 2.4038,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.653748035430908,
+      "rewards/margins": 0.7265360355377197,
+      "rewards/rejected": -3.380283832550049,
+      "step": 746
+    },
+    {
+      "epoch": 1.3560399636693914,
+      "grad_norm": 2.4417803287506104,
+      "learning_rate": 7.673520454747681e-06,
+      "logits/chosen": 0.04040508344769478,
+      "logits/rejected": 0.07281962782144547,
+      "logps/chosen": -77.71315002441406,
+      "logps/rejected": -94.82330322265625,
+      "loss": 1.6934,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4112367630004883,
+      "rewards/margins": 1.1407581567764282,
+      "rewards/rejected": -3.551994800567627,
+      "step": 747
+    },
+    {
+      "epoch": 1.357856494096276,
+      "grad_norm": 1.7434958219528198,
+      "learning_rate": 7.667891533457718e-06,
+      "logits/chosen": 0.1453830599784851,
+      "logits/rejected": 0.10198508203029633,
+      "logps/chosen": -79.32072448730469,
+      "logps/rejected": -88.78571319580078,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6976542472839355,
+      "rewards/margins": 0.8518580198287964,
+      "rewards/rejected": -3.5495121479034424,
+      "step": 748
+    },
+    {
+      "epoch": 1.3596730245231607,
+      "grad_norm": 2.364915609359741,
+      "learning_rate": 7.662258191459456e-06,
+      "logits/chosen": 0.07202528417110443,
+      "logits/rejected": 0.07270742952823639,
+      "logps/chosen": -82.5189208984375,
+      "logps/rejected": -87.38330841064453,
+      "loss": 2.2639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.583577871322632,
+      "rewards/margins": 0.5575169324874878,
+      "rewards/rejected": -3.141094923019409,
+      "step": 749
+    },
+    {
+      "epoch": 1.3614895549500454,
+      "grad_norm": 1.9622827768325806,
+      "learning_rate": 7.656620440240257e-06,
+      "logits/chosen": 0.0625351071357727,
+      "logits/rejected": 0.1307908296585083,
+      "logps/chosen": -84.89385223388672,
+      "logps/rejected": -92.00534057617188,
+      "loss": 2.0835,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.679396629333496,
+      "rewards/margins": 0.7283416986465454,
+      "rewards/rejected": -3.407738208770752,
+      "step": 750
+    },
+    {
+      "epoch": 1.36330608537693,
+      "grad_norm": 3.2049853801727295,
+      "learning_rate": 7.650978291296475e-06,
+      "logits/chosen": 0.06182331219315529,
+      "logits/rejected": 0.08293959498405457,
+      "logps/chosen": -77.00846099853516,
+      "logps/rejected": -85.75076293945312,
+      "loss": 2.1026,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4312846660614014,
+      "rewards/margins": 0.7030216455459595,
+      "rewards/rejected": -3.134305953979492,
+      "step": 751
+    },
+    {
+      "epoch": 1.3651226158038148,
+      "grad_norm": 2.065100908279419,
+      "learning_rate": 7.645331756133438e-06,
+      "logits/chosen": 0.12972712516784668,
+      "logits/rejected": 0.12206097692251205,
+      "logps/chosen": -88.62772369384766,
+      "logps/rejected": -98.35625457763672,
+      "loss": 1.8117,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.59391450881958,
+      "rewards/margins": 0.8645298480987549,
+      "rewards/rejected": -3.458444833755493,
+      "step": 752
+    },
+    {
+      "epoch": 1.3669391462306995,
+      "grad_norm": 2.078524112701416,
+      "learning_rate": 7.639680846265411e-06,
+      "logits/chosen": 0.09852910041809082,
+      "logits/rejected": 0.12210299074649811,
+      "logps/chosen": -81.41858673095703,
+      "logps/rejected": -92.6720199584961,
+      "loss": 1.9148,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4295332431793213,
+      "rewards/margins": 0.8047889471054077,
+      "rewards/rejected": -3.2343220710754395,
+      "step": 753
+    },
+    {
+      "epoch": 1.368755676657584,
+      "grad_norm": 1.991837739944458,
+      "learning_rate": 7.634025573215585e-06,
+      "logits/chosen": -0.02604215405881405,
+      "logits/rejected": 0.048631928861141205,
+      "logps/chosen": -95.89967346191406,
+      "logps/rejected": -109.60630798339844,
+      "loss": 1.9067,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.491032361984253,
+      "rewards/margins": 1.0193102359771729,
+      "rewards/rejected": -3.5103423595428467,
+      "step": 754
+    },
+    {
+      "epoch": 1.3705722070844686,
+      "grad_norm": 2.560757875442505,
+      "learning_rate": 7.628365948516047e-06,
+      "logits/chosen": 0.12844571471214294,
+      "logits/rejected": 0.11145608872175217,
+      "logps/chosen": -81.85939025878906,
+      "logps/rejected": -92.21177673339844,
+      "loss": 2.195,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6527202129364014,
+      "rewards/margins": 0.7286384105682373,
+      "rewards/rejected": -3.3813586235046387,
+      "step": 755
+    },
+    {
+      "epoch": 1.3723887375113533,
+      "grad_norm": 2.5189437866210938,
+      "learning_rate": 7.622701983707756e-06,
+      "logits/chosen": 0.049444377422332764,
+      "logits/rejected": 0.07628104835748672,
+      "logps/chosen": -80.98538970947266,
+      "logps/rejected": -89.72940063476562,
+      "loss": 2.1411,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7696704864501953,
+      "rewards/margins": 0.7333941459655762,
+      "rewards/rejected": -3.5030641555786133,
+      "step": 756
+    },
+    {
+      "epoch": 1.374205267938238,
+      "grad_norm": 2.5741689205169678,
+      "learning_rate": 7.617033690340523e-06,
+      "logits/chosen": 0.02790415659546852,
+      "logits/rejected": 0.0068018268793821335,
+      "logps/chosen": -83.76559448242188,
+      "logps/rejected": -91.61640167236328,
+      "loss": 2.203,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.734015941619873,
+      "rewards/margins": 0.6714599132537842,
+      "rewards/rejected": -3.405475616455078,
+      "step": 757
+    },
+    {
+      "epoch": 1.3760217983651226,
+      "grad_norm": 1.7192715406417847,
+      "learning_rate": 7.611361079972984e-06,
+      "logits/chosen": 0.1048116460442543,
+      "logits/rejected": 0.13254055380821228,
+      "logps/chosen": -67.2003402709961,
+      "logps/rejected": -91.35997009277344,
+      "loss": 1.5709,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.3325514793395996,
+      "rewards/margins": 1.2846351861953735,
+      "rewards/rejected": -3.617187023162842,
+      "step": 758
+    },
+    {
+      "epoch": 1.3778383287920073,
+      "grad_norm": 2.5107131004333496,
+      "learning_rate": 7.605684164172581e-06,
+      "logits/chosen": 0.04970414191484451,
+      "logits/rejected": 0.06727412343025208,
+      "logps/chosen": -78.22857666015625,
+      "logps/rejected": -91.56071472167969,
+      "loss": 2.2132,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.5088820457458496,
+      "rewards/margins": 0.8137573003768921,
+      "rewards/rejected": -3.322639226913452,
+      "step": 759
+    },
+    {
+      "epoch": 1.379654859218892,
+      "grad_norm": 2.327336311340332,
+      "learning_rate": 7.600002954515532e-06,
+      "logits/chosen": 0.09795573353767395,
+      "logits/rejected": 0.04695986211299896,
+      "logps/chosen": -85.50299835205078,
+      "logps/rejected": -92.48783111572266,
+      "loss": 1.9836,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.55703067779541,
+      "rewards/margins": 0.7251678109169006,
+      "rewards/rejected": -3.282198429107666,
+      "step": 760
+    },
+    {
+      "epoch": 1.3814713896457766,
+      "grad_norm": 2.2488017082214355,
+      "learning_rate": 7.5943174625868146e-06,
+      "logits/chosen": 0.00528271310031414,
+      "logits/rejected": 0.05780588090419769,
+      "logps/chosen": -90.95349884033203,
+      "logps/rejected": -98.3586654663086,
+      "loss": 2.4998,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.74594783782959,
+      "rewards/margins": 0.4672420024871826,
+      "rewards/rejected": -3.2131900787353516,
+      "step": 761
+    },
+    {
+      "epoch": 1.3832879200726613,
+      "grad_norm": 2.4573490619659424,
+      "learning_rate": 7.588627699980134e-06,
+      "logits/chosen": 0.023196734488010406,
+      "logits/rejected": 0.07495579123497009,
+      "logps/chosen": -77.51380157470703,
+      "logps/rejected": -95.95264434814453,
+      "loss": 1.8657,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.519115447998047,
+      "rewards/margins": 1.0751680135726929,
+      "rewards/rejected": -3.59428334236145,
+      "step": 762
+    },
+    {
+      "epoch": 1.3851044504995458,
+      "grad_norm": 2.718414783477783,
+      "learning_rate": 7.582933678297911e-06,
+      "logits/chosen": -0.006191038526594639,
+      "logits/rejected": 0.08378598839044571,
+      "logps/chosen": -85.86323547363281,
+      "logps/rejected": -98.59344482421875,
+      "loss": 2.3292,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.69970703125,
+      "rewards/margins": 0.7494994401931763,
+      "rewards/rejected": -3.4492063522338867,
+      "step": 763
+    },
+    {
+      "epoch": 1.3869209809264305,
+      "grad_norm": 3.296757936477661,
+      "learning_rate": 7.577235409151242e-06,
+      "logits/chosen": 0.03207925334572792,
+      "logits/rejected": 0.11028114706277847,
+      "logps/chosen": -77.22196960449219,
+      "logps/rejected": -87.56900024414062,
+      "loss": 2.5069,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.9215140342712402,
+      "rewards/margins": 0.6305817365646362,
+      "rewards/rejected": -3.552095890045166,
+      "step": 764
+    },
+    {
+      "epoch": 1.3887375113533151,
+      "grad_norm": 2.6371073722839355,
+      "learning_rate": 7.5715329041598955e-06,
+      "logits/chosen": 0.06983280181884766,
+      "logits/rejected": 0.0031841248273849487,
+      "logps/chosen": -93.26805877685547,
+      "logps/rejected": -100.43709564208984,
+      "loss": 2.1158,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.602299928665161,
+      "rewards/margins": 0.8267786502838135,
+      "rewards/rejected": -3.4290785789489746,
+      "step": 765
+    },
+    {
+      "epoch": 1.3905540417801998,
+      "grad_norm": 2.557405471801758,
+      "learning_rate": 7.565826174952267e-06,
+      "logits/chosen": 0.0866529792547226,
+      "logits/rejected": 0.11220981180667877,
+      "logps/chosen": -81.10444641113281,
+      "logps/rejected": -89.5517807006836,
+      "loss": 2.0158,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.469362258911133,
+      "rewards/margins": 0.6984151601791382,
+      "rewards/rejected": -3.1677775382995605,
+      "step": 766
+    },
+    {
+      "epoch": 1.3923705722070845,
+      "grad_norm": 2.013252019882202,
+      "learning_rate": 7.560115233165375e-06,
+      "logits/chosen": 0.10826343297958374,
+      "logits/rejected": 0.09555835276842117,
+      "logps/chosen": -90.81183624267578,
+      "logps/rejected": -99.46356201171875,
+      "loss": 1.9386,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6625754833221436,
+      "rewards/margins": 0.9187412858009338,
+      "rewards/rejected": -3.5813164710998535,
+      "step": 767
+    },
+    {
+      "epoch": 1.3941871026339692,
+      "grad_norm": 2.1441266536712646,
+      "learning_rate": 7.554400090444822e-06,
+      "logits/chosen": 0.04958781227469444,
+      "logits/rejected": 0.10418076813220978,
+      "logps/chosen": -73.46155548095703,
+      "logps/rejected": -83.24677276611328,
+      "loss": 1.9933,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4550204277038574,
+      "rewards/margins": 0.7676887512207031,
+      "rewards/rejected": -3.2227089405059814,
+      "step": 768
+    },
+    {
+      "epoch": 1.3960036330608538,
+      "grad_norm": 2.401198387145996,
+      "learning_rate": 7.54868075844478e-06,
+      "logits/chosen": 0.10376621782779694,
+      "logits/rejected": 0.1655123233795166,
+      "logps/chosen": -80.3701171875,
+      "logps/rejected": -95.8973159790039,
+      "loss": 1.784,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2211532592773438,
+      "rewards/margins": 1.070261836051941,
+      "rewards/rejected": -3.291415214538574,
+      "step": 769
+    },
+    {
+      "epoch": 1.3978201634877383,
+      "grad_norm": 1.93374502658844,
+      "learning_rate": 7.5429572488279615e-06,
+      "logits/chosen": 0.081766776740551,
+      "logits/rejected": 0.1179564967751503,
+      "logps/chosen": -73.5924301147461,
+      "logps/rejected": -85.84019470214844,
+      "loss": 2.2637,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3363659381866455,
+      "rewards/margins": 0.6945525407791138,
+      "rewards/rejected": -3.030918598175049,
+      "step": 770
+    },
+    {
+      "epoch": 1.3996366939146232,
+      "grad_norm": 2.4809932708740234,
+      "learning_rate": 7.5372295732656e-06,
+      "logits/chosen": 0.11731807887554169,
+      "logits/rejected": 0.05732544884085655,
+      "logps/chosen": -89.25398254394531,
+      "logps/rejected": -90.94881439208984,
+      "loss": 2.412,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.435483694076538,
+      "rewards/margins": 0.5316743850708008,
+      "rewards/rejected": -2.967158317565918,
+      "step": 771
+    },
+    {
+      "epoch": 1.4014532243415077,
+      "grad_norm": 2.622253894805908,
+      "learning_rate": 7.531497743437424e-06,
+      "logits/chosen": -0.02307087928056717,
+      "logits/rejected": 0.034644536674022675,
+      "logps/chosen": -79.53402709960938,
+      "logps/rejected": -95.38764953613281,
+      "loss": 2.1841,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5010921955108643,
+      "rewards/margins": 0.7504414916038513,
+      "rewards/rejected": -3.2515335083007812,
+      "step": 772
+    },
+    {
+      "epoch": 1.4032697547683923,
+      "grad_norm": 2.341421365737915,
+      "learning_rate": 7.525761771031632e-06,
+      "logits/chosen": 0.09784074872732162,
+      "logits/rejected": 0.1302230954170227,
+      "logps/chosen": -78.1878433227539,
+      "logps/rejected": -88.39629364013672,
+      "loss": 2.064,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.63604736328125,
+      "rewards/margins": 0.6544671654701233,
+      "rewards/rejected": -3.2905144691467285,
+      "step": 773
+    },
+    {
+      "epoch": 1.405086285195277,
+      "grad_norm": 2.4410409927368164,
+      "learning_rate": 7.520021667744869e-06,
+      "logits/chosen": 0.03553074598312378,
+      "logits/rejected": 0.024142932146787643,
+      "logps/chosen": -92.7217025756836,
+      "logps/rejected": -99.62353515625,
+      "loss": 2.077,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5202903747558594,
+      "rewards/margins": 0.9078298211097717,
+      "rewards/rejected": -3.4281198978424072,
+      "step": 774
+    },
+    {
+      "epoch": 1.4069028156221617,
+      "grad_norm": 2.1428725719451904,
+      "learning_rate": 7.514277445282206e-06,
+      "logits/chosen": 0.12562477588653564,
+      "logits/rejected": 0.11905661970376968,
+      "logps/chosen": -85.4616470336914,
+      "logps/rejected": -101.02755737304688,
+      "loss": 1.8298,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.236478328704834,
+      "rewards/margins": 0.9414302110671997,
+      "rewards/rejected": -3.177908420562744,
+      "step": 775
+    },
+    {
+      "epoch": 1.4087193460490464,
+      "grad_norm": 2.2605643272399902,
+      "learning_rate": 7.508529115357115e-06,
+      "logits/chosen": 0.08080364018678665,
+      "logits/rejected": 0.14329738914966583,
+      "logps/chosen": -86.23857116699219,
+      "logps/rejected": -94.77734375,
+      "loss": 2.1025,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3885529041290283,
+      "rewards/margins": 0.8040337562561035,
+      "rewards/rejected": -3.1925864219665527,
+      "step": 776
+    },
+    {
+      "epoch": 1.410535876475931,
+      "grad_norm": 2.1767449378967285,
+      "learning_rate": 7.502776689691439e-06,
+      "logits/chosen": 0.061047837138175964,
+      "logits/rejected": 0.049486950039863586,
+      "logps/chosen": -79.15450286865234,
+      "logps/rejected": -89.98368072509766,
+      "loss": 2.032,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.129742383956909,
+      "rewards/margins": 0.7515033483505249,
+      "rewards/rejected": -2.8812458515167236,
+      "step": 777
+    },
+    {
+      "epoch": 1.4123524069028157,
+      "grad_norm": 2.2512662410736084,
+      "learning_rate": 7.4970201800153765e-06,
+      "logits/chosen": 0.08744536340236664,
+      "logits/rejected": 0.10895106941461563,
+      "logps/chosen": -68.41326141357422,
+      "logps/rejected": -82.72618103027344,
+      "loss": 2.0954,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.360013008117676,
+      "rewards/margins": 0.856154203414917,
+      "rewards/rejected": -3.2161672115325928,
+      "step": 778
+    },
+    {
+      "epoch": 1.4141689373297002,
+      "grad_norm": 2.1528515815734863,
+      "learning_rate": 7.491259598067452e-06,
+      "logits/chosen": 0.10324890911579132,
+      "logits/rejected": 0.017273597419261932,
+      "logps/chosen": -101.93807983398438,
+      "logps/rejected": -110.01669311523438,
+      "loss": 1.8012,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2577784061431885,
+      "rewards/margins": 0.8829382658004761,
+      "rewards/rejected": -3.140716552734375,
+      "step": 779
+    },
+    {
+      "epoch": 1.4159854677565848,
+      "grad_norm": 2.24306321144104,
+      "learning_rate": 7.485494955594495e-06,
+      "logits/chosen": 0.0638672262430191,
+      "logits/rejected": 0.08847285807132721,
+      "logps/chosen": -83.1998291015625,
+      "logps/rejected": -87.86129760742188,
+      "loss": 2.215,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.975947380065918,
+      "rewards/margins": 0.5703001022338867,
+      "rewards/rejected": -2.546247720718384,
+      "step": 780
+    },
+    {
+      "epoch": 1.4178019981834695,
+      "grad_norm": 2.103116035461426,
+      "learning_rate": 7.479726264351619e-06,
+      "logits/chosen": 0.09146749973297119,
+      "logits/rejected": 0.12321461737155914,
+      "logps/chosen": -81.99695587158203,
+      "logps/rejected": -89.16120910644531,
+      "loss": 1.9022,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.159066677093506,
+      "rewards/margins": 0.8253484964370728,
+      "rewards/rejected": -2.984415054321289,
+      "step": 781
+    },
+    {
+      "epoch": 1.4196185286103542,
+      "grad_norm": 2.1531925201416016,
+      "learning_rate": 7.473953536102185e-06,
+      "logits/chosen": 0.1091499999165535,
+      "logits/rejected": 0.13049989938735962,
+      "logps/chosen": -81.74616241455078,
+      "logps/rejected": -94.27207946777344,
+      "loss": 1.9222,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.445281982421875,
+      "rewards/margins": 0.8645297288894653,
+      "rewards/rejected": -3.30981183052063,
+      "step": 782
+    },
+    {
+      "epoch": 1.4214350590372389,
+      "grad_norm": 2.489908218383789,
+      "learning_rate": 7.468176782617793e-06,
+      "logits/chosen": -0.04253482073545456,
+      "logits/rejected": 0.10589614510536194,
+      "logps/chosen": -83.19894409179688,
+      "logps/rejected": -99.28282928466797,
+      "loss": 1.7606,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2363641262054443,
+      "rewards/margins": 1.0474493503570557,
+      "rewards/rejected": -3.2838134765625,
+      "step": 783
+    },
+    {
+      "epoch": 1.4232515894641236,
+      "grad_norm": 2.4771764278411865,
+      "learning_rate": 7.462396015678249e-06,
+      "logits/chosen": 0.02688070759177208,
+      "logits/rejected": 0.05978942662477493,
+      "logps/chosen": -75.48037719726562,
+      "logps/rejected": -87.0803451538086,
+      "loss": 1.892,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3419995307922363,
+      "rewards/margins": 0.8640164732933044,
+      "rewards/rejected": -3.2060160636901855,
+      "step": 784
+    },
+    {
+      "epoch": 1.4250681198910082,
+      "grad_norm": 2.3507192134857178,
+      "learning_rate": 7.456611247071544e-06,
+      "logits/chosen": 0.13474830985069275,
+      "logits/rejected": 0.08549933135509491,
+      "logps/chosen": -78.67109680175781,
+      "logps/rejected": -80.96075439453125,
+      "loss": 2.1411,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.477113723754883,
+      "rewards/margins": 0.652525007724762,
+      "rewards/rejected": -3.129638671875,
+      "step": 785
+    },
+    {
+      "epoch": 1.4268846503178927,
+      "grad_norm": 2.009307861328125,
+      "learning_rate": 7.450822488593829e-06,
+      "logits/chosen": 0.04329455643892288,
+      "logits/rejected": 0.05247782543301582,
+      "logps/chosen": -89.56121063232422,
+      "logps/rejected": -99.5736083984375,
+      "loss": 1.8014,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2546510696411133,
+      "rewards/margins": 0.9092133045196533,
+      "rewards/rejected": -3.1638646125793457,
+      "step": 786
+    },
+    {
+      "epoch": 1.4287011807447776,
+      "grad_norm": 2.3687586784362793,
+      "learning_rate": 7.445029752049387e-06,
+      "logits/chosen": 0.04104599729180336,
+      "logits/rejected": 0.08001314103603363,
+      "logps/chosen": -77.9354248046875,
+      "logps/rejected": -92.67184448242188,
+      "loss": 1.7508,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2342448234558105,
+      "rewards/margins": 0.9252514839172363,
+      "rewards/rejected": -3.159496307373047,
+      "step": 787
+    },
+    {
+      "epoch": 1.430517711171662,
+      "grad_norm": 1.8127901554107666,
+      "learning_rate": 7.439233049250622e-06,
+      "logits/chosen": 0.07227849960327148,
+      "logits/rejected": 0.0907898098230362,
+      "logps/chosen": -71.49815368652344,
+      "logps/rejected": -82.32070922851562,
+      "loss": 1.8207,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.33886456489563,
+      "rewards/margins": 0.906674325466156,
+      "rewards/rejected": -3.2455389499664307,
+      "step": 788
+    },
+    {
+      "epoch": 1.4323342415985467,
+      "grad_norm": 2.0310990810394287,
+      "learning_rate": 7.433432392018013e-06,
+      "logits/chosen": 0.07254546135663986,
+      "logits/rejected": 0.08750150352716446,
+      "logps/chosen": -87.65977478027344,
+      "logps/rejected": -105.3295669555664,
+      "loss": 1.781,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.4844467639923096,
+      "rewards/margins": 0.8295308351516724,
+      "rewards/rejected": -3.3139777183532715,
+      "step": 789
+    },
+    {
+      "epoch": 1.4341507720254314,
+      "grad_norm": 2.093871831893921,
+      "learning_rate": 7.427627792180116e-06,
+      "logits/chosen": 0.09880155324935913,
+      "logits/rejected": 0.15118218958377838,
+      "logps/chosen": -75.81531524658203,
+      "logps/rejected": -88.76158142089844,
+      "loss": 2.2499,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6412038803100586,
+      "rewards/margins": 0.7693830132484436,
+      "rewards/rejected": -3.4105868339538574,
+      "step": 790
+    },
+    {
+      "epoch": 1.435967302452316,
+      "grad_norm": 2.393105983734131,
+      "learning_rate": 7.4218192615735185e-06,
+      "logits/chosen": 0.08802710473537445,
+      "logits/rejected": 0.04610452428460121,
+      "logps/chosen": -99.03189849853516,
+      "logps/rejected": -103.37187194824219,
+      "loss": 1.8155,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3743648529052734,
+      "rewards/margins": 0.8601830005645752,
+      "rewards/rejected": -3.2345480918884277,
+      "step": 791
+    },
+    {
+      "epoch": 1.4377838328792008,
+      "grad_norm": 2.148087978363037,
+      "learning_rate": 7.416006812042827e-06,
+      "logits/chosen": 0.05752583593130112,
+      "logits/rejected": 0.034685466438531876,
+      "logps/chosen": -89.9745101928711,
+      "logps/rejected": -97.91671752929688,
+      "loss": 1.9892,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.483428716659546,
+      "rewards/margins": 0.7583959102630615,
+      "rewards/rejected": -3.2418243885040283,
+      "step": 792
+    },
+    {
+      "epoch": 1.4396003633060854,
+      "grad_norm": 2.605093479156494,
+      "learning_rate": 7.410190455440638e-06,
+      "logits/chosen": 0.07939667254686356,
+      "logits/rejected": 0.05258120596408844,
+      "logps/chosen": -81.38958740234375,
+      "logps/rejected": -92.0197982788086,
+      "loss": 1.8609,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4127330780029297,
+      "rewards/margins": 0.8712365627288818,
+      "rewards/rejected": -3.2839698791503906,
+      "step": 793
+    },
+    {
+      "epoch": 1.44141689373297,
+      "grad_norm": 2.3762154579162598,
+      "learning_rate": 7.4043702036275136e-06,
+      "logits/chosen": 0.04353059083223343,
+      "logits/rejected": 0.08619395643472672,
+      "logps/chosen": -83.58119201660156,
+      "logps/rejected": -94.14823150634766,
+      "loss": 2.4321,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.5087859630584717,
+      "rewards/margins": 0.4951530396938324,
+      "rewards/rejected": -3.003939151763916,
+      "step": 794
+    },
+    {
+      "epoch": 1.4432334241598546,
+      "grad_norm": 2.350267171859741,
+      "learning_rate": 7.398546068471961e-06,
+      "logits/chosen": 0.18240118026733398,
+      "logits/rejected": 0.14964665472507477,
+      "logps/chosen": -79.20641326904297,
+      "logps/rejected": -86.13677978515625,
+      "loss": 2.1999,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4007363319396973,
+      "rewards/margins": 0.6775516867637634,
+      "rewards/rejected": -3.0782880783081055,
+      "step": 795
+    },
+    {
+      "epoch": 1.4450499545867395,
+      "grad_norm": 2.183232545852661,
+      "learning_rate": 7.392718061850407e-06,
+      "logits/chosen": 0.06714559346437454,
+      "logits/rejected": 0.046004436910152435,
+      "logps/chosen": -95.4114990234375,
+      "logps/rejected": -101.2634048461914,
+      "loss": 1.9802,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4638829231262207,
+      "rewards/margins": 0.7256942987442017,
+      "rewards/rejected": -3.189577579498291,
+      "step": 796
+    },
+    {
+      "epoch": 1.446866485013624,
+      "grad_norm": 2.2636830806732178,
+      "learning_rate": 7.3868861956471725e-06,
+      "logits/chosen": 0.025870440527796745,
+      "logits/rejected": 0.05159129202365875,
+      "logps/chosen": -80.84700775146484,
+      "logps/rejected": -88.84320831298828,
+      "loss": 1.7361,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.451986312866211,
+      "rewards/margins": 0.9137101173400879,
+      "rewards/rejected": -3.3656961917877197,
+      "step": 797
+    },
+    {
+      "epoch": 1.4486830154405086,
+      "grad_norm": 2.3395354747772217,
+      "learning_rate": 7.381050481754447e-06,
+      "logits/chosen": 0.06643021106719971,
+      "logits/rejected": 0.1107374057173729,
+      "logps/chosen": -88.24702453613281,
+      "logps/rejected": -97.00286865234375,
+      "loss": 2.3738,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.6666600704193115,
+      "rewards/margins": 0.5626463294029236,
+      "rewards/rejected": -3.22930645942688,
+      "step": 798
+    },
+    {
+      "epoch": 1.4504995458673933,
+      "grad_norm": 2.936711549758911,
+      "learning_rate": 7.375210932072265e-06,
+      "logits/chosen": 0.04520021751523018,
+      "logits/rejected": 0.06952523440122604,
+      "logps/chosen": -82.35380554199219,
+      "logps/rejected": -89.27215576171875,
+      "loss": 2.4222,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5209689140319824,
+      "rewards/margins": 0.5571924448013306,
+      "rewards/rejected": -3.0781612396240234,
+      "step": 799
+    },
+    {
+      "epoch": 1.452316076294278,
+      "grad_norm": 2.342952013015747,
+      "learning_rate": 7.36936755850849e-06,
+      "logits/chosen": 0.016199974343180656,
+      "logits/rejected": 0.07721687853336334,
+      "logps/chosen": -83.52679443359375,
+      "logps/rejected": -95.04873657226562,
+      "loss": 1.8844,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.487191677093506,
+      "rewards/margins": 0.9364355206489563,
+      "rewards/rejected": -3.4236273765563965,
+      "step": 800
+    },
+    {
+      "epoch": 1.4541326067211626,
+      "grad_norm": 2.611804246902466,
+      "learning_rate": 7.363520372978774e-06,
+      "logits/chosen": 0.0884896069765091,
+      "logits/rejected": 0.12684395909309387,
+      "logps/chosen": -73.71459197998047,
+      "logps/rejected": -78.7441635131836,
+      "loss": 2.5897,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.3413586616516113,
+      "rewards/margins": 0.3767651617527008,
+      "rewards/rejected": -2.7181236743927,
+      "step": 801
+    },
+    {
+      "epoch": 1.4559491371480473,
+      "grad_norm": 2.2446863651275635,
+      "learning_rate": 7.357669387406548e-06,
+      "logits/chosen": 0.04801754653453827,
+      "logits/rejected": 0.14114932715892792,
+      "logps/chosen": -72.67373657226562,
+      "logps/rejected": -88.53012084960938,
+      "loss": 1.8134,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4397833347320557,
+      "rewards/margins": 0.9794785976409912,
+      "rewards/rejected": -3.419261932373047,
+      "step": 802
+    },
+    {
+      "epoch": 1.457765667574932,
+      "grad_norm": 2.2070155143737793,
+      "learning_rate": 7.351814613722991e-06,
+      "logits/chosen": 0.09963800758123398,
+      "logits/rejected": 0.09044703096151352,
+      "logps/chosen": -93.37407684326172,
+      "logps/rejected": -101.1263656616211,
+      "loss": 2.0345,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4976069927215576,
+      "rewards/margins": 0.7321794033050537,
+      "rewards/rejected": -3.2297863960266113,
+      "step": 803
+    },
+    {
+      "epoch": 1.4595821980018164,
+      "grad_norm": 2.339240789413452,
+      "learning_rate": 7.3459560638670035e-06,
+      "logits/chosen": 0.06331755220890045,
+      "logits/rejected": 0.09320802986621857,
+      "logps/chosen": -84.83074951171875,
+      "logps/rejected": -90.84583282470703,
+      "loss": 1.7825,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.2360713481903076,
+      "rewards/margins": 0.8683581352233887,
+      "rewards/rejected": -3.104429244995117,
+      "step": 804
+    },
+    {
+      "epoch": 1.461398728428701,
+      "grad_norm": 2.2252860069274902,
+      "learning_rate": 7.3400937497851906e-06,
+      "logits/chosen": 0.06259419769048691,
+      "logits/rejected": 0.05668449029326439,
+      "logps/chosen": -76.97444915771484,
+      "logps/rejected": -95.10333251953125,
+      "loss": 1.9808,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.475806713104248,
+      "rewards/margins": 0.9982489943504333,
+      "rewards/rejected": -3.4740560054779053,
+      "step": 805
+    },
+    {
+      "epoch": 1.4632152588555858,
+      "grad_norm": 2.056358814239502,
+      "learning_rate": 7.334227683431832e-06,
+      "logits/chosen": 0.0795949399471283,
+      "logits/rejected": 0.11871220171451569,
+      "logps/chosen": -82.24653625488281,
+      "logps/rejected": -104.25836944580078,
+      "loss": 1.8753,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.397820234298706,
+      "rewards/margins": 1.0821467638015747,
+      "rewards/rejected": -3.479966878890991,
+      "step": 806
+    },
+    {
+      "epoch": 1.4650317892824705,
+      "grad_norm": 2.7462871074676514,
+      "learning_rate": 7.328357876768858e-06,
+      "logits/chosen": 0.08270560950040817,
+      "logits/rejected": 0.11420280486345291,
+      "logps/chosen": -76.36333465576172,
+      "logps/rejected": -86.0474853515625,
+      "loss": 2.3324,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.573794364929199,
+      "rewards/margins": 0.7034318447113037,
+      "rewards/rejected": -3.277225971221924,
+      "step": 807
+    },
+    {
+      "epoch": 1.4668483197093551,
+      "grad_norm": 2.5223159790039062,
+      "learning_rate": 7.322484341765824e-06,
+      "logits/chosen": 0.17336739599704742,
+      "logits/rejected": 0.15756890177726746,
+      "logps/chosen": -82.3572998046875,
+      "logps/rejected": -94.081787109375,
+      "loss": 2.0898,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3327016830444336,
+      "rewards/margins": 0.8286820650100708,
+      "rewards/rejected": -3.161383867263794,
+      "step": 808
+    },
+    {
+      "epoch": 1.4686648501362398,
+      "grad_norm": 2.4014031887054443,
+      "learning_rate": 7.316607090399894e-06,
+      "logits/chosen": 0.0774412751197815,
+      "logits/rejected": 0.016062507405877113,
+      "logps/chosen": -79.8453598022461,
+      "logps/rejected": -80.36003875732422,
+      "loss": 2.0743,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6673779487609863,
+      "rewards/margins": 0.6888259649276733,
+      "rewards/rejected": -3.356204032897949,
+      "step": 809
+    },
+    {
+      "epoch": 1.4704813805631245,
+      "grad_norm": 2.3294990062713623,
+      "learning_rate": 7.310726134655807e-06,
+      "logits/chosen": 0.03737390413880348,
+      "logits/rejected": 0.1001349613070488,
+      "logps/chosen": -81.04705810546875,
+      "logps/rejected": -96.8980712890625,
+      "loss": 1.7474,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.27017879486084,
+      "rewards/margins": 1.0432392358779907,
+      "rewards/rejected": -3.313417673110962,
+      "step": 810
+    },
+    {
+      "epoch": 1.472297910990009,
+      "grad_norm": 2.287733793258667,
+      "learning_rate": 7.30484148652585e-06,
+      "logits/chosen": 0.0469367690384388,
+      "logits/rejected": 0.06735274195671082,
+      "logps/chosen": -80.7060546875,
+      "logps/rejected": -93.6329345703125,
+      "loss": 2.0516,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.463186740875244,
+      "rewards/margins": 0.7543611526489258,
+      "rewards/rejected": -3.21754789352417,
+      "step": 811
+    },
+    {
+      "epoch": 1.4741144414168939,
+      "grad_norm": 3.1462066173553467,
+      "learning_rate": 7.298953158009852e-06,
+      "logits/chosen": 0.09930360317230225,
+      "logits/rejected": 0.13193194568157196,
+      "logps/chosen": -82.47021484375,
+      "logps/rejected": -94.59288024902344,
+      "loss": 1.9513,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.323047161102295,
+      "rewards/margins": 0.8602778911590576,
+      "rewards/rejected": -3.1833250522613525,
+      "step": 812
+    },
+    {
+      "epoch": 1.4759309718437783,
+      "grad_norm": 2.2134501934051514,
+      "learning_rate": 7.293061161115138e-06,
+      "logits/chosen": 0.04882989823818207,
+      "logits/rejected": 0.07393565028905869,
+      "logps/chosen": -69.4496841430664,
+      "logps/rejected": -79.9421157836914,
+      "loss": 1.9357,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.233206272125244,
+      "rewards/margins": 0.7889159321784973,
+      "rewards/rejected": -3.0221221446990967,
+      "step": 813
+    },
+    {
+      "epoch": 1.477747502270663,
+      "grad_norm": 2.3149659633636475,
+      "learning_rate": 7.287165507856513e-06,
+      "logits/chosen": 0.05155353993177414,
+      "logits/rejected": 0.036280907690525055,
+      "logps/chosen": -87.41694641113281,
+      "logps/rejected": -97.0438461303711,
+      "loss": 1.9108,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.27425217628479,
+      "rewards/margins": 0.8441011309623718,
+      "rewards/rejected": -3.1183536052703857,
+      "step": 814
+    },
+    {
+      "epoch": 1.4795640326975477,
+      "grad_norm": 2.4460105895996094,
+      "learning_rate": 7.281266210256244e-06,
+      "logits/chosen": 0.06415215134620667,
+      "logits/rejected": 0.08172871172428131,
+      "logps/chosen": -94.13727569580078,
+      "logps/rejected": -108.59651184082031,
+      "loss": 1.9764,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.476318120956421,
+      "rewards/margins": 0.8187392354011536,
+      "rewards/rejected": -3.2950570583343506,
+      "step": 815
+    },
+    {
+      "epoch": 1.4813805631244323,
+      "grad_norm": 2.5813546180725098,
+      "learning_rate": 7.275363280344023e-06,
+      "logits/chosen": 0.10900135338306427,
+      "logits/rejected": 0.06012497842311859,
+      "logps/chosen": -86.26004028320312,
+      "logps/rejected": -92.2495346069336,
+      "loss": 2.1791,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4611434936523438,
+      "rewards/margins": 0.6075427532196045,
+      "rewards/rejected": -3.0686862468719482,
+      "step": 816
+    },
+    {
+      "epoch": 1.483197093551317,
+      "grad_norm": 2.3455967903137207,
+      "learning_rate": 7.269456730156954e-06,
+      "logits/chosen": 0.08659522235393524,
+      "logits/rejected": 0.1016351729631424,
+      "logps/chosen": -75.75826263427734,
+      "logps/rejected": -85.91856384277344,
+      "loss": 2.0077,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4886691570281982,
+      "rewards/margins": 0.9004117846488953,
+      "rewards/rejected": -3.3890810012817383,
+      "step": 817
+    },
+    {
+      "epoch": 1.4850136239782017,
+      "grad_norm": 2.612858533859253,
+      "learning_rate": 7.26354657173952e-06,
+      "logits/chosen": 0.08592739701271057,
+      "logits/rejected": 0.15154355764389038,
+      "logps/chosen": -86.36177062988281,
+      "logps/rejected": -95.97001647949219,
+      "loss": 2.1468,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6705503463745117,
+      "rewards/margins": 0.7056388854980469,
+      "rewards/rejected": -3.376189708709717,
+      "step": 818
+    },
+    {
+      "epoch": 1.4868301544050864,
+      "grad_norm": 2.39699387550354,
+      "learning_rate": 7.257632817143566e-06,
+      "logits/chosen": 0.03663618862628937,
+      "logits/rejected": 0.06003079563379288,
+      "logps/chosen": -73.82808685302734,
+      "logps/rejected": -86.11414337158203,
+      "loss": 2.0372,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.251981258392334,
+      "rewards/margins": 0.8644815683364868,
+      "rewards/rejected": -3.1164627075195312,
+      "step": 819
+    },
+    {
+      "epoch": 1.4886466848319708,
+      "grad_norm": 2.1887619495391846,
+      "learning_rate": 7.251715478428263e-06,
+      "logits/chosen": -0.026794860139489174,
+      "logits/rejected": 0.11405564099550247,
+      "logps/chosen": -96.14276123046875,
+      "logps/rejected": -116.06658172607422,
+      "loss": 1.786,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.527817964553833,
+      "rewards/margins": 1.0386499166488647,
+      "rewards/rejected": -3.5664680004119873,
+      "step": 820
+    },
+    {
+      "epoch": 1.4904632152588557,
+      "grad_norm": 4.235897064208984,
+      "learning_rate": 7.2457945676600975e-06,
+      "logits/chosen": 0.10090241581201553,
+      "logits/rejected": 0.10542615503072739,
+      "logps/chosen": -78.00686645507812,
+      "logps/rejected": -86.51849365234375,
+      "loss": 2.1653,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6344261169433594,
+      "rewards/margins": 0.6512963771820068,
+      "rewards/rejected": -3.285722494125366,
+      "step": 821
+    },
+    {
+      "epoch": 1.4922797456857402,
+      "grad_norm": 2.189833402633667,
+      "learning_rate": 7.239870096912835e-06,
+      "logits/chosen": 0.0326794758439064,
+      "logits/rejected": 0.0896943062543869,
+      "logps/chosen": -89.783447265625,
+      "logps/rejected": -102.53228759765625,
+      "loss": 2.0422,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5607566833496094,
+      "rewards/margins": 1.009847640991211,
+      "rewards/rejected": -3.5706043243408203,
+      "step": 822
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.8665003260811674e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-822/training_args.bin b/checkpoint-822/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-822/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-822/zero_to_fp32.py b/checkpoint-822/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-822/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-84/README.md b/checkpoint-84/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-84/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-84/adapter_config.json b/checkpoint-84/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..154797ebe6197329495f8f587101cda9b9d46ce8
--- /dev/null
+++ b/checkpoint-84/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "o_proj",
+    "k_proj",
+    "up_proj",
+    "gate_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-84/adapter_model.safetensors b/checkpoint-84/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..ec1793e4177220cb6fc3396a023e5d5827d31b36
--- /dev/null
+++ b/checkpoint-84/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55162de64ad006819ed48c4133cb0d41cd73c8588ff99e0437eeb5cd44b911e7
+size 828527688
diff --git a/checkpoint-84/latest b/checkpoint-84/latest
new file mode 100644
index 0000000000000000000000000000000000000000..a0fad656e1c5da45033f74e6726c7ea1b32841a7
--- /dev/null
+++ b/checkpoint-84/latest
@@ -0,0 +1 @@
+global_step84
\ No newline at end of file
diff --git a/checkpoint-84/rng_state_0.pth b/checkpoint-84/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..35252259eb09d8de259231f63f19e786e44bc7b0
--- /dev/null
+++ b/checkpoint-84/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08282b46825aa78d10fe10e3fea89555c5b5a691b261a3ddfd58fcb58370edff
+size 15984
diff --git a/checkpoint-84/rng_state_1.pth b/checkpoint-84/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..eeb7d8df6ed170dd98dba8737bc9dd038af61afd
--- /dev/null
+++ b/checkpoint-84/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbab71d98a3a9a92df82a6bba463947327c3a1bcf35cd9f4f46114641fc42dd9
+size 15984
diff --git a/checkpoint-84/rng_state_2.pth b/checkpoint-84/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..e144a445ffd57fbb5be9b5131f17149bde6c4ff5
--- /dev/null
+++ b/checkpoint-84/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:caac82d57d878d30219a4f9ec289a97ff90c53afc160b968f251b3fd3454b8d8
+size 15984
diff --git a/checkpoint-84/rng_state_3.pth b/checkpoint-84/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a10f35268ac2a0cb68abc0b78ba5b150b0f29d78
--- /dev/null
+++ b/checkpoint-84/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19762d2d370222b01817da11bbaa6665d542293373186d66f754e7246bb861ed
+size 15984
diff --git a/checkpoint-84/rng_state_4.pth b/checkpoint-84/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..e6ef21562e384e0889ec2400e8f84b6b0bc59035
--- /dev/null
+++ b/checkpoint-84/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00c7508b346a7d3c5c23392845f1d013331114ade778794b76e919cb3ed5d33e
+size 15984
diff --git a/checkpoint-84/rng_state_5.pth b/checkpoint-84/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..126662e5f97dd3f9cd7fb87e70843d3d5532dde3
--- /dev/null
+++ b/checkpoint-84/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b89de7d14dd20a191f56b74c816ef8b7fe5c171e31efbeadbf321c4539ed68c3
+size 15984
diff --git a/checkpoint-84/rng_state_6.pth b/checkpoint-84/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d4e6b27cc99b0fa8e6bbf967892f9304b444d81d
--- /dev/null
+++ b/checkpoint-84/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c71152053553e6e22d670fbc4fd7550bf8a046b54cad7b71869787986a6a42c
+size 15984
diff --git a/checkpoint-84/rng_state_7.pth b/checkpoint-84/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..14e214a61e5311f2b7edf2200ec0365ed2dcc5e1
--- /dev/null
+++ b/checkpoint-84/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b67db12a26a26ffe03d9afc84a43857eb2e5b2fec2dd189653b415f74208190
+size 15984
diff --git a/checkpoint-84/special_tokens_map.json b/checkpoint-84/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-84/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-84/tokenizer.json b/checkpoint-84/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-84/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-84/tokenizer_config.json b/checkpoint-84/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e1adcc0c23bace7262ee80c9597c65c5cf6eb99b
--- /dev/null
+++ b/checkpoint-84/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2048,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-84/trainer_state.json b/checkpoint-84/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..acb1a4a76a9aa644c03450dc228b70fbf1cba7c0
--- /dev/null
+++ b/checkpoint-84/trainer_state.json
@@ -0,0 +1,153 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 84,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 16.365549600718158,
+      "learning_rate": 9e-06,
+      "logits/chosen": 0.43800339102745056,
+      "logits/rejected": 0.4281293749809265,
+      "logps/chosen": -324.7308349609375,
+      "logps/rejected": -337.3280029296875,
+      "loss": 0.6876,
+      "rewards/accuracies": 0.4374999701976776,
+      "rewards/chosen": 0.009319745004177094,
+      "rewards/margins": 0.012683698907494545,
+      "rewards/rejected": -0.0033639525063335896,
+      "step": 10
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 6.549199651973184,
+      "learning_rate": 8.64047538607195e-06,
+      "logits/chosen": 0.4200649559497833,
+      "logits/rejected": 0.41489219665527344,
+      "logps/chosen": -321.9690246582031,
+      "logps/rejected": -355.4017639160156,
+      "loss": 0.6962,
+      "rewards/accuracies": 0.48125001788139343,
+      "rewards/chosen": 0.034620046615600586,
+      "rewards/margins": -0.00358342076651752,
+      "rewards/rejected": 0.038203466683626175,
+      "step": 20
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 7.086531400587909,
+      "learning_rate": 7.62573262972983e-06,
+      "logits/chosen": 0.4370676875114441,
+      "logits/rejected": 0.4306566119194031,
+      "logps/chosen": -315.1405944824219,
+      "logps/rejected": -343.5477600097656,
+      "loss": 0.6646,
+      "rewards/accuracies": 0.6187500357627869,
+      "rewards/chosen": 0.1420479416847229,
+      "rewards/margins": 0.06437937170267105,
+      "rewards/rejected": 0.07766857743263245,
+      "step": 30
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 5.431551347783104,
+      "learning_rate": 6.135932223670029e-06,
+      "logits/chosen": 0.4090261161327362,
+      "logits/rejected": 0.40613797307014465,
+      "logps/chosen": -317.56817626953125,
+      "logps/rejected": -354.5175476074219,
+      "loss": 0.5202,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.44594669342041016,
+      "rewards/margins": 0.39714840054512024,
+      "rewards/rejected": 0.04879830405116081,
+      "step": 40
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 5.871421809973983,
+      "learning_rate": 4.4355778300251406e-06,
+      "logits/chosen": 0.4284099042415619,
+      "logits/rejected": 0.4222380816936493,
+      "logps/chosen": -319.2432861328125,
+      "logps/rejected": -343.3741760253906,
+      "loss": 0.4481,
+      "rewards/accuracies": 0.9124999642372131,
+      "rewards/chosen": 0.6467863321304321,
+      "rewards/margins": 0.6407225131988525,
+      "rewards/rejected": 0.0060638319700956345,
+      "step": 50
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 5.194288258129358,
+      "learning_rate": 2.8265555015932123e-06,
+      "logits/chosen": 0.4114301800727844,
+      "logits/rejected": 0.4095650017261505,
+      "logps/chosen": -312.0003662109375,
+      "logps/rejected": -348.30206298828125,
+      "loss": 0.3622,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.7869285345077515,
+      "rewards/margins": 0.9796440601348877,
+      "rewards/rejected": -0.19271555542945862,
+      "step": 60
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 3.569068921757812,
+      "learning_rate": 1.5945359205624398e-06,
+      "logits/chosen": 0.4047119617462158,
+      "logits/rejected": 0.4038873612880707,
+      "logps/chosen": -307.7536926269531,
+      "logps/rejected": -354.0033264160156,
+      "loss": 0.2611,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0587913990020752,
+      "rewards/margins": 1.4121346473693848,
+      "rewards/rejected": -0.35334333777427673,
+      "step": 70
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 4.489026775368698,
+      "learning_rate": 9.58255563091781e-07,
+      "logits/chosen": 0.3864372968673706,
+      "logits/rejected": 0.3842785954475403,
+      "logps/chosen": -312.03338623046875,
+      "logps/rejected": -350.1846923828125,
+      "loss": 0.2398,
+      "rewards/accuracies": 0.9937499761581421,
+      "rewards/chosen": 1.117773413658142,
+      "rewards/margins": 1.5565866231918335,
+      "rewards/rejected": -0.43881309032440186,
+      "step": 80
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 84,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 115,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 216736934658048.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-84/training_args.bin b/checkpoint-84/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..689f92ab91fc248999cca8078aa5f2d327b997ef
--- /dev/null
+++ b/checkpoint-84/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:540287b777c24cb36855ecf7a73f9ff8f66586431561f6b8a7857beb5a25ff53
+size 7544
diff --git a/checkpoint-84/zero_to_fp32.py b/checkpoint-84/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-84/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-959/README.md b/checkpoint-959/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..0f9aab4a1929dc2edc85b114897a9aacc5f5896a
--- /dev/null
+++ b/checkpoint-959/README.md
@@ -0,0 +1,202 @@
+---
+base_model: /raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.12.0
\ No newline at end of file
diff --git a/checkpoint-959/adapter_config.json b/checkpoint-959/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..994c81ac161b8498397cc448c5fb937b736e979a
--- /dev/null
+++ b/checkpoint-959/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/raid/HUB_LLM/080225_vi_test_llama33_70b_instruct/checkpoint-2568/SLERP-1.5/",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/checkpoint-959/adapter_model.safetensors b/checkpoint-959/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..d8b5f236f74df29c001f12deadf73a51ab0030b2
--- /dev/null
+++ b/checkpoint-959/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35f977ad492599e0c3e3391ba8cf8620d32356ee42e64915ff03d2896df68eb9
+size 207244392
diff --git a/checkpoint-959/latest b/checkpoint-959/latest
new file mode 100644
index 0000000000000000000000000000000000000000..3431edb873f740b3bd0e81441fe13a02c05fe4b0
--- /dev/null
+++ b/checkpoint-959/latest
@@ -0,0 +1 @@
+global_step958
\ No newline at end of file
diff --git a/checkpoint-959/rng_state_0.pth b/checkpoint-959/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b346349ce12dd5a17d4b91ed2a5722bb52550950
--- /dev/null
+++ b/checkpoint-959/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
+size 15984
diff --git a/checkpoint-959/rng_state_1.pth b/checkpoint-959/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..68f3c6994456cb8d0592a5375d99503c8924b1c4
--- /dev/null
+++ b/checkpoint-959/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73
+size 15984
diff --git a/checkpoint-959/rng_state_2.pth b/checkpoint-959/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..be044f6ceeed587d30e80c2f72d5aa19fdc9947b
--- /dev/null
+++ b/checkpoint-959/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b
+size 15984
diff --git a/checkpoint-959/rng_state_3.pth b/checkpoint-959/rng_state_3.pth
new file mode 100644
index 0000000000000000000000000000000000000000..fc825249656a9b858782542bd3f4386250f1dfe0
--- /dev/null
+++ b/checkpoint-959/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc
+size 15984
diff --git a/checkpoint-959/rng_state_4.pth b/checkpoint-959/rng_state_4.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d30f52a44be563c152ae09db6ae934da6da0d3ed
--- /dev/null
+++ b/checkpoint-959/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972
+size 15984
diff --git a/checkpoint-959/rng_state_5.pth b/checkpoint-959/rng_state_5.pth
new file mode 100644
index 0000000000000000000000000000000000000000..c8715d27ab23ae545d58039cf949cc44ecc1da5e
--- /dev/null
+++ b/checkpoint-959/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991
+size 15984
diff --git a/checkpoint-959/rng_state_6.pth b/checkpoint-959/rng_state_6.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1ed791b6ef76eadf0b0c55a5733411771e2ae027
--- /dev/null
+++ b/checkpoint-959/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa
+size 15984
diff --git a/checkpoint-959/rng_state_7.pth b/checkpoint-959/rng_state_7.pth
new file mode 100644
index 0000000000000000000000000000000000000000..800c3bbbc5edf7db01a8316069d439c5fb8d8c30
--- /dev/null
+++ b/checkpoint-959/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773
+size 15984
diff --git a/checkpoint-959/special_tokens_map.json b/checkpoint-959/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/checkpoint-959/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-959/tokenizer.json b/checkpoint-959/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/checkpoint-959/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-959/tokenizer_config.json b/checkpoint-959/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..37a03f8e380245ea011851fe292f21ccd12ec3ee
--- /dev/null
+++ b/checkpoint-959/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2500,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoint-959/trainer_state.json b/checkpoint-959/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d3613bc787e669054b6b3538c63acd3cade7ec83
--- /dev/null
+++ b/checkpoint-959/trainer_state.json
@@ -0,0 +1,14418 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.7411444141689372,
+  "eval_steps": 500,
+  "global_step": 959,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018165304268846503,
+      "grad_norm": 0.7654335498809814,
+      "learning_rate": 9.99999541185824e-06,
+      "logits/chosen": 0.10746052116155624,
+      "logits/rejected": 0.16293610632419586,
+      "logps/chosen": -51.70099639892578,
+      "logps/rejected": -56.85424041748047,
+      "loss": 2.7726,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036330608537693005,
+      "grad_norm": 0.8493374586105347,
+      "learning_rate": 9.999981647442309e-06,
+      "logits/chosen": 0.12589210271835327,
+      "logits/rejected": 0.18243370950222015,
+      "logps/chosen": -61.092987060546875,
+      "logps/rejected": -67.26304626464844,
+      "loss": 2.7516,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": 0.005276113282889128,
+      "rewards/margins": 0.010965153574943542,
+      "rewards/rejected": -0.00568903936073184,
+      "step": 2
+    },
+    {
+      "epoch": 0.005449591280653951,
+      "grad_norm": 0.739442765712738,
+      "learning_rate": 9.999958706780277e-06,
+      "logits/chosen": 0.0766456350684166,
+      "logits/rejected": 0.1422545611858368,
+      "logps/chosen": -55.764503479003906,
+      "logps/rejected": -57.38707733154297,
+      "loss": 2.7902,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0027439936529845,
+      "rewards/margins": -0.008403200656175613,
+      "rewards/rejected": 0.005659207701683044,
+      "step": 3
+    },
+    {
+      "epoch": 0.007266121707538601,
+      "grad_norm": 0.7261953949928284,
+      "learning_rate": 9.999926589918927e-06,
+      "logits/chosen": 0.1812177300453186,
+      "logits/rejected": 0.13825736939907074,
+      "logps/chosen": -56.98471450805664,
+      "logps/rejected": -55.47361373901367,
+      "loss": 2.7537,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": 0.006436044350266457,
+      "rewards/margins": 0.009899044409394264,
+      "rewards/rejected": -0.003463000524789095,
+      "step": 4
+    },
+    {
+      "epoch": 0.009082652134423252,
+      "grad_norm": 0.850686252117157,
+      "learning_rate": 9.999885296923748e-06,
+      "logits/chosen": 0.17714114487171173,
+      "logits/rejected": 0.12910494208335876,
+      "logps/chosen": -71.72655487060547,
+      "logps/rejected": -70.75139617919922,
+      "loss": 2.79,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.008413795381784439,
+      "rewards/margins": -0.007964953780174255,
+      "rewards/rejected": -0.0004488405538722873,
+      "step": 5
+    },
+    {
+      "epoch": 0.010899182561307902,
+      "grad_norm": 0.6627741456031799,
+      "learning_rate": 9.999834827878942e-06,
+      "logits/chosen": 0.13277305662631989,
+      "logits/rejected": 0.20749863982200623,
+      "logps/chosen": -47.56520080566406,
+      "logps/rejected": -53.18217468261719,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001662571681663394,
+      "rewards/margins": 0.00014669005759060383,
+      "rewards/rejected": 0.0015158820897340775,
+      "step": 6
+    },
+    {
+      "epoch": 0.012715712988192553,
+      "grad_norm": 0.8335347175598145,
+      "learning_rate": 9.999775182887429e-06,
+      "logits/chosen": 0.07216031849384308,
+      "logits/rejected": 0.13223227858543396,
+      "logps/chosen": -73.06997680664062,
+      "logps/rejected": -75.58594512939453,
+      "loss": 2.7716,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.004659444559365511,
+      "rewards/margins": 0.0013355333358049393,
+      "rewards/rejected": -0.005994977429509163,
+      "step": 7
+    },
+    {
+      "epoch": 0.014532243415077202,
+      "grad_norm": 0.7128610014915466,
+      "learning_rate": 9.999706362070833e-06,
+      "logits/chosen": 0.11368857324123383,
+      "logits/rejected": 0.15805509686470032,
+      "logps/chosen": -56.45945358276367,
+      "logps/rejected": -56.9697151184082,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.0038012717850506306,
+      "rewards/margins": 0.009716177359223366,
+      "rewards/rejected": -0.013517449609935284,
+      "step": 8
+    },
+    {
+      "epoch": 0.01634877384196185,
+      "grad_norm": 0.77040034532547,
+      "learning_rate": 9.99962836556949e-06,
+      "logits/chosen": 0.1818978637456894,
+      "logits/rejected": 0.20750640332698822,
+      "logps/chosen": -59.184532165527344,
+      "logps/rejected": -61.59263610839844,
+      "loss": 2.7688,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.009461460635066032,
+      "rewards/margins": 0.002394758863374591,
+      "rewards/rejected": -0.011856218799948692,
+      "step": 9
+    },
+    {
+      "epoch": 0.018165304268846504,
+      "grad_norm": 0.7393640875816345,
+      "learning_rate": 9.99954119354245e-06,
+      "logits/chosen": 0.14076584577560425,
+      "logits/rejected": 0.22711633145809174,
+      "logps/chosen": -56.34330749511719,
+      "logps/rejected": -60.51960372924805,
+      "loss": 2.7827,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.004167273174971342,
+      "rewards/margins": -0.00429012393578887,
+      "rewards/rejected": 0.00012284982949495316,
+      "step": 10
+    },
+    {
+      "epoch": 0.019981834695731154,
+      "grad_norm": 0.7271163463592529,
+      "learning_rate": 9.999444846167473e-06,
+      "logits/chosen": 0.20141208171844482,
+      "logits/rejected": 0.2038215547800064,
+      "logps/chosen": -57.712646484375,
+      "logps/rejected": -54.343231201171875,
+      "loss": 2.7565,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0009536671568639576,
+      "rewards/margins": 0.00866839662194252,
+      "rewards/rejected": -0.007714730221778154,
+      "step": 11
+    },
+    {
+      "epoch": 0.021798365122615803,
+      "grad_norm": 0.9451335668563843,
+      "learning_rate": 9.999339323641027e-06,
+      "logits/chosen": 0.16046015918254852,
+      "logits/rejected": 0.1963367462158203,
+      "logps/chosen": -51.579856872558594,
+      "logps/rejected": -54.97206115722656,
+      "loss": 2.7732,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.005096627864986658,
+      "rewards/margins": 0.00017735245637595654,
+      "rewards/rejected": -0.005273980088531971,
+      "step": 12
+    },
+    {
+      "epoch": 0.023614895549500452,
+      "grad_norm": 0.8279874324798584,
+      "learning_rate": 9.99922462617829e-06,
+      "logits/chosen": 0.12593263387680054,
+      "logits/rejected": 0.11708654463291168,
+      "logps/chosen": -62.25929641723633,
+      "logps/rejected": -60.633296966552734,
+      "loss": 2.7693,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.013154934160411358,
+      "rewards/margins": 0.002298446139320731,
+      "rewards/rejected": -0.015453380532562733,
+      "step": 13
+    },
+    {
+      "epoch": 0.025431425976385105,
+      "grad_norm": 0.7863137125968933,
+      "learning_rate": 9.999100754013152e-06,
+      "logits/chosen": 0.12714676558971405,
+      "logits/rejected": 0.18583568930625916,
+      "logps/chosen": -59.119529724121094,
+      "logps/rejected": -65.85352325439453,
+      "loss": 2.7715,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.01445689145475626,
+      "rewards/margins": 0.001161783467978239,
+      "rewards/rejected": -0.015618674457073212,
+      "step": 14
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.7623695731163025,
+      "learning_rate": 9.998967707398207e-06,
+      "logits/chosen": 0.09560943394899368,
+      "logits/rejected": 0.14794519543647766,
+      "logps/chosen": -47.649497985839844,
+      "logps/rejected": -57.21004867553711,
+      "loss": 2.7405,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.0010593307670205832,
+      "rewards/margins": 0.0166107639670372,
+      "rewards/rejected": -0.017670094966888428,
+      "step": 15
+    },
+    {
+      "epoch": 0.029064486830154404,
+      "grad_norm": 0.7020726203918457,
+      "learning_rate": 9.998825486604765e-06,
+      "logits/chosen": 0.19466270506381989,
+      "logits/rejected": 0.21006342768669128,
+      "logps/chosen": -50.8302116394043,
+      "logps/rejected": -58.202754974365234,
+      "loss": 2.7628,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -0.009497416205704212,
+      "rewards/margins": 0.005540419369935989,
+      "rewards/rejected": -0.015037836506962776,
+      "step": 16
+    },
+    {
+      "epoch": 0.030881017257039057,
+      "grad_norm": 0.7085703611373901,
+      "learning_rate": 9.998674091922833e-06,
+      "logits/chosen": 0.06517614424228668,
+      "logits/rejected": 0.15166127681732178,
+      "logps/chosen": -45.203216552734375,
+      "logps/rejected": -51.454952239990234,
+      "loss": 2.7568,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.009879402816295624,
+      "rewards/margins": 0.00827767699956894,
+      "rewards/rejected": -0.018157079815864563,
+      "step": 17
+    },
+    {
+      "epoch": 0.0326975476839237,
+      "grad_norm": 0.8115389347076416,
+      "learning_rate": 9.998513523661136e-06,
+      "logits/chosen": 0.09045977890491486,
+      "logits/rejected": 0.1498938351869583,
+      "logps/chosen": -60.35250473022461,
+      "logps/rejected": -62.11174774169922,
+      "loss": 2.7888,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.01872912421822548,
+      "rewards/margins": -0.0073302448727190495,
+      "rewards/rejected": -0.011398878879845142,
+      "step": 18
+    },
+    {
+      "epoch": 0.03451407811080836,
+      "grad_norm": 0.8222002983093262,
+      "learning_rate": 9.998343782147098e-06,
+      "logits/chosen": 0.08091418445110321,
+      "logits/rejected": 0.1225643903017044,
+      "logps/chosen": -54.25471115112305,
+      "logps/rejected": -57.91140365600586,
+      "loss": 2.7652,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.018515082076191902,
+      "rewards/margins": 0.004185312893241644,
+      "rewards/rejected": -0.022700395435094833,
+      "step": 19
+    },
+    {
+      "epoch": 0.03633060853769301,
+      "grad_norm": 0.862041711807251,
+      "learning_rate": 9.998164867726853e-06,
+      "logits/chosen": 0.04738985374569893,
+      "logits/rejected": 0.11436720937490463,
+      "logps/chosen": -62.22935104370117,
+      "logps/rejected": -67.67117309570312,
+      "loss": 2.7607,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.01734255626797676,
+      "rewards/margins": 0.006629224866628647,
+      "rewards/rejected": -0.023971781134605408,
+      "step": 20
+    },
+    {
+      "epoch": 0.03814713896457766,
+      "grad_norm": 0.8169613480567932,
+      "learning_rate": 9.997976780765237e-06,
+      "logits/chosen": 0.14641642570495605,
+      "logits/rejected": 0.17384907603263855,
+      "logps/chosen": -61.21760559082031,
+      "logps/rejected": -61.094486236572266,
+      "loss": 2.7752,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": -0.021789425984025,
+      "rewards/margins": -0.000787546974606812,
+      "rewards/rejected": -0.02100187912583351,
+      "step": 21
+    },
+    {
+      "epoch": 0.03996366939146231,
+      "grad_norm": 0.7457932829856873,
+      "learning_rate": 9.997779521645793e-06,
+      "logits/chosen": 0.16773203015327454,
+      "logits/rejected": 0.16435235738754272,
+      "logps/chosen": -55.192623138427734,
+      "logps/rejected": -53.05280303955078,
+      "loss": 2.7661,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.021701142191886902,
+      "rewards/margins": 0.0036970973014831543,
+      "rewards/rejected": -0.025398239493370056,
+      "step": 22
+    },
+    {
+      "epoch": 0.04178019981834696,
+      "grad_norm": 0.8263576030731201,
+      "learning_rate": 9.997573090770766e-06,
+      "logits/chosen": 0.10863066464662552,
+      "logits/rejected": 0.1277829110622406,
+      "logps/chosen": -61.80084228515625,
+      "logps/rejected": -64.94878387451172,
+      "loss": 2.7526,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.019380860030651093,
+      "rewards/margins": 0.010501865297555923,
+      "rewards/rejected": -0.029882723465561867,
+      "step": 23
+    },
+    {
+      "epoch": 0.043596730245231606,
+      "grad_norm": 0.7587623000144958,
+      "learning_rate": 9.997357488561107e-06,
+      "logits/chosen": 0.17992480099201202,
+      "logits/rejected": 0.16478824615478516,
+      "logps/chosen": -62.946449279785156,
+      "logps/rejected": -61.98965072631836,
+      "loss": 2.7534,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.02564322203397751,
+      "rewards/margins": 0.010079940780997276,
+      "rewards/rejected": -0.035723160952329636,
+      "step": 24
+    },
+    {
+      "epoch": 0.045413260672116255,
+      "grad_norm": 0.7733523845672607,
+      "learning_rate": 9.997132715456464e-06,
+      "logits/chosen": 0.1268949955701828,
+      "logits/rejected": 0.14968058466911316,
+      "logps/chosen": -53.57306671142578,
+      "logps/rejected": -58.07795715332031,
+      "loss": 2.7602,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.030423898249864578,
+      "rewards/margins": 0.006796791218221188,
+      "rewards/rejected": -0.03722068667411804,
+      "step": 25
+    },
+    {
+      "epoch": 0.047229791099000905,
+      "grad_norm": 0.8851524591445923,
+      "learning_rate": 9.99689877191519e-06,
+      "logits/chosen": 0.17443379759788513,
+      "logits/rejected": 0.1787741333246231,
+      "logps/chosen": -58.031558990478516,
+      "logps/rejected": -59.206520080566406,
+      "loss": 2.762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.02750026062130928,
+      "rewards/margins": 0.006129855290055275,
+      "rewards/rejected": -0.033630117774009705,
+      "step": 26
+    },
+    {
+      "epoch": 0.04904632152588556,
+      "grad_norm": 0.8524475693702698,
+      "learning_rate": 9.996655658414331e-06,
+      "logits/chosen": 0.18196682631969452,
+      "logits/rejected": 0.237998366355896,
+      "logps/chosen": -63.18816375732422,
+      "logps/rejected": -67.31481170654297,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.03693533316254616,
+      "rewards/margins": 0.011266030371189117,
+      "rewards/rejected": -0.048201363533735275,
+      "step": 27
+    },
+    {
+      "epoch": 0.05086285195277021,
+      "grad_norm": 0.8785350918769836,
+      "learning_rate": 9.996403375449647e-06,
+      "logits/chosen": 0.12056512385606766,
+      "logits/rejected": 0.17535904049873352,
+      "logps/chosen": -62.035221099853516,
+      "logps/rejected": -69.88546752929688,
+      "loss": 2.743,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.03874595835804939,
+      "rewards/margins": 0.015791552141308784,
+      "rewards/rejected": -0.054537512362003326,
+      "step": 28
+    },
+    {
+      "epoch": 0.05267938237965486,
+      "grad_norm": 0.9492508172988892,
+      "learning_rate": 9.996141923535582e-06,
+      "logits/chosen": 0.12200314551591873,
+      "logits/rejected": 0.13907021284103394,
+      "logps/chosen": -65.31945037841797,
+      "logps/rejected": -66.59127807617188,
+      "loss": 2.7641,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.0383562371134758,
+      "rewards/margins": 0.005150892771780491,
+      "rewards/rejected": -0.043507132679224014,
+      "step": 29
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.8084173202514648,
+      "learning_rate": 9.995871303205282e-06,
+      "logits/chosen": 0.07616369426250458,
+      "logits/rejected": 0.2001558393239975,
+      "logps/chosen": -51.103309631347656,
+      "logps/rejected": -62.28483581542969,
+      "loss": 2.7519,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.04615113511681557,
+      "rewards/margins": 0.011578184552490711,
+      "rewards/rejected": -0.057729318737983704,
+      "step": 30
+    },
+    {
+      "epoch": 0.05631244323342416,
+      "grad_norm": 0.943800151348114,
+      "learning_rate": 9.995591515010589e-06,
+      "logits/chosen": 0.04337490350008011,
+      "logits/rejected": 0.15334263443946838,
+      "logps/chosen": -57.265567779541016,
+      "logps/rejected": -68.5506820678711,
+      "loss": 2.7253,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.04982424154877663,
+      "rewards/margins": 0.024712499231100082,
+      "rewards/rejected": -0.07453674077987671,
+      "step": 31
+    },
+    {
+      "epoch": 0.05812897366030881,
+      "grad_norm": 0.9372840523719788,
+      "learning_rate": 9.99530255952204e-06,
+      "logits/chosen": 0.13595089316368103,
+      "logits/rejected": 0.14287832379341125,
+      "logps/chosen": -61.515377044677734,
+      "logps/rejected": -63.1092529296875,
+      "loss": 2.7571,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.05759892612695694,
+      "rewards/margins": 0.009150844067335129,
+      "rewards/rejected": -0.06674977391958237,
+      "step": 32
+    },
+    {
+      "epoch": 0.05994550408719346,
+      "grad_norm": 0.8408157825469971,
+      "learning_rate": 9.995004437328866e-06,
+      "logits/chosen": 0.18553151190280914,
+      "logits/rejected": 0.14024314284324646,
+      "logps/chosen": -61.73942565917969,
+      "logps/rejected": -53.5772819519043,
+      "loss": 2.7627,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06591347604990005,
+      "rewards/margins": 0.006169732194393873,
+      "rewards/rejected": -0.07208321243524551,
+      "step": 33
+    },
+    {
+      "epoch": 0.061762034514078114,
+      "grad_norm": 0.8856032490730286,
+      "learning_rate": 9.994697149038989e-06,
+      "logits/chosen": 0.08410881459712982,
+      "logits/rejected": 0.13991469144821167,
+      "logps/chosen": -60.161231994628906,
+      "logps/rejected": -60.03232192993164,
+      "loss": 2.7437,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.06631935387849808,
+      "rewards/margins": 0.016240764409303665,
+      "rewards/rejected": -0.08256012946367264,
+      "step": 34
+    },
+    {
+      "epoch": 0.06357856494096276,
+      "grad_norm": 0.9037399291992188,
+      "learning_rate": 9.994380695279025e-06,
+      "logits/chosen": 0.06669525057077408,
+      "logits/rejected": 0.16171438992023468,
+      "logps/chosen": -59.874202728271484,
+      "logps/rejected": -67.9597396850586,
+      "loss": 2.7221,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.06999661773443222,
+      "rewards/margins": 0.027104372158646584,
+      "rewards/rejected": -0.09710099548101425,
+      "step": 35
+    },
+    {
+      "epoch": 0.0653950953678474,
+      "grad_norm": 0.8819396495819092,
+      "learning_rate": 9.994055076694276e-06,
+      "logits/chosen": 0.18232542276382446,
+      "logits/rejected": 0.2102084755897522,
+      "logps/chosen": -59.298095703125,
+      "logps/rejected": -63.70309066772461,
+      "loss": 2.7179,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.08072888106107712,
+      "rewards/margins": 0.029247857630252838,
+      "rewards/rejected": -0.10997673869132996,
+      "step": 36
+    },
+    {
+      "epoch": 0.06721162579473206,
+      "grad_norm": 0.9046475291252136,
+      "learning_rate": 9.993720293948739e-06,
+      "logits/chosen": 0.13181668519973755,
+      "logits/rejected": 0.1316901445388794,
+      "logps/chosen": -71.46318054199219,
+      "logps/rejected": -76.40208435058594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.08365033566951752,
+      "rewards/margins": 0.01496690884232521,
+      "rewards/rejected": -0.09861725568771362,
+      "step": 37
+    },
+    {
+      "epoch": 0.06902815622161672,
+      "grad_norm": 0.9267721772193909,
+      "learning_rate": 9.993376347725091e-06,
+      "logits/chosen": 0.14555476605892181,
+      "logits/rejected": 0.21660488843917847,
+      "logps/chosen": -61.28938293457031,
+      "logps/rejected": -67.28329467773438,
+      "loss": 2.7007,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.07440298050642014,
+      "rewards/margins": 0.03913095220923424,
+      "rewards/rejected": -0.11353392899036407,
+      "step": 38
+    },
+    {
+      "epoch": 0.07084468664850137,
+      "grad_norm": 0.8545824885368347,
+      "learning_rate": 9.993023238724696e-06,
+      "logits/chosen": 0.09474823623895645,
+      "logits/rejected": 0.09304340183734894,
+      "logps/chosen": -61.27484130859375,
+      "logps/rejected": -62.003597259521484,
+      "loss": 2.7406,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.09001085162162781,
+      "rewards/margins": 0.017711879685521126,
+      "rewards/rejected": -0.10772272944450378,
+      "step": 39
+    },
+    {
+      "epoch": 0.07266121707538602,
+      "grad_norm": 0.9086014628410339,
+      "learning_rate": 9.99266096766761e-06,
+      "logits/chosen": 0.17910577356815338,
+      "logits/rejected": 0.14881345629692078,
+      "logps/chosen": -59.22669219970703,
+      "logps/rejected": -55.103187561035156,
+      "loss": 2.7422,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.1086670309305191,
+      "rewards/margins": 0.01710200309753418,
+      "rewards/rejected": -0.12576903402805328,
+      "step": 40
+    },
+    {
+      "epoch": 0.07447774750227067,
+      "grad_norm": 0.9802166223526001,
+      "learning_rate": 9.992289535292565e-06,
+      "logits/chosen": 0.18225271999835968,
+      "logits/rejected": 0.20810523629188538,
+      "logps/chosen": -55.57660675048828,
+      "logps/rejected": -59.129417419433594,
+      "loss": 2.7505,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.13455544412136078,
+      "rewards/margins": 0.012644169852137566,
+      "rewards/rejected": -0.1471996009349823,
+      "step": 41
+    },
+    {
+      "epoch": 0.07629427792915532,
+      "grad_norm": 0.919211208820343,
+      "learning_rate": 9.991908942356977e-06,
+      "logits/chosen": 0.07866425067186356,
+      "logits/rejected": 0.07479682564735413,
+      "logps/chosen": -66.11737060546875,
+      "logps/rejected": -64.02122497558594,
+      "loss": 2.7855,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.15568655729293823,
+      "rewards/margins": -0.0022164471447467804,
+      "rewards/rejected": -0.15347009897232056,
+      "step": 42
+    },
+    {
+      "epoch": 0.07811080835603997,
+      "grad_norm": 0.9612410068511963,
+      "learning_rate": 9.991519189636937e-06,
+      "logits/chosen": 0.11137107014656067,
+      "logits/rejected": 0.14462286233901978,
+      "logps/chosen": -67.43673706054688,
+      "logps/rejected": -74.29210662841797,
+      "loss": 2.7203,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.16468609869480133,
+      "rewards/margins": 0.032481495290994644,
+      "rewards/rejected": -0.19716759026050568,
+      "step": 43
+    },
+    {
+      "epoch": 0.07992733878292461,
+      "grad_norm": 0.8044790625572205,
+      "learning_rate": 9.991120277927224e-06,
+      "logits/chosen": 0.08509679138660431,
+      "logits/rejected": 0.11362393945455551,
+      "logps/chosen": -56.867591857910156,
+      "logps/rejected": -55.66028594970703,
+      "loss": 2.7302,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.15751589834690094,
+      "rewards/margins": 0.02398722618818283,
+      "rewards/rejected": -0.18150311708450317,
+      "step": 44
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.8599545955657959,
+      "learning_rate": 9.990712208041284e-06,
+      "logits/chosen": 0.1391746699810028,
+      "logits/rejected": 0.1746281236410141,
+      "logps/chosen": -62.22108459472656,
+      "logps/rejected": -58.50814437866211,
+      "loss": 2.7288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.16528920829296112,
+      "rewards/margins": 0.02449135296046734,
+      "rewards/rejected": -0.1897805631160736,
+      "step": 45
+    },
+    {
+      "epoch": 0.08356039963669391,
+      "grad_norm": 0.9419183731079102,
+      "learning_rate": 9.990294980811244e-06,
+      "logits/chosen": 0.1693490892648697,
+      "logits/rejected": 0.18133510649204254,
+      "logps/chosen": -60.769744873046875,
+      "logps/rejected": -65.61067199707031,
+      "loss": 2.7003,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.16450145840644836,
+      "rewards/margins": 0.04409575089812279,
+      "rewards/rejected": -0.20859721302986145,
+      "step": 46
+    },
+    {
+      "epoch": 0.08537693006357856,
+      "grad_norm": 0.9839057326316833,
+      "learning_rate": 9.989868597087907e-06,
+      "logits/chosen": 0.1363808959722519,
+      "logits/rejected": 0.22261787950992584,
+      "logps/chosen": -65.37456512451172,
+      "logps/rejected": -72.04705047607422,
+      "loss": 2.6636,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.18397849798202515,
+      "rewards/margins": 0.06703396141529083,
+      "rewards/rejected": -0.2510124444961548,
+      "step": 47
+    },
+    {
+      "epoch": 0.08719346049046321,
+      "grad_norm": 0.8818926811218262,
+      "learning_rate": 9.989433057740738e-06,
+      "logits/chosen": 0.1580447256565094,
+      "logits/rejected": 0.18051129579544067,
+      "logps/chosen": -56.6561279296875,
+      "logps/rejected": -64.66617584228516,
+      "loss": 2.6658,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.2094140350818634,
+      "rewards/margins": 0.05971873924136162,
+      "rewards/rejected": -0.2691327929496765,
+      "step": 48
+    },
+    {
+      "epoch": 0.08900999091734786,
+      "grad_norm": 0.8838489651679993,
+      "learning_rate": 9.98898836365788e-06,
+      "logits/chosen": 0.11484000086784363,
+      "logits/rejected": 0.12583914399147034,
+      "logps/chosen": -65.76107788085938,
+      "logps/rejected": -75.4742660522461,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.19882163405418396,
+      "rewards/margins": 0.0690067932009697,
+      "rewards/rejected": -0.26782843470573425,
+      "step": 49
+    },
+    {
+      "epoch": 0.09082652134423251,
+      "grad_norm": 0.9035817980766296,
+      "learning_rate": 9.988534515746141e-06,
+      "logits/chosen": 0.12592823803424835,
+      "logits/rejected": 0.18080386519432068,
+      "logps/chosen": -60.42765808105469,
+      "logps/rejected": -65.03330993652344,
+      "loss": 2.68,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.25037306547164917,
+      "rewards/margins": 0.05723651498556137,
+      "rewards/rejected": -0.30760958790779114,
+      "step": 50
+    },
+    {
+      "epoch": 0.09264305177111716,
+      "grad_norm": 1.0728137493133545,
+      "learning_rate": 9.988071514930998e-06,
+      "logits/chosen": 0.0955006331205368,
+      "logits/rejected": 0.15643228590488434,
+      "logps/chosen": -56.28706359863281,
+      "logps/rejected": -70.3702392578125,
+      "loss": 2.6104,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2517072558403015,
+      "rewards/margins": 0.11176257580518723,
+      "rewards/rejected": -0.36346977949142456,
+      "step": 51
+    },
+    {
+      "epoch": 0.09445958219800181,
+      "grad_norm": 0.9143723845481873,
+      "learning_rate": 9.987599362156587e-06,
+      "logits/chosen": 0.1018611341714859,
+      "logits/rejected": 0.21036753058433533,
+      "logps/chosen": -55.815738677978516,
+      "logps/rejected": -66.23766326904297,
+      "loss": 2.6207,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.2496231347322464,
+      "rewards/margins": 0.09438168257474899,
+      "rewards/rejected": -0.3440048098564148,
+      "step": 52
+    },
+    {
+      "epoch": 0.09627611262488647,
+      "grad_norm": 0.8731315732002258,
+      "learning_rate": 9.987118058385712e-06,
+      "logits/chosen": 0.12294681370258331,
+      "logits/rejected": 0.12897568941116333,
+      "logps/chosen": -63.27684020996094,
+      "logps/rejected": -65.21257019042969,
+      "loss": 2.7093,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.28397974371910095,
+      "rewards/margins": 0.05316353589296341,
+      "rewards/rejected": -0.33714330196380615,
+      "step": 53
+    },
+    {
+      "epoch": 0.09809264305177112,
+      "grad_norm": 0.951884388923645,
+      "learning_rate": 9.986627604599835e-06,
+      "logits/chosen": 0.1588897705078125,
+      "logits/rejected": 0.10302369296550751,
+      "logps/chosen": -62.071083068847656,
+      "logps/rejected": -59.575416564941406,
+      "loss": 2.7725,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.31246036291122437,
+      "rewards/margins": 0.011688929051160812,
+      "rewards/rejected": -0.3241492807865143,
+      "step": 54
+    },
+    {
+      "epoch": 0.09990917347865577,
+      "grad_norm": 1.0379350185394287,
+      "learning_rate": 9.986128001799077e-06,
+      "logits/chosen": 0.15824833512306213,
+      "logits/rejected": 0.16385456919670105,
+      "logps/chosen": -79.65828704833984,
+      "logps/rejected": -82.06480407714844,
+      "loss": 2.6826,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.36005648970603943,
+      "rewards/margins": 0.07091177999973297,
+      "rewards/rejected": -0.4309682846069336,
+      "step": 55
+    },
+    {
+      "epoch": 0.10172570390554042,
+      "grad_norm": 0.8740183115005493,
+      "learning_rate": 9.985619251002214e-06,
+      "logits/chosen": 0.0996306911110878,
+      "logits/rejected": 0.14288735389709473,
+      "logps/chosen": -59.8499755859375,
+      "logps/rejected": -63.10956573486328,
+      "loss": 2.664,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3242360055446625,
+      "rewards/margins": 0.06764136999845505,
+      "rewards/rejected": -0.39187735319137573,
+      "step": 56
+    },
+    {
+      "epoch": 0.10354223433242507,
+      "grad_norm": 0.9942138195037842,
+      "learning_rate": 9.985101353246676e-06,
+      "logits/chosen": 0.12070289999246597,
+      "logits/rejected": 0.1866157352924347,
+      "logps/chosen": -70.04698181152344,
+      "logps/rejected": -76.2674560546875,
+      "loss": 2.6165,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.4166935980319977,
+      "rewards/margins": 0.10780875384807587,
+      "rewards/rejected": -0.5245023965835571,
+      "step": 57
+    },
+    {
+      "epoch": 0.10535876475930972,
+      "grad_norm": 0.9078152775764465,
+      "learning_rate": 9.98457430958855e-06,
+      "logits/chosen": 0.16699416935443878,
+      "logits/rejected": 0.216594398021698,
+      "logps/chosen": -59.81795883178711,
+      "logps/rejected": -64.79395294189453,
+      "loss": 2.551,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.39265742897987366,
+      "rewards/margins": 0.14742323756217957,
+      "rewards/rejected": -0.5400806665420532,
+      "step": 58
+    },
+    {
+      "epoch": 0.10717529518619437,
+      "grad_norm": 0.9305158257484436,
+      "learning_rate": 9.984038121102569e-06,
+      "logits/chosen": 0.16001635789871216,
+      "logits/rejected": 0.18797104060649872,
+      "logps/chosen": -58.087158203125,
+      "logps/rejected": -59.003414154052734,
+      "loss": 2.6345,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.3626652956008911,
+      "rewards/margins": 0.08553728461265564,
+      "rewards/rejected": -0.44820258021354675,
+      "step": 59
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.9690614342689514,
+      "learning_rate": 9.983492788882113e-06,
+      "logits/chosen": 0.161406010389328,
+      "logits/rejected": 0.14847029745578766,
+      "logps/chosen": -61.01350021362305,
+      "logps/rejected": -60.993080139160156,
+      "loss": 2.7099,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.4803884029388428,
+      "rewards/margins": 0.0831567794084549,
+      "rewards/rejected": -0.5635451078414917,
+      "step": 60
+    },
+    {
+      "epoch": 0.11080835603996367,
+      "grad_norm": 0.971172034740448,
+      "learning_rate": 9.98293831403921e-06,
+      "logits/chosen": 0.14095918834209442,
+      "logits/rejected": 0.16142162680625916,
+      "logps/chosen": -62.31782150268555,
+      "logps/rejected": -64.44837188720703,
+      "loss": 2.6373,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.47784310579299927,
+      "rewards/margins": 0.09006088227033615,
+      "rewards/rejected": -0.567903995513916,
+      "step": 61
+    },
+    {
+      "epoch": 0.11262488646684832,
+      "grad_norm": 0.9984462857246399,
+      "learning_rate": 9.982374697704532e-06,
+      "logits/chosen": 0.08945554494857788,
+      "logits/rejected": 0.17623476684093475,
+      "logps/chosen": -58.84153366088867,
+      "logps/rejected": -72.59091186523438,
+      "loss": 2.5465,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.41726160049438477,
+      "rewards/margins": 0.16527526080608368,
+      "rewards/rejected": -0.5825368165969849,
+      "step": 62
+    },
+    {
+      "epoch": 0.11444141689373297,
+      "grad_norm": 1.007622241973877,
+      "learning_rate": 9.981801941027388e-06,
+      "logits/chosen": 0.16932496428489685,
+      "logits/rejected": 0.24257409572601318,
+      "logps/chosen": -61.62455749511719,
+      "logps/rejected": -69.40538024902344,
+      "loss": 2.5876,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.4484747052192688,
+      "rewards/margins": 0.11513397097587585,
+      "rewards/rejected": -0.563608705997467,
+      "step": 63
+    },
+    {
+      "epoch": 0.11625794732061762,
+      "grad_norm": 1.021275281906128,
+      "learning_rate": 9.981220045175731e-06,
+      "logits/chosen": 0.13468854129314423,
+      "logits/rejected": 0.10648790001869202,
+      "logps/chosen": -56.91660690307617,
+      "logps/rejected": -54.84038162231445,
+      "loss": 2.7096,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.47126585245132446,
+      "rewards/margins": 0.05656271427869797,
+      "rewards/rejected": -0.527828574180603,
+      "step": 64
+    },
+    {
+      "epoch": 0.11807447774750227,
+      "grad_norm": 0.9580955505371094,
+      "learning_rate": 9.980629011336149e-06,
+      "logits/chosen": 0.10542559623718262,
+      "logits/rejected": 0.18956655263900757,
+      "logps/chosen": -63.21677017211914,
+      "logps/rejected": -68.92955017089844,
+      "loss": 2.5447,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.4920822083950043,
+      "rewards/margins": 0.1548418551683426,
+      "rewards/rejected": -0.6469241380691528,
+      "step": 65
+    },
+    {
+      "epoch": 0.11989100817438691,
+      "grad_norm": 0.9601484537124634,
+      "learning_rate": 9.98002884071386e-06,
+      "logits/chosen": 0.1211993545293808,
+      "logits/rejected": 0.15335297584533691,
+      "logps/chosen": -65.79328918457031,
+      "logps/rejected": -71.8165283203125,
+      "loss": 2.6013,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.4890082776546478,
+      "rewards/margins": 0.16498278081417084,
+      "rewards/rejected": -0.6539911031723022,
+      "step": 66
+    },
+    {
+      "epoch": 0.12170753860127158,
+      "grad_norm": 0.9475951194763184,
+      "learning_rate": 9.97941953453272e-06,
+      "logits/chosen": 0.129757359623909,
+      "logits/rejected": 0.19597335159778595,
+      "logps/chosen": -61.52772903442383,
+      "logps/rejected": -62.8316764831543,
+      "loss": 2.6008,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6176788210868835,
+      "rewards/margins": 0.11390677094459534,
+      "rewards/rejected": -0.7315855622291565,
+      "step": 67
+    },
+    {
+      "epoch": 0.12352406902815623,
+      "grad_norm": 1.0577987432479858,
+      "learning_rate": 9.978801094035207e-06,
+      "logits/chosen": 0.1363590657711029,
+      "logits/rejected": 0.1551840454339981,
+      "logps/chosen": -71.82952117919922,
+      "logps/rejected": -78.26856994628906,
+      "loss": 2.6714,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.6073123216629028,
+      "rewards/margins": 0.08929078280925751,
+      "rewards/rejected": -0.6966031193733215,
+      "step": 68
+    },
+    {
+      "epoch": 0.12534059945504086,
+      "grad_norm": 1.2286291122436523,
+      "learning_rate": 9.978173520482429e-06,
+      "logits/chosen": 0.12522992491722107,
+      "logits/rejected": 0.1818617582321167,
+      "logps/chosen": -60.64283752441406,
+      "logps/rejected": -67.53813934326172,
+      "loss": 2.6884,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -0.5357421040534973,
+      "rewards/margins": 0.0799265205860138,
+      "rewards/rejected": -0.6156685948371887,
+      "step": 69
+    },
+    {
+      "epoch": 0.1271571298819255,
+      "grad_norm": 1.0266982316970825,
+      "learning_rate": 9.97753681515412e-06,
+      "logits/chosen": 0.17851999402046204,
+      "logits/rejected": 0.25888925790786743,
+      "logps/chosen": -71.65196990966797,
+      "logps/rejected": -82.34913635253906,
+      "loss": 2.5226,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7787651419639587,
+      "rewards/margins": 0.1905478835105896,
+      "rewards/rejected": -0.9693130254745483,
+      "step": 70
+    },
+    {
+      "epoch": 0.12897366030881016,
+      "grad_norm": 0.999660313129425,
+      "learning_rate": 9.976890979348628e-06,
+      "logits/chosen": 0.18746232986450195,
+      "logits/rejected": 0.2187044620513916,
+      "logps/chosen": -69.84027099609375,
+      "logps/rejected": -73.04399108886719,
+      "loss": 2.6945,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.6969642639160156,
+      "rewards/margins": 0.08156725764274597,
+      "rewards/rejected": -0.778531551361084,
+      "step": 71
+    },
+    {
+      "epoch": 0.1307901907356948,
+      "grad_norm": 1.0596684217453003,
+      "learning_rate": 9.976236014382934e-06,
+      "logits/chosen": 0.17854923009872437,
+      "logits/rejected": 0.18649883568286896,
+      "logps/chosen": -63.992881774902344,
+      "logps/rejected": -64.36544799804688,
+      "loss": 2.7689,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.7436442375183105,
+      "rewards/margins": 0.06337633728981018,
+      "rewards/rejected": -0.8070206046104431,
+      "step": 72
+    },
+    {
+      "epoch": 0.13260672116257946,
+      "grad_norm": 1.0307612419128418,
+      "learning_rate": 9.975571921592618e-06,
+      "logits/chosen": 0.0964425653219223,
+      "logits/rejected": 0.14382749795913696,
+      "logps/chosen": -66.61788940429688,
+      "logps/rejected": -70.80413055419922,
+      "loss": 2.5859,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.6208910942077637,
+      "rewards/margins": 0.17540723085403442,
+      "rewards/rejected": -0.7962983250617981,
+      "step": 73
+    },
+    {
+      "epoch": 0.1344232515894641,
+      "grad_norm": 0.9487422108650208,
+      "learning_rate": 9.97489870233188e-06,
+      "logits/chosen": 0.06293762475252151,
+      "logits/rejected": 0.19688080251216888,
+      "logps/chosen": -55.884613037109375,
+      "logps/rejected": -64.30046081542969,
+      "loss": 2.4062,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6415266990661621,
+      "rewards/margins": 0.2741628885269165,
+      "rewards/rejected": -0.9156895875930786,
+      "step": 74
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 1.223663330078125,
+      "learning_rate": 9.974216357973539e-06,
+      "logits/chosen": 0.10017126798629761,
+      "logits/rejected": 0.10504551976919174,
+      "logps/chosen": -83.83055114746094,
+      "logps/rejected": -81.8952407836914,
+      "loss": 2.8556,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8103373050689697,
+      "rewards/margins": 0.027550537139177322,
+      "rewards/rejected": -0.8378878831863403,
+      "step": 75
+    },
+    {
+      "epoch": 0.13805631244323344,
+      "grad_norm": 1.0121136903762817,
+      "learning_rate": 9.973524889909007e-06,
+      "logits/chosen": 0.07612155377864838,
+      "logits/rejected": 0.15954618155956268,
+      "logps/chosen": -65.6831283569336,
+      "logps/rejected": -75.42146301269531,
+      "loss": 2.5576,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6988077759742737,
+      "rewards/margins": 0.1909356415271759,
+      "rewards/rejected": -0.889743447303772,
+      "step": 76
+    },
+    {
+      "epoch": 0.13987284287011809,
+      "grad_norm": 1.0312719345092773,
+      "learning_rate": 9.97282429954831e-06,
+      "logits/chosen": 0.07015375047922134,
+      "logits/rejected": 0.09143385291099548,
+      "logps/chosen": -59.444358825683594,
+      "logps/rejected": -61.26897430419922,
+      "loss": 2.7845,
+      "rewards/accuracies": 0.453125,
+      "rewards/chosen": -0.7886172533035278,
+      "rewards/margins": 0.038636498153209686,
+      "rewards/rejected": -0.8272536993026733,
+      "step": 77
+    },
+    {
+      "epoch": 0.14168937329700274,
+      "grad_norm": 1.0625450611114502,
+      "learning_rate": 9.972114588320073e-06,
+      "logits/chosen": 0.07196499407291412,
+      "logits/rejected": 0.12630172073841095,
+      "logps/chosen": -75.5164566040039,
+      "logps/rejected": -80.74308776855469,
+      "loss": 2.5834,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.6821954250335693,
+      "rewards/margins": 0.16858401894569397,
+      "rewards/rejected": -0.8507794737815857,
+      "step": 78
+    },
+    {
+      "epoch": 0.14350590372388738,
+      "grad_norm": 0.9069045186042786,
+      "learning_rate": 9.97139575767152e-06,
+      "logits/chosen": 0.10926744341850281,
+      "logits/rejected": 0.14418405294418335,
+      "logps/chosen": -53.721046447753906,
+      "logps/rejected": -58.17475891113281,
+      "loss": 2.4955,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.6156648993492126,
+      "rewards/margins": 0.19551396369934082,
+      "rewards/rejected": -0.8111788034439087,
+      "step": 79
+    },
+    {
+      "epoch": 0.14532243415077203,
+      "grad_norm": 1.1463656425476074,
+      "learning_rate": 9.970667809068476e-06,
+      "logits/chosen": 0.09196805953979492,
+      "logits/rejected": 0.14393256604671478,
+      "logps/chosen": -77.11197662353516,
+      "logps/rejected": -87.39771270751953,
+      "loss": 2.5763,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8568902611732483,
+      "rewards/margins": 0.15524335205554962,
+      "rewards/rejected": -1.0121335983276367,
+      "step": 80
+    },
+    {
+      "epoch": 0.14713896457765668,
+      "grad_norm": 1.0046662092208862,
+      "learning_rate": 9.969930743995351e-06,
+      "logits/chosen": 0.2063535749912262,
+      "logits/rejected": 0.20602768659591675,
+      "logps/chosen": -59.890933990478516,
+      "logps/rejected": -59.117794036865234,
+      "loss": 2.6829,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.7329096794128418,
+      "rewards/margins": 0.0983649268746376,
+      "rewards/rejected": -0.8312745094299316,
+      "step": 81
+    },
+    {
+      "epoch": 0.14895549500454133,
+      "grad_norm": 1.0280438661575317,
+      "learning_rate": 9.969184563955152e-06,
+      "logits/chosen": 0.17393875122070312,
+      "logits/rejected": 0.16783203184604645,
+      "logps/chosen": -72.82754516601562,
+      "logps/rejected": -70.6113510131836,
+      "loss": 2.6277,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7506997585296631,
+      "rewards/margins": 0.13392671942710876,
+      "rewards/rejected": -0.8846263885498047,
+      "step": 82
+    },
+    {
+      "epoch": 0.15077202543142598,
+      "grad_norm": 0.9356668591499329,
+      "learning_rate": 9.968429270469467e-06,
+      "logits/chosen": 0.09279462695121765,
+      "logits/rejected": 0.1681107133626938,
+      "logps/chosen": -62.300445556640625,
+      "logps/rejected": -68.70709991455078,
+      "loss": 2.4821,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.7132639288902283,
+      "rewards/margins": 0.2519262731075287,
+      "rewards/rejected": -0.9651902318000793,
+      "step": 83
+    },
+    {
+      "epoch": 0.15258855585831063,
+      "grad_norm": 1.0442771911621094,
+      "learning_rate": 9.967664865078472e-06,
+      "logits/chosen": 0.17844007909297943,
+      "logits/rejected": 0.21383100748062134,
+      "logps/chosen": -61.43730163574219,
+      "logps/rejected": -64.16680908203125,
+      "loss": 2.6289,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.7574411034584045,
+      "rewards/margins": 0.13140378892421722,
+      "rewards/rejected": -0.888844907283783,
+      "step": 84
+    },
+    {
+      "epoch": 0.15440508628519528,
+      "grad_norm": 1.137267827987671,
+      "learning_rate": 9.966891349340922e-06,
+      "logits/chosen": 0.19703873991966248,
+      "logits/rejected": 0.17027492821216583,
+      "logps/chosen": -68.07546997070312,
+      "logps/rejected": -63.9476203918457,
+      "loss": 2.7702,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -0.8362983465194702,
+      "rewards/margins": 0.05476854741573334,
+      "rewards/rejected": -0.8910670280456543,
+      "step": 85
+    },
+    {
+      "epoch": 0.15622161671207993,
+      "grad_norm": 1.0395056009292603,
+      "learning_rate": 9.966108724834151e-06,
+      "logits/chosen": 0.10182631760835648,
+      "logits/rejected": 0.11173731088638306,
+      "logps/chosen": -64.47720336914062,
+      "logps/rejected": -75.30056762695312,
+      "loss": 2.5889,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8160837292671204,
+      "rewards/margins": 0.2798304557800293,
+      "rewards/rejected": -1.0959142446517944,
+      "step": 86
+    },
+    {
+      "epoch": 0.15803814713896458,
+      "grad_norm": 1.1256901025772095,
+      "learning_rate": 9.965316993154069e-06,
+      "logits/chosen": 0.1232331395149231,
+      "logits/rejected": 0.19064000248908997,
+      "logps/chosen": -64.3778305053711,
+      "logps/rejected": -76.91246032714844,
+      "loss": 2.5522,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8341800570487976,
+      "rewards/margins": 0.2616519033908844,
+      "rewards/rejected": -1.0958319902420044,
+      "step": 87
+    },
+    {
+      "epoch": 0.15985467756584923,
+      "grad_norm": 1.20338773727417,
+      "learning_rate": 9.964516155915152e-06,
+      "logits/chosen": 0.12199988961219788,
+      "logits/rejected": 0.12264345586299896,
+      "logps/chosen": -81.56141662597656,
+      "logps/rejected": -80.49986267089844,
+      "loss": 2.7432,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -0.7930806875228882,
+      "rewards/margins": 0.06459490954875946,
+      "rewards/rejected": -0.8576756119728088,
+      "step": 88
+    },
+    {
+      "epoch": 0.16167120799273388,
+      "grad_norm": 1.005759835243225,
+      "learning_rate": 9.963706214750446e-06,
+      "logits/chosen": 0.07040335237979889,
+      "logits/rejected": 0.10741756111383438,
+      "logps/chosen": -65.43824768066406,
+      "logps/rejected": -74.47370147705078,
+      "loss": 2.5472,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.8203200697898865,
+      "rewards/margins": 0.24586233496665955,
+      "rewards/rejected": -1.0661823749542236,
+      "step": 89
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.9761494994163513,
+      "learning_rate": 9.962887171311563e-06,
+      "logits/chosen": 0.1630188375711441,
+      "logits/rejected": 0.190489262342453,
+      "logps/chosen": -57.84222412109375,
+      "logps/rejected": -64.116455078125,
+      "loss": 2.5339,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8240211009979248,
+      "rewards/margins": 0.21757929027080536,
+      "rewards/rejected": -1.0416003465652466,
+      "step": 90
+    },
+    {
+      "epoch": 0.16530426884650318,
+      "grad_norm": 1.002192497253418,
+      "learning_rate": 9.962059027268676e-06,
+      "logits/chosen": 0.08669686317443848,
+      "logits/rejected": 0.13606388866901398,
+      "logps/chosen": -63.68013000488281,
+      "logps/rejected": -72.30770111083984,
+      "loss": 2.5165,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.7919780611991882,
+      "rewards/margins": 0.27081233263015747,
+      "rewards/rejected": -1.0627902746200562,
+      "step": 91
+    },
+    {
+      "epoch": 0.16712079927338783,
+      "grad_norm": 0.9397479295730591,
+      "learning_rate": 9.961221784310514e-06,
+      "logits/chosen": 0.14396092295646667,
+      "logits/rejected": 0.1820589005947113,
+      "logps/chosen": -57.551353454589844,
+      "logps/rejected": -66.83124542236328,
+      "loss": 2.4254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7751535177230835,
+      "rewards/margins": 0.26002568006515503,
+      "rewards/rejected": -1.0351792573928833,
+      "step": 92
+    },
+    {
+      "epoch": 0.16893732970027248,
+      "grad_norm": 1.2053430080413818,
+      "learning_rate": 9.96037544414436e-06,
+      "logits/chosen": 0.16130733489990234,
+      "logits/rejected": 0.1852559745311737,
+      "logps/chosen": -73.8154067993164,
+      "logps/rejected": -81.3624267578125,
+      "loss": 2.7543,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.8622196316719055,
+      "rewards/margins": 0.11067891120910645,
+      "rewards/rejected": -0.9728984236717224,
+      "step": 93
+    },
+    {
+      "epoch": 0.17075386012715713,
+      "grad_norm": 1.1681262254714966,
+      "learning_rate": 9.959520008496054e-06,
+      "logits/chosen": 0.17866984009742737,
+      "logits/rejected": 0.20266617834568024,
+      "logps/chosen": -61.40083312988281,
+      "logps/rejected": -65.13737487792969,
+      "loss": 2.7134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9223482012748718,
+      "rewards/margins": 0.08481010049581528,
+      "rewards/rejected": -1.0071581602096558,
+      "step": 94
+    },
+    {
+      "epoch": 0.17257039055404177,
+      "grad_norm": 1.0269131660461426,
+      "learning_rate": 9.95865547910997e-06,
+      "logits/chosen": 0.11446906626224518,
+      "logits/rejected": 0.17379909753799438,
+      "logps/chosen": -62.823944091796875,
+      "logps/rejected": -69.5682601928711,
+      "loss": 2.5031,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8894773125648499,
+      "rewards/margins": 0.24369436502456665,
+      "rewards/rejected": -1.1331716775894165,
+      "step": 95
+    },
+    {
+      "epoch": 0.17438692098092642,
+      "grad_norm": 1.0400865077972412,
+      "learning_rate": 9.957781857749043e-06,
+      "logits/chosen": 0.16556067764759064,
+      "logits/rejected": 0.23536017537117004,
+      "logps/chosen": -61.09260940551758,
+      "logps/rejected": -63.444427490234375,
+      "loss": 2.7239,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.0037989616394043,
+      "rewards/margins": 0.07835513353347778,
+      "rewards/rejected": -1.0821542739868164,
+      "step": 96
+    },
+    {
+      "epoch": 0.17620345140781107,
+      "grad_norm": 1.1138463020324707,
+      "learning_rate": 9.956899146194732e-06,
+      "logits/chosen": 0.13205701112747192,
+      "logits/rejected": 0.1702471375465393,
+      "logps/chosen": -63.70584487915039,
+      "logps/rejected": -69.10890197753906,
+      "loss": 2.7076,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9300363659858704,
+      "rewards/margins": 0.17512086033821106,
+      "rewards/rejected": -1.1051572561264038,
+      "step": 97
+    },
+    {
+      "epoch": 0.17801998183469572,
+      "grad_norm": 0.9917119741439819,
+      "learning_rate": 9.956007346247043e-06,
+      "logits/chosen": 0.12265195697546005,
+      "logits/rejected": 0.18780440092086792,
+      "logps/chosen": -62.316497802734375,
+      "logps/rejected": -69.26319885253906,
+      "loss": 2.4996,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8957506418228149,
+      "rewards/margins": 0.2764662802219391,
+      "rewards/rejected": -1.1722170114517212,
+      "step": 98
+    },
+    {
+      "epoch": 0.17983651226158037,
+      "grad_norm": 1.00100576877594,
+      "learning_rate": 9.95510645972451e-06,
+      "logits/chosen": 0.16969357430934906,
+      "logits/rejected": 0.16122889518737793,
+      "logps/chosen": -60.42213439941406,
+      "logps/rejected": -63.85418701171875,
+      "loss": 2.5677,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8840566873550415,
+      "rewards/margins": 0.1611585170030594,
+      "rewards/rejected": -1.0452152490615845,
+      "step": 99
+    },
+    {
+      "epoch": 0.18165304268846502,
+      "grad_norm": 1.1503106355667114,
+      "learning_rate": 9.954196488464198e-06,
+      "logits/chosen": 0.19831174612045288,
+      "logits/rejected": 0.20889577269554138,
+      "logps/chosen": -65.17144775390625,
+      "logps/rejected": -69.9671859741211,
+      "loss": 2.7065,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.9322465062141418,
+      "rewards/margins": 0.1645345687866211,
+      "rewards/rejected": -1.0967810153961182,
+      "step": 100
+    },
+    {
+      "epoch": 0.18346957311534967,
+      "grad_norm": 1.416610836982727,
+      "learning_rate": 9.953277434321696e-06,
+      "logits/chosen": 0.08757522702217102,
+      "logits/rejected": 0.1615956872701645,
+      "logps/chosen": -72.49248504638672,
+      "logps/rejected": -80.52176666259766,
+      "loss": 2.5346,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8279029130935669,
+      "rewards/margins": 0.26418614387512207,
+      "rewards/rejected": -1.092089056968689,
+      "step": 101
+    },
+    {
+      "epoch": 0.18528610354223432,
+      "grad_norm": 0.9907792210578918,
+      "learning_rate": 9.952349299171117e-06,
+      "logits/chosen": 0.07832024991512299,
+      "logits/rejected": 0.16370174288749695,
+      "logps/chosen": -67.5047836303711,
+      "logps/rejected": -73.78684997558594,
+      "loss": 2.4325,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8129651546478271,
+      "rewards/margins": 0.3001910448074341,
+      "rewards/rejected": -1.1131561994552612,
+      "step": 102
+    },
+    {
+      "epoch": 0.18710263396911897,
+      "grad_norm": 1.1051058769226074,
+      "learning_rate": 9.95141208490509e-06,
+      "logits/chosen": 0.10540622472763062,
+      "logits/rejected": 0.2087182104587555,
+      "logps/chosen": -55.68672180175781,
+      "logps/rejected": -72.7021484375,
+      "loss": 2.3004,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -0.8526840209960938,
+      "rewards/margins": 0.3691489100456238,
+      "rewards/rejected": -1.2218331098556519,
+      "step": 103
+    },
+    {
+      "epoch": 0.18891916439600362,
+      "grad_norm": 1.0593968629837036,
+      "learning_rate": 9.950465793434759e-06,
+      "logits/chosen": 0.10488969832658768,
+      "logits/rejected": 0.13292263448238373,
+      "logps/chosen": -69.47488403320312,
+      "logps/rejected": -73.08143615722656,
+      "loss": 2.5388,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8357015252113342,
+      "rewards/margins": 0.2204672396183014,
+      "rewards/rejected": -1.0561686754226685,
+      "step": 104
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 1.119372010231018,
+      "learning_rate": 9.949510426689773e-06,
+      "logits/chosen": 0.06970994919538498,
+      "logits/rejected": 0.0831708014011383,
+      "logps/chosen": -66.97635650634766,
+      "logps/rejected": -67.95066833496094,
+      "loss": 2.6443,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -0.9023284912109375,
+      "rewards/margins": 0.1277586966753006,
+      "rewards/rejected": -1.0300871133804321,
+      "step": 105
+    },
+    {
+      "epoch": 0.19255222524977295,
+      "grad_norm": 1.2842986583709717,
+      "learning_rate": 9.948545986618295e-06,
+      "logits/chosen": 0.17146825790405273,
+      "logits/rejected": 0.1910206824541092,
+      "logps/chosen": -69.1142578125,
+      "logps/rejected": -70.71878814697266,
+      "loss": 2.8494,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0275481939315796,
+      "rewards/margins": 0.09755454212427139,
+      "rewards/rejected": -1.1251027584075928,
+      "step": 106
+    },
+    {
+      "epoch": 0.1943687556766576,
+      "grad_norm": 1.044800877571106,
+      "learning_rate": 9.947572475186984e-06,
+      "logits/chosen": 0.1206701397895813,
+      "logits/rejected": 0.17635390162467957,
+      "logps/chosen": -67.00174713134766,
+      "logps/rejected": -71.5347900390625,
+      "loss": 2.46,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.889815628528595,
+      "rewards/margins": 0.255011647939682,
+      "rewards/rejected": -1.1448272466659546,
+      "step": 107
+    },
+    {
+      "epoch": 0.19618528610354224,
+      "grad_norm": 1.1496058702468872,
+      "learning_rate": 9.946589894381002e-06,
+      "logits/chosen": 0.18972846865653992,
+      "logits/rejected": 0.12482471019029617,
+      "logps/chosen": -66.56938934326172,
+      "logps/rejected": -73.7935791015625,
+      "loss": 2.6472,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.8382111191749573,
+      "rewards/margins": 0.16597144305706024,
+      "rewards/rejected": -1.0041825771331787,
+      "step": 108
+    },
+    {
+      "epoch": 0.1980018165304269,
+      "grad_norm": 1.1357221603393555,
+      "learning_rate": 9.945598246204e-06,
+      "logits/chosen": 0.1766098588705063,
+      "logits/rejected": 0.16134579479694366,
+      "logps/chosen": -72.30979919433594,
+      "logps/rejected": -69.26260375976562,
+      "loss": 2.7037,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9316724538803101,
+      "rewards/margins": 0.0998368114233017,
+      "rewards/rejected": -1.0315091609954834,
+      "step": 109
+    },
+    {
+      "epoch": 0.19981834695731154,
+      "grad_norm": 1.1147487163543701,
+      "learning_rate": 9.94459753267812e-06,
+      "logits/chosen": 0.10100046545267105,
+      "logits/rejected": 0.12630510330200195,
+      "logps/chosen": -66.10484313964844,
+      "logps/rejected": -68.13362121582031,
+      "loss": 2.6224,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8918489217758179,
+      "rewards/margins": 0.165731742978096,
+      "rewards/rejected": -1.0575807094573975,
+      "step": 110
+    },
+    {
+      "epoch": 0.2016348773841962,
+      "grad_norm": 1.3498740196228027,
+      "learning_rate": 9.943587755843996e-06,
+      "logits/chosen": 0.19711939990520477,
+      "logits/rejected": 0.18357205390930176,
+      "logps/chosen": -64.5146484375,
+      "logps/rejected": -65.80925750732422,
+      "loss": 2.6102,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9744794368743896,
+      "rewards/margins": 0.16535742580890656,
+      "rewards/rejected": -1.1398367881774902,
+      "step": 111
+    },
+    {
+      "epoch": 0.20345140781108084,
+      "grad_norm": 1.0210968255996704,
+      "learning_rate": 9.942568917760733e-06,
+      "logits/chosen": 0.16077302396297455,
+      "logits/rejected": 0.22646722197532654,
+      "logps/chosen": -59.73281478881836,
+      "logps/rejected": -64.54032897949219,
+      "loss": 2.4897,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8548685312271118,
+      "rewards/margins": 0.19910478591918945,
+      "rewards/rejected": -1.0539731979370117,
+      "step": 112
+    },
+    {
+      "epoch": 0.2052679382379655,
+      "grad_norm": 1.160152554512024,
+      "learning_rate": 9.941541020505924e-06,
+      "logits/chosen": 0.10292509198188782,
+      "logits/rejected": 0.09458380937576294,
+      "logps/chosen": -77.44420623779297,
+      "logps/rejected": -85.09677124023438,
+      "loss": 2.5037,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.9961792230606079,
+      "rewards/margins": 0.2911871373653412,
+      "rewards/rejected": -1.287366509437561,
+      "step": 113
+    },
+    {
+      "epoch": 0.20708446866485014,
+      "grad_norm": 1.1742796897888184,
+      "learning_rate": 9.940504066175626e-06,
+      "logits/chosen": 0.10538715869188309,
+      "logits/rejected": 0.07242526113986969,
+      "logps/chosen": -70.37040710449219,
+      "logps/rejected": -68.23387908935547,
+      "loss": 2.6328,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.9170363545417786,
+      "rewards/margins": 0.14501458406448364,
+      "rewards/rejected": -1.0620509386062622,
+      "step": 114
+    },
+    {
+      "epoch": 0.2089009990917348,
+      "grad_norm": 1.1259719133377075,
+      "learning_rate": 9.939458056884375e-06,
+      "logits/chosen": 0.12264375388622284,
+      "logits/rejected": 0.16776585578918457,
+      "logps/chosen": -59.847049713134766,
+      "logps/rejected": -65.45285034179688,
+      "loss": 2.5752,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8872510194778442,
+      "rewards/margins": 0.2041575014591217,
+      "rewards/rejected": -1.0914084911346436,
+      "step": 115
+    },
+    {
+      "epoch": 0.21071752951861944,
+      "grad_norm": 2.1809544563293457,
+      "learning_rate": 9.938402994765163e-06,
+      "logits/chosen": 0.13642175495624542,
+      "logits/rejected": 0.11773751676082611,
+      "logps/chosen": -65.34564971923828,
+      "logps/rejected": -66.55570983886719,
+      "loss": 2.6792,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9432386755943298,
+      "rewards/margins": 0.12600603699684143,
+      "rewards/rejected": -1.0692447423934937,
+      "step": 116
+    },
+    {
+      "epoch": 0.2125340599455041,
+      "grad_norm": 1.4895256757736206,
+      "learning_rate": 9.937338881969444e-06,
+      "logits/chosen": 0.12349803745746613,
+      "logits/rejected": 0.1182522252202034,
+      "logps/chosen": -66.9234619140625,
+      "logps/rejected": -69.5535888671875,
+      "loss": 2.6029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.9096715450286865,
+      "rewards/margins": 0.23732726275920868,
+      "rewards/rejected": -1.1469988822937012,
+      "step": 117
+    },
+    {
+      "epoch": 0.21435059037238874,
+      "grad_norm": 1.0333324670791626,
+      "learning_rate": 9.93626572066713e-06,
+      "logits/chosen": 0.16722331941127777,
+      "logits/rejected": 0.21501797437667847,
+      "logps/chosen": -66.49773406982422,
+      "logps/rejected": -72.3947982788086,
+      "loss": 2.4687,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -0.9357748627662659,
+      "rewards/margins": 0.2686734199523926,
+      "rewards/rejected": -1.2044482231140137,
+      "step": 118
+    },
+    {
+      "epoch": 0.2161671207992734,
+      "grad_norm": 1.2127379179000854,
+      "learning_rate": 9.935183513046585e-06,
+      "logits/chosen": 0.10065922141075134,
+      "logits/rejected": 0.12764661014080048,
+      "logps/chosen": -73.05583953857422,
+      "logps/rejected": -77.18297576904297,
+      "loss": 2.6288,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -0.8347401022911072,
+      "rewards/margins": 0.2145780473947525,
+      "rewards/rejected": -1.0493181943893433,
+      "step": 119
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 1.4050469398498535,
+      "learning_rate": 9.93409226131462e-06,
+      "logits/chosen": 0.08468589186668396,
+      "logits/rejected": 0.179460808634758,
+      "logps/chosen": -66.12811279296875,
+      "logps/rejected": -74.2131118774414,
+      "loss": 2.4658,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -0.8611673712730408,
+      "rewards/margins": 0.20363186299800873,
+      "rewards/rejected": -1.064799189567566,
+      "step": 120
+    },
+    {
+      "epoch": 0.2198001816530427,
+      "grad_norm": 1.18356192111969,
+      "learning_rate": 9.932991967696484e-06,
+      "logits/chosen": 0.18308167159557343,
+      "logits/rejected": 0.12839025259017944,
+      "logps/chosen": -63.325035095214844,
+      "logps/rejected": -63.72959518432617,
+      "loss": 2.579,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9539028406143188,
+      "rewards/margins": 0.16736355423927307,
+      "rewards/rejected": -1.1212663650512695,
+      "step": 121
+    },
+    {
+      "epoch": 0.22161671207992734,
+      "grad_norm": 1.1829110383987427,
+      "learning_rate": 9.931882634435868e-06,
+      "logits/chosen": 0.10780126601457596,
+      "logits/rejected": 0.109224833548069,
+      "logps/chosen": -71.1589126586914,
+      "logps/rejected": -74.609375,
+      "loss": 2.6368,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9908429384231567,
+      "rewards/margins": 0.12795159220695496,
+      "rewards/rejected": -1.118794560432434,
+      "step": 122
+    },
+    {
+      "epoch": 0.22343324250681199,
+      "grad_norm": 1.1116541624069214,
+      "learning_rate": 9.930764263794898e-06,
+      "logits/chosen": 0.10946042090654373,
+      "logits/rejected": 0.11234283447265625,
+      "logps/chosen": -67.98091888427734,
+      "logps/rejected": -70.58726501464844,
+      "loss": 2.4417,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -0.871981143951416,
+      "rewards/margins": 0.22934575378894806,
+      "rewards/rejected": -1.1013269424438477,
+      "step": 123
+    },
+    {
+      "epoch": 0.22524977293369663,
+      "grad_norm": 1.1564345359802246,
+      "learning_rate": 9.929636858054124e-06,
+      "logits/chosen": 0.11665979772806168,
+      "logits/rejected": 0.15129488706588745,
+      "logps/chosen": -62.33778381347656,
+      "logps/rejected": -67.58541870117188,
+      "loss": 2.6023,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.8282152414321899,
+      "rewards/margins": 0.206033393740654,
+      "rewards/rejected": -1.0342485904693604,
+      "step": 124
+    },
+    {
+      "epoch": 0.22706630336058128,
+      "grad_norm": 1.149032711982727,
+      "learning_rate": 9.928500419512522e-06,
+      "logits/chosen": 0.02143971435725689,
+      "logits/rejected": 0.10582254827022552,
+      "logps/chosen": -66.31533813476562,
+      "logps/rejected": -76.4206771850586,
+      "loss": 2.4589,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.8935360908508301,
+      "rewards/margins": 0.3176972270011902,
+      "rewards/rejected": -1.2112332582473755,
+      "step": 125
+    },
+    {
+      "epoch": 0.22888283378746593,
+      "grad_norm": 1.1076140403747559,
+      "learning_rate": 9.927354950487491e-06,
+      "logits/chosen": 0.13772962987422943,
+      "logits/rejected": 0.12653110921382904,
+      "logps/chosen": -62.97186279296875,
+      "logps/rejected": -65.31454467773438,
+      "loss": 2.4839,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.9201483130455017,
+      "rewards/margins": 0.2287607342004776,
+      "rewards/rejected": -1.1489089727401733,
+      "step": 126
+    },
+    {
+      "epoch": 0.23069936421435058,
+      "grad_norm": 1.1476696729660034,
+      "learning_rate": 9.926200453314842e-06,
+      "logits/chosen": 0.11984287202358246,
+      "logits/rejected": 0.1570628136396408,
+      "logps/chosen": -72.74269104003906,
+      "logps/rejected": -77.52953338623047,
+      "loss": 2.5198,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9817119836807251,
+      "rewards/margins": 0.2599087059497833,
+      "rewards/rejected": -1.241620659828186,
+      "step": 127
+    },
+    {
+      "epoch": 0.23251589464123523,
+      "grad_norm": 1.0800604820251465,
+      "learning_rate": 9.925036930348793e-06,
+      "logits/chosen": 0.03499937057495117,
+      "logits/rejected": 0.16788874566555023,
+      "logps/chosen": -64.4096450805664,
+      "logps/rejected": -81.16387939453125,
+      "loss": 2.3547,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.040608286857605,
+      "rewards/margins": 0.48978835344314575,
+      "rewards/rejected": -1.530396580696106,
+      "step": 128
+    },
+    {
+      "epoch": 0.23433242506811988,
+      "grad_norm": 1.2677640914916992,
+      "learning_rate": 9.92386438396197e-06,
+      "logits/chosen": 0.1200772225856781,
+      "logits/rejected": 0.11345633864402771,
+      "logps/chosen": -68.35733032226562,
+      "logps/rejected": -71.72528839111328,
+      "loss": 2.631,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -0.9437850117683411,
+      "rewards/margins": 0.13758014142513275,
+      "rewards/rejected": -1.0813652276992798,
+      "step": 129
+    },
+    {
+      "epoch": 0.23614895549500453,
+      "grad_norm": 1.2041019201278687,
+      "learning_rate": 9.9226828165454e-06,
+      "logits/chosen": 0.07035915553569794,
+      "logits/rejected": 0.16790318489074707,
+      "logps/chosen": -73.99917602539062,
+      "logps/rejected": -83.07939147949219,
+      "loss": 2.4497,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0043660402297974,
+      "rewards/margins": 0.36712244153022766,
+      "rewards/rejected": -1.3714885711669922,
+      "step": 130
+    },
+    {
+      "epoch": 0.23796548592188918,
+      "grad_norm": 1.112176537513733,
+      "learning_rate": 9.921492230508507e-06,
+      "logits/chosen": 0.18427804112434387,
+      "logits/rejected": 0.17424902319908142,
+      "logps/chosen": -63.21034240722656,
+      "logps/rejected": -65.23414611816406,
+      "loss": 2.4612,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -0.9832345843315125,
+      "rewards/margins": 0.21530815958976746,
+      "rewards/rejected": -1.1985427141189575,
+      "step": 131
+    },
+    {
+      "epoch": 0.23978201634877383,
+      "grad_norm": 1.1540772914886475,
+      "learning_rate": 9.9202926282791e-06,
+      "logits/chosen": 0.14970409870147705,
+      "logits/rejected": 0.17991000413894653,
+      "logps/chosen": -76.66849517822266,
+      "logps/rejected": -85.59910583496094,
+      "loss": 2.4164,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0053367614746094,
+      "rewards/margins": 0.3513883352279663,
+      "rewards/rejected": -1.3567250967025757,
+      "step": 132
+    },
+    {
+      "epoch": 0.24159854677565848,
+      "grad_norm": 1.32212495803833,
+      "learning_rate": 9.919084012303378e-06,
+      "logits/chosen": 0.11897419393062592,
+      "logits/rejected": 0.15737299621105194,
+      "logps/chosen": -75.7364501953125,
+      "logps/rejected": -77.97785186767578,
+      "loss": 2.7243,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.139754295349121,
+      "rewards/margins": 0.17446394264698029,
+      "rewards/rejected": -1.3142181634902954,
+      "step": 133
+    },
+    {
+      "epoch": 0.24341507720254316,
+      "grad_norm": 1.4282020330429077,
+      "learning_rate": 9.917866385045918e-06,
+      "logits/chosen": 0.13694903254508972,
+      "logits/rejected": 0.2048657089471817,
+      "logps/chosen": -70.1098403930664,
+      "logps/rejected": -78.7418212890625,
+      "loss": 2.8086,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2937737703323364,
+      "rewards/margins": 0.2787961959838867,
+      "rewards/rejected": -1.5725698471069336,
+      "step": 134
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 1.1548171043395996,
+      "learning_rate": 9.916639748989677e-06,
+      "logits/chosen": 0.050287194550037384,
+      "logits/rejected": 0.10146909952163696,
+      "logps/chosen": -72.51569366455078,
+      "logps/rejected": -80.42554473876953,
+      "loss": 2.4774,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0650720596313477,
+      "rewards/margins": 0.2093387395143509,
+      "rewards/rejected": -1.2744107246398926,
+      "step": 135
+    },
+    {
+      "epoch": 0.24704813805631246,
+      "grad_norm": 1.1347655057907104,
+      "learning_rate": 9.915404106635979e-06,
+      "logits/chosen": 0.13506914675235748,
+      "logits/rejected": 0.18150334060192108,
+      "logps/chosen": -71.03006744384766,
+      "logps/rejected": -75.59716033935547,
+      "loss": 2.4399,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1068118810653687,
+      "rewards/margins": 0.3031606674194336,
+      "rewards/rejected": -1.4099724292755127,
+      "step": 136
+    },
+    {
+      "epoch": 0.2488646684831971,
+      "grad_norm": 1.1634560823440552,
+      "learning_rate": 9.914159460504512e-06,
+      "logits/chosen": 0.17681393027305603,
+      "logits/rejected": 0.2249586582183838,
+      "logps/chosen": -66.17890167236328,
+      "logps/rejected": -71.03226470947266,
+      "loss": 2.5281,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0759899616241455,
+      "rewards/margins": 0.21175377070903778,
+      "rewards/rejected": -1.2877436876296997,
+      "step": 137
+    },
+    {
+      "epoch": 0.2506811989100817,
+      "grad_norm": 2.4183638095855713,
+      "learning_rate": 9.912905813133325e-06,
+      "logits/chosen": 0.053306616842746735,
+      "logits/rejected": 0.08808214217424393,
+      "logps/chosen": -75.88124084472656,
+      "logps/rejected": -86.6006088256836,
+      "loss": 2.5718,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1421475410461426,
+      "rewards/margins": 0.36426350474357605,
+      "rewards/rejected": -1.506411075592041,
+      "step": 138
+    },
+    {
+      "epoch": 0.2524977293369664,
+      "grad_norm": 1.1398063898086548,
+      "learning_rate": 9.911643167078827e-06,
+      "logits/chosen": 0.046535998582839966,
+      "logits/rejected": 0.11657831072807312,
+      "logps/chosen": -62.68219757080078,
+      "logps/rejected": -74.52103424072266,
+      "loss": 2.5334,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.0171339511871338,
+      "rewards/margins": 0.3033108115196228,
+      "rewards/rejected": -1.3204445838928223,
+      "step": 139
+    },
+    {
+      "epoch": 0.254314259763851,
+      "grad_norm": 1.2559521198272705,
+      "learning_rate": 9.91037152491577e-06,
+      "logits/chosen": 0.07763661444187164,
+      "logits/rejected": 0.11103180050849915,
+      "logps/chosen": -68.87091064453125,
+      "logps/rejected": -79.08797454833984,
+      "loss": 2.454,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.017935872077942,
+      "rewards/margins": 0.2824122905731201,
+      "rewards/rejected": -1.300348162651062,
+      "step": 140
+    },
+    {
+      "epoch": 0.2561307901907357,
+      "grad_norm": 1.1818301677703857,
+      "learning_rate": 9.909090889237257e-06,
+      "logits/chosen": 0.07567673176527023,
+      "logits/rejected": 0.1396160125732422,
+      "logps/chosen": -62.44426345825195,
+      "logps/rejected": -61.77680206298828,
+      "loss": 2.4511,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -0.9497895240783691,
+      "rewards/margins": 0.23719964921474457,
+      "rewards/rejected": -1.1869890689849854,
+      "step": 141
+    },
+    {
+      "epoch": 0.2579473206176203,
+      "grad_norm": 1.1855460405349731,
+      "learning_rate": 9.907801262654725e-06,
+      "logits/chosen": 0.08448053896427155,
+      "logits/rejected": 0.1705108880996704,
+      "logps/chosen": -66.66730499267578,
+      "logps/rejected": -76.75447845458984,
+      "loss": 2.4476,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.105279803276062,
+      "rewards/margins": 0.2967475950717926,
+      "rewards/rejected": -1.4020274877548218,
+      "step": 142
+    },
+    {
+      "epoch": 0.259763851044505,
+      "grad_norm": 1.3652665615081787,
+      "learning_rate": 9.906502647797946e-06,
+      "logits/chosen": 0.09915038198232651,
+      "logits/rejected": 0.08110683411359787,
+      "logps/chosen": -66.32193756103516,
+      "logps/rejected": -80.20841217041016,
+      "loss": 2.5982,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -1.1550779342651367,
+      "rewards/margins": 0.36828702688217163,
+      "rewards/rejected": -1.5233650207519531,
+      "step": 143
+    },
+    {
+      "epoch": 0.2615803814713896,
+      "grad_norm": 1.4188799858093262,
+      "learning_rate": 9.905195047315024e-06,
+      "logits/chosen": 0.12019304931163788,
+      "logits/rejected": 0.11188551783561707,
+      "logps/chosen": -85.44612884521484,
+      "logps/rejected": -91.55411529541016,
+      "loss": 2.8244,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.171909213066101,
+      "rewards/margins": 0.16801074147224426,
+      "rewards/rejected": -1.3399198055267334,
+      "step": 144
+    },
+    {
+      "epoch": 0.2633969118982743,
+      "grad_norm": 1.307726502418518,
+      "learning_rate": 9.903878463872384e-06,
+      "logits/chosen": 0.12220380455255508,
+      "logits/rejected": 0.08115807920694351,
+      "logps/chosen": -71.56598663330078,
+      "logps/rejected": -71.203857421875,
+      "loss": 2.5832,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1082667112350464,
+      "rewards/margins": 0.18824435770511627,
+      "rewards/rejected": -1.296510934829712,
+      "step": 145
+    },
+    {
+      "epoch": 0.2652134423251589,
+      "grad_norm": 1.0912556648254395,
+      "learning_rate": 9.902552900154769e-06,
+      "logits/chosen": 0.14343701303005219,
+      "logits/rejected": 0.17181995511054993,
+      "logps/chosen": -65.52398681640625,
+      "logps/rejected": -72.67620086669922,
+      "loss": 2.3888,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0277646780014038,
+      "rewards/margins": 0.34770333766937256,
+      "rewards/rejected": -1.375468134880066,
+      "step": 146
+    },
+    {
+      "epoch": 0.2670299727520436,
+      "grad_norm": 1.375835657119751,
+      "learning_rate": 9.90121835886523e-06,
+      "logits/chosen": 0.06410901993513107,
+      "logits/rejected": 0.09152361750602722,
+      "logps/chosen": -78.27774047851562,
+      "logps/rejected": -82.27354431152344,
+      "loss": 2.5421,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.0331169366836548,
+      "rewards/margins": 0.20058104395866394,
+      "rewards/rejected": -1.2336980104446411,
+      "step": 147
+    },
+    {
+      "epoch": 0.2688465031789282,
+      "grad_norm": 1.1775563955307007,
+      "learning_rate": 9.899874842725136e-06,
+      "logits/chosen": 0.15871602296829224,
+      "logits/rejected": 0.12121336162090302,
+      "logps/chosen": -68.25074005126953,
+      "logps/rejected": -69.08786010742188,
+      "loss": 2.5082,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.1035441160202026,
+      "rewards/margins": 0.2747136950492859,
+      "rewards/rejected": -1.3782578706741333,
+      "step": 148
+    },
+    {
+      "epoch": 0.2706630336058129,
+      "grad_norm": 1.1836682558059692,
+      "learning_rate": 9.898522354474144e-06,
+      "logits/chosen": 0.10703336447477341,
+      "logits/rejected": 0.09559071063995361,
+      "logps/chosen": -67.44327545166016,
+      "logps/rejected": -66.1326904296875,
+      "loss": 2.4807,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0860165357589722,
+      "rewards/margins": 0.24197286367416382,
+      "rewards/rejected": -1.3279893398284912,
+      "step": 149
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 1.4430557489395142,
+      "learning_rate": 9.897160896870217e-06,
+      "logits/chosen": 0.1853700429201126,
+      "logits/rejected": 0.2392357438802719,
+      "logps/chosen": -69.39933776855469,
+      "logps/rejected": -76.33921813964844,
+      "loss": 2.6237,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2587306499481201,
+      "rewards/margins": 0.1618267297744751,
+      "rewards/rejected": -1.4205572605133057,
+      "step": 150
+    },
+    {
+      "epoch": 0.2742960944595822,
+      "grad_norm": 1.3623749017715454,
+      "learning_rate": 9.895790472689605e-06,
+      "logits/chosen": 0.15992893278598785,
+      "logits/rejected": 0.09628183394670486,
+      "logps/chosen": -70.02740478515625,
+      "logps/rejected": -66.36286926269531,
+      "loss": 2.7214,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.1345252990722656,
+      "rewards/margins": 0.1051594465970993,
+      "rewards/rejected": -1.2396849393844604,
+      "step": 151
+    },
+    {
+      "epoch": 0.2761126248864669,
+      "grad_norm": 1.3764369487762451,
+      "learning_rate": 9.894411084726837e-06,
+      "logits/chosen": 0.14009161293506622,
+      "logits/rejected": 0.13651950657367706,
+      "logps/chosen": -74.57782745361328,
+      "logps/rejected": -79.09630584716797,
+      "loss": 2.5094,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.1602882146835327,
+      "rewards/margins": 0.3822028338909149,
+      "rewards/rejected": -1.54249107837677,
+      "step": 152
+    },
+    {
+      "epoch": 0.2779291553133515,
+      "grad_norm": 1.3273645639419556,
+      "learning_rate": 9.893022735794728e-06,
+      "logits/chosen": 0.03166107460856438,
+      "logits/rejected": 0.059291813522577286,
+      "logps/chosen": -75.47296142578125,
+      "logps/rejected": -88.63102722167969,
+      "loss": 2.4789,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.10808265209198,
+      "rewards/margins": 0.3505283296108246,
+      "rewards/rejected": -1.4586111307144165,
+      "step": 153
+    },
+    {
+      "epoch": 0.27974568574023617,
+      "grad_norm": 1.1496751308441162,
+      "learning_rate": 9.891625428724365e-06,
+      "logits/chosen": 0.13897705078125,
+      "logits/rejected": 0.13928522169589996,
+      "logps/chosen": -65.34259796142578,
+      "logps/rejected": -68.99885559082031,
+      "loss": 2.2804,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.1391716003417969,
+      "rewards/margins": 0.3395899534225464,
+      "rewards/rejected": -1.4787613153457642,
+      "step": 154
+    },
+    {
+      "epoch": 0.2815622161671208,
+      "grad_norm": 1.3669211864471436,
+      "learning_rate": 9.890219166365097e-06,
+      "logits/chosen": 0.08065556734800339,
+      "logits/rejected": 0.14106576144695282,
+      "logps/chosen": -72.62825012207031,
+      "logps/rejected": -77.62796020507812,
+      "loss": 2.5929,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.0997321605682373,
+      "rewards/margins": 0.20812571048736572,
+      "rewards/rejected": -1.3078577518463135,
+      "step": 155
+    },
+    {
+      "epoch": 0.28337874659400547,
+      "grad_norm": 1.2757997512817383,
+      "learning_rate": 9.888803951584537e-06,
+      "logits/chosen": 0.0465204194188118,
+      "logits/rejected": 0.06447532027959824,
+      "logps/chosen": -66.84536743164062,
+      "logps/rejected": -78.97770690917969,
+      "loss": 2.3491,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.1228134632110596,
+      "rewards/margins": 0.4900767207145691,
+      "rewards/rejected": -1.612890362739563,
+      "step": 156
+    },
+    {
+      "epoch": 0.2851952770208901,
+      "grad_norm": 1.391892910003662,
+      "learning_rate": 9.887379787268558e-06,
+      "logits/chosen": 0.08740498870611191,
+      "logits/rejected": 0.11772032082080841,
+      "logps/chosen": -67.31288146972656,
+      "logps/rejected": -70.06088256835938,
+      "loss": 2.6179,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.2256510257720947,
+      "rewards/margins": 0.19449593126773834,
+      "rewards/rejected": -1.4201467037200928,
+      "step": 157
+    },
+    {
+      "epoch": 0.28701180744777477,
+      "grad_norm": 1.445900321006775,
+      "learning_rate": 9.885946676321279e-06,
+      "logits/chosen": 0.1524508148431778,
+      "logits/rejected": 0.21413244307041168,
+      "logps/chosen": -74.10218811035156,
+      "logps/rejected": -76.91876220703125,
+      "loss": 2.5262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.1993122100830078,
+      "rewards/margins": 0.33456599712371826,
+      "rewards/rejected": -1.5338780879974365,
+      "step": 158
+    },
+    {
+      "epoch": 0.2888283378746594,
+      "grad_norm": 1.361208438873291,
+      "learning_rate": 9.884504621665059e-06,
+      "logits/chosen": 0.1192079707980156,
+      "logits/rejected": 0.21478833258152008,
+      "logps/chosen": -75.69478607177734,
+      "logps/rejected": -84.40799713134766,
+      "loss": 2.3984,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2067720890045166,
+      "rewards/margins": 0.34302443265914917,
+      "rewards/rejected": -1.5497965812683105,
+      "step": 159
+    },
+    {
+      "epoch": 0.29064486830154407,
+      "grad_norm": 3.4042046070098877,
+      "learning_rate": 9.883053626240503e-06,
+      "logits/chosen": 0.13823604583740234,
+      "logits/rejected": 0.0957983061671257,
+      "logps/chosen": -85.02832794189453,
+      "logps/rejected": -86.334716796875,
+      "loss": 3.0326,
+      "rewards/accuracies": 0.484375,
+      "rewards/chosen": -1.1793922185897827,
+      "rewards/margins": 0.013544075191020966,
+      "rewards/rejected": -1.1929364204406738,
+      "step": 160
+    },
+    {
+      "epoch": 0.2924613987284287,
+      "grad_norm": 1.7824169397354126,
+      "learning_rate": 9.881593693006438e-06,
+      "logits/chosen": 0.09778247773647308,
+      "logits/rejected": 0.12810076773166656,
+      "logps/chosen": -78.78877258300781,
+      "logps/rejected": -78.63519287109375,
+      "loss": 3.1665,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.360141396522522,
+      "rewards/margins": 0.148757204413414,
+      "rewards/rejected": -1.5088986158370972,
+      "step": 161
+    },
+    {
+      "epoch": 0.29427792915531337,
+      "grad_norm": 1.4275903701782227,
+      "learning_rate": 9.880124824939927e-06,
+      "logits/chosen": 0.17759747803211212,
+      "logits/rejected": 0.14611241221427917,
+      "logps/chosen": -70.13676452636719,
+      "logps/rejected": -67.08554077148438,
+      "loss": 2.8382,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.1756031513214111,
+      "rewards/margins": 0.14519944787025452,
+      "rewards/rejected": -1.3208026885986328,
+      "step": 162
+    },
+    {
+      "epoch": 0.296094459582198,
+      "grad_norm": 1.3242970705032349,
+      "learning_rate": 9.878647025036245e-06,
+      "logits/chosen": 0.07917390763759613,
+      "logits/rejected": 0.16502070426940918,
+      "logps/chosen": -77.08013916015625,
+      "logps/rejected": -92.47212982177734,
+      "loss": 2.3183,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.2245774269104004,
+      "rewards/margins": 0.573800802230835,
+      "rewards/rejected": -1.7983782291412354,
+      "step": 163
+    },
+    {
+      "epoch": 0.29791099000908267,
+      "grad_norm": 1.160335659980774,
+      "learning_rate": 9.877160296308886e-06,
+      "logits/chosen": 0.06782057881355286,
+      "logits/rejected": 0.09083382785320282,
+      "logps/chosen": -61.769866943359375,
+      "logps/rejected": -66.22183990478516,
+      "loss": 2.3323,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.2067499160766602,
+      "rewards/margins": 0.36826351284980774,
+      "rewards/rejected": -1.575013279914856,
+      "step": 164
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 1.2561142444610596,
+      "learning_rate": 9.875664641789545e-06,
+      "logits/chosen": 0.10073137283325195,
+      "logits/rejected": 0.12745651602745056,
+      "logps/chosen": -68.94499206542969,
+      "logps/rejected": -76.13021850585938,
+      "loss": 2.4058,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2001702785491943,
+      "rewards/margins": 0.31210076808929443,
+      "rewards/rejected": -1.5122709274291992,
+      "step": 165
+    },
+    {
+      "epoch": 0.30154405086285196,
+      "grad_norm": 1.2163783311843872,
+      "learning_rate": 9.874160064528124e-06,
+      "logits/chosen": 0.1380203366279602,
+      "logits/rejected": 0.20783495903015137,
+      "logps/chosen": -62.44160461425781,
+      "logps/rejected": -70.53887176513672,
+      "loss": 2.2759,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.243611216545105,
+      "rewards/margins": 0.46271997690200806,
+      "rewards/rejected": -1.7063312530517578,
+      "step": 166
+    },
+    {
+      "epoch": 0.3033605812897366,
+      "grad_norm": 1.2518328428268433,
+      "learning_rate": 9.872646567592719e-06,
+      "logits/chosen": 0.13933810591697693,
+      "logits/rejected": 0.1436997652053833,
+      "logps/chosen": -69.45280456542969,
+      "logps/rejected": -78.73098754882812,
+      "loss": 2.368,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1905428171157837,
+      "rewards/margins": 0.42389774322509766,
+      "rewards/rejected": -1.6144405603408813,
+      "step": 167
+    },
+    {
+      "epoch": 0.30517711171662126,
+      "grad_norm": 1.2762094736099243,
+      "learning_rate": 9.871124154069613e-06,
+      "logits/chosen": 0.11822449415922165,
+      "logits/rejected": 0.13434451818466187,
+      "logps/chosen": -72.77944946289062,
+      "logps/rejected": -75.60210418701172,
+      "loss": 2.5178,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.304598331451416,
+      "rewards/margins": 0.21065138280391693,
+      "rewards/rejected": -1.5152498483657837,
+      "step": 168
+    },
+    {
+      "epoch": 0.3069936421435059,
+      "grad_norm": 1.7200556993484497,
+      "learning_rate": 9.86959282706327e-06,
+      "logits/chosen": 0.12234638631343842,
+      "logits/rejected": 0.123184435069561,
+      "logps/chosen": -84.78328704833984,
+      "logps/rejected": -82.90792083740234,
+      "loss": 2.7762,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.4255175590515137,
+      "rewards/margins": 0.3239808976650238,
+      "rewards/rejected": -1.7494984865188599,
+      "step": 169
+    },
+    {
+      "epoch": 0.30881017257039056,
+      "grad_norm": 1.2458773851394653,
+      "learning_rate": 9.868052589696337e-06,
+      "logits/chosen": 0.14292597770690918,
+      "logits/rejected": 0.1542571634054184,
+      "logps/chosen": -66.72957611083984,
+      "logps/rejected": -76.97203826904297,
+      "loss": 2.419,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.2561144828796387,
+      "rewards/margins": 0.39490899443626404,
+      "rewards/rejected": -1.6510233879089355,
+      "step": 170
+    },
+    {
+      "epoch": 0.3106267029972752,
+      "grad_norm": 1.192762017250061,
+      "learning_rate": 9.866503445109621e-06,
+      "logits/chosen": 0.12845008075237274,
+      "logits/rejected": 0.10619282722473145,
+      "logps/chosen": -64.939453125,
+      "logps/rejected": -70.89356231689453,
+      "loss": 2.3475,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.2126998901367188,
+      "rewards/margins": 0.39672625064849854,
+      "rewards/rejected": -1.6094262599945068,
+      "step": 171
+    },
+    {
+      "epoch": 0.31244323342415986,
+      "grad_norm": 1.6527279615402222,
+      "learning_rate": 9.864945396462101e-06,
+      "logits/chosen": 0.03876817971467972,
+      "logits/rejected": 0.01723310723900795,
+      "logps/chosen": -80.94947052001953,
+      "logps/rejected": -83.43637084960938,
+      "loss": 2.6719,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.203782320022583,
+      "rewards/margins": 0.2551845610141754,
+      "rewards/rejected": -1.458966851234436,
+      "step": 172
+    },
+    {
+      "epoch": 0.3142597638510445,
+      "grad_norm": 1.427216649055481,
+      "learning_rate": 9.86337844693091e-06,
+      "logits/chosen": 0.08040126413106918,
+      "logits/rejected": 0.13297931849956512,
+      "logps/chosen": -67.49502563476562,
+      "logps/rejected": -77.2923812866211,
+      "loss": 2.4931,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.3284540176391602,
+      "rewards/margins": 0.4811919927597046,
+      "rewards/rejected": -1.8096460103988647,
+      "step": 173
+    },
+    {
+      "epoch": 0.31607629427792916,
+      "grad_norm": 1.3090578317642212,
+      "learning_rate": 9.861802599711329e-06,
+      "logits/chosen": 0.109119713306427,
+      "logits/rejected": 0.07613471150398254,
+      "logps/chosen": -71.72999572753906,
+      "logps/rejected": -73.84215545654297,
+      "loss": 2.4972,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2370662689208984,
+      "rewards/margins": 0.27422747015953064,
+      "rewards/rejected": -1.5112937688827515,
+      "step": 174
+    },
+    {
+      "epoch": 0.3178928247048138,
+      "grad_norm": 1.351342797279358,
+      "learning_rate": 9.860217858016783e-06,
+      "logits/chosen": 0.10673967003822327,
+      "logits/rejected": 0.1354019045829773,
+      "logps/chosen": -70.85772705078125,
+      "logps/rejected": -81.281982421875,
+      "loss": 2.4456,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2566276788711548,
+      "rewards/margins": 0.34767431020736694,
+      "rewards/rejected": -1.604301929473877,
+      "step": 175
+    },
+    {
+      "epoch": 0.31970935513169846,
+      "grad_norm": 1.3725523948669434,
+      "learning_rate": 9.858624225078841e-06,
+      "logits/chosen": 0.12088489532470703,
+      "logits/rejected": 0.0771353617310524,
+      "logps/chosen": -74.0596694946289,
+      "logps/rejected": -71.7733154296875,
+      "loss": 2.5003,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1832196712493896,
+      "rewards/margins": 0.31638142466545105,
+      "rewards/rejected": -1.4996010065078735,
+      "step": 176
+    },
+    {
+      "epoch": 0.3215258855585831,
+      "grad_norm": 1.3968805074691772,
+      "learning_rate": 9.857021704147195e-06,
+      "logits/chosen": 0.11635589599609375,
+      "logits/rejected": 0.09500478953123093,
+      "logps/chosen": -75.86662292480469,
+      "logps/rejected": -75.28397369384766,
+      "loss": 2.5589,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.3443347215652466,
+      "rewards/margins": 0.22420868277549744,
+      "rewards/rejected": -1.5685434341430664,
+      "step": 177
+    },
+    {
+      "epoch": 0.32334241598546776,
+      "grad_norm": 1.3207515478134155,
+      "learning_rate": 9.855410298489663e-06,
+      "logits/chosen": 0.032826680690050125,
+      "logits/rejected": 0.0877794623374939,
+      "logps/chosen": -67.0349349975586,
+      "logps/rejected": -75.06578063964844,
+      "loss": 2.4595,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.2243653535842896,
+      "rewards/margins": 0.36731529235839844,
+      "rewards/rejected": -1.591680645942688,
+      "step": 178
+    },
+    {
+      "epoch": 0.32515894641235243,
+      "grad_norm": 1.3098173141479492,
+      "learning_rate": 9.853790011392186e-06,
+      "logits/chosen": 0.08737780898809433,
+      "logits/rejected": 0.08714289963245392,
+      "logps/chosen": -71.3626708984375,
+      "logps/rejected": -81.07239532470703,
+      "loss": 2.2606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.1984796524047852,
+      "rewards/margins": 0.4800388514995575,
+      "rewards/rejected": -1.678518533706665,
+      "step": 179
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 1.3656742572784424,
+      "learning_rate": 9.852160846158808e-06,
+      "logits/chosen": 0.08435464650392532,
+      "logits/rejected": 0.13674329221248627,
+      "logps/chosen": -69.021484375,
+      "logps/rejected": -79.02738952636719,
+      "loss": 2.4039,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4128899574279785,
+      "rewards/margins": 0.3882848620414734,
+      "rewards/rejected": -1.8011748790740967,
+      "step": 180
+    },
+    {
+      "epoch": 0.32879200726612173,
+      "grad_norm": 1.2837380170822144,
+      "learning_rate": 9.850522806111681e-06,
+      "logits/chosen": 0.20875662565231323,
+      "logits/rejected": 0.1668507307767868,
+      "logps/chosen": -68.29991149902344,
+      "logps/rejected": -67.59306335449219,
+      "loss": 2.5716,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -1.3281925916671753,
+      "rewards/margins": 0.21396151185035706,
+      "rewards/rejected": -1.54215407371521,
+      "step": 181
+    },
+    {
+      "epoch": 0.33060853769300635,
+      "grad_norm": 1.3804494142532349,
+      "learning_rate": 9.848875894591055e-06,
+      "logits/chosen": 0.12790340185165405,
+      "logits/rejected": 0.1156705766916275,
+      "logps/chosen": -75.76934814453125,
+      "logps/rejected": -76.44409942626953,
+      "loss": 2.2623,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.260763168334961,
+      "rewards/margins": 0.45292162895202637,
+      "rewards/rejected": -1.7136849164962769,
+      "step": 182
+    },
+    {
+      "epoch": 0.33242506811989103,
+      "grad_norm": 1.287442922592163,
+      "learning_rate": 9.847220114955269e-06,
+      "logits/chosen": 0.1627904772758484,
+      "logits/rejected": 0.1957186460494995,
+      "logps/chosen": -67.52108001708984,
+      "logps/rejected": -79.52376556396484,
+      "loss": 2.348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2697941064834595,
+      "rewards/margins": 0.5318787097930908,
+      "rewards/rejected": -1.8016728162765503,
+      "step": 183
+    },
+    {
+      "epoch": 0.33424159854677565,
+      "grad_norm": 1.3078449964523315,
+      "learning_rate": 9.845555470580746e-06,
+      "logits/chosen": 0.08075303584337234,
+      "logits/rejected": 0.08640636503696442,
+      "logps/chosen": -64.12767791748047,
+      "logps/rejected": -69.04762268066406,
+      "loss": 2.4969,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.2283066511154175,
+      "rewards/margins": 0.31595146656036377,
+      "rewards/rejected": -1.5442581176757812,
+      "step": 184
+    },
+    {
+      "epoch": 0.33605812897366033,
+      "grad_norm": 1.649794578552246,
+      "learning_rate": 9.843881964861985e-06,
+      "logits/chosen": 0.07760760188102722,
+      "logits/rejected": 0.11186876147985458,
+      "logps/chosen": -75.8207015991211,
+      "logps/rejected": -81.5796890258789,
+      "loss": 2.5595,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4034444093704224,
+      "rewards/margins": 0.33963268995285034,
+      "rewards/rejected": -1.743077039718628,
+      "step": 185
+    },
+    {
+      "epoch": 0.33787465940054495,
+      "grad_norm": 1.2912336587905884,
+      "learning_rate": 9.842199601211556e-06,
+      "logits/chosen": 0.133261039853096,
+      "logits/rejected": 0.1473626047372818,
+      "logps/chosen": -69.50003051757812,
+      "logps/rejected": -77.0198745727539,
+      "loss": 2.3544,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.2854596376419067,
+      "rewards/margins": 0.4699360728263855,
+      "rewards/rejected": -1.7553956508636475,
+      "step": 186
+    },
+    {
+      "epoch": 0.33969118982742963,
+      "grad_norm": 1.4907958507537842,
+      "learning_rate": 9.840508383060092e-06,
+      "logits/chosen": 0.06240752339363098,
+      "logits/rejected": 0.15645891427993774,
+      "logps/chosen": -66.49507141113281,
+      "logps/rejected": -72.68167114257812,
+      "loss": 2.5754,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.3408069610595703,
+      "rewards/margins": 0.2825961112976074,
+      "rewards/rejected": -1.6234029531478882,
+      "step": 187
+    },
+    {
+      "epoch": 0.34150772025431425,
+      "grad_norm": 1.528688669204712,
+      "learning_rate": 9.838808313856281e-06,
+      "logits/chosen": -0.010553614236414433,
+      "logits/rejected": -0.0162151250988245,
+      "logps/chosen": -87.83446502685547,
+      "logps/rejected": -82.62303161621094,
+      "loss": 2.6229,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.2400531768798828,
+      "rewards/margins": 0.21957488358020782,
+      "rewards/rejected": -1.4596279859542847,
+      "step": 188
+    },
+    {
+      "epoch": 0.34332425068119893,
+      "grad_norm": 1.5255874395370483,
+      "learning_rate": 9.83709939706686e-06,
+      "logits/chosen": 0.12232109159231186,
+      "logits/rejected": 0.08726370334625244,
+      "logps/chosen": -66.86070251464844,
+      "logps/rejected": -70.5438232421875,
+      "loss": 2.7017,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.433100700378418,
+      "rewards/margins": 0.20580635964870453,
+      "rewards/rejected": -1.6389069557189941,
+      "step": 189
+    },
+    {
+      "epoch": 0.34514078110808355,
+      "grad_norm": 1.4504551887512207,
+      "learning_rate": 9.835381636176604e-06,
+      "logits/chosen": 0.1462351530790329,
+      "logits/rejected": 0.13504080474376678,
+      "logps/chosen": -76.45362091064453,
+      "logps/rejected": -78.51449584960938,
+      "loss": 2.5936,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.356174349784851,
+      "rewards/margins": 0.21241840720176697,
+      "rewards/rejected": -1.5685927867889404,
+      "step": 190
+    },
+    {
+      "epoch": 0.3469573115349682,
+      "grad_norm": 1.4351036548614502,
+      "learning_rate": 9.833655034688336e-06,
+      "logits/chosen": 0.1534399539232254,
+      "logits/rejected": 0.178826704621315,
+      "logps/chosen": -66.26270294189453,
+      "logps/rejected": -70.225830078125,
+      "loss": 2.6442,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.3860033750534058,
+      "rewards/margins": 0.16981087625026703,
+      "rewards/rejected": -1.555814266204834,
+      "step": 191
+    },
+    {
+      "epoch": 0.34877384196185285,
+      "grad_norm": 1.295323133468628,
+      "learning_rate": 9.831919596122888e-06,
+      "logits/chosen": 0.11593925207853317,
+      "logits/rejected": 0.19400468468666077,
+      "logps/chosen": -66.74591064453125,
+      "logps/rejected": -72.82708740234375,
+      "loss": 2.3599,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.2114710807800293,
+      "rewards/margins": 0.34326374530792236,
+      "rewards/rejected": -1.5547348260879517,
+      "step": 192
+    },
+    {
+      "epoch": 0.3505903723887375,
+      "grad_norm": 1.452673077583313,
+      "learning_rate": 9.830175324019125e-06,
+      "logits/chosen": 0.13779595494270325,
+      "logits/rejected": 0.15601256489753723,
+      "logps/chosen": -73.34432220458984,
+      "logps/rejected": -76.34349822998047,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.409874677658081,
+      "rewards/margins": 0.3434732258319855,
+      "rewards/rejected": -1.7533478736877441,
+      "step": 193
+    },
+    {
+      "epoch": 0.35240690281562215,
+      "grad_norm": 1.3285003900527954,
+      "learning_rate": 9.828422221933924e-06,
+      "logits/chosen": 0.020087052136659622,
+      "logits/rejected": 0.07995946705341339,
+      "logps/chosen": -72.9058837890625,
+      "logps/rejected": -81.98945617675781,
+      "loss": 2.3612,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.2450451850891113,
+      "rewards/margins": 0.40142494440078735,
+      "rewards/rejected": -1.646470069885254,
+      "step": 194
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 1.2274519205093384,
+      "learning_rate": 9.826660293442158e-06,
+      "logits/chosen": 0.0241906326264143,
+      "logits/rejected": 0.09953958541154861,
+      "logps/chosen": -66.5189437866211,
+      "logps/rejected": -75.99092102050781,
+      "loss": 2.2288,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3973838090896606,
+      "rewards/margins": 0.5354034304618835,
+      "rewards/rejected": -1.9327871799468994,
+      "step": 195
+    },
+    {
+      "epoch": 0.35603996366939145,
+      "grad_norm": 1.3833938837051392,
+      "learning_rate": 9.824889542136714e-06,
+      "logits/chosen": 0.086525097489357,
+      "logits/rejected": 0.08451628684997559,
+      "logps/chosen": -74.15232849121094,
+      "logps/rejected": -79.37950897216797,
+      "loss": 2.4671,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.4214783906936646,
+      "rewards/margins": 0.3752206265926361,
+      "rewards/rejected": -1.7966989278793335,
+      "step": 196
+    },
+    {
+      "epoch": 0.3578564940962761,
+      "grad_norm": 1.6680957078933716,
+      "learning_rate": 9.823109971628459e-06,
+      "logits/chosen": 0.06370481848716736,
+      "logits/rejected": 0.12739142775535583,
+      "logps/chosen": -75.81684875488281,
+      "logps/rejected": -76.65689086914062,
+      "loss": 2.3705,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.3144075870513916,
+      "rewards/margins": 0.36334556341171265,
+      "rewards/rejected": -1.6777533292770386,
+      "step": 197
+    },
+    {
+      "epoch": 0.35967302452316074,
+      "grad_norm": 1.4894323348999023,
+      "learning_rate": 9.821321585546244e-06,
+      "logits/chosen": 0.10617184638977051,
+      "logits/rejected": 0.11465627700090408,
+      "logps/chosen": -73.0533218383789,
+      "logps/rejected": -83.94466400146484,
+      "loss": 2.2915,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4909334182739258,
+      "rewards/margins": 0.5887378454208374,
+      "rewards/rejected": -2.0796711444854736,
+      "step": 198
+    },
+    {
+      "epoch": 0.3614895549500454,
+      "grad_norm": 1.422853946685791,
+      "learning_rate": 9.819524387536905e-06,
+      "logits/chosen": 0.069038525223732,
+      "logits/rejected": 0.09677774459123611,
+      "logps/chosen": -85.49876403808594,
+      "logps/rejected": -89.743408203125,
+      "loss": 2.3,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.3645334243774414,
+      "rewards/margins": 0.5102630853652954,
+      "rewards/rejected": -1.8747965097427368,
+      "step": 199
+    },
+    {
+      "epoch": 0.36330608537693004,
+      "grad_norm": 1.3350121974945068,
+      "learning_rate": 9.81771838126524e-06,
+      "logits/chosen": 0.0018447795882821083,
+      "logits/rejected": 0.055721428245306015,
+      "logps/chosen": -71.9021224975586,
+      "logps/rejected": -81.6389389038086,
+      "loss": 2.2872,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.3360776901245117,
+      "rewards/margins": 0.46150827407836914,
+      "rewards/rejected": -1.7975859642028809,
+      "step": 200
+    },
+    {
+      "epoch": 0.3651226158038147,
+      "grad_norm": 1.458803653717041,
+      "learning_rate": 9.815903570414006e-06,
+      "logits/chosen": 0.059184275567531586,
+      "logits/rejected": 0.07923795282840729,
+      "logps/chosen": -80.88011932373047,
+      "logps/rejected": -86.74174499511719,
+      "loss": 2.4279,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.6620349884033203,
+      "rewards/margins": 0.38666611909866333,
+      "rewards/rejected": -2.048701047897339,
+      "step": 201
+    },
+    {
+      "epoch": 0.36693914623069934,
+      "grad_norm": 1.3583872318267822,
+      "learning_rate": 9.814079958683925e-06,
+      "logits/chosen": 0.11471173167228699,
+      "logits/rejected": 0.1523369550704956,
+      "logps/chosen": -71.39076232910156,
+      "logps/rejected": -78.50016021728516,
+      "loss": 2.4603,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4875473976135254,
+      "rewards/margins": 0.3805108070373535,
+      "rewards/rejected": -1.8680580854415894,
+      "step": 202
+    },
+    {
+      "epoch": 0.368755676657584,
+      "grad_norm": 1.6755985021591187,
+      "learning_rate": 9.812247549793656e-06,
+      "logits/chosen": 0.14959998428821564,
+      "logits/rejected": 0.1801329255104065,
+      "logps/chosen": -76.0824203491211,
+      "logps/rejected": -85.81806945800781,
+      "loss": 2.8134,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7176380157470703,
+      "rewards/margins": 0.28729167580604553,
+      "rewards/rejected": -2.004929780960083,
+      "step": 203
+    },
+    {
+      "epoch": 0.37057220708446864,
+      "grad_norm": 1.4551233053207397,
+      "learning_rate": 9.810406347479798e-06,
+      "logits/chosen": 0.08063512295484543,
+      "logits/rejected": 0.03579093888401985,
+      "logps/chosen": -87.19414520263672,
+      "logps/rejected": -89.355224609375,
+      "loss": 2.4092,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5363514423370361,
+      "rewards/margins": 0.4680227041244507,
+      "rewards/rejected": -2.0043740272521973,
+      "step": 204
+    },
+    {
+      "epoch": 0.3723887375113533,
+      "grad_norm": 1.4248968362808228,
+      "learning_rate": 9.808556355496885e-06,
+      "logits/chosen": 0.06655821204185486,
+      "logits/rejected": 0.050458114594221115,
+      "logps/chosen": -92.04095458984375,
+      "logps/rejected": -95.39706420898438,
+      "loss": 2.3507,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5707283020019531,
+      "rewards/margins": 0.388569712638855,
+      "rewards/rejected": -1.959298014640808,
+      "step": 205
+    },
+    {
+      "epoch": 0.37420526793823794,
+      "grad_norm": 1.3779215812683105,
+      "learning_rate": 9.806697577617371e-06,
+      "logits/chosen": 0.09702017903327942,
+      "logits/rejected": 0.13923662900924683,
+      "logps/chosen": -78.27027893066406,
+      "logps/rejected": -85.93509674072266,
+      "loss": 2.1888,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.4591946601867676,
+      "rewards/margins": 0.5624303817749023,
+      "rewards/rejected": -2.02162504196167,
+      "step": 206
+    },
+    {
+      "epoch": 0.3760217983651226,
+      "grad_norm": 1.5306010246276855,
+      "learning_rate": 9.804830017631631e-06,
+      "logits/chosen": 0.038323137909173965,
+      "logits/rejected": 0.04337020218372345,
+      "logps/chosen": -75.31045532226562,
+      "logps/rejected": -83.26338958740234,
+      "loss": 2.7312,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4613063335418701,
+      "rewards/margins": 0.3355714678764343,
+      "rewards/rejected": -1.7968778610229492,
+      "step": 207
+    },
+    {
+      "epoch": 0.37783832879200724,
+      "grad_norm": 1.629341959953308,
+      "learning_rate": 9.802953679347943e-06,
+      "logits/chosen": 0.1168309897184372,
+      "logits/rejected": 0.21053184568881989,
+      "logps/chosen": -70.42684936523438,
+      "logps/rejected": -87.77594757080078,
+      "loss": 2.7324,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.8266733884811401,
+      "rewards/margins": 0.5579056739807129,
+      "rewards/rejected": -2.3845791816711426,
+      "step": 208
+    },
+    {
+      "epoch": 0.3796548592188919,
+      "grad_norm": 1.5751092433929443,
+      "learning_rate": 9.801068566592486e-06,
+      "logits/chosen": 0.11355097591876984,
+      "logits/rejected": 0.11962890625,
+      "logps/chosen": -87.75645446777344,
+      "logps/rejected": -96.15601348876953,
+      "loss": 2.5237,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.4331413507461548,
+      "rewards/margins": 0.38925477862358093,
+      "rewards/rejected": -1.8223960399627686,
+      "step": 209
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 1.6638984680175781,
+      "learning_rate": 9.799174683209336e-06,
+      "logits/chosen": 0.0970507487654686,
+      "logits/rejected": 0.07422082126140594,
+      "logps/chosen": -85.44343566894531,
+      "logps/rejected": -90.49578857421875,
+      "loss": 2.5179,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6639773845672607,
+      "rewards/margins": 0.4513780176639557,
+      "rewards/rejected": -2.1153552532196045,
+      "step": 210
+    },
+    {
+      "epoch": 0.3832879200726612,
+      "grad_norm": 1.389356017112732,
+      "learning_rate": 9.79727203306045e-06,
+      "logits/chosen": 0.027732742950320244,
+      "logits/rejected": 0.0795917734503746,
+      "logps/chosen": -85.95429229736328,
+      "logps/rejected": -92.91676330566406,
+      "loss": 2.3892,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.4551334381103516,
+      "rewards/margins": 0.4000782370567322,
+      "rewards/rejected": -1.8552117347717285,
+      "step": 211
+    },
+    {
+      "epoch": 0.3851044504995459,
+      "grad_norm": 1.438284993171692,
+      "learning_rate": 9.79536062002566e-06,
+      "logits/chosen": 0.04227686673402786,
+      "logits/rejected": 0.09655077010393143,
+      "logps/chosen": -77.67405700683594,
+      "logps/rejected": -88.0655746459961,
+      "loss": 2.3246,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5603740215301514,
+      "rewards/margins": 0.5005100965499878,
+      "rewards/rejected": -2.060884475708008,
+      "step": 212
+    },
+    {
+      "epoch": 0.3869209809264305,
+      "grad_norm": 1.35500168800354,
+      "learning_rate": 9.793440448002676e-06,
+      "logits/chosen": 0.10956872254610062,
+      "logits/rejected": 0.10562983900308609,
+      "logps/chosen": -74.20452117919922,
+      "logps/rejected": -73.23489379882812,
+      "loss": 2.4585,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.399983286857605,
+      "rewards/margins": 0.2682442367076874,
+      "rewards/rejected": -1.6682274341583252,
+      "step": 213
+    },
+    {
+      "epoch": 0.3887375113533152,
+      "grad_norm": 1.3036773204803467,
+      "learning_rate": 9.791511520907056e-06,
+      "logits/chosen": 0.06179399788379669,
+      "logits/rejected": 0.052528850734233856,
+      "logps/chosen": -72.4264144897461,
+      "logps/rejected": -72.77079010009766,
+      "loss": 2.3441,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.391230821609497,
+      "rewards/margins": 0.3969228267669678,
+      "rewards/rejected": -1.7881536483764648,
+      "step": 214
+    },
+    {
+      "epoch": 0.3905540417801998,
+      "grad_norm": 1.4877872467041016,
+      "learning_rate": 9.789573842672223e-06,
+      "logits/chosen": 0.05735350027680397,
+      "logits/rejected": 0.09979183971881866,
+      "logps/chosen": -82.134033203125,
+      "logps/rejected": -96.38580322265625,
+      "loss": 2.2192,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7035160064697266,
+      "rewards/margins": 0.6622112989425659,
+      "rewards/rejected": -2.365727424621582,
+      "step": 215
+    },
+    {
+      "epoch": 0.3923705722070845,
+      "grad_norm": 1.247817039489746,
+      "learning_rate": 9.787627417249441e-06,
+      "logits/chosen": 0.0014538783580064774,
+      "logits/rejected": 0.03942735865712166,
+      "logps/chosen": -73.8500747680664,
+      "logps/rejected": -95.03350067138672,
+      "loss": 1.9498,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.4928035736083984,
+      "rewards/margins": 0.9987993836402893,
+      "rewards/rejected": -2.491603374481201,
+      "step": 216
+    },
+    {
+      "epoch": 0.3941871026339691,
+      "grad_norm": 1.3884788751602173,
+      "learning_rate": 9.785672248607807e-06,
+      "logits/chosen": 0.029594585299491882,
+      "logits/rejected": 0.11805769056081772,
+      "logps/chosen": -66.9087905883789,
+      "logps/rejected": -80.82840728759766,
+      "loss": 2.2455,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6260508298873901,
+      "rewards/margins": 0.5117952227592468,
+      "rewards/rejected": -2.137845993041992,
+      "step": 217
+    },
+    {
+      "epoch": 0.3960036330608538,
+      "grad_norm": 2.9690237045288086,
+      "learning_rate": 9.78370834073425e-06,
+      "logits/chosen": 0.11608768254518509,
+      "logits/rejected": 0.11099248379468918,
+      "logps/chosen": -73.23339080810547,
+      "logps/rejected": -73.60865783691406,
+      "loss": 2.6526,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.5958513021469116,
+      "rewards/margins": 0.20439797639846802,
+      "rewards/rejected": -1.8002492189407349,
+      "step": 218
+    },
+    {
+      "epoch": 0.3978201634877384,
+      "grad_norm": 1.4823459386825562,
+      "learning_rate": 9.781735697633526e-06,
+      "logits/chosen": 0.07910319417715073,
+      "logits/rejected": 0.1617601215839386,
+      "logps/chosen": -72.02169799804688,
+      "logps/rejected": -79.70098114013672,
+      "loss": 2.3689,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.722070336341858,
+      "rewards/margins": 0.4829583168029785,
+      "rewards/rejected": -2.205028772354126,
+      "step": 219
+    },
+    {
+      "epoch": 0.3996366939146231,
+      "grad_norm": 2.249466896057129,
+      "learning_rate": 9.779754323328192e-06,
+      "logits/chosen": 0.17587795853614807,
+      "logits/rejected": 0.13541430234909058,
+      "logps/chosen": -75.69686889648438,
+      "logps/rejected": -77.53881072998047,
+      "loss": 3.2204,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.03781795501709,
+      "rewards/margins": 0.1416795551776886,
+      "rewards/rejected": -2.179497480392456,
+      "step": 220
+    },
+    {
+      "epoch": 0.4014532243415077,
+      "grad_norm": 1.4283430576324463,
+      "learning_rate": 9.777764221858616e-06,
+      "logits/chosen": 0.1529032289981842,
+      "logits/rejected": 0.1623322069644928,
+      "logps/chosen": -69.98622131347656,
+      "logps/rejected": -77.31781005859375,
+      "loss": 2.3358,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6996090412139893,
+      "rewards/margins": 0.568400502204895,
+      "rewards/rejected": -2.2680094242095947,
+      "step": 221
+    },
+    {
+      "epoch": 0.4032697547683924,
+      "grad_norm": 1.4971157312393188,
+      "learning_rate": 9.775765397282963e-06,
+      "logits/chosen": 0.13248610496520996,
+      "logits/rejected": 0.13485054671764374,
+      "logps/chosen": -70.01846313476562,
+      "logps/rejected": -75.87899017333984,
+      "loss": 2.4037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6077332496643066,
+      "rewards/margins": 0.39039328694343567,
+      "rewards/rejected": -1.99812650680542,
+      "step": 222
+    },
+    {
+      "epoch": 0.405086285195277,
+      "grad_norm": 1.590364694595337,
+      "learning_rate": 9.773757853677182e-06,
+      "logits/chosen": 0.08200166374444962,
+      "logits/rejected": 0.06919535249471664,
+      "logps/chosen": -85.51278686523438,
+      "logps/rejected": -89.3447265625,
+      "loss": 2.6149,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.5626994371414185,
+      "rewards/margins": 0.30814388394355774,
+      "rewards/rejected": -1.8708434104919434,
+      "step": 223
+    },
+    {
+      "epoch": 0.4069028156221617,
+      "grad_norm": 1.41611647605896,
+      "learning_rate": 9.771741595135009e-06,
+      "logits/chosen": 0.057254984974861145,
+      "logits/rejected": 0.13574014604091644,
+      "logps/chosen": -73.97720336914062,
+      "logps/rejected": -84.72975158691406,
+      "loss": 2.3215,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6877132654190063,
+      "rewards/margins": 0.4473133981227875,
+      "rewards/rejected": -2.135026693344116,
+      "step": 224
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 1.7008063793182373,
+      "learning_rate": 9.769716625767939e-06,
+      "logits/chosen": 0.05822606012225151,
+      "logits/rejected": 0.06510132551193237,
+      "logps/chosen": -81.26387023925781,
+      "logps/rejected": -82.76637268066406,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.739980697631836,
+      "rewards/margins": 0.3888433277606964,
+      "rewards/rejected": -2.12882399559021,
+      "step": 225
+    },
+    {
+      "epoch": 0.410535876475931,
+      "grad_norm": 1.3915003538131714,
+      "learning_rate": 9.767682949705243e-06,
+      "logits/chosen": 0.08782866597175598,
+      "logits/rejected": 0.17832686007022858,
+      "logps/chosen": -67.25590515136719,
+      "logps/rejected": -78.19799041748047,
+      "loss": 2.3992,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.6588337421417236,
+      "rewards/margins": 0.4392687976360321,
+      "rewards/rejected": -2.098102331161499,
+      "step": 226
+    },
+    {
+      "epoch": 0.4123524069028156,
+      "grad_norm": 1.4107364416122437,
+      "learning_rate": 9.765640571093938e-06,
+      "logits/chosen": 0.14615394175052643,
+      "logits/rejected": 0.14398689568042755,
+      "logps/chosen": -66.720703125,
+      "logps/rejected": -72.72846221923828,
+      "loss": 2.488,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.740675449371338,
+      "rewards/margins": 0.34203463792800903,
+      "rewards/rejected": -2.0827102661132812,
+      "step": 227
+    },
+    {
+      "epoch": 0.4141689373297003,
+      "grad_norm": 1.438272476196289,
+      "learning_rate": 9.76358949409879e-06,
+      "logits/chosen": 0.1331941783428192,
+      "logits/rejected": 0.18831086158752441,
+      "logps/chosen": -77.92586517333984,
+      "logps/rejected": -81.89257049560547,
+      "loss": 2.466,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7596431970596313,
+      "rewards/margins": 0.3160095512866974,
+      "rewards/rejected": -2.075652837753296,
+      "step": 228
+    },
+    {
+      "epoch": 0.4159854677565849,
+      "grad_norm": 1.3581331968307495,
+      "learning_rate": 9.7615297229023e-06,
+      "logits/chosen": 0.13822412490844727,
+      "logits/rejected": 0.14220967888832092,
+      "logps/chosen": -64.66896057128906,
+      "logps/rejected": -76.72779846191406,
+      "loss": 2.282,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.5948115587234497,
+      "rewards/margins": 0.5508276224136353,
+      "rewards/rejected": -2.145639181137085,
+      "step": 229
+    },
+    {
+      "epoch": 0.4178019981834696,
+      "grad_norm": 1.4766261577606201,
+      "learning_rate": 9.759461261704705e-06,
+      "logits/chosen": 0.06772036850452423,
+      "logits/rejected": 0.1212601587176323,
+      "logps/chosen": -75.17322540283203,
+      "logps/rejected": -86.30448913574219,
+      "loss": 2.1345,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6930228471755981,
+      "rewards/margins": 0.607439398765564,
+      "rewards/rejected": -2.300462245941162,
+      "step": 230
+    },
+    {
+      "epoch": 0.4196185286103542,
+      "grad_norm": 1.4598506689071655,
+      "learning_rate": 9.757384114723954e-06,
+      "logits/chosen": 0.11245124042034149,
+      "logits/rejected": 0.17101560533046722,
+      "logps/chosen": -75.50772094726562,
+      "logps/rejected": -83.95561218261719,
+      "loss": 2.3606,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7730398178100586,
+      "rewards/margins": 0.4058808386325836,
+      "rewards/rejected": -2.1789209842681885,
+      "step": 231
+    },
+    {
+      "epoch": 0.4214350590372389,
+      "grad_norm": 1.5139139890670776,
+      "learning_rate": 9.755298286195712e-06,
+      "logits/chosen": 0.05479501932859421,
+      "logits/rejected": 0.12558911740779877,
+      "logps/chosen": -81.28207397460938,
+      "logps/rejected": -86.58873748779297,
+      "loss": 2.4171,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6220098733901978,
+      "rewards/margins": 0.49352455139160156,
+      "rewards/rejected": -2.1155343055725098,
+      "step": 232
+    },
+    {
+      "epoch": 0.4232515894641235,
+      "grad_norm": 1.6987231969833374,
+      "learning_rate": 9.753203780373348e-06,
+      "logits/chosen": 0.07917524874210358,
+      "logits/rejected": 0.14443910121917725,
+      "logps/chosen": -81.83525848388672,
+      "logps/rejected": -77.00749206542969,
+      "loss": 2.7548,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7713696956634521,
+      "rewards/margins": 0.20764464139938354,
+      "rewards/rejected": -1.9790143966674805,
+      "step": 233
+    },
+    {
+      "epoch": 0.4250681198910082,
+      "grad_norm": 1.6056586503982544,
+      "learning_rate": 9.751100601527922e-06,
+      "logits/chosen": 0.12120751291513443,
+      "logits/rejected": 0.21893832087516785,
+      "logps/chosen": -73.1818618774414,
+      "logps/rejected": -84.82020568847656,
+      "loss": 2.4369,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8145970106124878,
+      "rewards/margins": 0.5701674818992615,
+      "rewards/rejected": -2.3847644329071045,
+      "step": 234
+    },
+    {
+      "epoch": 0.4268846503178928,
+      "grad_norm": 1.4865500926971436,
+      "learning_rate": 9.748988753948183e-06,
+      "logits/chosen": 0.0659053698182106,
+      "logits/rejected": 0.09985598176717758,
+      "logps/chosen": -83.79960632324219,
+      "logps/rejected": -87.37390899658203,
+      "loss": 2.3437,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8071691989898682,
+      "rewards/margins": 0.4156024754047394,
+      "rewards/rejected": -2.2227721214294434,
+      "step": 235
+    },
+    {
+      "epoch": 0.4287011807447775,
+      "grad_norm": 1.7585958242416382,
+      "learning_rate": 9.746868241940554e-06,
+      "logits/chosen": 0.13636741042137146,
+      "logits/rejected": 0.16040681302547455,
+      "logps/chosen": -70.46460723876953,
+      "logps/rejected": -71.19532012939453,
+      "loss": 2.6303,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7820823192596436,
+      "rewards/margins": 0.37900543212890625,
+      "rewards/rejected": -2.16108775138855,
+      "step": 236
+    },
+    {
+      "epoch": 0.4305177111716621,
+      "grad_norm": 1.4949400424957275,
+      "learning_rate": 9.744739069829132e-06,
+      "logits/chosen": 0.16385243833065033,
+      "logits/rejected": 0.13090217113494873,
+      "logps/chosen": -74.92865753173828,
+      "logps/rejected": -78.68392181396484,
+      "loss": 2.3038,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.847611904144287,
+      "rewards/margins": 0.561863899230957,
+      "rewards/rejected": -2.409475564956665,
+      "step": 237
+    },
+    {
+      "epoch": 0.4323342415985468,
+      "grad_norm": 1.6149885654449463,
+      "learning_rate": 9.742601241955666e-06,
+      "logits/chosen": 0.10731178522109985,
+      "logits/rejected": 0.11118797957897186,
+      "logps/chosen": -87.08879089355469,
+      "logps/rejected": -90.8418960571289,
+      "loss": 2.2981,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.7146490812301636,
+      "rewards/margins": 0.4193970859050751,
+      "rewards/rejected": -2.1340463161468506,
+      "step": 238
+    },
+    {
+      "epoch": 0.43415077202543145,
+      "grad_norm": 1.4300076961517334,
+      "learning_rate": 9.740454762679562e-06,
+      "logits/chosen": 0.031154140830039978,
+      "logits/rejected": 0.13953763246536255,
+      "logps/chosen": -68.21051025390625,
+      "logps/rejected": -87.66563415527344,
+      "loss": 2.0885,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6824418306350708,
+      "rewards/margins": 0.6189282536506653,
+      "rewards/rejected": -2.3013699054718018,
+      "step": 239
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 1.3545743227005005,
+      "learning_rate": 9.738299636377863e-06,
+      "logits/chosen": 0.1132105141878128,
+      "logits/rejected": 0.1149899885058403,
+      "logps/chosen": -75.41773223876953,
+      "logps/rejected": -80.03020477294922,
+      "loss": 2.3774,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.5836522579193115,
+      "rewards/margins": 0.40351229906082153,
+      "rewards/rejected": -1.9871646165847778,
+      "step": 240
+    },
+    {
+      "epoch": 0.43778383287920075,
+      "grad_norm": 1.2866510152816772,
+      "learning_rate": 9.736135867445246e-06,
+      "logits/chosen": 0.08788580447435379,
+      "logits/rejected": 0.15736152231693268,
+      "logps/chosen": -70.54780578613281,
+      "logps/rejected": -86.70913696289062,
+      "loss": 1.9408,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7565429210662842,
+      "rewards/margins": 0.7674389481544495,
+      "rewards/rejected": -2.523982048034668,
+      "step": 241
+    },
+    {
+      "epoch": 0.4396003633060854,
+      "grad_norm": 1.5900717973709106,
+      "learning_rate": 9.733963460294016e-06,
+      "logits/chosen": 0.08901432901620865,
+      "logits/rejected": 0.09095099568367004,
+      "logps/chosen": -77.32083129882812,
+      "logps/rejected": -83.0787124633789,
+      "loss": 2.4779,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9288290739059448,
+      "rewards/margins": 0.3023950755596161,
+      "rewards/rejected": -2.2312240600585938,
+      "step": 242
+    },
+    {
+      "epoch": 0.44141689373297005,
+      "grad_norm": 1.664976716041565,
+      "learning_rate": 9.731782419354087e-06,
+      "logits/chosen": 0.0038250258658081293,
+      "logits/rejected": 0.05805446207523346,
+      "logps/chosen": -75.7310562133789,
+      "logps/rejected": -81.24979400634766,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8681721687316895,
+      "rewards/margins": 0.41048479080200195,
+      "rewards/rejected": -2.2786567211151123,
+      "step": 243
+    },
+    {
+      "epoch": 0.44323342415985467,
+      "grad_norm": 1.688614845275879,
+      "learning_rate": 9.729592749072981e-06,
+      "logits/chosen": 0.10514964163303375,
+      "logits/rejected": 0.08623237907886505,
+      "logps/chosen": -78.1123046875,
+      "logps/rejected": -85.77177429199219,
+      "loss": 2.4137,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.709516167640686,
+      "rewards/margins": 0.477593332529068,
+      "rewards/rejected": -2.1871094703674316,
+      "step": 244
+    },
+    {
+      "epoch": 0.44504995458673935,
+      "grad_norm": 1.603507399559021,
+      "learning_rate": 9.727394453915817e-06,
+      "logits/chosen": 0.06938113272190094,
+      "logits/rejected": 0.10225434601306915,
+      "logps/chosen": -72.41216278076172,
+      "logps/rejected": -82.71170043945312,
+      "loss": 2.3143,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8322726488113403,
+      "rewards/margins": 0.3731868863105774,
+      "rewards/rejected": -2.2054593563079834,
+      "step": 245
+    },
+    {
+      "epoch": 0.44686648501362397,
+      "grad_norm": 1.6047879457473755,
+      "learning_rate": 9.725187538365304e-06,
+      "logits/chosen": 0.11169447004795074,
+      "logits/rejected": 0.14944276213645935,
+      "logps/chosen": -75.95654296875,
+      "logps/rejected": -87.93280029296875,
+      "loss": 2.3278,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9458414316177368,
+      "rewards/margins": 0.5952720046043396,
+      "rewards/rejected": -2.5411133766174316,
+      "step": 246
+    },
+    {
+      "epoch": 0.44868301544050865,
+      "grad_norm": 1.5822384357452393,
+      "learning_rate": 9.722972006921725e-06,
+      "logits/chosen": 0.07633841782808304,
+      "logits/rejected": 0.13307756185531616,
+      "logps/chosen": -82.36216735839844,
+      "logps/rejected": -89.472900390625,
+      "loss": 2.3939,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.026167154312134,
+      "rewards/margins": 0.4467831552028656,
+      "rewards/rejected": -2.4729504585266113,
+      "step": 247
+    },
+    {
+      "epoch": 0.45049954586739327,
+      "grad_norm": 1.636837363243103,
+      "learning_rate": 9.720747864102935e-06,
+      "logits/chosen": 0.013166696764528751,
+      "logits/rejected": 0.10926786065101624,
+      "logps/chosen": -81.95793151855469,
+      "logps/rejected": -95.96204376220703,
+      "loss": 2.4935,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.1589713096618652,
+      "rewards/margins": 0.4789046049118042,
+      "rewards/rejected": -2.637876033782959,
+      "step": 248
+    },
+    {
+      "epoch": 0.45231607629427795,
+      "grad_norm": 1.6743932962417603,
+      "learning_rate": 9.718515114444347e-06,
+      "logits/chosen": 0.09870442003011703,
+      "logits/rejected": 0.14206278324127197,
+      "logps/chosen": -81.16842651367188,
+      "logps/rejected": -95.32533264160156,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.291147470474243,
+      "rewards/margins": 0.7547516822814941,
+      "rewards/rejected": -3.0458991527557373,
+      "step": 249
+    },
+    {
+      "epoch": 0.45413260672116257,
+      "grad_norm": 1.4136468172073364,
+      "learning_rate": 9.716273762498929e-06,
+      "logits/chosen": 0.11840160191059113,
+      "logits/rejected": 0.14921030402183533,
+      "logps/chosen": -70.83485412597656,
+      "logps/rejected": -73.80975341796875,
+      "loss": 2.4975,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8664369583129883,
+      "rewards/margins": 0.32539018988609314,
+      "rewards/rejected": -2.1918272972106934,
+      "step": 250
+    },
+    {
+      "epoch": 0.45594913714804725,
+      "grad_norm": 1.771864652633667,
+      "learning_rate": 9.714023812837185e-06,
+      "logits/chosen": 0.04426509141921997,
+      "logits/rejected": 0.08872814476490021,
+      "logps/chosen": -82.2516098022461,
+      "logps/rejected": -87.78474426269531,
+      "loss": 2.627,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.2238872051239014,
+      "rewards/margins": 0.48393067717552185,
+      "rewards/rejected": -2.707818031311035,
+      "step": 251
+    },
+    {
+      "epoch": 0.45776566757493187,
+      "grad_norm": 1.4489670991897583,
+      "learning_rate": 9.711765270047155e-06,
+      "logits/chosen": 0.025218207389116287,
+      "logits/rejected": 0.09768272191286087,
+      "logps/chosen": -70.79837036132812,
+      "logps/rejected": -86.96686553955078,
+      "loss": 2.1641,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8761096000671387,
+      "rewards/margins": 0.8277947306632996,
+      "rewards/rejected": -2.703904151916504,
+      "step": 252
+    },
+    {
+      "epoch": 0.45958219800181654,
+      "grad_norm": 1.7034775018692017,
+      "learning_rate": 9.709498138734405e-06,
+      "logits/chosen": 0.04030866175889969,
+      "logits/rejected": 0.08729197829961777,
+      "logps/chosen": -81.70675659179688,
+      "logps/rejected": -85.92189025878906,
+      "loss": 2.58,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.0707144737243652,
+      "rewards/margins": 0.3897002637386322,
+      "rewards/rejected": -2.4604148864746094,
+      "step": 253
+    },
+    {
+      "epoch": 0.46139872842870117,
+      "grad_norm": 1.8713371753692627,
+      "learning_rate": 9.707222423522004e-06,
+      "logits/chosen": 0.047953542321920395,
+      "logits/rejected": 0.02678016573190689,
+      "logps/chosen": -84.65204620361328,
+      "logps/rejected": -90.09396362304688,
+      "loss": 2.7899,
+      "rewards/accuracies": 0.515625,
+      "rewards/chosen": -2.103985071182251,
+      "rewards/margins": 0.22625797986984253,
+      "rewards/rejected": -2.330242872238159,
+      "step": 254
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 1.4706422090530396,
+      "learning_rate": 9.704938129050535e-06,
+      "logits/chosen": 0.04734738916158676,
+      "logits/rejected": 0.11658424139022827,
+      "logps/chosen": -78.08146667480469,
+      "logps/rejected": -96.80207061767578,
+      "loss": 2.1796,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9245433807373047,
+      "rewards/margins": 0.7155373096466064,
+      "rewards/rejected": -2.640080451965332,
+      "step": 255
+    },
+    {
+      "epoch": 0.46503178928247046,
+      "grad_norm": 1.4902199506759644,
+      "learning_rate": 9.702645259978072e-06,
+      "logits/chosen": 0.09310627728700638,
+      "logits/rejected": 0.1795577108860016,
+      "logps/chosen": -78.96179962158203,
+      "logps/rejected": -84.86495208740234,
+      "loss": 2.2113,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0413970947265625,
+      "rewards/margins": 0.4318653345108032,
+      "rewards/rejected": -2.4732625484466553,
+      "step": 256
+    },
+    {
+      "epoch": 0.46684831970935514,
+      "grad_norm": 1.3563650846481323,
+      "learning_rate": 9.700343820980172e-06,
+      "logits/chosen": 0.08617695420980453,
+      "logits/rejected": 0.1092975065112114,
+      "logps/chosen": -76.31070709228516,
+      "logps/rejected": -82.52798461914062,
+      "loss": 2.3793,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7843788862228394,
+      "rewards/margins": 0.5503235459327698,
+      "rewards/rejected": -2.334702491760254,
+      "step": 257
+    },
+    {
+      "epoch": 0.46866485013623976,
+      "grad_norm": 1.353379249572754,
+      "learning_rate": 9.698033816749874e-06,
+      "logits/chosen": 0.07846446335315704,
+      "logits/rejected": 0.15949462354183197,
+      "logps/chosen": -70.9232177734375,
+      "logps/rejected": -83.97347259521484,
+      "loss": 2.0527,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.7502186298370361,
+      "rewards/margins": 0.6754422783851624,
+      "rewards/rejected": -2.4256608486175537,
+      "step": 258
+    },
+    {
+      "epoch": 0.47048138056312444,
+      "grad_norm": 1.2879610061645508,
+      "learning_rate": 9.695715251997676e-06,
+      "logits/chosen": 0.041364431381225586,
+      "logits/rejected": 0.2137874960899353,
+      "logps/chosen": -74.40650177001953,
+      "logps/rejected": -99.26273345947266,
+      "loss": 1.8381,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -1.7944194078445435,
+      "rewards/margins": 1.1633189916610718,
+      "rewards/rejected": -2.9577386379241943,
+      "step": 259
+    },
+    {
+      "epoch": 0.47229791099000906,
+      "grad_norm": 1.5336003303527832,
+      "learning_rate": 9.693388131451536e-06,
+      "logits/chosen": 0.10320600867271423,
+      "logits/rejected": 0.20043231546878815,
+      "logps/chosen": -70.24287414550781,
+      "logps/rejected": -87.45311737060547,
+      "loss": 2.2716,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8973060846328735,
+      "rewards/margins": 0.7614614963531494,
+      "rewards/rejected": -2.6587674617767334,
+      "step": 260
+    },
+    {
+      "epoch": 0.47411444141689374,
+      "grad_norm": 1.50228750705719,
+      "learning_rate": 9.691052459856858e-06,
+      "logits/chosen": 0.10719013214111328,
+      "logits/rejected": 0.11425416171550751,
+      "logps/chosen": -77.69039154052734,
+      "logps/rejected": -83.77989196777344,
+      "loss": 2.3255,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7352136373519897,
+      "rewards/margins": 0.588873565196991,
+      "rewards/rejected": -2.324087142944336,
+      "step": 261
+    },
+    {
+      "epoch": 0.47593097184377836,
+      "grad_norm": 1.6516278982162476,
+      "learning_rate": 9.688708241976484e-06,
+      "logits/chosen": 0.0962933823466301,
+      "logits/rejected": 0.09487868845462799,
+      "logps/chosen": -83.95497131347656,
+      "logps/rejected": -86.93241882324219,
+      "loss": 2.5126,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0139267444610596,
+      "rewards/margins": 0.4219791889190674,
+      "rewards/rejected": -2.435905694961548,
+      "step": 262
+    },
+    {
+      "epoch": 0.47774750227066304,
+      "grad_norm": 1.4588408470153809,
+      "learning_rate": 9.686355482590679e-06,
+      "logits/chosen": 0.04918690025806427,
+      "logits/rejected": 0.1662694364786148,
+      "logps/chosen": -76.76988983154297,
+      "logps/rejected": -98.94943237304688,
+      "loss": 2.0243,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.8614740371704102,
+      "rewards/margins": 0.9306913614273071,
+      "rewards/rejected": -2.7921652793884277,
+      "step": 263
+    },
+    {
+      "epoch": 0.47956403269754766,
+      "grad_norm": 1.8992609977722168,
+      "learning_rate": 9.683994186497132e-06,
+      "logits/chosen": 0.039469510316848755,
+      "logits/rejected": 0.11771678924560547,
+      "logps/chosen": -78.4907455444336,
+      "logps/rejected": -97.9975814819336,
+      "loss": 2.6106,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1690123081207275,
+      "rewards/margins": 0.9005274772644043,
+      "rewards/rejected": -3.069540023803711,
+      "step": 264
+    },
+    {
+      "epoch": 0.48138056312443234,
+      "grad_norm": 2.196244239807129,
+      "learning_rate": 9.681624358510936e-06,
+      "logits/chosen": 0.11924441158771515,
+      "logits/rejected": 0.08125054091215134,
+      "logps/chosen": -67.94137573242188,
+      "logps/rejected": -70.6710433959961,
+      "loss": 2.1509,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.787060022354126,
+      "rewards/margins": 0.5431486368179321,
+      "rewards/rejected": -2.3302085399627686,
+      "step": 265
+    },
+    {
+      "epoch": 0.48319709355131696,
+      "grad_norm": 2.009992837905884,
+      "learning_rate": 9.679246003464585e-06,
+      "logits/chosen": 0.032514430582523346,
+      "logits/rejected": 0.04692292958498001,
+      "logps/chosen": -85.35857391357422,
+      "logps/rejected": -90.84698486328125,
+      "loss": 2.8119,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1302294731140137,
+      "rewards/margins": 0.3329467177391052,
+      "rewards/rejected": -2.4631760120391846,
+      "step": 266
+    },
+    {
+      "epoch": 0.48501362397820164,
+      "grad_norm": 1.535208821296692,
+      "learning_rate": 9.676859126207957e-06,
+      "logits/chosen": 0.03791799396276474,
+      "logits/rejected": 0.0628521591424942,
+      "logps/chosen": -73.64868927001953,
+      "logps/rejected": -80.9854965209961,
+      "loss": 2.2386,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9628286361694336,
+      "rewards/margins": 0.6735512018203735,
+      "rewards/rejected": -2.6363797187805176,
+      "step": 267
+    },
+    {
+      "epoch": 0.4868301544050863,
+      "grad_norm": 1.9440377950668335,
+      "learning_rate": 9.674463731608309e-06,
+      "logits/chosen": 0.0922878235578537,
+      "logits/rejected": 0.08414338529109955,
+      "logps/chosen": -80.34213256835938,
+      "logps/rejected": -87.45000457763672,
+      "loss": 2.8564,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.194643020629883,
+      "rewards/margins": 0.5086687207221985,
+      "rewards/rejected": -2.7033114433288574,
+      "step": 268
+    },
+    {
+      "epoch": 0.48864668483197093,
+      "grad_norm": 1.57882559299469,
+      "learning_rate": 9.672059824550268e-06,
+      "logits/chosen": 0.08647017180919647,
+      "logits/rejected": 0.07438144087791443,
+      "logps/chosen": -89.79737091064453,
+      "logps/rejected": -95.89274597167969,
+      "loss": 2.1801,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9004266262054443,
+      "rewards/margins": 0.5936228632926941,
+      "rewards/rejected": -2.494049549102783,
+      "step": 269
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 1.47967529296875,
+      "learning_rate": 9.669647409935822e-06,
+      "logits/chosen": 0.0663951188325882,
+      "logits/rejected": 0.12090058624744415,
+      "logps/chosen": -86.87987518310547,
+      "logps/rejected": -90.49571228027344,
+      "loss": 2.342,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.043788194656372,
+      "rewards/margins": 0.4767158329486847,
+      "rewards/rejected": -2.5205039978027344,
+      "step": 270
+    },
+    {
+      "epoch": 0.49227974568574023,
+      "grad_norm": 1.5860496759414673,
+      "learning_rate": 9.667226492684302e-06,
+      "logits/chosen": 0.06286406517028809,
+      "logits/rejected": 0.1226339116692543,
+      "logps/chosen": -91.13972473144531,
+      "logps/rejected": -105.41304016113281,
+      "loss": 2.2264,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.8733948469161987,
+      "rewards/margins": 0.6329715251922607,
+      "rewards/rejected": -2.50636625289917,
+      "step": 271
+    },
+    {
+      "epoch": 0.4940962761126249,
+      "grad_norm": 1.6157667636871338,
+      "learning_rate": 9.66479707773238e-06,
+      "logits/chosen": 0.07893703132867813,
+      "logits/rejected": 0.1179615929722786,
+      "logps/chosen": -77.89990234375,
+      "logps/rejected": -87.70220947265625,
+      "loss": 2.4723,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.136741876602173,
+      "rewards/margins": 0.4580468237400055,
+      "rewards/rejected": -2.5947885513305664,
+      "step": 272
+    },
+    {
+      "epoch": 0.49591280653950953,
+      "grad_norm": 1.5408835411071777,
+      "learning_rate": 9.662359170034058e-06,
+      "logits/chosen": 0.0976145789027214,
+      "logits/rejected": 0.12423861026763916,
+      "logps/chosen": -76.74501037597656,
+      "logps/rejected": -82.18941497802734,
+      "loss": 2.5565,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.039276361465454,
+      "rewards/margins": 0.3934023678302765,
+      "rewards/rejected": -2.432678461074829,
+      "step": 273
+    },
+    {
+      "epoch": 0.4977293369663942,
+      "grad_norm": 1.44657564163208,
+      "learning_rate": 9.659912774560654e-06,
+      "logits/chosen": 0.04748811572790146,
+      "logits/rejected": 0.12317924201488495,
+      "logps/chosen": -75.47578430175781,
+      "logps/rejected": -88.56098937988281,
+      "loss": 2.1908,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.050949811935425,
+      "rewards/margins": 0.564163863658905,
+      "rewards/rejected": -2.6151139736175537,
+      "step": 274
+    },
+    {
+      "epoch": 0.49954586739327883,
+      "grad_norm": 2.1769776344299316,
+      "learning_rate": 9.65745789630079e-06,
+      "logits/chosen": 0.11112834513187408,
+      "logits/rejected": 0.10816515237092972,
+      "logps/chosen": -83.5405044555664,
+      "logps/rejected": -83.06329345703125,
+      "loss": 3.0939,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.142024040222168,
+      "rewards/margins": 0.1820124089717865,
+      "rewards/rejected": -2.3240363597869873,
+      "step": 275
+    },
+    {
+      "epoch": 0.5013623978201635,
+      "grad_norm": 1.7329221963882446,
+      "learning_rate": 9.654994540260396e-06,
+      "logits/chosen": 0.0653618574142456,
+      "logits/rejected": 0.08004368096590042,
+      "logps/chosen": -80.83209228515625,
+      "logps/rejected": -82.97142028808594,
+      "loss": 2.7161,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.851841688156128,
+      "rewards/margins": 0.2948168218135834,
+      "rewards/rejected": -2.146658420562744,
+      "step": 276
+    },
+    {
+      "epoch": 0.5031789282470481,
+      "grad_norm": 1.592657208442688,
+      "learning_rate": 9.65252271146268e-06,
+      "logits/chosen": 0.09880789369344711,
+      "logits/rejected": 0.14229761064052582,
+      "logps/chosen": -67.1727294921875,
+      "logps/rejected": -75.03417205810547,
+      "loss": 2.5366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8714643716812134,
+      "rewards/margins": 0.37937116622924805,
+      "rewards/rejected": -2.250835657119751,
+      "step": 277
+    },
+    {
+      "epoch": 0.5049954586739328,
+      "grad_norm": 1.6487712860107422,
+      "learning_rate": 9.650042414948133e-06,
+      "logits/chosen": 0.13465353846549988,
+      "logits/rejected": 0.12865689396858215,
+      "logps/chosen": -76.4417724609375,
+      "logps/rejected": -78.6947021484375,
+      "loss": 2.6085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.875291347503662,
+      "rewards/margins": 0.35175251960754395,
+      "rewards/rejected": -2.227043867111206,
+      "step": 278
+    },
+    {
+      "epoch": 0.5068119891008175,
+      "grad_norm": 1.6523009538650513,
+      "learning_rate": 9.64755365577451e-06,
+      "logits/chosen": 0.04238567873835564,
+      "logits/rejected": 0.07994347810745239,
+      "logps/chosen": -76.92097473144531,
+      "logps/rejected": -83.20886993408203,
+      "loss": 2.3156,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9759610891342163,
+      "rewards/margins": 0.5398927330970764,
+      "rewards/rejected": -2.5158536434173584,
+      "step": 279
+    },
+    {
+      "epoch": 0.508628519527702,
+      "grad_norm": 1.7999261617660522,
+      "learning_rate": 9.645056439016827e-06,
+      "logits/chosen": 0.07349395751953125,
+      "logits/rejected": 0.07518415153026581,
+      "logps/chosen": -77.56079864501953,
+      "logps/rejected": -84.97645568847656,
+      "loss": 2.8085,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9408115148544312,
+      "rewards/margins": 0.3470652997493744,
+      "rewards/rejected": -2.287877082824707,
+      "step": 280
+    },
+    {
+      "epoch": 0.5104450499545867,
+      "grad_norm": 1.5684200525283813,
+      "learning_rate": 9.642550769767342e-06,
+      "logits/chosen": 0.16188879311084747,
+      "logits/rejected": 0.12772323191165924,
+      "logps/chosen": -89.29315185546875,
+      "logps/rejected": -94.35065460205078,
+      "loss": 2.2314,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8958841562271118,
+      "rewards/margins": 0.5102108716964722,
+      "rewards/rejected": -2.406095027923584,
+      "step": 281
+    },
+    {
+      "epoch": 0.5122615803814714,
+      "grad_norm": 1.7878178358078003,
+      "learning_rate": 9.640036653135548e-06,
+      "logits/chosen": 0.060573749244213104,
+      "logits/rejected": 0.13457715511322021,
+      "logps/chosen": -68.9404525756836,
+      "logps/rejected": -74.77693176269531,
+      "loss": 2.4359,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.889530897140503,
+      "rewards/margins": 0.45911547541618347,
+      "rewards/rejected": -2.348646402359009,
+      "step": 282
+    },
+    {
+      "epoch": 0.5140781108083561,
+      "grad_norm": 1.4741288423538208,
+      "learning_rate": 9.637514094248172e-06,
+      "logits/chosen": 0.10433132946491241,
+      "logits/rejected": 0.1439315229654312,
+      "logps/chosen": -74.89447021484375,
+      "logps/rejected": -86.58551788330078,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9090807437896729,
+      "rewards/margins": 0.6733560562133789,
+      "rewards/rejected": -2.5824368000030518,
+      "step": 283
+    },
+    {
+      "epoch": 0.5158946412352406,
+      "grad_norm": 1.2864018678665161,
+      "learning_rate": 9.634983098249146e-06,
+      "logits/chosen": 0.10891089588403702,
+      "logits/rejected": 0.11755162477493286,
+      "logps/chosen": -66.90985107421875,
+      "logps/rejected": -71.8125228881836,
+      "loss": 2.1538,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.7779250144958496,
+      "rewards/margins": 0.5625147223472595,
+      "rewards/rejected": -2.340439796447754,
+      "step": 284
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 1.465747594833374,
+      "learning_rate": 9.632443670299616e-06,
+      "logits/chosen": 0.08224496245384216,
+      "logits/rejected": 0.12130744755268097,
+      "logps/chosen": -75.4281997680664,
+      "logps/rejected": -85.0781021118164,
+      "loss": 2.2988,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7217226028442383,
+      "rewards/margins": 0.4750482439994812,
+      "rewards/rejected": -2.196770668029785,
+      "step": 285
+    },
+    {
+      "epoch": 0.51952770208901,
+      "grad_norm": 1.492859959602356,
+      "learning_rate": 9.629895815577915e-06,
+      "logits/chosen": 0.06619664281606674,
+      "logits/rejected": 0.13152630627155304,
+      "logps/chosen": -96.65383911132812,
+      "logps/rejected": -111.93521881103516,
+      "loss": 2.2831,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.7611618041992188,
+      "rewards/margins": 0.5886337757110596,
+      "rewards/rejected": -2.349795341491699,
+      "step": 286
+    },
+    {
+      "epoch": 0.5213442325158947,
+      "grad_norm": 1.5534065961837769,
+      "learning_rate": 9.627339539279564e-06,
+      "logits/chosen": 0.06637927144765854,
+      "logits/rejected": 0.09107412397861481,
+      "logps/chosen": -71.92534637451172,
+      "logps/rejected": -82.98391723632812,
+      "loss": 2.5101,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.7867075204849243,
+      "rewards/margins": 0.39493298530578613,
+      "rewards/rejected": -2.181640625,
+      "step": 287
+    },
+    {
+      "epoch": 0.5231607629427792,
+      "grad_norm": 1.6778221130371094,
+      "learning_rate": 9.624774846617254e-06,
+      "logits/chosen": 0.14700329303741455,
+      "logits/rejected": 0.12778782844543457,
+      "logps/chosen": -65.2364730834961,
+      "logps/rejected": -74.39017486572266,
+      "loss": 2.5236,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6942007541656494,
+      "rewards/margins": 0.35334131121635437,
+      "rewards/rejected": -2.047542095184326,
+      "step": 288
+    },
+    {
+      "epoch": 0.5249772933696639,
+      "grad_norm": 1.462215542793274,
+      "learning_rate": 9.622201742820839e-06,
+      "logits/chosen": 0.11467991769313812,
+      "logits/rejected": 0.1028795838356018,
+      "logps/chosen": -67.85939025878906,
+      "logps/rejected": -74.3462142944336,
+      "loss": 2.3891,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7194863557815552,
+      "rewards/margins": 0.3846975564956665,
+      "rewards/rejected": -2.1041836738586426,
+      "step": 289
+    },
+    {
+      "epoch": 0.5267938237965486,
+      "grad_norm": 1.7591735124588013,
+      "learning_rate": 9.619620233137328e-06,
+      "logits/chosen": 0.08407986164093018,
+      "logits/rejected": 0.12201236933469772,
+      "logps/chosen": -82.57340240478516,
+      "logps/rejected": -86.94192504882812,
+      "loss": 2.4569,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8536537885665894,
+      "rewards/margins": 0.38806283473968506,
+      "rewards/rejected": -2.2417166233062744,
+      "step": 290
+    },
+    {
+      "epoch": 0.5286103542234333,
+      "grad_norm": 1.5584774017333984,
+      "learning_rate": 9.617030322830868e-06,
+      "logits/chosen": 0.1126309484243393,
+      "logits/rejected": 0.08907752484083176,
+      "logps/chosen": -73.36901092529297,
+      "logps/rejected": -76.65506744384766,
+      "loss": 2.4021,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8091729879379272,
+      "rewards/margins": 0.3603411316871643,
+      "rewards/rejected": -2.1695141792297363,
+      "step": 291
+    },
+    {
+      "epoch": 0.5304268846503178,
+      "grad_norm": 1.487938642501831,
+      "learning_rate": 9.614432017182736e-06,
+      "logits/chosen": 0.09620118141174316,
+      "logits/rejected": 0.08514149487018585,
+      "logps/chosen": -88.66265106201172,
+      "logps/rejected": -92.26091766357422,
+      "loss": 2.3636,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.5980674028396606,
+      "rewards/margins": 0.4888036549091339,
+      "rewards/rejected": -2.0868711471557617,
+      "step": 292
+    },
+    {
+      "epoch": 0.5322434150772025,
+      "grad_norm": 1.445408582687378,
+      "learning_rate": 9.611825321491331e-06,
+      "logits/chosen": 0.10014252364635468,
+      "logits/rejected": 0.12857215106487274,
+      "logps/chosen": -82.46649932861328,
+      "logps/rejected": -92.84205627441406,
+      "loss": 2.343,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.4620139598846436,
+      "rewards/margins": 0.5616316199302673,
+      "rewards/rejected": -2.0236456394195557,
+      "step": 293
+    },
+    {
+      "epoch": 0.5340599455040872,
+      "grad_norm": 1.262193202972412,
+      "learning_rate": 9.609210241072158e-06,
+      "logits/chosen": 0.09868282079696655,
+      "logits/rejected": 0.16074486076831818,
+      "logps/chosen": -67.0407485961914,
+      "logps/rejected": -82.36201477050781,
+      "loss": 2.1163,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.5362765789031982,
+      "rewards/margins": 0.7091963887214661,
+      "rewards/rejected": -2.2454731464385986,
+      "step": 294
+    },
+    {
+      "epoch": 0.5358764759309719,
+      "grad_norm": 1.601891279220581,
+      "learning_rate": 9.606586781257822e-06,
+      "logits/chosen": 0.1236480325460434,
+      "logits/rejected": 0.12187935411930084,
+      "logps/chosen": -78.2726058959961,
+      "logps/rejected": -84.49217987060547,
+      "loss": 2.5075,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.779215931892395,
+      "rewards/margins": 0.38987410068511963,
+      "rewards/rejected": -2.1690900325775146,
+      "step": 295
+    },
+    {
+      "epoch": 0.5376930063578564,
+      "grad_norm": 1.4211548566818237,
+      "learning_rate": 9.603954947398016e-06,
+      "logits/chosen": 0.18375667929649353,
+      "logits/rejected": 0.21945703029632568,
+      "logps/chosen": -70.29829406738281,
+      "logps/rejected": -78.42284393310547,
+      "loss": 2.1959,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.7448753118515015,
+      "rewards/margins": 0.5106840133666992,
+      "rewards/rejected": -2.2555594444274902,
+      "step": 296
+    },
+    {
+      "epoch": 0.5395095367847411,
+      "grad_norm": 1.458617925643921,
+      "learning_rate": 9.601314744859504e-06,
+      "logits/chosen": 0.07360847294330597,
+      "logits/rejected": 0.1596718430519104,
+      "logps/chosen": -72.28546142578125,
+      "logps/rejected": -87.13935089111328,
+      "loss": 2.3538,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8007893562316895,
+      "rewards/margins": 0.5433144569396973,
+      "rewards/rejected": -2.3441038131713867,
+      "step": 297
+    },
+    {
+      "epoch": 0.5413260672116258,
+      "grad_norm": 1.6526339054107666,
+      "learning_rate": 9.598666179026123e-06,
+      "logits/chosen": 0.12013350427150726,
+      "logits/rejected": 0.10237178206443787,
+      "logps/chosen": -83.06564331054688,
+      "logps/rejected": -85.60771942138672,
+      "loss": 2.4531,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.911988615989685,
+      "rewards/margins": 0.35885563492774963,
+      "rewards/rejected": -2.2708444595336914,
+      "step": 298
+    },
+    {
+      "epoch": 0.5431425976385105,
+      "grad_norm": 1.9808340072631836,
+      "learning_rate": 9.596009255298755e-06,
+      "logits/chosen": 0.062342576682567596,
+      "logits/rejected": 0.09053834527730942,
+      "logps/chosen": -96.33489990234375,
+      "logps/rejected": -93.43024444580078,
+      "loss": 2.8549,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.845227599143982,
+      "rewards/margins": 0.22291362285614014,
+      "rewards/rejected": -2.068141222000122,
+      "step": 299
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 1.6276236772537231,
+      "learning_rate": 9.593343979095334e-06,
+      "logits/chosen": 0.2073422521352768,
+      "logits/rejected": 0.13107003271579742,
+      "logps/chosen": -72.45758056640625,
+      "logps/rejected": -77.42770385742188,
+      "loss": 2.4644,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9407453536987305,
+      "rewards/margins": 0.3134301006793976,
+      "rewards/rejected": -2.2541751861572266,
+      "step": 300
+    },
+    {
+      "epoch": 0.5467756584922797,
+      "grad_norm": 1.5001753568649292,
+      "learning_rate": 9.590670355850819e-06,
+      "logits/chosen": 0.15097060799598694,
+      "logits/rejected": 0.16254279017448425,
+      "logps/chosen": -70.44524383544922,
+      "logps/rejected": -70.64158630371094,
+      "loss": 2.4159,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.6921097040176392,
+      "rewards/margins": 0.3336741626262665,
+      "rewards/rejected": -2.0257837772369385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5485921889191644,
+      "grad_norm": 1.8360233306884766,
+      "learning_rate": 9.587988391017198e-06,
+      "logits/chosen": 0.14593760669231415,
+      "logits/rejected": 0.12311654537916183,
+      "logps/chosen": -78.32576751708984,
+      "logps/rejected": -88.25840759277344,
+      "loss": 2.5826,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.6591644287109375,
+      "rewards/margins": 0.5071319341659546,
+      "rewards/rejected": -2.1662964820861816,
+      "step": 302
+    },
+    {
+      "epoch": 0.5504087193460491,
+      "grad_norm": 1.4822838306427002,
+      "learning_rate": 9.585298090063459e-06,
+      "logits/chosen": 0.20818498730659485,
+      "logits/rejected": 0.16436657309532166,
+      "logps/chosen": -67.59429931640625,
+      "logps/rejected": -77.14763641357422,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6798239946365356,
+      "rewards/margins": 0.5954271554946899,
+      "rewards/rejected": -2.2752511501312256,
+      "step": 303
+    },
+    {
+      "epoch": 0.5522252497729337,
+      "grad_norm": 1.6118305921554565,
+      "learning_rate": 9.582599458475598e-06,
+      "logits/chosen": 0.0827561467885971,
+      "logits/rejected": 0.09151773154735565,
+      "logps/chosen": -73.28964233398438,
+      "logps/rejected": -76.90730285644531,
+      "loss": 2.3792,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.8056440353393555,
+      "rewards/margins": 0.39551618695259094,
+      "rewards/rejected": -2.201160192489624,
+      "step": 304
+    },
+    {
+      "epoch": 0.5540417801998183,
+      "grad_norm": 1.4609856605529785,
+      "learning_rate": 9.579892501756593e-06,
+      "logits/chosen": 0.05332394689321518,
+      "logits/rejected": 0.1590057611465454,
+      "logps/chosen": -77.51653289794922,
+      "logps/rejected": -95.51261138916016,
+      "loss": 2.2105,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.740310549736023,
+      "rewards/margins": 0.6771562695503235,
+      "rewards/rejected": -2.4174671173095703,
+      "step": 305
+    },
+    {
+      "epoch": 0.555858310626703,
+      "grad_norm": 1.7181960344314575,
+      "learning_rate": 9.5771772254264e-06,
+      "logits/chosen": 0.05865276977419853,
+      "logits/rejected": 0.03500773757696152,
+      "logps/chosen": -83.72650146484375,
+      "logps/rejected": -93.69804382324219,
+      "loss": 2.1687,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6025928258895874,
+      "rewards/margins": 0.7339029312133789,
+      "rewards/rejected": -2.3364956378936768,
+      "step": 306
+    },
+    {
+      "epoch": 0.5576748410535877,
+      "grad_norm": 1.497698187828064,
+      "learning_rate": 9.57445363502194e-06,
+      "logits/chosen": 0.1225174218416214,
+      "logits/rejected": 0.09580346196889877,
+      "logps/chosen": -76.27323150634766,
+      "logps/rejected": -76.90716552734375,
+      "loss": 2.2116,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.715336799621582,
+      "rewards/margins": 0.4648086428642273,
+      "rewards/rejected": -2.180145263671875,
+      "step": 307
+    },
+    {
+      "epoch": 0.5594913714804723,
+      "grad_norm": 1.3957064151763916,
+      "learning_rate": 9.571721736097089e-06,
+      "logits/chosen": 0.07596514374017715,
+      "logits/rejected": 0.14832191169261932,
+      "logps/chosen": -84.65274810791016,
+      "logps/rejected": -105.76422119140625,
+      "loss": 1.9942,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.6643587350845337,
+      "rewards/margins": 0.7884883284568787,
+      "rewards/rejected": -2.4528470039367676,
+      "step": 308
+    },
+    {
+      "epoch": 0.5613079019073569,
+      "grad_norm": 1.6824349164962769,
+      "learning_rate": 9.568981534222664e-06,
+      "logits/chosen": 0.049865882843732834,
+      "logits/rejected": 0.023348212242126465,
+      "logps/chosen": -80.3697509765625,
+      "logps/rejected": -86.07074737548828,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.009049654006958,
+      "rewards/margins": 0.5470253825187683,
+      "rewards/rejected": -2.556075096130371,
+      "step": 309
+    },
+    {
+      "epoch": 0.5631244323342416,
+      "grad_norm": 1.8752086162567139,
+      "learning_rate": 9.566233034986413e-06,
+      "logits/chosen": 0.07250591367483139,
+      "logits/rejected": 0.12809628248214722,
+      "logps/chosen": -74.84794616699219,
+      "logps/rejected": -81.79740905761719,
+      "loss": 2.6286,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.75887131690979,
+      "rewards/margins": 0.33979111909866333,
+      "rewards/rejected": -2.0986623764038086,
+      "step": 310
+    },
+    {
+      "epoch": 0.5649409627611263,
+      "grad_norm": 1.7387233972549438,
+      "learning_rate": 9.563476243993008e-06,
+      "logits/chosen": 0.130618155002594,
+      "logits/rejected": 0.12952059507369995,
+      "logps/chosen": -80.75495910644531,
+      "logps/rejected": -90.4281234741211,
+      "loss": 2.4426,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.836004614830017,
+      "rewards/margins": 0.4418962597846985,
+      "rewards/rejected": -2.2779006958007812,
+      "step": 311
+    },
+    {
+      "epoch": 0.5667574931880109,
+      "grad_norm": 1.5147444009780884,
+      "learning_rate": 9.56071116686402e-06,
+      "logits/chosen": 0.10193713754415512,
+      "logits/rejected": 0.22481802105903625,
+      "logps/chosen": -73.21926879882812,
+      "logps/rejected": -81.27864837646484,
+      "loss": 2.578,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.6358015537261963,
+      "rewards/margins": 0.38279739022254944,
+      "rewards/rejected": -2.018598794937134,
+      "step": 312
+    },
+    {
+      "epoch": 0.5685740236148955,
+      "grad_norm": 1.6253665685653687,
+      "learning_rate": 9.557937809237927e-06,
+      "logits/chosen": 0.09468917548656464,
+      "logits/rejected": 0.09415112435817719,
+      "logps/chosen": -80.62995147705078,
+      "logps/rejected": -86.65946197509766,
+      "loss": 2.3854,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8635404109954834,
+      "rewards/margins": 0.4438764154911041,
+      "rewards/rejected": -2.3074169158935547,
+      "step": 313
+    },
+    {
+      "epoch": 0.5703905540417802,
+      "grad_norm": 1.7893344163894653,
+      "learning_rate": 9.555156176770087e-06,
+      "logits/chosen": 0.15863659977912903,
+      "logits/rejected": 0.09485571831464767,
+      "logps/chosen": -76.10442352294922,
+      "logps/rejected": -72.73162078857422,
+      "loss": 2.6795,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.935206413269043,
+      "rewards/margins": 0.22100940346717834,
+      "rewards/rejected": -2.1562156677246094,
+      "step": 314
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 1.696327805519104,
+      "learning_rate": 9.552366275132733e-06,
+      "logits/chosen": 0.07012113183736801,
+      "logits/rejected": 0.13891686499118805,
+      "logps/chosen": -78.08012390136719,
+      "logps/rejected": -83.05044555664062,
+      "loss": 2.6181,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.7869051694869995,
+      "rewards/margins": 0.23424415290355682,
+      "rewards/rejected": -2.0211493968963623,
+      "step": 315
+    },
+    {
+      "epoch": 0.5740236148955495,
+      "grad_norm": 1.6044729948043823,
+      "learning_rate": 9.54956811001496e-06,
+      "logits/chosen": 0.11298641562461853,
+      "logits/rejected": 0.2595204710960388,
+      "logps/chosen": -75.5998306274414,
+      "logps/rejected": -85.31849670410156,
+      "loss": 2.5849,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.5757699012756348,
+      "rewards/margins": 0.35041162371635437,
+      "rewards/rejected": -1.9261815547943115,
+      "step": 316
+    },
+    {
+      "epoch": 0.5758401453224341,
+      "grad_norm": 1.7399548292160034,
+      "learning_rate": 9.546761687122715e-06,
+      "logits/chosen": 0.14933931827545166,
+      "logits/rejected": 0.1868455857038498,
+      "logps/chosen": -68.05965423583984,
+      "logps/rejected": -82.37442779541016,
+      "loss": 2.3208,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8474600315093994,
+      "rewards/margins": 0.5766161680221558,
+      "rewards/rejected": -2.4240763187408447,
+      "step": 317
+    },
+    {
+      "epoch": 0.5776566757493188,
+      "grad_norm": 1.597495675086975,
+      "learning_rate": 9.54394701217878e-06,
+      "logits/chosen": 0.04729313403367996,
+      "logits/rejected": 0.05815067142248154,
+      "logps/chosen": -97.36541748046875,
+      "logps/rejected": -104.02650451660156,
+      "loss": 2.2106,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8631618022918701,
+      "rewards/margins": 0.6027428507804871,
+      "rewards/rejected": -2.465904474258423,
+      "step": 318
+    },
+    {
+      "epoch": 0.5794732061762035,
+      "grad_norm": 1.8646149635314941,
+      "learning_rate": 9.541124090922771e-06,
+      "logits/chosen": 0.1399674415588379,
+      "logits/rejected": 0.11107950657606125,
+      "logps/chosen": -82.39740753173828,
+      "logps/rejected": -76.80902099609375,
+      "loss": 2.6435,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.7670300006866455,
+      "rewards/margins": 0.15345275402069092,
+      "rewards/rejected": -1.920482873916626,
+      "step": 319
+    },
+    {
+      "epoch": 0.5812897366030881,
+      "grad_norm": 1.502447247505188,
+      "learning_rate": 9.538292929111114e-06,
+      "logits/chosen": 0.10672347247600555,
+      "logits/rejected": 0.08290571719408035,
+      "logps/chosen": -74.72683715820312,
+      "logps/rejected": -86.37158203125,
+      "loss": 2.2137,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8041073083877563,
+      "rewards/margins": 0.596272885799408,
+      "rewards/rejected": -2.4003803730010986,
+      "step": 320
+    },
+    {
+      "epoch": 0.5831062670299727,
+      "grad_norm": 1.7287745475769043,
+      "learning_rate": 9.535453532517039e-06,
+      "logits/chosen": 0.13742896914482117,
+      "logits/rejected": 0.07607734203338623,
+      "logps/chosen": -86.82475280761719,
+      "logps/rejected": -77.8934555053711,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7918696403503418,
+      "rewards/margins": 0.2591831088066101,
+      "rewards/rejected": -2.0510525703430176,
+      "step": 321
+    },
+    {
+      "epoch": 0.5849227974568574,
+      "grad_norm": 1.5914596319198608,
+      "learning_rate": 9.532605906930575e-06,
+      "logits/chosen": 0.11301672458648682,
+      "logits/rejected": 0.1944851577281952,
+      "logps/chosen": -68.9898452758789,
+      "logps/rejected": -73.17451477050781,
+      "loss": 2.4565,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.6999026536941528,
+      "rewards/margins": 0.30462026596069336,
+      "rewards/rejected": -2.0045228004455566,
+      "step": 322
+    },
+    {
+      "epoch": 0.5867393278837421,
+      "grad_norm": 1.6072031259536743,
+      "learning_rate": 9.529750058158522e-06,
+      "logits/chosen": 0.07092760503292084,
+      "logits/rejected": 0.08035591244697571,
+      "logps/chosen": -81.5555419921875,
+      "logps/rejected": -81.93560028076172,
+      "loss": 2.2729,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.6834638118743896,
+      "rewards/margins": 0.48882579803466797,
+      "rewards/rejected": -2.1722893714904785,
+      "step": 323
+    },
+    {
+      "epoch": 0.5885558583106267,
+      "grad_norm": 1.6433221101760864,
+      "learning_rate": 9.526885992024453e-06,
+      "logits/chosen": 0.13823899626731873,
+      "logits/rejected": 0.13610433042049408,
+      "logps/chosen": -78.9625244140625,
+      "logps/rejected": -83.2666015625,
+      "loss": 2.2075,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.6939194202423096,
+      "rewards/margins": 0.5755473971366882,
+      "rewards/rejected": -2.2694668769836426,
+      "step": 324
+    },
+    {
+      "epoch": 0.5903723887375113,
+      "grad_norm": 1.7674918174743652,
+      "learning_rate": 9.524013714368702e-06,
+      "logits/chosen": 0.1704932153224945,
+      "logits/rejected": 0.1073535829782486,
+      "logps/chosen": -72.47895050048828,
+      "logps/rejected": -73.0809326171875,
+      "loss": 2.5473,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8390110731124878,
+      "rewards/margins": 0.34467947483062744,
+      "rewards/rejected": -2.1836905479431152,
+      "step": 325
+    },
+    {
+      "epoch": 0.592188919164396,
+      "grad_norm": 1.8219250440597534,
+      "learning_rate": 9.521133231048338e-06,
+      "logits/chosen": 0.07941028475761414,
+      "logits/rejected": 0.13045310974121094,
+      "logps/chosen": -80.73494720458984,
+      "logps/rejected": -91.58990478515625,
+      "loss": 2.4866,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8594518899917603,
+      "rewards/margins": 0.4554288685321808,
+      "rewards/rejected": -2.314880847930908,
+      "step": 326
+    },
+    {
+      "epoch": 0.5940054495912807,
+      "grad_norm": 1.3753328323364258,
+      "learning_rate": 9.51824454793717e-06,
+      "logits/chosen": 0.08879546821117401,
+      "logits/rejected": 0.04693777486681938,
+      "logps/chosen": -82.9569091796875,
+      "logps/rejected": -91.44571685791016,
+      "loss": 2.2602,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.549119472503662,
+      "rewards/margins": 0.5145683288574219,
+      "rewards/rejected": -2.063688039779663,
+      "step": 327
+    },
+    {
+      "epoch": 0.5958219800181653,
+      "grad_norm": 1.4296562671661377,
+      "learning_rate": 9.515347670925728e-06,
+      "logits/chosen": 0.15614314377307892,
+      "logits/rejected": 0.1598319411277771,
+      "logps/chosen": -75.60345458984375,
+      "logps/rejected": -80.81770324707031,
+      "loss": 2.3946,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.556883454322815,
+      "rewards/margins": 0.4525718092918396,
+      "rewards/rejected": -2.0094552040100098,
+      "step": 328
+    },
+    {
+      "epoch": 0.59763851044505,
+      "grad_norm": 1.7932195663452148,
+      "learning_rate": 9.512442605921245e-06,
+      "logits/chosen": 0.05863601714372635,
+      "logits/rejected": 0.0788806602358818,
+      "logps/chosen": -75.95397186279297,
+      "logps/rejected": -84.50177001953125,
+      "loss": 2.5085,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9159862995147705,
+      "rewards/margins": 0.4600725769996643,
+      "rewards/rejected": -2.376059055328369,
+      "step": 329
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 1.6558383703231812,
+      "learning_rate": 9.509529358847655e-06,
+      "logits/chosen": 0.08205496519804001,
+      "logits/rejected": 0.13091425597667694,
+      "logps/chosen": -82.73474884033203,
+      "logps/rejected": -94.02273559570312,
+      "loss": 2.4419,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.083285093307495,
+      "rewards/margins": 0.5300815105438232,
+      "rewards/rejected": -2.6133666038513184,
+      "step": 330
+    },
+    {
+      "epoch": 0.6012715712988193,
+      "grad_norm": 1.7888520956039429,
+      "learning_rate": 9.506607935645579e-06,
+      "logits/chosen": 0.11793217062950134,
+      "logits/rejected": 0.19568441808223724,
+      "logps/chosen": -79.86200714111328,
+      "logps/rejected": -89.25471496582031,
+      "loss": 2.4273,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9273384809494019,
+      "rewards/margins": 0.5115458965301514,
+      "rewards/rejected": -2.4388844966888428,
+      "step": 331
+    },
+    {
+      "epoch": 0.6030881017257039,
+      "grad_norm": 2.0436160564422607,
+      "learning_rate": 9.503678342272306e-06,
+      "logits/chosen": 0.043473344296216965,
+      "logits/rejected": 0.15554016828536987,
+      "logps/chosen": -75.2901611328125,
+      "logps/rejected": -83.20555877685547,
+      "loss": 2.6941,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.725769281387329,
+      "rewards/margins": 0.2789250314235687,
+      "rewards/rejected": -2.0046942234039307,
+      "step": 332
+    },
+    {
+      "epoch": 0.6049046321525886,
+      "grad_norm": 1.665578842163086,
+      "learning_rate": 9.500740584701785e-06,
+      "logits/chosen": 0.17119848728179932,
+      "logits/rejected": 0.14128939807415009,
+      "logps/chosen": -83.1322021484375,
+      "logps/rejected": -93.38603210449219,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.8975954055786133,
+      "rewards/margins": 0.7810046672821045,
+      "rewards/rejected": -2.6786000728607178,
+      "step": 333
+    },
+    {
+      "epoch": 0.6067211625794732,
+      "grad_norm": 1.9258419275283813,
+      "learning_rate": 9.497794668924617e-06,
+      "logits/chosen": 0.028591612353920937,
+      "logits/rejected": 0.08894480764865875,
+      "logps/chosen": -78.12230682373047,
+      "logps/rejected": -90.00184631347656,
+      "loss": 2.4006,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.8113291263580322,
+      "rewards/margins": 0.5277000665664673,
+      "rewards/rejected": -2.339029550552368,
+      "step": 334
+    },
+    {
+      "epoch": 0.6085376930063578,
+      "grad_norm": 1.4878523349761963,
+      "learning_rate": 9.494840600948038e-06,
+      "logits/chosen": 0.03548199310898781,
+      "logits/rejected": 0.09170147776603699,
+      "logps/chosen": -73.15607452392578,
+      "logps/rejected": -83.77317810058594,
+      "loss": 2.3301,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.6781296730041504,
+      "rewards/margins": 0.4735565185546875,
+      "rewards/rejected": -2.151685953140259,
+      "step": 335
+    },
+    {
+      "epoch": 0.6103542234332425,
+      "grad_norm": 1.8042774200439453,
+      "learning_rate": 9.491878386795906e-06,
+      "logits/chosen": 0.046132348477840424,
+      "logits/rejected": 0.0721711814403534,
+      "logps/chosen": -81.01045227050781,
+      "logps/rejected": -89.28679656982422,
+      "loss": 2.5593,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.012852191925049,
+      "rewards/margins": 0.44549623131752014,
+      "rewards/rejected": -2.458348512649536,
+      "step": 336
+    },
+    {
+      "epoch": 0.6121707538601272,
+      "grad_norm": 2.048952102661133,
+      "learning_rate": 9.488908032508691e-06,
+      "logits/chosen": 0.10774732381105423,
+      "logits/rejected": 0.11935572326183319,
+      "logps/chosen": -91.25210571289062,
+      "logps/rejected": -90.42224884033203,
+      "loss": 2.8282,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0127718448638916,
+      "rewards/margins": 0.16889013350009918,
+      "rewards/rejected": -2.18166184425354,
+      "step": 337
+    },
+    {
+      "epoch": 0.6139872842870118,
+      "grad_norm": 1.6447219848632812,
+      "learning_rate": 9.485929544143462e-06,
+      "logits/chosen": 0.12260966747999191,
+      "logits/rejected": 0.1517799198627472,
+      "logps/chosen": -72.49166870117188,
+      "logps/rejected": -82.3083724975586,
+      "loss": 2.4029,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.624878168106079,
+      "rewards/margins": 0.5487539768218994,
+      "rewards/rejected": -2.1736321449279785,
+      "step": 338
+    },
+    {
+      "epoch": 0.6158038147138964,
+      "grad_norm": 1.599530577659607,
+      "learning_rate": 9.482942927773876e-06,
+      "logits/chosen": 0.09051798284053802,
+      "logits/rejected": 0.11202570050954819,
+      "logps/chosen": -84.42019653320312,
+      "logps/rejected": -92.61677551269531,
+      "loss": 2.2269,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6424546241760254,
+      "rewards/margins": 0.5715836882591248,
+      "rewards/rejected": -2.214038133621216,
+      "step": 339
+    },
+    {
+      "epoch": 0.6176203451407811,
+      "grad_norm": 1.5411081314086914,
+      "learning_rate": 9.479948189490164e-06,
+      "logits/chosen": 0.06563656777143478,
+      "logits/rejected": 0.17349205911159515,
+      "logps/chosen": -70.98219299316406,
+      "logps/rejected": -78.88846588134766,
+      "loss": 2.3328,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.824373722076416,
+      "rewards/margins": 0.46550822257995605,
+      "rewards/rejected": -2.289881706237793,
+      "step": 340
+    },
+    {
+      "epoch": 0.6194368755676658,
+      "grad_norm": 1.8855030536651611,
+      "learning_rate": 9.476945335399122e-06,
+      "logits/chosen": 0.11399642378091812,
+      "logits/rejected": 0.12690946459770203,
+      "logps/chosen": -85.86933898925781,
+      "logps/rejected": -89.75601196289062,
+      "loss": 2.4269,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9560539722442627,
+      "rewards/margins": 0.38427016139030457,
+      "rewards/rejected": -2.3403239250183105,
+      "step": 341
+    },
+    {
+      "epoch": 0.6212534059945504,
+      "grad_norm": 1.9973480701446533,
+      "learning_rate": 9.473934371624087e-06,
+      "logits/chosen": 0.029105912894010544,
+      "logits/rejected": 0.12763622403144836,
+      "logps/chosen": -80.68119049072266,
+      "logps/rejected": -93.90747833251953,
+      "loss": 2.4709,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.906477928161621,
+      "rewards/margins": 0.6422742009162903,
+      "rewards/rejected": -2.5487518310546875,
+      "step": 342
+    },
+    {
+      "epoch": 0.623069936421435,
+      "grad_norm": 1.7752224206924438,
+      "learning_rate": 9.47091530430494e-06,
+      "logits/chosen": 0.14096824824810028,
+      "logits/rejected": 0.1774113029241562,
+      "logps/chosen": -71.80216217041016,
+      "logps/rejected": -79.65301513671875,
+      "loss": 2.2678,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.5735952854156494,
+      "rewards/margins": 0.5823659300804138,
+      "rewards/rejected": -2.155961036682129,
+      "step": 343
+    },
+    {
+      "epoch": 0.6248864668483197,
+      "grad_norm": 1.855377197265625,
+      "learning_rate": 9.467888139598086e-06,
+      "logits/chosen": 0.10637074708938599,
+      "logits/rejected": 0.0798158049583435,
+      "logps/chosen": -79.01347351074219,
+      "logps/rejected": -80.77239990234375,
+      "loss": 2.7867,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9598251581192017,
+      "rewards/margins": 0.3184160590171814,
+      "rewards/rejected": -2.2782411575317383,
+      "step": 344
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 1.423040747642517,
+      "learning_rate": 9.464852883676441e-06,
+      "logits/chosen": 0.1304859220981598,
+      "logits/rejected": 0.1759231686592102,
+      "logps/chosen": -74.09733581542969,
+      "logps/rejected": -89.27589416503906,
+      "loss": 2.2777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7697277069091797,
+      "rewards/margins": 0.635611355304718,
+      "rewards/rejected": -2.405339002609253,
+      "step": 345
+    },
+    {
+      "epoch": 0.628519527702089,
+      "grad_norm": 1.6246038675308228,
+      "learning_rate": 9.461809542729421e-06,
+      "logits/chosen": 0.03750050812959671,
+      "logits/rejected": 0.10337980091571808,
+      "logps/chosen": -81.89110565185547,
+      "logps/rejected": -95.36811828613281,
+      "loss": 2.0438,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7299752235412598,
+      "rewards/margins": 0.7404756546020508,
+      "rewards/rejected": -2.4704508781433105,
+      "step": 346
+    },
+    {
+      "epoch": 0.6303360581289736,
+      "grad_norm": 1.6618752479553223,
+      "learning_rate": 9.458758122962926e-06,
+      "logits/chosen": 0.05359608680009842,
+      "logits/rejected": 0.10455545783042908,
+      "logps/chosen": -85.34078216552734,
+      "logps/rejected": -90.46200561523438,
+      "loss": 2.3299,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9331045150756836,
+      "rewards/margins": 0.4755082130432129,
+      "rewards/rejected": -2.4086129665374756,
+      "step": 347
+    },
+    {
+      "epoch": 0.6321525885558583,
+      "grad_norm": 1.5805696249008179,
+      "learning_rate": 9.455698630599332e-06,
+      "logits/chosen": 0.10048776119947433,
+      "logits/rejected": 0.1140337884426117,
+      "logps/chosen": -80.25875854492188,
+      "logps/rejected": -94.44998168945312,
+      "loss": 2.3439,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8421945571899414,
+      "rewards/margins": 0.5977468490600586,
+      "rewards/rejected": -2.43994140625,
+      "step": 348
+    },
+    {
+      "epoch": 0.633969118982743,
+      "grad_norm": 1.37093985080719,
+      "learning_rate": 9.452631071877478e-06,
+      "logits/chosen": 0.11764326691627502,
+      "logits/rejected": 0.10735289752483368,
+      "logps/chosen": -72.02367401123047,
+      "logps/rejected": -77.9522705078125,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.6527085304260254,
+      "rewards/margins": 0.6065118908882141,
+      "rewards/rejected": -2.259220600128174,
+      "step": 349
+    },
+    {
+      "epoch": 0.6357856494096276,
+      "grad_norm": 2.0214192867279053,
+      "learning_rate": 9.449555453052652e-06,
+      "logits/chosen": 0.13177426159381866,
+      "logits/rejected": 0.12408209592103958,
+      "logps/chosen": -76.62931823730469,
+      "logps/rejected": -81.16517639160156,
+      "loss": 2.9175,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": -1.9575048685073853,
+      "rewards/margins": 0.21441945433616638,
+      "rewards/rejected": -2.171924352645874,
+      "step": 350
+    },
+    {
+      "epoch": 0.6376021798365122,
+      "grad_norm": 1.6041687726974487,
+      "learning_rate": 9.446471780396573e-06,
+      "logits/chosen": 0.18574532866477966,
+      "logits/rejected": 0.22683225572109222,
+      "logps/chosen": -73.95652770996094,
+      "logps/rejected": -80.29524993896484,
+      "loss": 2.3833,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8684748411178589,
+      "rewards/margins": 0.37268272042274475,
+      "rewards/rejected": -2.2411575317382812,
+      "step": 351
+    },
+    {
+      "epoch": 0.6394187102633969,
+      "grad_norm": 1.7410838603973389,
+      "learning_rate": 9.443380060197387e-06,
+      "logits/chosen": 0.07876043766736984,
+      "logits/rejected": 0.14103996753692627,
+      "logps/chosen": -74.64432525634766,
+      "logps/rejected": -83.6999740600586,
+      "loss": 2.5558,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.029916763305664,
+      "rewards/margins": 0.4088842272758484,
+      "rewards/rejected": -2.438800811767578,
+      "step": 352
+    },
+    {
+      "epoch": 0.6412352406902816,
+      "grad_norm": 1.87971031665802,
+      "learning_rate": 9.440280298759653e-06,
+      "logits/chosen": 0.13997013866901398,
+      "logits/rejected": 0.1322249174118042,
+      "logps/chosen": -88.78776550292969,
+      "logps/rejected": -88.22732543945312,
+      "loss": 2.6547,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9983258247375488,
+      "rewards/margins": 0.2458382397890091,
+      "rewards/rejected": -2.244164228439331,
+      "step": 353
+    },
+    {
+      "epoch": 0.6430517711171662,
+      "grad_norm": 1.7056363821029663,
+      "learning_rate": 9.437172502404318e-06,
+      "logits/chosen": 0.07248476892709732,
+      "logits/rejected": 0.13019773364067078,
+      "logps/chosen": -78.4591064453125,
+      "logps/rejected": -80.49126434326172,
+      "loss": 2.3809,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9740041494369507,
+      "rewards/margins": 0.4853326082229614,
+      "rewards/rejected": -2.459336757659912,
+      "step": 354
+    },
+    {
+      "epoch": 0.6448683015440508,
+      "grad_norm": 1.435718059539795,
+      "learning_rate": 9.434056677468726e-06,
+      "logits/chosen": 0.09164869040250778,
+      "logits/rejected": 0.14243285357952118,
+      "logps/chosen": -77.83367156982422,
+      "logps/rejected": -86.18836212158203,
+      "loss": 2.006,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.7355122566223145,
+      "rewards/margins": 0.6869419813156128,
+      "rewards/rejected": -2.422454357147217,
+      "step": 355
+    },
+    {
+      "epoch": 0.6466848319709355,
+      "grad_norm": 1.6553188562393188,
+      "learning_rate": 9.430932830306587e-06,
+      "logits/chosen": 0.04967673122882843,
+      "logits/rejected": 0.1456151008605957,
+      "logps/chosen": -73.50204467773438,
+      "logps/rejected": -85.31363677978516,
+      "loss": 2.4388,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.978384256362915,
+      "rewards/margins": 0.48918718099594116,
+      "rewards/rejected": -2.467571496963501,
+      "step": 356
+    },
+    {
+      "epoch": 0.6485013623978202,
+      "grad_norm": 1.8443480730056763,
+      "learning_rate": 9.427800967287963e-06,
+      "logits/chosen": 0.06455090641975403,
+      "logits/rejected": 0.12608012557029724,
+      "logps/chosen": -75.49232482910156,
+      "logps/rejected": -77.49136352539062,
+      "loss": 2.706,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -1.8730812072753906,
+      "rewards/margins": 0.2989741861820221,
+      "rewards/rejected": -2.17205548286438,
+      "step": 357
+    },
+    {
+      "epoch": 0.6503178928247049,
+      "grad_norm": 1.6912306547164917,
+      "learning_rate": 9.424661094799273e-06,
+      "logits/chosen": 0.11878645420074463,
+      "logits/rejected": 0.13628609478473663,
+      "logps/chosen": -73.09883880615234,
+      "logps/rejected": -84.19624328613281,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.928421139717102,
+      "rewards/margins": 0.5017052292823792,
+      "rewards/rejected": -2.430126190185547,
+      "step": 358
+    },
+    {
+      "epoch": 0.6521344232515894,
+      "grad_norm": 1.3643461465835571,
+      "learning_rate": 9.421513219243262e-06,
+      "logits/chosen": 0.07683826237916946,
+      "logits/rejected": 0.15765298902988434,
+      "logps/chosen": -76.0871353149414,
+      "logps/rejected": -97.47781372070312,
+      "loss": 1.7844,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -1.8300602436065674,
+      "rewards/margins": 0.911116361618042,
+      "rewards/rejected": -2.7411766052246094,
+      "step": 359
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 1.6327749490737915,
+      "learning_rate": 9.418357347038999e-06,
+      "logits/chosen": 0.1078951433300972,
+      "logits/rejected": 0.12233921140432358,
+      "logps/chosen": -75.89913177490234,
+      "logps/rejected": -78.4587173461914,
+      "loss": 2.4976,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.7992844581604004,
+      "rewards/margins": 0.311392605304718,
+      "rewards/rejected": -2.1106772422790527,
+      "step": 360
+    },
+    {
+      "epoch": 0.6557674841053588,
+      "grad_norm": 1.463025689125061,
+      "learning_rate": 9.415193484621852e-06,
+      "logits/chosen": 0.057331383228302,
+      "logits/rejected": 0.1563551127910614,
+      "logps/chosen": -80.7889633178711,
+      "logps/rejected": -93.21602630615234,
+      "loss": 2.1398,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.744189739227295,
+      "rewards/margins": 0.576555073261261,
+      "rewards/rejected": -2.320744752883911,
+      "step": 361
+    },
+    {
+      "epoch": 0.6575840145322435,
+      "grad_norm": 1.743695616722107,
+      "learning_rate": 9.412021638443491e-06,
+      "logits/chosen": 0.03781123086810112,
+      "logits/rejected": 0.154897540807724,
+      "logps/chosen": -77.72047424316406,
+      "logps/rejected": -88.80599975585938,
+      "loss": 2.2361,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9849369525909424,
+      "rewards/margins": 0.6562294363975525,
+      "rewards/rejected": -2.6411664485931396,
+      "step": 362
+    },
+    {
+      "epoch": 0.659400544959128,
+      "grad_norm": 1.5648279190063477,
+      "learning_rate": 9.408841814971862e-06,
+      "logits/chosen": 0.10401102900505066,
+      "logits/rejected": 0.12699122726917267,
+      "logps/chosen": -83.10671997070312,
+      "logps/rejected": -86.8663101196289,
+      "loss": 2.4009,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7048468589782715,
+      "rewards/margins": 0.4090174734592438,
+      "rewards/rejected": -2.1138644218444824,
+      "step": 363
+    },
+    {
+      "epoch": 0.6612170753860127,
+      "grad_norm": 1.3385239839553833,
+      "learning_rate": 9.405654020691178e-06,
+      "logits/chosen": 0.06100422143936157,
+      "logits/rejected": 0.10872650146484375,
+      "logps/chosen": -76.72837829589844,
+      "logps/rejected": -87.63470458984375,
+      "loss": 2.141,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.747226595878601,
+      "rewards/margins": 0.615075945854187,
+      "rewards/rejected": -2.362302780151367,
+      "step": 364
+    },
+    {
+      "epoch": 0.6630336058128974,
+      "grad_norm": 1.4060734510421753,
+      "learning_rate": 9.402458262101906e-06,
+      "logits/chosen": 0.10638861358165741,
+      "logits/rejected": 0.16951681673526764,
+      "logps/chosen": -77.35757446289062,
+      "logps/rejected": -92.03012084960938,
+      "loss": 2.0511,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8359463214874268,
+      "rewards/margins": 0.7444422841072083,
+      "rewards/rejected": -2.5803885459899902,
+      "step": 365
+    },
+    {
+      "epoch": 0.6648501362397821,
+      "grad_norm": 1.4371014833450317,
+      "learning_rate": 9.399254545720757e-06,
+      "logits/chosen": 0.0383220911026001,
+      "logits/rejected": 0.05856206640601158,
+      "logps/chosen": -82.40064239501953,
+      "logps/rejected": -92.1117935180664,
+      "loss": 2.1756,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8580785989761353,
+      "rewards/margins": 0.6475991010665894,
+      "rewards/rejected": -2.5056777000427246,
+      "step": 366
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.6673945188522339,
+      "learning_rate": 9.396042878080661e-06,
+      "logits/chosen": 0.14657820761203766,
+      "logits/rejected": 0.19021111726760864,
+      "logps/chosen": -73.56724548339844,
+      "logps/rejected": -78.85279846191406,
+      "loss": 2.4018,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9079951047897339,
+      "rewards/margins": 0.3709837794303894,
+      "rewards/rejected": -2.2789790630340576,
+      "step": 367
+    },
+    {
+      "epoch": 0.6684831970935513,
+      "grad_norm": 1.7402448654174805,
+      "learning_rate": 9.392823265730775e-06,
+      "logits/chosen": 0.16515754163265228,
+      "logits/rejected": 0.12460774928331375,
+      "logps/chosen": -69.89266204833984,
+      "logps/rejected": -74.53837585449219,
+      "loss": 2.4262,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.0627622604370117,
+      "rewards/margins": 0.5317320823669434,
+      "rewards/rejected": -2.594494342803955,
+      "step": 368
+    },
+    {
+      "epoch": 0.670299727520436,
+      "grad_norm": 1.5290364027023315,
+      "learning_rate": 9.389595715236446e-06,
+      "logits/chosen": 0.0954434722661972,
+      "logits/rejected": 0.16226956248283386,
+      "logps/chosen": -77.17019653320312,
+      "logps/rejected": -91.07398986816406,
+      "loss": 2.2511,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.241170644760132,
+      "rewards/margins": 0.6319360136985779,
+      "rewards/rejected": -2.8731067180633545,
+      "step": 369
+    },
+    {
+      "epoch": 0.6721162579473207,
+      "grad_norm": 1.4633573293685913,
+      "learning_rate": 9.386360233179206e-06,
+      "logits/chosen": 0.06517557799816132,
+      "logits/rejected": 0.07250035554170609,
+      "logps/chosen": -79.23770141601562,
+      "logps/rejected": -88.50403594970703,
+      "loss": 2.3858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.751698613166809,
+      "rewards/margins": 0.49529415369033813,
+      "rewards/rejected": -2.246993064880371,
+      "step": 370
+    },
+    {
+      "epoch": 0.6739327883742052,
+      "grad_norm": 2.007215976715088,
+      "learning_rate": 9.383116826156775e-06,
+      "logits/chosen": 0.13584929704666138,
+      "logits/rejected": 0.10950647294521332,
+      "logps/chosen": -80.75408935546875,
+      "logps/rejected": -79.3903579711914,
+      "loss": 2.6967,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.2421724796295166,
+      "rewards/margins": 0.34328368306159973,
+      "rewards/rejected": -2.585456132888794,
+      "step": 371
+    },
+    {
+      "epoch": 0.6757493188010899,
+      "grad_norm": 1.7323477268218994,
+      "learning_rate": 9.37986550078302e-06,
+      "logits/chosen": 0.011626070365309715,
+      "logits/rejected": 0.06207559257745743,
+      "logps/chosen": -75.51494598388672,
+      "logps/rejected": -85.83218383789062,
+      "loss": 2.4608,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9569413661956787,
+      "rewards/margins": 0.5160585641860962,
+      "rewards/rejected": -2.4729998111724854,
+      "step": 372
+    },
+    {
+      "epoch": 0.6775658492279746,
+      "grad_norm": 1.6365752220153809,
+      "learning_rate": 9.376606263687959e-06,
+      "logits/chosen": 0.10213632136583328,
+      "logits/rejected": 0.08355780690908432,
+      "logps/chosen": -79.01370239257812,
+      "logps/rejected": -81.9417953491211,
+      "loss": 2.4701,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.064467668533325,
+      "rewards/margins": 0.4157037138938904,
+      "rewards/rejected": -2.4801712036132812,
+      "step": 373
+    },
+    {
+      "epoch": 0.6793823796548593,
+      "grad_norm": 1.5918922424316406,
+      "learning_rate": 9.373339121517748e-06,
+      "logits/chosen": 0.09486684203147888,
+      "logits/rejected": 0.07311725616455078,
+      "logps/chosen": -82.15351867675781,
+      "logps/rejected": -91.7690658569336,
+      "loss": 2.1297,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0153114795684814,
+      "rewards/margins": 0.669613242149353,
+      "rewards/rejected": -2.684924840927124,
+      "step": 374
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 1.6511566638946533,
+      "learning_rate": 9.370064080934654e-06,
+      "logits/chosen": 0.1406637728214264,
+      "logits/rejected": 0.20172299444675446,
+      "logps/chosen": -69.41023254394531,
+      "logps/rejected": -74.65138244628906,
+      "loss": 2.4704,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.061182737350464,
+      "rewards/margins": 0.3090137243270874,
+      "rewards/rejected": -2.3701963424682617,
+      "step": 375
+    },
+    {
+      "epoch": 0.6830154405086285,
+      "grad_norm": 1.770624041557312,
+      "learning_rate": 9.366781148617056e-06,
+      "logits/chosen": 0.10267248749732971,
+      "logits/rejected": 0.07900385558605194,
+      "logps/chosen": -77.27003479003906,
+      "logps/rejected": -88.09040832519531,
+      "loss": 2.2813,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.132927656173706,
+      "rewards/margins": 0.6629016399383545,
+      "rewards/rejected": -2.7958290576934814,
+      "step": 376
+    },
+    {
+      "epoch": 0.6848319709355132,
+      "grad_norm": 1.5172028541564941,
+      "learning_rate": 9.363490331259426e-06,
+      "logits/chosen": 0.05240853130817413,
+      "logits/rejected": 0.10445387661457062,
+      "logps/chosen": -75.5218276977539,
+      "logps/rejected": -85.49366760253906,
+      "loss": 2.1442,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9302880764007568,
+      "rewards/margins": 0.5776917338371277,
+      "rewards/rejected": -2.5079798698425293,
+      "step": 377
+    },
+    {
+      "epoch": 0.6866485013623979,
+      "grad_norm": 1.6540950536727905,
+      "learning_rate": 9.360191635572313e-06,
+      "logits/chosen": 0.14195458590984344,
+      "logits/rejected": 0.0978107899427414,
+      "logps/chosen": -85.00608825683594,
+      "logps/rejected": -85.1456298828125,
+      "loss": 2.423,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.8765841722488403,
+      "rewards/margins": 0.49441370368003845,
+      "rewards/rejected": -2.370997905731201,
+      "step": 378
+    },
+    {
+      "epoch": 0.6884650317892824,
+      "grad_norm": 1.6463801860809326,
+      "learning_rate": 9.356885068282334e-06,
+      "logits/chosen": 0.13197994232177734,
+      "logits/rejected": 0.09924699366092682,
+      "logps/chosen": -86.94219970703125,
+      "logps/rejected": -86.85704803466797,
+      "loss": 2.4801,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9674016237258911,
+      "rewards/margins": 0.4530283510684967,
+      "rewards/rejected": -2.4204299449920654,
+      "step": 379
+    },
+    {
+      "epoch": 0.6902815622161671,
+      "grad_norm": 1.5428895950317383,
+      "learning_rate": 9.353570636132151e-06,
+      "logits/chosen": 0.10434838384389877,
+      "logits/rejected": 0.12619757652282715,
+      "logps/chosen": -78.56607818603516,
+      "logps/rejected": -86.33320617675781,
+      "loss": 2.2711,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9667280912399292,
+      "rewards/margins": 0.6350463628768921,
+      "rewards/rejected": -2.601774215698242,
+      "step": 380
+    },
+    {
+      "epoch": 0.6920980926430518,
+      "grad_norm": 1.6211251020431519,
+      "learning_rate": 9.350248345880471e-06,
+      "logits/chosen": 0.14081251621246338,
+      "logits/rejected": 0.1453506350517273,
+      "logps/chosen": -74.38184356689453,
+      "logps/rejected": -80.61524200439453,
+      "loss": 2.3863,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.0189671516418457,
+      "rewards/margins": 0.4528267979621887,
+      "rewards/rejected": -2.4717938899993896,
+      "step": 381
+    },
+    {
+      "epoch": 0.6939146230699365,
+      "grad_norm": 2.6346240043640137,
+      "learning_rate": 9.346918204302022e-06,
+      "logits/chosen": 0.0519830696284771,
+      "logits/rejected": 0.04284125566482544,
+      "logps/chosen": -82.38153839111328,
+      "logps/rejected": -86.35488891601562,
+      "loss": 2.5434,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9824391603469849,
+      "rewards/margins": 0.4156208634376526,
+      "rewards/rejected": -2.3980600833892822,
+      "step": 382
+    },
+    {
+      "epoch": 0.695731153496821,
+      "grad_norm": 1.4762026071548462,
+      "learning_rate": 9.343580218187544e-06,
+      "logits/chosen": 0.02196469157934189,
+      "logits/rejected": 0.04618150740861893,
+      "logps/chosen": -72.66182708740234,
+      "logps/rejected": -88.68983459472656,
+      "loss": 2.1699,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9227774143218994,
+      "rewards/margins": 0.9200201630592346,
+      "rewards/rejected": -2.8427975177764893,
+      "step": 383
+    },
+    {
+      "epoch": 0.6975476839237057,
+      "grad_norm": 1.9675711393356323,
+      "learning_rate": 9.340234394343768e-06,
+      "logits/chosen": 0.12956203520298004,
+      "logits/rejected": 0.16079677641391754,
+      "logps/chosen": -87.38492584228516,
+      "logps/rejected": -87.80662536621094,
+      "loss": 2.3138,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.144991397857666,
+      "rewards/margins": 0.4893474578857422,
+      "rewards/rejected": -2.634338855743408,
+      "step": 384
+    },
+    {
+      "epoch": 0.6993642143505904,
+      "grad_norm": 1.9116860628128052,
+      "learning_rate": 9.336880739593415e-06,
+      "logits/chosen": 0.06013559550046921,
+      "logits/rejected": 0.09889352321624756,
+      "logps/chosen": -86.55718231201172,
+      "logps/rejected": -91.01994323730469,
+      "loss": 2.4864,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9528815746307373,
+      "rewards/margins": 0.44100096821784973,
+      "rewards/rejected": -2.3938825130462646,
+      "step": 385
+    },
+    {
+      "epoch": 0.701180744777475,
+      "grad_norm": 2.0582289695739746,
+      "learning_rate": 9.33351926077517e-06,
+      "logits/chosen": 0.11538423597812653,
+      "logits/rejected": 0.17568480968475342,
+      "logps/chosen": -76.24241638183594,
+      "logps/rejected": -87.1357650756836,
+      "loss": 2.6878,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -2.102504253387451,
+      "rewards/margins": 0.3671773076057434,
+      "rewards/rejected": -2.469681739807129,
+      "step": 386
+    },
+    {
+      "epoch": 0.7029972752043597,
+      "grad_norm": 1.6962640285491943,
+      "learning_rate": 9.330149964743674e-06,
+      "logits/chosen": 0.12425235658884048,
+      "logits/rejected": 0.1852879822254181,
+      "logps/chosen": -82.67887115478516,
+      "logps/rejected": -98.66133117675781,
+      "loss": 2.343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.182525634765625,
+      "rewards/margins": 0.6602460145950317,
+      "rewards/rejected": -2.842771530151367,
+      "step": 387
+    },
+    {
+      "epoch": 0.7048138056312443,
+      "grad_norm": 1.6582266092300415,
+      "learning_rate": 9.326772858369506e-06,
+      "logits/chosen": 0.14438432455062866,
+      "logits/rejected": 0.0922938883304596,
+      "logps/chosen": -78.87794494628906,
+      "logps/rejected": -84.929443359375,
+      "loss": 2.3036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.988234281539917,
+      "rewards/margins": 0.4737466275691986,
+      "rewards/rejected": -2.4619810581207275,
+      "step": 388
+    },
+    {
+      "epoch": 0.706630336058129,
+      "grad_norm": 1.6689919233322144,
+      "learning_rate": 9.323387948539176e-06,
+      "logits/chosen": 0.1282673329114914,
+      "logits/rejected": 0.13633723556995392,
+      "logps/chosen": -70.22606658935547,
+      "logps/rejected": -79.26126861572266,
+      "loss": 2.241,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9162399768829346,
+      "rewards/margins": 0.5345122814178467,
+      "rewards/rejected": -2.4507524967193604,
+      "step": 389
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 1.5861207246780396,
+      "learning_rate": 9.319995242155102e-06,
+      "logits/chosen": 0.13024169206619263,
+      "logits/rejected": 0.14390775561332703,
+      "logps/chosen": -87.00923919677734,
+      "logps/rejected": -91.24594116210938,
+      "loss": 2.284,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9127002954483032,
+      "rewards/margins": 0.5265774130821228,
+      "rewards/rejected": -2.4392776489257812,
+      "step": 390
+    },
+    {
+      "epoch": 0.7102633969118983,
+      "grad_norm": 2.0118792057037354,
+      "learning_rate": 9.316594746135608e-06,
+      "logits/chosen": 0.14648675918579102,
+      "logits/rejected": 0.11786539107561111,
+      "logps/chosen": -82.68962097167969,
+      "logps/rejected": -88.26126861572266,
+      "loss": 2.7281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.9463748931884766,
+      "rewards/margins": 0.20282456278800964,
+      "rewards/rejected": -2.1491994857788086,
+      "step": 391
+    },
+    {
+      "epoch": 0.7120799273387829,
+      "grad_norm": 1.8176332712173462,
+      "learning_rate": 9.313186467414892e-06,
+      "logits/chosen": 0.12084021419286728,
+      "logits/rejected": 0.1433698982000351,
+      "logps/chosen": -73.46707153320312,
+      "logps/rejected": -82.46024322509766,
+      "loss": 2.5247,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.0802624225616455,
+      "rewards/margins": 0.41601985692977905,
+      "rewards/rejected": -2.4962823390960693,
+      "step": 392
+    },
+    {
+      "epoch": 0.7138964577656676,
+      "grad_norm": 1.7207351922988892,
+      "learning_rate": 9.30977041294303e-06,
+      "logits/chosen": 0.11983273923397064,
+      "logits/rejected": 0.16106371581554413,
+      "logps/chosen": -76.09751892089844,
+      "logps/rejected": -85.20323944091797,
+      "loss": 2.5284,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -1.9018669128417969,
+      "rewards/margins": 0.4894491136074066,
+      "rewards/rejected": -2.3913159370422363,
+      "step": 393
+    },
+    {
+      "epoch": 0.7157129881925522,
+      "grad_norm": 1.6489367485046387,
+      "learning_rate": 9.306346589685956e-06,
+      "logits/chosen": 0.14845696091651917,
+      "logits/rejected": 0.10356368124485016,
+      "logps/chosen": -83.94526672363281,
+      "logps/rejected": -86.53378295898438,
+      "loss": 2.3394,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9089899063110352,
+      "rewards/margins": 0.4619474411010742,
+      "rewards/rejected": -2.3709375858306885,
+      "step": 394
+    },
+    {
+      "epoch": 0.7175295186194369,
+      "grad_norm": 1.792077898979187,
+      "learning_rate": 9.302915004625435e-06,
+      "logits/chosen": 0.06936248391866684,
+      "logits/rejected": 0.13349927961826324,
+      "logps/chosen": -91.82075500488281,
+      "logps/rejected": -98.7391357421875,
+      "loss": 2.3875,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9371806383132935,
+      "rewards/margins": 0.6781354546546936,
+      "rewards/rejected": -2.6153156757354736,
+      "step": 395
+    },
+    {
+      "epoch": 0.7193460490463215,
+      "grad_norm": 2.049694538116455,
+      "learning_rate": 9.29947566475907e-06,
+      "logits/chosen": 0.08474650233983994,
+      "logits/rejected": 0.13380834460258484,
+      "logps/chosen": -84.59528350830078,
+      "logps/rejected": -96.24267578125,
+      "loss": 2.7001,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9375882148742676,
+      "rewards/margins": 0.4607384204864502,
+      "rewards/rejected": -2.3983266353607178,
+      "step": 396
+    },
+    {
+      "epoch": 0.7211625794732062,
+      "grad_norm": 1.9235875606536865,
+      "learning_rate": 9.296028577100271e-06,
+      "logits/chosen": 0.06255945563316345,
+      "logits/rejected": 0.17074811458587646,
+      "logps/chosen": -68.99498748779297,
+      "logps/rejected": -84.45899963378906,
+      "loss": 2.347,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9344629049301147,
+      "rewards/margins": 0.688963770866394,
+      "rewards/rejected": -2.6234264373779297,
+      "step": 397
+    },
+    {
+      "epoch": 0.7229791099000908,
+      "grad_norm": 1.6617276668548584,
+      "learning_rate": 9.292573748678254e-06,
+      "logits/chosen": 0.11962493509054184,
+      "logits/rejected": 0.11647717654705048,
+      "logps/chosen": -85.3626708984375,
+      "logps/rejected": -95.70562744140625,
+      "loss": 2.2229,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.8968594074249268,
+      "rewards/margins": 0.6918852925300598,
+      "rewards/rejected": -2.588744640350342,
+      "step": 398
+    },
+    {
+      "epoch": 0.7247956403269755,
+      "grad_norm": 1.4650444984436035,
+      "learning_rate": 9.289111186538013e-06,
+      "logits/chosen": 0.06390775740146637,
+      "logits/rejected": 0.11076060682535172,
+      "logps/chosen": -69.9189453125,
+      "logps/rejected": -81.5618667602539,
+      "loss": 2.2271,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.857581377029419,
+      "rewards/margins": 0.5540103316307068,
+      "rewards/rejected": -2.4115917682647705,
+      "step": 399
+    },
+    {
+      "epoch": 0.7266121707538601,
+      "grad_norm": 1.5728726387023926,
+      "learning_rate": 9.285640897740316e-06,
+      "logits/chosen": 0.08816932141780853,
+      "logits/rejected": 0.1330798715353012,
+      "logps/chosen": -78.04735565185547,
+      "logps/rejected": -87.89193725585938,
+      "loss": 2.3595,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.823161244392395,
+      "rewards/margins": 0.4568220376968384,
+      "rewards/rejected": -2.2799830436706543,
+      "step": 400
+    },
+    {
+      "epoch": 0.7284287011807448,
+      "grad_norm": 1.6291779279708862,
+      "learning_rate": 9.282162889361686e-06,
+      "logits/chosen": 0.17718347907066345,
+      "logits/rejected": 0.20603135228157043,
+      "logps/chosen": -67.99884033203125,
+      "logps/rejected": -75.93153381347656,
+      "loss": 2.3002,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.671633243560791,
+      "rewards/margins": 0.45543500781059265,
+      "rewards/rejected": -2.127068281173706,
+      "step": 401
+    },
+    {
+      "epoch": 0.7302452316076294,
+      "grad_norm": 1.5066276788711548,
+      "learning_rate": 9.278677168494388e-06,
+      "logits/chosen": 0.16840124130249023,
+      "logits/rejected": 0.17102479934692383,
+      "logps/chosen": -78.1556625366211,
+      "logps/rejected": -84.91495513916016,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.9279460906982422,
+      "rewards/margins": 0.6659258008003235,
+      "rewards/rejected": -2.593871831893921,
+      "step": 402
+    },
+    {
+      "epoch": 0.7320617620345141,
+      "grad_norm": 1.71995210647583,
+      "learning_rate": 9.275183742246412e-06,
+      "logits/chosen": -0.0005522281862795353,
+      "logits/rejected": 0.1412215530872345,
+      "logps/chosen": -73.37832641601562,
+      "logps/rejected": -92.71226501464844,
+      "loss": 2.3315,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6880759000778198,
+      "rewards/margins": 0.5844776630401611,
+      "rewards/rejected": -2.2725534439086914,
+      "step": 403
+    },
+    {
+      "epoch": 0.7338782924613987,
+      "grad_norm": 1.676321029663086,
+      "learning_rate": 9.271682617741466e-06,
+      "logits/chosen": 0.15319044888019562,
+      "logits/rejected": 0.14503052830696106,
+      "logps/chosen": -75.85136413574219,
+      "logps/rejected": -82.61698150634766,
+      "loss": 2.3528,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.819403052330017,
+      "rewards/margins": 0.3919827342033386,
+      "rewards/rejected": -2.211385726928711,
+      "step": 404
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 1.4811294078826904,
+      "learning_rate": 9.268173802118949e-06,
+      "logits/chosen": 0.07122528553009033,
+      "logits/rejected": 0.12146291136741638,
+      "logps/chosen": -76.78916931152344,
+      "logps/rejected": -85.33805084228516,
+      "loss": 2.1505,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.9886473417282104,
+      "rewards/margins": 0.5911746025085449,
+      "rewards/rejected": -2.579822301864624,
+      "step": 405
+    },
+    {
+      "epoch": 0.737511353315168,
+      "grad_norm": 1.629601001739502,
+      "learning_rate": 9.264657302533947e-06,
+      "logits/chosen": 0.11274963617324829,
+      "logits/rejected": 0.0651293620467186,
+      "logps/chosen": -82.86985778808594,
+      "logps/rejected": -88.68803405761719,
+      "loss": 2.1873,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.8538885116577148,
+      "rewards/margins": 0.6726438403129578,
+      "rewards/rejected": -2.5265324115753174,
+      "step": 406
+    },
+    {
+      "epoch": 0.7393278837420527,
+      "grad_norm": 1.5687483549118042,
+      "learning_rate": 9.261133126157218e-06,
+      "logits/chosen": 0.06054290384054184,
+      "logits/rejected": 0.1543959081172943,
+      "logps/chosen": -68.5821304321289,
+      "logps/rejected": -80.544921875,
+      "loss": 2.3587,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.7550634145736694,
+      "rewards/margins": 0.46927812695503235,
+      "rewards/rejected": -2.224341630935669,
+      "step": 407
+    },
+    {
+      "epoch": 0.7411444141689373,
+      "grad_norm": 1.849908471107483,
+      "learning_rate": 9.257601280175167e-06,
+      "logits/chosen": 0.15177126228809357,
+      "logits/rejected": 0.1485670655965805,
+      "logps/chosen": -72.81680297851562,
+      "logps/rejected": -80.58684539794922,
+      "loss": 2.444,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9739688634872437,
+      "rewards/margins": 0.6166737079620361,
+      "rewards/rejected": -2.5906424522399902,
+      "step": 408
+    },
+    {
+      "epoch": 0.742960944595822,
+      "grad_norm": 1.779342770576477,
+      "learning_rate": 9.254061771789847e-06,
+      "logits/chosen": 0.05093669891357422,
+      "logits/rejected": 0.1048060953617096,
+      "logps/chosen": -84.48141479492188,
+      "logps/rejected": -91.49285125732422,
+      "loss": 2.5324,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9952929019927979,
+      "rewards/margins": 0.3900759220123291,
+      "rewards/rejected": -2.385368585586548,
+      "step": 409
+    },
+    {
+      "epoch": 0.7447774750227066,
+      "grad_norm": 1.71133291721344,
+      "learning_rate": 9.25051460821893e-06,
+      "logits/chosen": 0.1308833658695221,
+      "logits/rejected": 0.08994461596012115,
+      "logps/chosen": -74.09046936035156,
+      "logps/rejected": -80.80438995361328,
+      "loss": 2.1895,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.8457674980163574,
+      "rewards/margins": 0.5731581449508667,
+      "rewards/rejected": -2.4189257621765137,
+      "step": 410
+    },
+    {
+      "epoch": 0.7465940054495913,
+      "grad_norm": 1.6980112791061401,
+      "learning_rate": 9.2469597966957e-06,
+      "logits/chosen": 0.05856658145785332,
+      "logits/rejected": 0.08371179550886154,
+      "logps/chosen": -78.58245849609375,
+      "logps/rejected": -88.30054473876953,
+      "loss": 2.46,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9803493022918701,
+      "rewards/margins": 0.6115507483482361,
+      "rewards/rejected": -2.591899871826172,
+      "step": 411
+    },
+    {
+      "epoch": 0.7484105358764759,
+      "grad_norm": 1.8350476026535034,
+      "learning_rate": 9.243397344469037e-06,
+      "logits/chosen": 0.05285171419382095,
+      "logits/rejected": 0.13240401446819305,
+      "logps/chosen": -73.79881286621094,
+      "logps/rejected": -86.56880187988281,
+      "loss": 2.2929,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.060075283050537,
+      "rewards/margins": 0.54813152551651,
+      "rewards/rejected": -2.6082065105438232,
+      "step": 412
+    },
+    {
+      "epoch": 0.7502270663033606,
+      "grad_norm": 1.4403142929077148,
+      "learning_rate": 9.239827258803402e-06,
+      "logits/chosen": 0.07327421009540558,
+      "logits/rejected": 0.18387822806835175,
+      "logps/chosen": -68.3246078491211,
+      "logps/rejected": -81.33067321777344,
+      "loss": 1.8794,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -1.8658640384674072,
+      "rewards/margins": 0.7600168585777283,
+      "rewards/rejected": -2.625880718231201,
+      "step": 413
+    },
+    {
+      "epoch": 0.7520435967302452,
+      "grad_norm": 1.883155345916748,
+      "learning_rate": 9.23624954697882e-06,
+      "logits/chosen": 0.10411994159221649,
+      "logits/rejected": 0.06234448403120041,
+      "logps/chosen": -78.31743621826172,
+      "logps/rejected": -80.0100326538086,
+      "loss": 2.5646,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.934099555015564,
+      "rewards/margins": 0.28737103939056396,
+      "rewards/rejected": -2.221470832824707,
+      "step": 414
+    },
+    {
+      "epoch": 0.7538601271571299,
+      "grad_norm": 1.5926934480667114,
+      "learning_rate": 9.232664216290868e-06,
+      "logits/chosen": 0.09589770436286926,
+      "logits/rejected": 0.041689179837703705,
+      "logps/chosen": -77.89552307128906,
+      "logps/rejected": -84.02517700195312,
+      "loss": 2.2284,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.787453532218933,
+      "rewards/margins": 0.49444466829299927,
+      "rewards/rejected": -2.281898260116577,
+      "step": 415
+    },
+    {
+      "epoch": 0.7556766575840145,
+      "grad_norm": 1.511832356452942,
+      "learning_rate": 9.229071274050663e-06,
+      "logits/chosen": 0.07688678801059723,
+      "logits/rejected": 0.14106512069702148,
+      "logps/chosen": -78.53334045410156,
+      "logps/rejected": -86.92496490478516,
+      "loss": 2.2211,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0686798095703125,
+      "rewards/margins": 0.5378819704055786,
+      "rewards/rejected": -2.6065618991851807,
+      "step": 416
+    },
+    {
+      "epoch": 0.7574931880108992,
+      "grad_norm": 1.7858117818832397,
+      "learning_rate": 9.225470727584835e-06,
+      "logits/chosen": 0.14854730665683746,
+      "logits/rejected": 0.06465649604797363,
+      "logps/chosen": -87.31832122802734,
+      "logps/rejected": -79.3746566772461,
+      "loss": 2.4446,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -1.9429931640625,
+      "rewards/margins": 0.38861486315727234,
+      "rewards/rejected": -2.3316078186035156,
+      "step": 417
+    },
+    {
+      "epoch": 0.7593097184377838,
+      "grad_norm": 1.9230687618255615,
+      "learning_rate": 9.221862584235527e-06,
+      "logits/chosen": 0.07233145087957382,
+      "logits/rejected": 0.13825634121894836,
+      "logps/chosen": -81.53103637695312,
+      "logps/rejected": -90.54696655273438,
+      "loss": 2.7171,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.065859794616699,
+      "rewards/margins": 0.30683329701423645,
+      "rewards/rejected": -2.3726933002471924,
+      "step": 418
+    },
+    {
+      "epoch": 0.7611262488646685,
+      "grad_norm": 2.063098192214966,
+      "learning_rate": 9.218246851360374e-06,
+      "logits/chosen": 0.10360075533390045,
+      "logits/rejected": 0.14677830040454865,
+      "logps/chosen": -85.08120727539062,
+      "logps/rejected": -89.35393524169922,
+      "loss": 2.4292,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.300565719604492,
+      "rewards/margins": 0.4277467727661133,
+      "rewards/rejected": -2.7283124923706055,
+      "step": 419
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 2.4728786945343018,
+      "learning_rate": 9.214623536332483e-06,
+      "logits/chosen": 0.08115474879741669,
+      "logits/rejected": 0.07649822533130646,
+      "logps/chosen": -83.79824829101562,
+      "logps/rejected": -87.47264099121094,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.0838184356689453,
+      "rewards/margins": 0.41770505905151367,
+      "rewards/rejected": -2.501523494720459,
+      "step": 420
+    },
+    {
+      "epoch": 0.7647593097184378,
+      "grad_norm": 1.7078697681427002,
+      "learning_rate": 9.210992646540425e-06,
+      "logits/chosen": 0.08155008405447006,
+      "logits/rejected": 0.08723931759595871,
+      "logps/chosen": -74.4200210571289,
+      "logps/rejected": -87.49038696289062,
+      "loss": 2.2064,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9076974391937256,
+      "rewards/margins": 0.6807244420051575,
+      "rewards/rejected": -2.5884220600128174,
+      "step": 421
+    },
+    {
+      "epoch": 0.7665758401453224,
+      "grad_norm": 1.5644993782043457,
+      "learning_rate": 9.207354189388214e-06,
+      "logits/chosen": 0.16107802093029022,
+      "logits/rejected": 0.15951679646968842,
+      "logps/chosen": -74.85030364990234,
+      "logps/rejected": -81.25472259521484,
+      "loss": 2.2389,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.0082590579986572,
+      "rewards/margins": 0.5728211402893066,
+      "rewards/rejected": -2.581080198287964,
+      "step": 422
+    },
+    {
+      "epoch": 0.7683923705722071,
+      "grad_norm": 1.6519557237625122,
+      "learning_rate": 9.203708172295299e-06,
+      "logits/chosen": 0.08725707978010178,
+      "logits/rejected": 0.11393023282289505,
+      "logps/chosen": -79.56071472167969,
+      "logps/rejected": -90.27279663085938,
+      "loss": 2.2853,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9312469959259033,
+      "rewards/margins": 0.6229196786880493,
+      "rewards/rejected": -2.554166793823242,
+      "step": 423
+    },
+    {
+      "epoch": 0.7702089009990918,
+      "grad_norm": 1.668648600578308,
+      "learning_rate": 9.200054602696544e-06,
+      "logits/chosen": 0.1369432657957077,
+      "logits/rejected": 0.08549812436103821,
+      "logps/chosen": -80.69041442871094,
+      "logps/rejected": -85.47969055175781,
+      "loss": 2.3022,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.7405030727386475,
+      "rewards/margins": 0.4421282112598419,
+      "rewards/rejected": -2.182631492614746,
+      "step": 424
+    },
+    {
+      "epoch": 0.7720254314259763,
+      "grad_norm": 1.7834622859954834,
+      "learning_rate": 9.196393488042213e-06,
+      "logits/chosen": 0.06559979915618896,
+      "logits/rejected": 0.05396304652094841,
+      "logps/chosen": -76.51896667480469,
+      "logps/rejected": -88.19717407226562,
+      "loss": 2.5216,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.050562620162964,
+      "rewards/margins": 0.5811472535133362,
+      "rewards/rejected": -2.631709575653076,
+      "step": 425
+    },
+    {
+      "epoch": 0.773841961852861,
+      "grad_norm": 1.5034058094024658,
+      "learning_rate": 9.192724835797956e-06,
+      "logits/chosen": 0.10820607095956802,
+      "logits/rejected": 0.14427171647548676,
+      "logps/chosen": -81.0079116821289,
+      "logps/rejected": -95.6849136352539,
+      "loss": 2.0472,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.19939923286438,
+      "rewards/margins": 0.8476728796958923,
+      "rewards/rejected": -3.047071933746338,
+      "step": 426
+    },
+    {
+      "epoch": 0.7756584922797457,
+      "grad_norm": 1.7205973863601685,
+      "learning_rate": 9.189048653444796e-06,
+      "logits/chosen": 0.08522580564022064,
+      "logits/rejected": 0.09135682880878448,
+      "logps/chosen": -77.4021224975586,
+      "logps/rejected": -85.47643280029297,
+      "loss": 2.2618,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.002034902572632,
+      "rewards/margins": 0.5705874562263489,
+      "rewards/rejected": -2.572622776031494,
+      "step": 427
+    },
+    {
+      "epoch": 0.7774750227066304,
+      "grad_norm": 2.131197214126587,
+      "learning_rate": 9.185364948479109e-06,
+      "logits/chosen": 0.05514610558748245,
+      "logits/rejected": 0.0597330704331398,
+      "logps/chosen": -86.83938598632812,
+      "logps/rejected": -94.71990966796875,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.088331699371338,
+      "rewards/margins": 0.47913864254951477,
+      "rewards/rejected": -2.567470073699951,
+      "step": 428
+    },
+    {
+      "epoch": 0.779291553133515,
+      "grad_norm": 3.0228030681610107,
+      "learning_rate": 9.181673728412605e-06,
+      "logits/chosen": 0.09502460807561874,
+      "logits/rejected": 0.10438862442970276,
+      "logps/chosen": -77.20478820800781,
+      "logps/rejected": -80.07164764404297,
+      "loss": 2.5924,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.107954978942871,
+      "rewards/margins": 0.4795181155204773,
+      "rewards/rejected": -2.5874733924865723,
+      "step": 429
+    },
+    {
+      "epoch": 0.7811080835603996,
+      "grad_norm": 2.630490779876709,
+      "learning_rate": 9.17797500077233e-06,
+      "logits/chosen": 0.053769052028656006,
+      "logits/rejected": 0.1357191652059555,
+      "logps/chosen": -74.01277160644531,
+      "logps/rejected": -86.05644989013672,
+      "loss": 2.6208,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0179696083068848,
+      "rewards/margins": 0.3262583613395691,
+      "rewards/rejected": -2.3442280292510986,
+      "step": 430
+    },
+    {
+      "epoch": 0.7829246139872843,
+      "grad_norm": 1.7841664552688599,
+      "learning_rate": 9.174268773100636e-06,
+      "logits/chosen": 0.09468546509742737,
+      "logits/rejected": 0.16804733872413635,
+      "logps/chosen": -80.07379913330078,
+      "logps/rejected": -87.17518615722656,
+      "loss": 2.4676,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.27384877204895,
+      "rewards/margins": 0.4068894386291504,
+      "rewards/rejected": -2.6807379722595215,
+      "step": 431
+    },
+    {
+      "epoch": 0.784741144414169,
+      "grad_norm": 1.7030746936798096,
+      "learning_rate": 9.170555052955158e-06,
+      "logits/chosen": 0.10446357727050781,
+      "logits/rejected": 0.1235620528459549,
+      "logps/chosen": -89.81584930419922,
+      "logps/rejected": -94.43669128417969,
+      "loss": 2.3366,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.98858642578125,
+      "rewards/margins": 0.5402320623397827,
+      "rewards/rejected": -2.5288188457489014,
+      "step": 432
+    },
+    {
+      "epoch": 0.7865576748410535,
+      "grad_norm": 1.4258465766906738,
+      "learning_rate": 9.166833847908825e-06,
+      "logits/chosen": 0.10827025771141052,
+      "logits/rejected": 0.10875076055526733,
+      "logps/chosen": -69.29400634765625,
+      "logps/rejected": -80.72003936767578,
+      "loss": 2.2501,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9478650093078613,
+      "rewards/margins": 0.5940475463867188,
+      "rewards/rejected": -2.54191255569458,
+      "step": 433
+    },
+    {
+      "epoch": 0.7883742052679382,
+      "grad_norm": 1.8069814443588257,
+      "learning_rate": 9.163105165549819e-06,
+      "logits/chosen": 0.08856840431690216,
+      "logits/rejected": 0.07843751460313797,
+      "logps/chosen": -79.93275451660156,
+      "logps/rejected": -83.75475311279297,
+      "loss": 2.5214,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.0315518379211426,
+      "rewards/margins": 0.45457565784454346,
+      "rewards/rejected": -2.4861276149749756,
+      "step": 434
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 1.7881503105163574,
+      "learning_rate": 9.159369013481574e-06,
+      "logits/chosen": 0.08166562020778656,
+      "logits/rejected": 0.08240097761154175,
+      "logps/chosen": -92.98298645019531,
+      "logps/rejected": -96.15763092041016,
+      "loss": 2.053,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4435133934020996,
+      "rewards/margins": 0.7140956521034241,
+      "rewards/rejected": -3.157609462738037,
+      "step": 435
+    },
+    {
+      "epoch": 0.7920072661217076,
+      "grad_norm": 1.7574496269226074,
+      "learning_rate": 9.155625399322754e-06,
+      "logits/chosen": 0.08114133030176163,
+      "logits/rejected": 0.11279730498790741,
+      "logps/chosen": -74.07096862792969,
+      "logps/rejected": -80.94386291503906,
+      "loss": 2.4289,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2497410774230957,
+      "rewards/margins": 0.4424628019332886,
+      "rewards/rejected": -2.6922037601470947,
+      "step": 436
+    },
+    {
+      "epoch": 0.7938237965485921,
+      "grad_norm": 1.4814083576202393,
+      "learning_rate": 9.15187433070724e-06,
+      "logits/chosen": 0.026073571294546127,
+      "logits/rejected": 0.1382063925266266,
+      "logps/chosen": -72.39479064941406,
+      "logps/rejected": -86.76889038085938,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9130635261535645,
+      "rewards/margins": 0.6362709999084473,
+      "rewards/rejected": -2.549334764480591,
+      "step": 437
+    },
+    {
+      "epoch": 0.7956403269754768,
+      "grad_norm": 1.6441129446029663,
+      "learning_rate": 9.148115815284113e-06,
+      "logits/chosen": 0.07432619482278824,
+      "logits/rejected": 0.10215617716312408,
+      "logps/chosen": -77.05310821533203,
+      "logps/rejected": -80.81717681884766,
+      "loss": 2.338,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9012683629989624,
+      "rewards/margins": 0.43776583671569824,
+      "rewards/rejected": -2.339034080505371,
+      "step": 438
+    },
+    {
+      "epoch": 0.7974568574023615,
+      "grad_norm": 1.8839560747146606,
+      "learning_rate": 9.144349860717643e-06,
+      "logits/chosen": 0.07576426863670349,
+      "logits/rejected": 0.08868670463562012,
+      "logps/chosen": -74.13190460205078,
+      "logps/rejected": -80.07232666015625,
+      "loss": 2.3485,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9554617404937744,
+      "rewards/margins": 0.531810998916626,
+      "rewards/rejected": -2.4872727394104004,
+      "step": 439
+    },
+    {
+      "epoch": 0.7992733878292462,
+      "grad_norm": 1.826811671257019,
+      "learning_rate": 9.140576474687263e-06,
+      "logits/chosen": 0.09120994061231613,
+      "logits/rejected": 0.08987519890069962,
+      "logps/chosen": -75.1444320678711,
+      "logps/rejected": -78.58987426757812,
+      "loss": 2.6513,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158261299133301,
+      "rewards/margins": 0.28016677498817444,
+      "rewards/rejected": -2.4384284019470215,
+      "step": 440
+    },
+    {
+      "epoch": 0.8010899182561307,
+      "grad_norm": 1.8647096157073975,
+      "learning_rate": 9.13679566488757e-06,
+      "logits/chosen": 0.12438184767961502,
+      "logits/rejected": 0.13951222598552704,
+      "logps/chosen": -83.79389953613281,
+      "logps/rejected": -90.33230590820312,
+      "loss": 2.3956,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0185930728912354,
+      "rewards/margins": 0.43685024976730347,
+      "rewards/rejected": -2.4554433822631836,
+      "step": 441
+    },
+    {
+      "epoch": 0.8029064486830154,
+      "grad_norm": 1.7610801458358765,
+      "learning_rate": 9.133007439028288e-06,
+      "logits/chosen": 0.08218151330947876,
+      "logits/rejected": 0.072816863656044,
+      "logps/chosen": -86.35633087158203,
+      "logps/rejected": -84.08189392089844,
+      "loss": 2.243,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.8836033344268799,
+      "rewards/margins": 0.516740083694458,
+      "rewards/rejected": -2.400343418121338,
+      "step": 442
+    },
+    {
+      "epoch": 0.8047229791099001,
+      "grad_norm": 1.6049703359603882,
+      "learning_rate": 9.129211804834271e-06,
+      "logits/chosen": 0.059631846845149994,
+      "logits/rejected": 0.09744230657815933,
+      "logps/chosen": -76.90062713623047,
+      "logps/rejected": -82.29356384277344,
+      "loss": 2.2574,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.9436204433441162,
+      "rewards/margins": 0.47444695234298706,
+      "rewards/rejected": -2.418067216873169,
+      "step": 443
+    },
+    {
+      "epoch": 0.8065395095367848,
+      "grad_norm": 1.9819791316986084,
+      "learning_rate": 9.12540877004548e-06,
+      "logits/chosen": 0.07371871173381805,
+      "logits/rejected": 0.058843065053224564,
+      "logps/chosen": -78.42516326904297,
+      "logps/rejected": -82.23822784423828,
+      "loss": 2.6959,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.273909330368042,
+      "rewards/margins": 0.3514346182346344,
+      "rewards/rejected": -2.6253440380096436,
+      "step": 444
+    },
+    {
+      "epoch": 0.8083560399636693,
+      "grad_norm": 1.9540653228759766,
+      "learning_rate": 9.12159834241696e-06,
+      "logits/chosen": 0.13155074417591095,
+      "logits/rejected": 0.11775672435760498,
+      "logps/chosen": -69.17151641845703,
+      "logps/rejected": -77.9190902709961,
+      "loss": 2.6036,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.073258638381958,
+      "rewards/margins": 0.5955159664154053,
+      "rewards/rejected": -2.6687746047973633,
+      "step": 445
+    },
+    {
+      "epoch": 0.810172570390554,
+      "grad_norm": 1.9634339809417725,
+      "learning_rate": 9.117780529718843e-06,
+      "logits/chosen": 0.05265370383858681,
+      "logits/rejected": 0.09651105850934982,
+      "logps/chosen": -78.77818298339844,
+      "logps/rejected": -84.63584899902344,
+      "loss": 2.2157,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.084688186645508,
+      "rewards/margins": 0.541061282157898,
+      "rewards/rejected": -2.625749349594116,
+      "step": 446
+    },
+    {
+      "epoch": 0.8119891008174387,
+      "grad_norm": 1.9086997509002686,
+      "learning_rate": 9.113955339736309e-06,
+      "logits/chosen": 0.08473093807697296,
+      "logits/rejected": 0.047919195145368576,
+      "logps/chosen": -83.44226837158203,
+      "logps/rejected": -87.24287414550781,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9542481899261475,
+      "rewards/margins": 0.4577428698539734,
+      "rewards/rejected": -2.4119908809661865,
+      "step": 447
+    },
+    {
+      "epoch": 0.8138056312443234,
+      "grad_norm": 1.6967483758926392,
+      "learning_rate": 9.11012278026959e-06,
+      "logits/chosen": 0.062431350350379944,
+      "logits/rejected": 0.10980932414531708,
+      "logps/chosen": -90.98541259765625,
+      "logps/rejected": -102.51634979248047,
+      "loss": 2.3339,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1866049766540527,
+      "rewards/margins": 0.6549904346466064,
+      "rewards/rejected": -2.841595411300659,
+      "step": 448
+    },
+    {
+      "epoch": 0.815622161671208,
+      "grad_norm": 2.295504331588745,
+      "learning_rate": 9.106282859133936e-06,
+      "logits/chosen": 0.04481218010187149,
+      "logits/rejected": 0.002619542181491852,
+      "logps/chosen": -81.81380462646484,
+      "logps/rejected": -80.60113525390625,
+      "loss": 2.8324,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.1727848052978516,
+      "rewards/margins": 0.2777697443962097,
+      "rewards/rejected": -2.450554609298706,
+      "step": 449
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 1.7951968908309937,
+      "learning_rate": 9.102435584159623e-06,
+      "logits/chosen": 0.029065577313303947,
+      "logits/rejected": 0.11651361733675003,
+      "logps/chosen": -78.08231353759766,
+      "logps/rejected": -88.59742736816406,
+      "loss": 2.5562,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.2762022018432617,
+      "rewards/margins": 0.4410257041454315,
+      "rewards/rejected": -2.7172276973724365,
+      "step": 450
+    },
+    {
+      "epoch": 0.8192552225249773,
+      "grad_norm": 1.6301803588867188,
+      "learning_rate": 9.098580963191908e-06,
+      "logits/chosen": 0.06471782922744751,
+      "logits/rejected": 0.11030608415603638,
+      "logps/chosen": -78.3387451171875,
+      "logps/rejected": -91.82540130615234,
+      "loss": 2.1826,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.885233998298645,
+      "rewards/margins": 0.6971657872200012,
+      "rewards/rejected": -2.582399845123291,
+      "step": 451
+    },
+    {
+      "epoch": 0.821071752951862,
+      "grad_norm": 1.705962061882019,
+      "learning_rate": 9.094719004091039e-06,
+      "logits/chosen": 0.12591060996055603,
+      "logits/rejected": 0.1107010543346405,
+      "logps/chosen": -76.82783508300781,
+      "logps/rejected": -84.23365020751953,
+      "loss": 2.2843,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.04795241355896,
+      "rewards/margins": 0.5103356242179871,
+      "rewards/rejected": -2.558288335800171,
+      "step": 452
+    },
+    {
+      "epoch": 0.8228882833787466,
+      "grad_norm": 1.6535886526107788,
+      "learning_rate": 9.090849714732217e-06,
+      "logits/chosen": 0.11322569847106934,
+      "logits/rejected": 0.13092264533042908,
+      "logps/chosen": -80.16563415527344,
+      "logps/rejected": -82.96965026855469,
+      "loss": 2.3556,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9074076414108276,
+      "rewards/margins": 0.39431121945381165,
+      "rewards/rejected": -2.3017189502716064,
+      "step": 453
+    },
+    {
+      "epoch": 0.8247048138056312,
+      "grad_norm": 1.7194557189941406,
+      "learning_rate": 9.086973103005602e-06,
+      "logits/chosen": 0.1069367378950119,
+      "logits/rejected": 0.06237747147679329,
+      "logps/chosen": -72.59986877441406,
+      "logps/rejected": -77.62615203857422,
+      "loss": 2.2768,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.023085355758667,
+      "rewards/margins": 0.6052648425102234,
+      "rewards/rejected": -2.6283504962921143,
+      "step": 454
+    },
+    {
+      "epoch": 0.8265213442325159,
+      "grad_norm": 1.7921446561813354,
+      "learning_rate": 9.08308917681628e-06,
+      "logits/chosen": 0.12843580543994904,
+      "logits/rejected": 0.08002308011054993,
+      "logps/chosen": -80.6107177734375,
+      "logps/rejected": -80.92134857177734,
+      "loss": 2.3881,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9404933452606201,
+      "rewards/margins": 0.37478840351104736,
+      "rewards/rejected": -2.315281629562378,
+      "step": 455
+    },
+    {
+      "epoch": 0.8283378746594006,
+      "grad_norm": 2.3353819847106934,
+      "learning_rate": 9.07919794408425e-06,
+      "logits/chosen": 0.05452323704957962,
+      "logits/rejected": 0.12883196771144867,
+      "logps/chosen": -83.56001281738281,
+      "logps/rejected": -95.23291015625,
+      "loss": 2.8503,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.2597994804382324,
+      "rewards/margins": 0.4092620015144348,
+      "rewards/rejected": -2.6690614223480225,
+      "step": 456
+    },
+    {
+      "epoch": 0.8301544050862852,
+      "grad_norm": 1.6829804182052612,
+      "learning_rate": 9.075299412744417e-06,
+      "logits/chosen": 0.1223950982093811,
+      "logits/rejected": 0.10596577078104019,
+      "logps/chosen": -77.75115203857422,
+      "logps/rejected": -83.63490295410156,
+      "loss": 2.1639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.026258945465088,
+      "rewards/margins": 0.6107547879219055,
+      "rewards/rejected": -2.6370139122009277,
+      "step": 457
+    },
+    {
+      "epoch": 0.8319709355131698,
+      "grad_norm": 2.053755283355713,
+      "learning_rate": 9.07139359074656e-06,
+      "logits/chosen": 0.08021271228790283,
+      "logits/rejected": 0.11617676913738251,
+      "logps/chosen": -75.00261688232422,
+      "logps/rejected": -87.25222778320312,
+      "loss": 2.4048,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3207621574401855,
+      "rewards/margins": 0.5956254005432129,
+      "rewards/rejected": -2.9163873195648193,
+      "step": 458
+    },
+    {
+      "epoch": 0.8337874659400545,
+      "grad_norm": 2.1006064414978027,
+      "learning_rate": 9.067480486055333e-06,
+      "logits/chosen": 0.1115679144859314,
+      "logits/rejected": 0.08167213946580887,
+      "logps/chosen": -81.78447723388672,
+      "logps/rejected": -82.21318817138672,
+      "loss": 2.8725,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.165053367614746,
+      "rewards/margins": 0.2777022123336792,
+      "rewards/rejected": -2.442755699157715,
+      "step": 459
+    },
+    {
+      "epoch": 0.8356039963669392,
+      "grad_norm": 1.6339340209960938,
+      "learning_rate": 9.063560106650238e-06,
+      "logits/chosen": 0.0966312363743782,
+      "logits/rejected": 0.09943810105323792,
+      "logps/chosen": -71.0906982421875,
+      "logps/rejected": -76.22811126708984,
+      "loss": 2.2922,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.0007636547088623,
+      "rewards/margins": 0.47423097491264343,
+      "rewards/rejected": -2.474994659423828,
+      "step": 460
+    },
+    {
+      "epoch": 0.8374205267938238,
+      "grad_norm": 1.8504972457885742,
+      "learning_rate": 9.059632460525613e-06,
+      "logits/chosen": 0.09517084062099457,
+      "logits/rejected": 0.14127981662750244,
+      "logps/chosen": -79.06185913085938,
+      "logps/rejected": -89.11494445800781,
+      "loss": 2.3288,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.0016026496887207,
+      "rewards/margins": 0.5282158255577087,
+      "rewards/rejected": -2.529818534851074,
+      "step": 461
+    },
+    {
+      "epoch": 0.8392370572207084,
+      "grad_norm": 1.8151304721832275,
+      "learning_rate": 9.055697555690607e-06,
+      "logits/chosen": 0.08868349343538284,
+      "logits/rejected": 0.14090785384178162,
+      "logps/chosen": -77.26272583007812,
+      "logps/rejected": -85.14620971679688,
+      "loss": 2.3694,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.087087869644165,
+      "rewards/margins": 0.5032299757003784,
+      "rewards/rejected": -2.590317964553833,
+      "step": 462
+    },
+    {
+      "epoch": 0.8410535876475931,
+      "grad_norm": 1.8596280813217163,
+      "learning_rate": 9.051755400169182e-06,
+      "logits/chosen": 0.06272133439779282,
+      "logits/rejected": 0.12342572212219238,
+      "logps/chosen": -75.46456909179688,
+      "logps/rejected": -81.9759292602539,
+      "loss": 2.5167,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.0463178157806396,
+      "rewards/margins": 0.3146142363548279,
+      "rewards/rejected": -2.3609323501586914,
+      "step": 463
+    },
+    {
+      "epoch": 0.8428701180744778,
+      "grad_norm": 1.7658417224884033,
+      "learning_rate": 9.047806002000075e-06,
+      "logits/chosen": -0.007259421981871128,
+      "logits/rejected": 0.09611339122056961,
+      "logps/chosen": -73.88439178466797,
+      "logps/rejected": -80.77063751220703,
+      "loss": 2.4938,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9998183250427246,
+      "rewards/margins": 0.4115346670150757,
+      "rewards/rejected": -2.4113528728485107,
+      "step": 464
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 1.6596916913986206,
+      "learning_rate": 9.043849369236799e-06,
+      "logits/chosen": 0.03223409131169319,
+      "logits/rejected": 0.05010119825601578,
+      "logps/chosen": -66.32562255859375,
+      "logps/rejected": -69.36509704589844,
+      "loss": 2.5881,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -1.909837245941162,
+      "rewards/margins": 0.2923263907432556,
+      "rewards/rejected": -2.2021636962890625,
+      "step": 465
+    },
+    {
+      "epoch": 0.846503178928247,
+      "grad_norm": 1.6832820177078247,
+      "learning_rate": 9.039885509947616e-06,
+      "logits/chosen": 0.06287454813718796,
+      "logits/rejected": 0.12746769189834595,
+      "logps/chosen": -76.6998519897461,
+      "logps/rejected": -87.36117553710938,
+      "loss": 2.3914,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.029526710510254,
+      "rewards/margins": 0.48344865441322327,
+      "rewards/rejected": -2.5129752159118652,
+      "step": 466
+    },
+    {
+      "epoch": 0.8483197093551317,
+      "grad_norm": 1.6972366571426392,
+      "learning_rate": 9.035914432215527e-06,
+      "logits/chosen": 0.1179896891117096,
+      "logits/rejected": 0.14000467956066132,
+      "logps/chosen": -75.3287353515625,
+      "logps/rejected": -78.60227966308594,
+      "loss": 2.4194,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.088515043258667,
+      "rewards/margins": 0.42544490098953247,
+      "rewards/rejected": -2.513960123062134,
+      "step": 467
+    },
+    {
+      "epoch": 0.8501362397820164,
+      "grad_norm": 2.5104973316192627,
+      "learning_rate": 9.031936144138247e-06,
+      "logits/chosen": 0.08423591405153275,
+      "logits/rejected": 0.1387752890586853,
+      "logps/chosen": -78.6176528930664,
+      "logps/rejected": -83.52703094482422,
+      "loss": 2.7262,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.1089184284210205,
+      "rewards/margins": 0.44402289390563965,
+      "rewards/rejected": -2.55294132232666,
+      "step": 468
+    },
+    {
+      "epoch": 0.851952770208901,
+      "grad_norm": 1.944931149482727,
+      "learning_rate": 9.027950653828202e-06,
+      "logits/chosen": 0.12252221256494522,
+      "logits/rejected": 0.14959384500980377,
+      "logps/chosen": -72.63184356689453,
+      "logps/rejected": -70.89137268066406,
+      "loss": 2.7323,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.9612714052200317,
+      "rewards/margins": 0.24262896180152893,
+      "rewards/rejected": -2.2039003372192383,
+      "step": 469
+    },
+    {
+      "epoch": 0.8537693006357856,
+      "grad_norm": 1.6699979305267334,
+      "learning_rate": 9.0239579694125e-06,
+      "logits/chosen": 0.11827167868614197,
+      "logits/rejected": 0.11933022737503052,
+      "logps/chosen": -70.45188903808594,
+      "logps/rejected": -76.2336654663086,
+      "loss": 2.457,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -1.854758381843567,
+      "rewards/margins": 0.47020024061203003,
+      "rewards/rejected": -2.3249588012695312,
+      "step": 470
+    },
+    {
+      "epoch": 0.8555858310626703,
+      "grad_norm": 1.5790536403656006,
+      "learning_rate": 9.019958099032919e-06,
+      "logits/chosen": 0.04856352508068085,
+      "logits/rejected": 0.11418008804321289,
+      "logps/chosen": -75.7929458618164,
+      "logps/rejected": -86.55818176269531,
+      "loss": 2.2809,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.003899335861206,
+      "rewards/margins": 0.5960561037063599,
+      "rewards/rejected": -2.5999553203582764,
+      "step": 471
+    },
+    {
+      "epoch": 0.857402361489555,
+      "grad_norm": 1.3594144582748413,
+      "learning_rate": 9.015951050845891e-06,
+      "logits/chosen": 0.10341258347034454,
+      "logits/rejected": 0.12394269555807114,
+      "logps/chosen": -78.8187255859375,
+      "logps/rejected": -84.04349517822266,
+      "loss": 1.9536,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9372440576553345,
+      "rewards/margins": 0.6742948889732361,
+      "rewards/rejected": -2.611538887023926,
+      "step": 472
+    },
+    {
+      "epoch": 0.8592188919164396,
+      "grad_norm": 1.7592689990997314,
+      "learning_rate": 9.011936833022485e-06,
+      "logits/chosen": 0.10970023274421692,
+      "logits/rejected": 0.09040558338165283,
+      "logps/chosen": -80.59556579589844,
+      "logps/rejected": -84.2996826171875,
+      "loss": 2.5658,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -1.935511827468872,
+      "rewards/margins": 0.35643547773361206,
+      "rewards/rejected": -2.29194712638855,
+      "step": 473
+    },
+    {
+      "epoch": 0.8610354223433242,
+      "grad_norm": 1.6418087482452393,
+      "learning_rate": 9.00791545374839e-06,
+      "logits/chosen": 0.16755647957324982,
+      "logits/rejected": 0.1745520681142807,
+      "logps/chosen": -82.31010437011719,
+      "logps/rejected": -85.66128540039062,
+      "loss": 2.1854,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -1.9519617557525635,
+      "rewards/margins": 0.5059336423873901,
+      "rewards/rejected": -2.457895278930664,
+      "step": 474
+    },
+    {
+      "epoch": 0.8628519527702089,
+      "grad_norm": 1.4616092443466187,
+      "learning_rate": 9.0038869212239e-06,
+      "logits/chosen": 0.03978392109274864,
+      "logits/rejected": 0.1235240027308464,
+      "logps/chosen": -77.78251647949219,
+      "logps/rejected": -85.66397094726562,
+      "loss": 2.1995,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.8721034526824951,
+      "rewards/margins": 0.45375847816467285,
+      "rewards/rejected": -2.325861930847168,
+      "step": 475
+    },
+    {
+      "epoch": 0.8646684831970936,
+      "grad_norm": 1.838180422782898,
+      "learning_rate": 8.99985124366389e-06,
+      "logits/chosen": 0.08277668058872223,
+      "logits/rejected": 0.035900432616472244,
+      "logps/chosen": -83.23967742919922,
+      "logps/rejected": -83.16535949707031,
+      "loss": 2.7347,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -2.2339279651641846,
+      "rewards/margins": 0.16711921989917755,
+      "rewards/rejected": -2.4010472297668457,
+      "step": 476
+    },
+    {
+      "epoch": 0.8664850136239782,
+      "grad_norm": 1.6746636629104614,
+      "learning_rate": 8.995808429297815e-06,
+      "logits/chosen": 0.14687396585941315,
+      "logits/rejected": 0.09570137411355972,
+      "logps/chosen": -78.79898071289062,
+      "logps/rejected": -77.9619369506836,
+      "loss": 2.4449,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -1.8334993124008179,
+      "rewards/margins": 0.3442971706390381,
+      "rewards/rejected": -2.1777963638305664,
+      "step": 477
+    },
+    {
+      "epoch": 0.8683015440508629,
+      "grad_norm": 1.7042535543441772,
+      "learning_rate": 8.991758486369675e-06,
+      "logits/chosen": 0.047281138598918915,
+      "logits/rejected": -0.023666318506002426,
+      "logps/chosen": -83.798583984375,
+      "logps/rejected": -91.88795471191406,
+      "loss": 2.3072,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -1.9728612899780273,
+      "rewards/margins": 0.5659723281860352,
+      "rewards/rejected": -2.5388338565826416,
+      "step": 478
+    },
+    {
+      "epoch": 0.8701180744777475,
+      "grad_norm": 1.5552817583084106,
+      "learning_rate": 8.987701423138007e-06,
+      "logits/chosen": 0.0772751122713089,
+      "logits/rejected": 0.08708472549915314,
+      "logps/chosen": -72.42391204833984,
+      "logps/rejected": -78.93070220947266,
+      "loss": 2.2815,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9803509712219238,
+      "rewards/margins": 0.5386512279510498,
+      "rewards/rejected": -2.5190021991729736,
+      "step": 479
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.6372044086456299,
+      "learning_rate": 8.983637247875872e-06,
+      "logits/chosen": -0.03352706879377365,
+      "logits/rejected": 0.01581352949142456,
+      "logps/chosen": -81.29138946533203,
+      "logps/rejected": -89.1744155883789,
+      "loss": 2.09,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -1.9441754817962646,
+      "rewards/margins": 0.6137264370918274,
+      "rewards/rejected": -2.5579018592834473,
+      "step": 480
+    },
+    {
+      "epoch": 0.8737511353315168,
+      "grad_norm": 1.6375274658203125,
+      "learning_rate": 8.979565968870831e-06,
+      "logits/chosen": 0.07507088780403137,
+      "logits/rejected": 0.1289597451686859,
+      "logps/chosen": -74.78109741210938,
+      "logps/rejected": -84.55204772949219,
+      "loss": 2.4566,
+      "rewards/accuracies": 0.578125,
+      "rewards/chosen": -1.9041385650634766,
+      "rewards/margins": 0.42814433574676514,
+      "rewards/rejected": -2.332282781600952,
+      "step": 481
+    },
+    {
+      "epoch": 0.8755676657584015,
+      "grad_norm": 1.5601258277893066,
+      "learning_rate": 8.975487594424927e-06,
+      "logits/chosen": 0.012452262453734875,
+      "logits/rejected": 0.06930352002382278,
+      "logps/chosen": -83.01705169677734,
+      "logps/rejected": -92.38286590576172,
+      "loss": 2.2736,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.013011932373047,
+      "rewards/margins": 0.5085774064064026,
+      "rewards/rejected": -2.5215890407562256,
+      "step": 482
+    },
+    {
+      "epoch": 0.8773841961852861,
+      "grad_norm": 1.8685195446014404,
+      "learning_rate": 8.971402132854677e-06,
+      "logits/chosen": 0.09415426105260849,
+      "logits/rejected": 0.12719042599201202,
+      "logps/chosen": -79.85440063476562,
+      "logps/rejected": -82.80875396728516,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.114530086517334,
+      "rewards/margins": 0.46328115463256836,
+      "rewards/rejected": -2.5778112411499023,
+      "step": 483
+    },
+    {
+      "epoch": 0.8792007266121707,
+      "grad_norm": 1.8812834024429321,
+      "learning_rate": 8.967309592491052e-06,
+      "logits/chosen": 0.13547101616859436,
+      "logits/rejected": 0.11872326582670212,
+      "logps/chosen": -74.42122650146484,
+      "logps/rejected": -81.9700927734375,
+      "loss": 2.5016,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.1828627586364746,
+      "rewards/margins": 0.44510167837142944,
+      "rewards/rejected": -2.627964496612549,
+      "step": 484
+    },
+    {
+      "epoch": 0.8810172570390554,
+      "grad_norm": 1.510517954826355,
+      "learning_rate": 8.963209981679451e-06,
+      "logits/chosen": 0.03274242952466011,
+      "logits/rejected": 0.07188954204320908,
+      "logps/chosen": -82.55563354492188,
+      "logps/rejected": -100.33609771728516,
+      "loss": 1.9177,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9487042427062988,
+      "rewards/margins": 0.8189151287078857,
+      "rewards/rejected": -2.7676191329956055,
+      "step": 485
+    },
+    {
+      "epoch": 0.8828337874659401,
+      "grad_norm": 1.5973646640777588,
+      "learning_rate": 8.959103308779696e-06,
+      "logits/chosen": 0.0365552082657814,
+      "logits/rejected": 0.051129020750522614,
+      "logps/chosen": -72.4110107421875,
+      "logps/rejected": -83.10322570800781,
+      "loss": 2.1547,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9375314712524414,
+      "rewards/margins": 0.6991211175918579,
+      "rewards/rejected": -2.6366524696350098,
+      "step": 486
+    },
+    {
+      "epoch": 0.8846503178928247,
+      "grad_norm": 1.6707032918930054,
+      "learning_rate": 8.954989582166009e-06,
+      "logits/chosen": 0.0635693296790123,
+      "logits/rejected": 0.009854275733232498,
+      "logps/chosen": -87.12749481201172,
+      "logps/rejected": -88.9491195678711,
+      "loss": 2.2043,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.9257593154907227,
+      "rewards/margins": 0.5429801344871521,
+      "rewards/rejected": -2.4687397480010986,
+      "step": 487
+    },
+    {
+      "epoch": 0.8864668483197093,
+      "grad_norm": 1.4422773122787476,
+      "learning_rate": 8.95086881022699e-06,
+      "logits/chosen": -0.024460218846797943,
+      "logits/rejected": 0.14137038588523865,
+      "logps/chosen": -74.60330200195312,
+      "logps/rejected": -95.01730346679688,
+      "loss": 2.0246,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.054253339767456,
+      "rewards/margins": 0.8260326981544495,
+      "rewards/rejected": -2.88028621673584,
+      "step": 488
+    },
+    {
+      "epoch": 0.888283378746594,
+      "grad_norm": 1.8121236562728882,
+      "learning_rate": 8.946741001365616e-06,
+      "logits/chosen": 0.05854415148496628,
+      "logits/rejected": 0.13758361339569092,
+      "logps/chosen": -74.46614074707031,
+      "logps/rejected": -83.96118927001953,
+      "loss": 2.5066,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.105980157852173,
+      "rewards/margins": 0.41478782892227173,
+      "rewards/rejected": -2.520768165588379,
+      "step": 489
+    },
+    {
+      "epoch": 0.8900999091734787,
+      "grad_norm": 2.023728370666504,
+      "learning_rate": 8.942606163999205e-06,
+      "logits/chosen": 0.04619833081960678,
+      "logits/rejected": 0.09330146014690399,
+      "logps/chosen": -75.35594940185547,
+      "logps/rejected": -89.18836975097656,
+      "loss": 2.2655,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.9396051168441772,
+      "rewards/margins": 0.6726698875427246,
+      "rewards/rejected": -2.6122751235961914,
+      "step": 490
+    },
+    {
+      "epoch": 0.8919164396003633,
+      "grad_norm": 1.559735894203186,
+      "learning_rate": 8.938464306559412e-06,
+      "logits/chosen": 0.093504399061203,
+      "logits/rejected": 0.09722252935171127,
+      "logps/chosen": -80.77140808105469,
+      "logps/rejected": -87.86373901367188,
+      "loss": 2.1676,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.1073269844055176,
+      "rewards/margins": 0.5740070939064026,
+      "rewards/rejected": -2.6813340187072754,
+      "step": 491
+    },
+    {
+      "epoch": 0.8937329700272479,
+      "grad_norm": 1.6771574020385742,
+      "learning_rate": 8.934315437492203e-06,
+      "logits/chosen": 0.06368491798639297,
+      "logits/rejected": 0.04900998994708061,
+      "logps/chosen": -78.2313003540039,
+      "logps/rejected": -92.83306121826172,
+      "loss": 2.3435,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.110978364944458,
+      "rewards/margins": 0.6751725673675537,
+      "rewards/rejected": -2.786151170730591,
+      "step": 492
+    },
+    {
+      "epoch": 0.8955495004541326,
+      "grad_norm": 1.651183843612671,
+      "learning_rate": 8.930159565257846e-06,
+      "logits/chosen": 0.08393608778715134,
+      "logits/rejected": 0.0475153923034668,
+      "logps/chosen": -77.8431396484375,
+      "logps/rejected": -88.03681945800781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0274362564086914,
+      "rewards/margins": 0.8075671195983887,
+      "rewards/rejected": -2.835003137588501,
+      "step": 493
+    },
+    {
+      "epoch": 0.8973660308810173,
+      "grad_norm": 2.262662172317505,
+      "learning_rate": 8.925996698330887e-06,
+      "logits/chosen": 0.014226208440959454,
+      "logits/rejected": 0.09588178992271423,
+      "logps/chosen": -88.56185913085938,
+      "logps/rejected": -95.1131591796875,
+      "loss": 2.5853,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4549288749694824,
+      "rewards/margins": 0.5275144577026367,
+      "rewards/rejected": -2.982443332672119,
+      "step": 494
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 1.6070181131362915,
+      "learning_rate": 8.92182684520014e-06,
+      "logits/chosen": 0.11615607142448425,
+      "logits/rejected": 0.16645964980125427,
+      "logps/chosen": -78.97079467773438,
+      "logps/rejected": -90.49947357177734,
+      "loss": 2.1072,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.175706386566162,
+      "rewards/margins": 0.7435849905014038,
+      "rewards/rejected": -2.9192914962768555,
+      "step": 495
+    },
+    {
+      "epoch": 0.9009990917347865,
+      "grad_norm": 1.8343369960784912,
+      "learning_rate": 8.917650014368658e-06,
+      "logits/chosen": 0.12583515048027039,
+      "logits/rejected": 0.07308925688266754,
+      "logps/chosen": -83.38356018066406,
+      "logps/rejected": -90.73155975341797,
+      "loss": 2.3116,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3124871253967285,
+      "rewards/margins": 0.5869801640510559,
+      "rewards/rejected": -2.8994674682617188,
+      "step": 496
+    },
+    {
+      "epoch": 0.9028156221616712,
+      "grad_norm": 1.7175630331039429,
+      "learning_rate": 8.913466214353728e-06,
+      "logits/chosen": 0.14971572160720825,
+      "logits/rejected": 0.17372727394104004,
+      "logps/chosen": -86.66060638427734,
+      "logps/rejected": -94.42353820800781,
+      "loss": 2.3541,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.943577766418457,
+      "rewards/margins": 0.5955328345298767,
+      "rewards/rejected": -2.5391108989715576,
+      "step": 497
+    },
+    {
+      "epoch": 0.9046321525885559,
+      "grad_norm": 1.754708170890808,
+      "learning_rate": 8.909275453686845e-06,
+      "logits/chosen": 0.07779194414615631,
+      "logits/rejected": 0.09394712746143341,
+      "logps/chosen": -71.64151000976562,
+      "logps/rejected": -78.53131103515625,
+      "loss": 2.0436,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.0102486610412598,
+      "rewards/margins": 0.6668623685836792,
+      "rewards/rejected": -2.6771109104156494,
+      "step": 498
+    },
+    {
+      "epoch": 0.9064486830154405,
+      "grad_norm": 1.9303216934204102,
+      "learning_rate": 8.905077740913701e-06,
+      "logits/chosen": 0.15853236615657806,
+      "logits/rejected": 0.11452697217464447,
+      "logps/chosen": -81.46524047851562,
+      "logps/rejected": -89.1716537475586,
+      "loss": 2.2477,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.264355182647705,
+      "rewards/margins": 0.6477693915367126,
+      "rewards/rejected": -2.9121243953704834,
+      "step": 499
+    },
+    {
+      "epoch": 0.9082652134423251,
+      "grad_norm": 2.1294379234313965,
+      "learning_rate": 8.900873084594164e-06,
+      "logits/chosen": 0.029480352997779846,
+      "logits/rejected": 0.09827219694852829,
+      "logps/chosen": -95.93231964111328,
+      "logps/rejected": -106.21436309814453,
+      "loss": 2.6346,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.3455631732940674,
+      "rewards/margins": 0.5977038145065308,
+      "rewards/rejected": -2.9432668685913086,
+      "step": 500
+    },
+    {
+      "epoch": 0.9100817438692098,
+      "grad_norm": 1.7347708940505981,
+      "learning_rate": 8.896661493302258e-06,
+      "logits/chosen": 0.08531993627548218,
+      "logits/rejected": 0.14993277192115784,
+      "logps/chosen": -79.09967041015625,
+      "logps/rejected": -90.50935363769531,
+      "loss": 2.2969,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -1.9908722639083862,
+      "rewards/margins": 0.5710910558700562,
+      "rewards/rejected": -2.5619633197784424,
+      "step": 501
+    },
+    {
+      "epoch": 0.9118982742960945,
+      "grad_norm": 2.065664529800415,
+      "learning_rate": 8.892442975626152e-06,
+      "logits/chosen": 0.18976512551307678,
+      "logits/rejected": 0.13936059176921844,
+      "logps/chosen": -82.80276489257812,
+      "logps/rejected": -84.57154083251953,
+      "loss": 2.5968,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4863569736480713,
+      "rewards/margins": 0.3651903569698334,
+      "rewards/rejected": -2.8515477180480957,
+      "step": 502
+    },
+    {
+      "epoch": 0.9137148047229791,
+      "grad_norm": 2.054955005645752,
+      "learning_rate": 8.888217540168139e-06,
+      "logits/chosen": 0.01604822278022766,
+      "logits/rejected": 0.1370609700679779,
+      "logps/chosen": -74.58015441894531,
+      "logps/rejected": -96.80826568603516,
+      "loss": 2.169,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.397087574005127,
+      "rewards/margins": 0.9368999004364014,
+      "rewards/rejected": -3.333987236022949,
+      "step": 503
+    },
+    {
+      "epoch": 0.9155313351498637,
+      "grad_norm": 2.179532527923584,
+      "learning_rate": 8.883985195544617e-06,
+      "logits/chosen": 0.03752445429563522,
+      "logits/rejected": 0.024428365752100945,
+      "logps/chosen": -83.81852722167969,
+      "logps/rejected": -88.1905288696289,
+      "loss": 2.5361,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.239609956741333,
+      "rewards/margins": 0.4188295900821686,
+      "rewards/rejected": -2.6584396362304688,
+      "step": 504
+    },
+    {
+      "epoch": 0.9173478655767484,
+      "grad_norm": 1.9171282052993774,
+      "learning_rate": 8.879745950386075e-06,
+      "logits/chosen": 0.17986616492271423,
+      "logits/rejected": 0.20335282385349274,
+      "logps/chosen": -72.679931640625,
+      "logps/rejected": -71.41309356689453,
+      "loss": 2.5452,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.261596918106079,
+      "rewards/margins": 0.3737923502922058,
+      "rewards/rejected": -2.6353893280029297,
+      "step": 505
+    },
+    {
+      "epoch": 0.9191643960036331,
+      "grad_norm": 2.3437929153442383,
+      "learning_rate": 8.87549981333707e-06,
+      "logits/chosen": 0.09010382741689682,
+      "logits/rejected": 0.040441811084747314,
+      "logps/chosen": -81.56139373779297,
+      "logps/rejected": -83.21439361572266,
+      "loss": 2.74,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2574033737182617,
+      "rewards/margins": 0.28719204664230347,
+      "rewards/rejected": -2.544595241546631,
+      "step": 506
+    },
+    {
+      "epoch": 0.9209809264305178,
+      "grad_norm": 2.4240329265594482,
+      "learning_rate": 8.871246793056215e-06,
+      "logits/chosen": 0.02888420596718788,
+      "logits/rejected": 0.11542786657810211,
+      "logps/chosen": -84.60144805908203,
+      "logps/rejected": -100.05682373046875,
+      "loss": 2.7052,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.863104820251465,
+      "rewards/margins": 0.41941094398498535,
+      "rewards/rejected": -3.28251576423645,
+      "step": 507
+    },
+    {
+      "epoch": 0.9227974568574023,
+      "grad_norm": 1.9247711896896362,
+      "learning_rate": 8.866986898216157e-06,
+      "logits/chosen": 0.06140238791704178,
+      "logits/rejected": 0.05623817816376686,
+      "logps/chosen": -77.40584564208984,
+      "logps/rejected": -86.71392822265625,
+      "loss": 2.2445,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.4979231357574463,
+      "rewards/margins": 0.6161274313926697,
+      "rewards/rejected": -3.11405086517334,
+      "step": 508
+    },
+    {
+      "epoch": 0.924613987284287,
+      "grad_norm": 1.854956030845642,
+      "learning_rate": 8.862720137503568e-06,
+      "logits/chosen": 0.06755004823207855,
+      "logits/rejected": 0.03762562945485115,
+      "logps/chosen": -86.03059387207031,
+      "logps/rejected": -90.12970733642578,
+      "loss": 2.1459,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.3017396926879883,
+      "rewards/margins": 0.6131142973899841,
+      "rewards/rejected": -2.914853572845459,
+      "step": 509
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.8044302463531494,
+      "learning_rate": 8.858446519619113e-06,
+      "logits/chosen": 0.08446178585290909,
+      "logits/rejected": 0.08818987011909485,
+      "logps/chosen": -77.58124542236328,
+      "logps/rejected": -85.02752685546875,
+      "loss": 2.1311,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2378087043762207,
+      "rewards/margins": 0.7067832350730896,
+      "rewards/rejected": -2.944591999053955,
+      "step": 510
+    },
+    {
+      "epoch": 0.9282470481380564,
+      "grad_norm": 1.4898866415023804,
+      "learning_rate": 8.854166053277443e-06,
+      "logits/chosen": 0.1039918065071106,
+      "logits/rejected": 0.1396111100912094,
+      "logps/chosen": -87.59841918945312,
+      "logps/rejected": -101.05659484863281,
+      "loss": 2.1092,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1167173385620117,
+      "rewards/margins": 0.7099149227142334,
+      "rewards/rejected": -2.826632499694824,
+      "step": 511
+    },
+    {
+      "epoch": 0.9300635785649409,
+      "grad_norm": 1.8009706735610962,
+      "learning_rate": 8.849878747207175e-06,
+      "logits/chosen": 0.11744043976068497,
+      "logits/rejected": 0.07621707767248154,
+      "logps/chosen": -78.37222290039062,
+      "logps/rejected": -74.32958221435547,
+      "loss": 2.6399,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.4002914428710938,
+      "rewards/margins": 0.21068021655082703,
+      "rewards/rejected": -2.610971450805664,
+      "step": 512
+    },
+    {
+      "epoch": 0.9318801089918256,
+      "grad_norm": 1.6368399858474731,
+      "learning_rate": 8.845584610150871e-06,
+      "logits/chosen": 0.09938757866621017,
+      "logits/rejected": 0.12675486505031586,
+      "logps/chosen": -82.24690246582031,
+      "logps/rejected": -92.60913848876953,
+      "loss": 1.9078,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.0825133323669434,
+      "rewards/margins": 0.8327144384384155,
+      "rewards/rejected": -2.9152278900146484,
+      "step": 513
+    },
+    {
+      "epoch": 0.9336966394187103,
+      "grad_norm": 1.8117483854293823,
+      "learning_rate": 8.841283650865027e-06,
+      "logits/chosen": 0.1355382651090622,
+      "logits/rejected": 0.14869986474514008,
+      "logps/chosen": -85.03866577148438,
+      "logps/rejected": -90.55638885498047,
+      "loss": 2.1243,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3315932750701904,
+      "rewards/margins": 0.6421911120414734,
+      "rewards/rejected": -2.9737846851348877,
+      "step": 514
+    },
+    {
+      "epoch": 0.935513169845595,
+      "grad_norm": 1.785927414894104,
+      "learning_rate": 8.836975878120046e-06,
+      "logits/chosen": 0.11557039618492126,
+      "logits/rejected": 0.11716655641794205,
+      "logps/chosen": -83.99951171875,
+      "logps/rejected": -92.36119079589844,
+      "loss": 2.4256,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.439833402633667,
+      "rewards/margins": 0.38893458247184753,
+      "rewards/rejected": -2.8287675380706787,
+      "step": 515
+    },
+    {
+      "epoch": 0.9373297002724795,
+      "grad_norm": 3.526102066040039,
+      "learning_rate": 8.832661300700228e-06,
+      "logits/chosen": 0.10778439044952393,
+      "logits/rejected": 0.18004637956619263,
+      "logps/chosen": -84.20158386230469,
+      "logps/rejected": -91.33497619628906,
+      "loss": 2.3568,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2032201290130615,
+      "rewards/margins": 0.5167797207832336,
+      "rewards/rejected": -2.7200000286102295,
+      "step": 516
+    },
+    {
+      "epoch": 0.9391462306993642,
+      "grad_norm": 1.7708752155303955,
+      "learning_rate": 8.828339927403745e-06,
+      "logits/chosen": 0.0862819030880928,
+      "logits/rejected": 0.16991934180259705,
+      "logps/chosen": -87.4437026977539,
+      "logps/rejected": -96.60844421386719,
+      "loss": 2.3041,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.230825185775757,
+      "rewards/margins": 0.5197271108627319,
+      "rewards/rejected": -2.7505524158477783,
+      "step": 517
+    },
+    {
+      "epoch": 0.9409627611262489,
+      "grad_norm": 2.1336302757263184,
+      "learning_rate": 8.824011767042631e-06,
+      "logits/chosen": 0.12070942670106888,
+      "logits/rejected": 0.19643370807170868,
+      "logps/chosen": -75.97718811035156,
+      "logps/rejected": -83.26056671142578,
+      "loss": 2.3468,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3768177032470703,
+      "rewards/margins": 0.5026894211769104,
+      "rewards/rejected": -2.879507064819336,
+      "step": 518
+    },
+    {
+      "epoch": 0.9427792915531336,
+      "grad_norm": 2.6361515522003174,
+      "learning_rate": 8.819676828442758e-06,
+      "logits/chosen": 0.04481849446892738,
+      "logits/rejected": -0.011357773095369339,
+      "logps/chosen": -84.09026336669922,
+      "logps/rejected": -87.71636199951172,
+      "loss": 2.8427,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.4249284267425537,
+      "rewards/margins": 0.41216185688972473,
+      "rewards/rejected": -2.837090015411377,
+      "step": 519
+    },
+    {
+      "epoch": 0.9445958219800181,
+      "grad_norm": 2.1068665981292725,
+      "learning_rate": 8.815335120443822e-06,
+      "logits/chosen": 0.1595858335494995,
+      "logits/rejected": 0.14017102122306824,
+      "logps/chosen": -94.00933837890625,
+      "logps/rejected": -96.91961669921875,
+      "loss": 2.6559,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.4731483459472656,
+      "rewards/margins": 0.3141450881958008,
+      "rewards/rejected": -2.7872931957244873,
+      "step": 520
+    },
+    {
+      "epoch": 0.9464123524069028,
+      "grad_norm": 1.7320570945739746,
+      "learning_rate": 8.810986651899322e-06,
+      "logits/chosen": 0.10371927917003632,
+      "logits/rejected": 0.17988254129886627,
+      "logps/chosen": -78.32308197021484,
+      "logps/rejected": -88.74859619140625,
+      "loss": 2.1866,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3973731994628906,
+      "rewards/margins": 0.7199669480323792,
+      "rewards/rejected": -3.117340087890625,
+      "step": 521
+    },
+    {
+      "epoch": 0.9482288828337875,
+      "grad_norm": 1.6995161771774292,
+      "learning_rate": 8.80663143167654e-06,
+      "logits/chosen": 0.14279219508171082,
+      "logits/rejected": 0.1058904379606247,
+      "logps/chosen": -87.79972839355469,
+      "logps/rejected": -94.35545349121094,
+      "loss": 2.1217,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.979785680770874,
+      "rewards/margins": 0.6210839152336121,
+      "rewards/rejected": -2.600869655609131,
+      "step": 522
+    },
+    {
+      "epoch": 0.9500454132606722,
+      "grad_norm": 2.1860764026641846,
+      "learning_rate": 8.80226946865653e-06,
+      "logits/chosen": 0.11601082980632782,
+      "logits/rejected": 0.15693408250808716,
+      "logps/chosen": -88.86433410644531,
+      "logps/rejected": -92.43590545654297,
+      "loss": 2.5629,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.2875983715057373,
+      "rewards/margins": 0.39402255415916443,
+      "rewards/rejected": -2.6816210746765137,
+      "step": 523
+    },
+    {
+      "epoch": 0.9518619436875567,
+      "grad_norm": 1.829032063484192,
+      "learning_rate": 8.797900771734094e-06,
+      "logits/chosen": 0.12328370660543442,
+      "logits/rejected": 0.042389824986457825,
+      "logps/chosen": -77.34864807128906,
+      "logps/rejected": -78.61405944824219,
+      "loss": 2.5327,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.2514185905456543,
+      "rewards/margins": 0.3621898591518402,
+      "rewards/rejected": -2.6136088371276855,
+      "step": 524
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 2.235381841659546,
+      "learning_rate": 8.793525349817765e-06,
+      "logits/chosen": 0.12495981156826019,
+      "logits/rejected": 0.13291439414024353,
+      "logps/chosen": -76.7927474975586,
+      "logps/rejected": -86.1208267211914,
+      "loss": 2.5188,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3426284790039062,
+      "rewards/margins": 0.46077418327331543,
+      "rewards/rejected": -2.8034026622772217,
+      "step": 525
+    },
+    {
+      "epoch": 0.9554950045413261,
+      "grad_norm": 1.8071073293685913,
+      "learning_rate": 8.78914321182979e-06,
+      "logits/chosen": 0.13614074885845184,
+      "logits/rejected": 0.1179000735282898,
+      "logps/chosen": -74.44164276123047,
+      "logps/rejected": -77.69221496582031,
+      "loss": 2.6082,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.0528411865234375,
+      "rewards/margins": 0.29493850469589233,
+      "rewards/rejected": -2.3477797508239746,
+      "step": 526
+    },
+    {
+      "epoch": 0.9573115349682108,
+      "grad_norm": 1.8623837232589722,
+      "learning_rate": 8.784754366706115e-06,
+      "logits/chosen": 0.02768833190202713,
+      "logits/rejected": 0.09584817290306091,
+      "logps/chosen": -77.13914489746094,
+      "logps/rejected": -86.53968048095703,
+      "loss": 2.2244,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.199755907058716,
+      "rewards/margins": 0.6271941661834717,
+      "rewards/rejected": -2.8269503116607666,
+      "step": 527
+    },
+    {
+      "epoch": 0.9591280653950953,
+      "grad_norm": 1.5299558639526367,
+      "learning_rate": 8.780358823396352e-06,
+      "logits/chosen": 0.11699292808771133,
+      "logits/rejected": 0.10016686469316483,
+      "logps/chosen": -86.1636962890625,
+      "logps/rejected": -87.16045379638672,
+      "loss": 2.0645,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.163857936859131,
+      "rewards/margins": 0.697452187538147,
+      "rewards/rejected": -2.8613100051879883,
+      "step": 528
+    },
+    {
+      "epoch": 0.96094459582198,
+      "grad_norm": 1.6236025094985962,
+      "learning_rate": 8.775956590863785e-06,
+      "logits/chosen": 0.11622033268213272,
+      "logits/rejected": 0.06629584729671478,
+      "logps/chosen": -76.42916870117188,
+      "logps/rejected": -80.24000549316406,
+      "loss": 1.8904,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.102374315261841,
+      "rewards/margins": 0.7230857014656067,
+      "rewards/rejected": -2.825460195541382,
+      "step": 529
+    },
+    {
+      "epoch": 0.9627611262488647,
+      "grad_norm": 1.6119427680969238,
+      "learning_rate": 8.771547678085332e-06,
+      "logits/chosen": 0.07009143382310867,
+      "logits/rejected": 0.16355563700199127,
+      "logps/chosen": -72.83528137207031,
+      "logps/rejected": -87.72441101074219,
+      "loss": 2.0136,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1697566509246826,
+      "rewards/margins": 0.8312336206436157,
+      "rewards/rejected": -3.000990390777588,
+      "step": 530
+    },
+    {
+      "epoch": 0.9645776566757494,
+      "grad_norm": 2.0324013233184814,
+      "learning_rate": 8.767132094051534e-06,
+      "logits/chosen": 0.14837129414081573,
+      "logits/rejected": 0.0700330063700676,
+      "logps/chosen": -84.65350341796875,
+      "logps/rejected": -91.84485626220703,
+      "loss": 2.2551,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5343966484069824,
+      "rewards/margins": 0.6457280516624451,
+      "rewards/rejected": -3.180124521255493,
+      "step": 531
+    },
+    {
+      "epoch": 0.9663941871026339,
+      "grad_norm": 1.4526044130325317,
+      "learning_rate": 8.762709847766532e-06,
+      "logits/chosen": 0.013620391488075256,
+      "logits/rejected": 0.10376289486885071,
+      "logps/chosen": -70.80522155761719,
+      "logps/rejected": -84.26114654541016,
+      "loss": 1.8607,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.249685287475586,
+      "rewards/margins": 0.8104004859924316,
+      "rewards/rejected": -3.0600855350494385,
+      "step": 532
+    },
+    {
+      "epoch": 0.9682107175295186,
+      "grad_norm": 2.0312576293945312,
+      "learning_rate": 8.758280948248059e-06,
+      "logits/chosen": 0.0965358167886734,
+      "logits/rejected": 0.10093791037797928,
+      "logps/chosen": -81.83265686035156,
+      "logps/rejected": -88.52181243896484,
+      "loss": 2.6128,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4901068210601807,
+      "rewards/margins": 0.30889907479286194,
+      "rewards/rejected": -2.7990057468414307,
+      "step": 533
+    },
+    {
+      "epoch": 0.9700272479564033,
+      "grad_norm": 2.547189712524414,
+      "learning_rate": 8.753845404527413e-06,
+      "logits/chosen": 0.020555848255753517,
+      "logits/rejected": 0.0945032387971878,
+      "logps/chosen": -72.16207122802734,
+      "logps/rejected": -86.27986907958984,
+      "loss": 2.0608,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.148913860321045,
+      "rewards/margins": 0.7298768162727356,
+      "rewards/rejected": -2.8787906169891357,
+      "step": 534
+    },
+    {
+      "epoch": 0.971843778383288,
+      "grad_norm": 1.996645450592041,
+      "learning_rate": 8.74940322564944e-06,
+      "logits/chosen": 0.0345986932516098,
+      "logits/rejected": 0.08187264204025269,
+      "logps/chosen": -78.76763916015625,
+      "logps/rejected": -91.45555114746094,
+      "loss": 2.4587,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.4772286415100098,
+      "rewards/margins": 0.4609605073928833,
+      "rewards/rejected": -2.9381890296936035,
+      "step": 535
+    },
+    {
+      "epoch": 0.9736603088101726,
+      "grad_norm": 1.8145204782485962,
+      "learning_rate": 8.744954420672514e-06,
+      "logits/chosen": 0.03380711376667023,
+      "logits/rejected": 0.09039584547281265,
+      "logps/chosen": -79.83094787597656,
+      "logps/rejected": -90.87137603759766,
+      "loss": 2.3589,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.1433308124542236,
+      "rewards/margins": 0.5286959409713745,
+      "rewards/rejected": -2.6720268726348877,
+      "step": 536
+    },
+    {
+      "epoch": 0.9754768392370572,
+      "grad_norm": 1.9745460748672485,
+      "learning_rate": 8.740498998668523e-06,
+      "logits/chosen": 0.07233883440494537,
+      "logits/rejected": 0.17097605764865875,
+      "logps/chosen": -72.3458251953125,
+      "logps/rejected": -83.47319793701172,
+      "loss": 2.3469,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199948310852051,
+      "rewards/margins": 0.47777751088142395,
+      "rewards/rejected": -2.6777257919311523,
+      "step": 537
+    },
+    {
+      "epoch": 0.9772933696639419,
+      "grad_norm": 1.665255069732666,
+      "learning_rate": 8.736036968722851e-06,
+      "logits/chosen": 0.09596951305866241,
+      "logits/rejected": 0.06625551730394363,
+      "logps/chosen": -80.11689758300781,
+      "logps/rejected": -88.4573974609375,
+      "loss": 2.1889,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.1831865310668945,
+      "rewards/margins": 0.6106584668159485,
+      "rewards/rejected": -2.7938451766967773,
+      "step": 538
+    },
+    {
+      "epoch": 0.9791099000908265,
+      "grad_norm": 1.5587108135223389,
+      "learning_rate": 8.73156833993435e-06,
+      "logits/chosen": 0.07950340211391449,
+      "logits/rejected": 0.0903201699256897,
+      "logps/chosen": -78.9182357788086,
+      "logps/rejected": -92.290283203125,
+      "loss": 2.048,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.12032413482666,
+      "rewards/margins": 0.7009568214416504,
+      "rewards/rejected": -2.8212809562683105,
+      "step": 539
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 1.7177670001983643,
+      "learning_rate": 8.727093121415338e-06,
+      "logits/chosen": 0.06052964925765991,
+      "logits/rejected": 0.14043518900871277,
+      "logps/chosen": -79.04943084716797,
+      "logps/rejected": -88.881591796875,
+      "loss": 2.2089,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.2680561542510986,
+      "rewards/margins": 0.5380589962005615,
+      "rewards/rejected": -2.806114912033081,
+      "step": 540
+    },
+    {
+      "epoch": 0.9827429609445958,
+      "grad_norm": 1.7630436420440674,
+      "learning_rate": 8.722611322291558e-06,
+      "logits/chosen": 0.02032918483018875,
+      "logits/rejected": 0.11728623509407043,
+      "logps/chosen": -78.1409912109375,
+      "logps/rejected": -89.3685073852539,
+      "loss": 2.3303,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.356957197189331,
+      "rewards/margins": 0.5139226317405701,
+      "rewards/rejected": -2.870880126953125,
+      "step": 541
+    },
+    {
+      "epoch": 0.9845594913714805,
+      "grad_norm": 1.852418065071106,
+      "learning_rate": 8.718122951702183e-06,
+      "logits/chosen": 0.0439969077706337,
+      "logits/rejected": 0.16671347618103027,
+      "logps/chosen": -69.96346282958984,
+      "logps/rejected": -85.24410247802734,
+      "loss": 2.1393,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3295838832855225,
+      "rewards/margins": 0.6544574499130249,
+      "rewards/rejected": -2.984041213989258,
+      "step": 542
+    },
+    {
+      "epoch": 0.9863760217983651,
+      "grad_norm": 1.9028044939041138,
+      "learning_rate": 8.713628018799782e-06,
+      "logits/chosen": 0.04875154793262482,
+      "logits/rejected": -0.009283583611249924,
+      "logps/chosen": -75.80145263671875,
+      "logps/rejected": -78.19249725341797,
+      "loss": 2.4923,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.241586208343506,
+      "rewards/margins": 0.4672737419605255,
+      "rewards/rejected": -2.708860158920288,
+      "step": 543
+    },
+    {
+      "epoch": 0.9881925522252498,
+      "grad_norm": 1.5648458003997803,
+      "learning_rate": 8.709126532750304e-06,
+      "logits/chosen": 0.02982347458600998,
+      "logits/rejected": 0.07835812866687775,
+      "logps/chosen": -75.35140991210938,
+      "logps/rejected": -91.11531829833984,
+      "loss": 2.0945,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.062386989593506,
+      "rewards/margins": 0.6832989454269409,
+      "rewards/rejected": -2.7456860542297363,
+      "step": 544
+    },
+    {
+      "epoch": 0.9900090826521344,
+      "grad_norm": 1.6214131116867065,
+      "learning_rate": 8.704618502733064e-06,
+      "logits/chosen": 0.09551770985126495,
+      "logits/rejected": 0.11310344189405441,
+      "logps/chosen": -80.15961456298828,
+      "logps/rejected": -93.38126373291016,
+      "loss": 2.1642,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.255985736846924,
+      "rewards/margins": 0.7712126970291138,
+      "rewards/rejected": -3.027198553085327,
+      "step": 545
+    },
+    {
+      "epoch": 0.9918256130790191,
+      "grad_norm": 1.762953519821167,
+      "learning_rate": 8.70010393794072e-06,
+      "logits/chosen": 0.0971008688211441,
+      "logits/rejected": 0.16590046882629395,
+      "logps/chosen": -86.24911499023438,
+      "logps/rejected": -102.59831237792969,
+      "loss": 2.0951,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.3542394638061523,
+      "rewards/margins": 0.7857018709182739,
+      "rewards/rejected": -3.139941453933716,
+      "step": 546
+    },
+    {
+      "epoch": 0.9936421435059037,
+      "grad_norm": 1.6033780574798584,
+      "learning_rate": 8.695582847579254e-06,
+      "logits/chosen": 0.15147003531455994,
+      "logits/rejected": 0.13297875225543976,
+      "logps/chosen": -78.5172348022461,
+      "logps/rejected": -86.49285888671875,
+      "loss": 2.1101,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2025821208953857,
+      "rewards/margins": 0.7471145391464233,
+      "rewards/rejected": -2.9496963024139404,
+      "step": 547
+    },
+    {
+      "epoch": 0.9954586739327884,
+      "grad_norm": 1.4905331134796143,
+      "learning_rate": 8.69105524086796e-06,
+      "logits/chosen": 0.07559307664632797,
+      "logits/rejected": 0.09286531805992126,
+      "logps/chosen": -74.92215728759766,
+      "logps/rejected": -78.23614501953125,
+      "loss": 2.2745,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.24694561958313,
+      "rewards/margins": 0.5010396838188171,
+      "rewards/rejected": -2.747985363006592,
+      "step": 548
+    },
+    {
+      "epoch": 0.997275204359673,
+      "grad_norm": 1.823449730873108,
+      "learning_rate": 8.686521127039414e-06,
+      "logits/chosen": 0.03763032704591751,
+      "logits/rejected": 0.06642220914363861,
+      "logps/chosen": -79.37518310546875,
+      "logps/rejected": -87.17401885986328,
+      "loss": 2.5954,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.158803701400757,
+      "rewards/margins": 0.4884320795536041,
+      "rewards/rejected": -2.647235870361328,
+      "step": 549
+    },
+    {
+      "epoch": 0.9990917347865577,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.681980515339464e-06,
+      "logits/chosen": 0.08186414837837219,
+      "logits/rejected": 0.08158639073371887,
+      "logps/chosen": -84.89871215820312,
+      "logps/rejected": -85.87324523925781,
+      "loss": 2.7544,
+      "rewards/accuracies": 0.546875,
+      "rewards/chosen": -2.1884236335754395,
+      "rewards/margins": 0.2923518121242523,
+      "rewards/rejected": -2.4807753562927246,
+      "step": 550
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.311372995376587,
+      "learning_rate": 8.677433415027209e-06,
+      "logits/chosen": 0.18737269937992096,
+      "logits/rejected": 0.11028257012367249,
+      "logps/chosen": -76.3554916381836,
+      "logps/rejected": -86.39812469482422,
+      "loss": 1.1517,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1043050289154053,
+      "rewards/margins": 0.7949569225311279,
+      "rewards/rejected": -2.8992621898651123,
+      "step": 551
+    },
+    {
+      "epoch": 1.0018165304268847,
+      "grad_norm": 1.6554900407791138,
+      "learning_rate": 8.672879835374976e-06,
+      "logits/chosen": 0.11259499937295914,
+      "logits/rejected": 0.07095417380332947,
+      "logps/chosen": -84.55628967285156,
+      "logps/rejected": -88.55287170410156,
+      "loss": 2.1275,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.2375316619873047,
+      "rewards/margins": 0.5453992486000061,
+      "rewards/rejected": -2.782930850982666,
+      "step": 552
+    },
+    {
+      "epoch": 1.0036330608537694,
+      "grad_norm": 1.6807721853256226,
+      "learning_rate": 8.668319785668307e-06,
+      "logits/chosen": 0.14847862720489502,
+      "logits/rejected": 0.1863957643508911,
+      "logps/chosen": -87.27869415283203,
+      "logps/rejected": -89.04036712646484,
+      "loss": 2.3209,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.087535858154297,
+      "rewards/margins": 0.5243443846702576,
+      "rewards/rejected": -2.61188006401062,
+      "step": 553
+    },
+    {
+      "epoch": 1.005449591280654,
+      "grad_norm": 1.8480761051177979,
+      "learning_rate": 8.66375327520594e-06,
+      "logits/chosen": 0.07236558198928833,
+      "logits/rejected": 0.13397395610809326,
+      "logps/chosen": -85.43444061279297,
+      "logps/rejected": -101.47248840332031,
+      "loss": 1.7106,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.157266616821289,
+      "rewards/margins": 0.9501473903656006,
+      "rewards/rejected": -3.1074140071868896,
+      "step": 554
+    },
+    {
+      "epoch": 1.0072661217075387,
+      "grad_norm": 1.586917757987976,
+      "learning_rate": 8.659180313299783e-06,
+      "logits/chosen": 0.05113527923822403,
+      "logits/rejected": 0.14182163774967194,
+      "logps/chosen": -78.31108856201172,
+      "logps/rejected": -90.46333312988281,
+      "loss": 1.8965,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -1.945315957069397,
+      "rewards/margins": 0.8104385137557983,
+      "rewards/rejected": -2.755754232406616,
+      "step": 555
+    },
+    {
+      "epoch": 1.0090826521344232,
+      "grad_norm": 1.406339406967163,
+      "learning_rate": 8.6546009092749e-06,
+      "logits/chosen": -0.01422208547592163,
+      "logits/rejected": 0.10859756916761398,
+      "logps/chosen": -75.02066040039062,
+      "logps/rejected": -87.69715118408203,
+      "loss": 2.027,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -1.9996552467346191,
+      "rewards/margins": 0.7365851402282715,
+      "rewards/rejected": -2.7362406253814697,
+      "step": 556
+    },
+    {
+      "epoch": 1.0108991825613078,
+      "grad_norm": 1.5511010885238647,
+      "learning_rate": 8.650015072469496e-06,
+      "logits/chosen": 0.07154582440853119,
+      "logits/rejected": 0.11479673534631729,
+      "logps/chosen": -88.33872985839844,
+      "logps/rejected": -96.75527954101562,
+      "loss": 1.966,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2190096378326416,
+      "rewards/margins": 0.861441969871521,
+      "rewards/rejected": -3.080451250076294,
+      "step": 557
+    },
+    {
+      "epoch": 1.0127157129881925,
+      "grad_norm": 1.5951570272445679,
+      "learning_rate": 8.645422812234888e-06,
+      "logits/chosen": 0.17342886328697205,
+      "logits/rejected": 0.11909954994916916,
+      "logps/chosen": -79.63985443115234,
+      "logps/rejected": -83.41281127929688,
+      "loss": 2.1115,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3469128608703613,
+      "rewards/margins": 0.7317838668823242,
+      "rewards/rejected": -3.0786967277526855,
+      "step": 558
+    },
+    {
+      "epoch": 1.0145322434150772,
+      "grad_norm": 1.891517162322998,
+      "learning_rate": 8.640824137935498e-06,
+      "logits/chosen": 0.14266839623451233,
+      "logits/rejected": 0.14209085702896118,
+      "logps/chosen": -84.77491760253906,
+      "logps/rejected": -87.46759796142578,
+      "loss": 2.2691,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.022730827331543,
+      "rewards/margins": 0.42655885219573975,
+      "rewards/rejected": -2.449289560317993,
+      "step": 559
+    },
+    {
+      "epoch": 1.0163487738419619,
+      "grad_norm": 1.7919303178787231,
+      "learning_rate": 8.636219058948823e-06,
+      "logits/chosen": 0.01973957009613514,
+      "logits/rejected": 0.12187746167182922,
+      "logps/chosen": -78.26383209228516,
+      "logps/rejected": -92.714599609375,
+      "loss": 1.974,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2774088382720947,
+      "rewards/margins": 0.8670985698699951,
+      "rewards/rejected": -3.1445071697235107,
+      "step": 560
+    },
+    {
+      "epoch": 1.0181653042688465,
+      "grad_norm": 1.5417025089263916,
+      "learning_rate": 8.631607584665415e-06,
+      "logits/chosen": 0.07740732282400131,
+      "logits/rejected": 0.17052005231380463,
+      "logps/chosen": -83.54193115234375,
+      "logps/rejected": -95.1084976196289,
+      "loss": 1.9963,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.1598939895629883,
+      "rewards/margins": 0.802270233631134,
+      "rewards/rejected": -2.9621646404266357,
+      "step": 561
+    },
+    {
+      "epoch": 1.0199818346957312,
+      "grad_norm": 1.9338072538375854,
+      "learning_rate": 8.62698972448888e-06,
+      "logits/chosen": 0.09191437065601349,
+      "logits/rejected": 0.1776769608259201,
+      "logps/chosen": -82.68082427978516,
+      "logps/rejected": -96.89566802978516,
+      "loss": 2.0399,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3505523204803467,
+      "rewards/margins": 0.7893310785293579,
+      "rewards/rejected": -3.1398837566375732,
+      "step": 562
+    },
+    {
+      "epoch": 1.021798365122616,
+      "grad_norm": 1.609601378440857,
+      "learning_rate": 8.622365487835836e-06,
+      "logits/chosen": 0.13105043768882751,
+      "logits/rejected": 0.138211190700531,
+      "logps/chosen": -77.19727325439453,
+      "logps/rejected": -84.1854476928711,
+      "loss": 2.3184,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.13655948638916,
+      "rewards/margins": 0.5744473934173584,
+      "rewards/rejected": -2.7110068798065186,
+      "step": 563
+    },
+    {
+      "epoch": 1.0236148955495004,
+      "grad_norm": 2.1222517490386963,
+      "learning_rate": 8.617734884135906e-06,
+      "logits/chosen": 0.14744582772254944,
+      "logits/rejected": 0.053739145398139954,
+      "logps/chosen": -81.11271667480469,
+      "logps/rejected": -81.40655517578125,
+      "loss": 2.2,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.419705390930176,
+      "rewards/margins": 0.7198007702827454,
+      "rewards/rejected": -3.1395058631896973,
+      "step": 564
+    },
+    {
+      "epoch": 1.025431425976385,
+      "grad_norm": 1.8075231313705444,
+      "learning_rate": 8.613097922831699e-06,
+      "logits/chosen": 0.07210052758455276,
+      "logits/rejected": 0.07022904604673386,
+      "logps/chosen": -88.72612762451172,
+      "logps/rejected": -98.2815933227539,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1057255268096924,
+      "rewards/margins": 0.8747395873069763,
+      "rewards/rejected": -2.9804654121398926,
+      "step": 565
+    },
+    {
+      "epoch": 1.0272479564032697,
+      "grad_norm": 1.3381479978561401,
+      "learning_rate": 8.608454613378783e-06,
+      "logits/chosen": 0.08187520503997803,
+      "logits/rejected": 0.1098201796412468,
+      "logps/chosen": -77.1299819946289,
+      "logps/rejected": -89.75099182128906,
+      "loss": 1.8687,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3585519790649414,
+      "rewards/margins": 0.8574986457824707,
+      "rewards/rejected": -3.216050386428833,
+      "step": 566
+    },
+    {
+      "epoch": 1.0290644868301544,
+      "grad_norm": 1.6325740814208984,
+      "learning_rate": 8.603804965245678e-06,
+      "logits/chosen": 0.12077341973781586,
+      "logits/rejected": 0.09103713929653168,
+      "logps/chosen": -74.03280639648438,
+      "logps/rejected": -79.61197662353516,
+      "loss": 2.0884,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.430152654647827,
+      "rewards/margins": 0.6819649934768677,
+      "rewards/rejected": -3.1121177673339844,
+      "step": 567
+    },
+    {
+      "epoch": 1.030881017257039,
+      "grad_norm": 1.782224178314209,
+      "learning_rate": 8.599148987913829e-06,
+      "logits/chosen": -0.008054995909333229,
+      "logits/rejected": 0.08848709613084793,
+      "logps/chosen": -88.39682006835938,
+      "logps/rejected": -98.20311737060547,
+      "loss": 2.3203,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.21012282371521,
+      "rewards/margins": 0.5451631546020508,
+      "rewards/rejected": -2.75528621673584,
+      "step": 568
+    },
+    {
+      "epoch": 1.0326975476839237,
+      "grad_norm": 1.7691792249679565,
+      "learning_rate": 8.594486690877577e-06,
+      "logits/chosen": 0.05038870871067047,
+      "logits/rejected": -0.0020756814628839493,
+      "logps/chosen": -88.35281372070312,
+      "logps/rejected": -87.42422485351562,
+      "loss": 2.1789,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6035892963409424,
+      "rewards/margins": 0.559741735458374,
+      "rewards/rejected": -3.1633315086364746,
+      "step": 569
+    },
+    {
+      "epoch": 1.0345140781108084,
+      "grad_norm": 1.7002201080322266,
+      "learning_rate": 8.589818083644161e-06,
+      "logits/chosen": 0.030099213123321533,
+      "logits/rejected": 0.10112117975950241,
+      "logps/chosen": -74.28582763671875,
+      "logps/rejected": -85.93130493164062,
+      "loss": 2.2068,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.324796676635742,
+      "rewards/margins": 0.5959895849227905,
+      "rewards/rejected": -2.9207863807678223,
+      "step": 570
+    },
+    {
+      "epoch": 1.036330608537693,
+      "grad_norm": 1.5611207485198975,
+      "learning_rate": 8.585143175733686e-06,
+      "logits/chosen": 0.09465142339468002,
+      "logits/rejected": 0.13904502987861633,
+      "logps/chosen": -86.5047378540039,
+      "logps/rejected": -95.09537506103516,
+      "loss": 1.9628,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5564661026000977,
+      "rewards/margins": 0.8365517258644104,
+      "rewards/rejected": -3.3930182456970215,
+      "step": 571
+    },
+    {
+      "epoch": 1.0381471389645776,
+      "grad_norm": 1.7271509170532227,
+      "learning_rate": 8.5804619766791e-06,
+      "logits/chosen": 0.08665300905704498,
+      "logits/rejected": 0.10898350924253464,
+      "logps/chosen": -85.17437744140625,
+      "logps/rejected": -94.50398254394531,
+      "loss": 2.2717,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3690195083618164,
+      "rewards/margins": 0.6290315389633179,
+      "rewards/rejected": -2.9980509281158447,
+      "step": 572
+    },
+    {
+      "epoch": 1.0399636693914622,
+      "grad_norm": 1.7832714319229126,
+      "learning_rate": 8.575774496026184e-06,
+      "logits/chosen": 0.033456090837717056,
+      "logits/rejected": 0.09503711760044098,
+      "logps/chosen": -87.27224731445312,
+      "logps/rejected": -94.98229217529297,
+      "loss": 2.1781,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.509322166442871,
+      "rewards/margins": 0.5943110585212708,
+      "rewards/rejected": -3.103632926940918,
+      "step": 573
+    },
+    {
+      "epoch": 1.041780199818347,
+      "grad_norm": 1.8239610195159912,
+      "learning_rate": 8.571080743333528e-06,
+      "logits/chosen": 0.11436322331428528,
+      "logits/rejected": 0.06905262917280197,
+      "logps/chosen": -81.12911987304688,
+      "logps/rejected": -88.73529052734375,
+      "loss": 2.1358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2652292251586914,
+      "rewards/margins": 0.6765917539596558,
+      "rewards/rejected": -2.9418206214904785,
+      "step": 574
+    },
+    {
+      "epoch": 1.0435967302452316,
+      "grad_norm": 1.9338902235031128,
+      "learning_rate": 8.566380728172512e-06,
+      "logits/chosen": 0.11198662221431732,
+      "logits/rejected": 0.11845988035202026,
+      "logps/chosen": -82.55985260009766,
+      "logps/rejected": -94.05873107910156,
+      "loss": 1.9609,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3074758052825928,
+      "rewards/margins": 0.8309231400489807,
+      "rewards/rejected": -3.1383988857269287,
+      "step": 575
+    },
+    {
+      "epoch": 1.0454132606721163,
+      "grad_norm": 1.6932177543640137,
+      "learning_rate": 8.561674460127281e-06,
+      "logits/chosen": 0.0954650342464447,
+      "logits/rejected": 0.139107346534729,
+      "logps/chosen": -83.95409393310547,
+      "logps/rejected": -98.31000518798828,
+      "loss": 2.311,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.476616859436035,
+      "rewards/margins": 0.7193690538406372,
+      "rewards/rejected": -3.195985794067383,
+      "step": 576
+    },
+    {
+      "epoch": 1.047229791099001,
+      "grad_norm": 1.6585469245910645,
+      "learning_rate": 8.556961948794738e-06,
+      "logits/chosen": 0.050676412880420685,
+      "logits/rejected": 0.13941077888011932,
+      "logps/chosen": -91.46603393554688,
+      "logps/rejected": -111.38130187988281,
+      "loss": 1.9574,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.630396604537964,
+      "rewards/margins": 0.8525002002716064,
+      "rewards/rejected": -3.4828968048095703,
+      "step": 577
+    },
+    {
+      "epoch": 1.0490463215258856,
+      "grad_norm": 1.8747247457504272,
+      "learning_rate": 8.552243203784514e-06,
+      "logits/chosen": 0.06109774485230446,
+      "logits/rejected": 0.03656052052974701,
+      "logps/chosen": -84.25686645507812,
+      "logps/rejected": -99.20081329345703,
+      "loss": 2.0244,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4541397094726562,
+      "rewards/margins": 1.0175232887268066,
+      "rewards/rejected": -3.471662759780884,
+      "step": 578
+    },
+    {
+      "epoch": 1.0508628519527703,
+      "grad_norm": 1.714175820350647,
+      "learning_rate": 8.547518234718954e-06,
+      "logits/chosen": 0.07959345728158951,
+      "logits/rejected": 0.17609870433807373,
+      "logps/chosen": -79.58773040771484,
+      "logps/rejected": -98.85990142822266,
+      "loss": 1.7738,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.431654930114746,
+      "rewards/margins": 1.0072871446609497,
+      "rewards/rejected": -3.4389421939849854,
+      "step": 579
+    },
+    {
+      "epoch": 1.052679382379655,
+      "grad_norm": 1.6469632387161255,
+      "learning_rate": 8.542787051233088e-06,
+      "logits/chosen": 0.06128916144371033,
+      "logits/rejected": 0.11901555210351944,
+      "logps/chosen": -90.99638366699219,
+      "logps/rejected": -104.79571533203125,
+      "loss": 1.8789,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.3995859622955322,
+      "rewards/margins": 0.9405049085617065,
+      "rewards/rejected": -3.340090751647949,
+      "step": 580
+    },
+    {
+      "epoch": 1.0544959128065394,
+      "grad_norm": 1.8093998432159424,
+      "learning_rate": 8.538049662974625e-06,
+      "logits/chosen": 0.136922687292099,
+      "logits/rejected": 0.06489068269729614,
+      "logps/chosen": -84.12126159667969,
+      "logps/rejected": -85.51757049560547,
+      "loss": 2.2372,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7452549934387207,
+      "rewards/margins": 0.6496641635894775,
+      "rewards/rejected": -3.394918918609619,
+      "step": 581
+    },
+    {
+      "epoch": 1.056312443233424,
+      "grad_norm": 1.761516809463501,
+      "learning_rate": 8.533306079603928e-06,
+      "logits/chosen": 0.02220803126692772,
+      "logits/rejected": 0.03564752638339996,
+      "logps/chosen": -91.6943359375,
+      "logps/rejected": -105.33677673339844,
+      "loss": 1.831,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.1495776176452637,
+      "rewards/margins": 0.8695197701454163,
+      "rewards/rejected": -3.019097328186035,
+      "step": 582
+    },
+    {
+      "epoch": 1.0581289736603088,
+      "grad_norm": 1.8054577112197876,
+      "learning_rate": 8.52855631079398e-06,
+      "logits/chosen": 0.044406406581401825,
+      "logits/rejected": 0.13649103045463562,
+      "logps/chosen": -83.07669830322266,
+      "logps/rejected": -99.42033386230469,
+      "loss": 2.1295,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.674705743789673,
+      "rewards/margins": 0.8526960015296936,
+      "rewards/rejected": -3.52740216255188,
+      "step": 583
+    },
+    {
+      "epoch": 1.0599455040871935,
+      "grad_norm": 2.092067241668701,
+      "learning_rate": 8.523800366230397e-06,
+      "logits/chosen": 0.11083705723285675,
+      "logits/rejected": 0.13336583971977234,
+      "logps/chosen": -82.55166625976562,
+      "logps/rejected": -88.78050994873047,
+      "loss": 2.783,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.7093071937561035,
+      "rewards/margins": 0.34429922699928284,
+      "rewards/rejected": -3.0536062717437744,
+      "step": 584
+    },
+    {
+      "epoch": 1.0617620345140781,
+      "grad_norm": 2.0747568607330322,
+      "learning_rate": 8.519038255611372e-06,
+      "logits/chosen": 0.06060079485177994,
+      "logits/rejected": 0.06341060996055603,
+      "logps/chosen": -82.93423461914062,
+      "logps/rejected": -85.8175048828125,
+      "loss": 2.1046,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6574935913085938,
+      "rewards/margins": 0.7763374447822571,
+      "rewards/rejected": -3.433830738067627,
+      "step": 585
+    },
+    {
+      "epoch": 1.0635785649409628,
+      "grad_norm": 2.1435482501983643,
+      "learning_rate": 8.51426998864768e-06,
+      "logits/chosen": 0.044051673263311386,
+      "logits/rejected": 0.08015372604131699,
+      "logps/chosen": -86.49735260009766,
+      "logps/rejected": -97.29399108886719,
+      "loss": 2.1365,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.620311737060547,
+      "rewards/margins": 0.6845528483390808,
+      "rewards/rejected": -3.3048648834228516,
+      "step": 586
+    },
+    {
+      "epoch": 1.0653950953678475,
+      "grad_norm": 1.65945303440094,
+      "learning_rate": 8.509495575062647e-06,
+      "logits/chosen": 0.15405744314193726,
+      "logits/rejected": 0.12232419848442078,
+      "logps/chosen": -73.09071350097656,
+      "logps/rejected": -74.24645233154297,
+      "loss": 2.1912,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3465700149536133,
+      "rewards/margins": 0.6234737634658813,
+      "rewards/rejected": -2.970043420791626,
+      "step": 587
+    },
+    {
+      "epoch": 1.067211625794732,
+      "grad_norm": 2.166276216506958,
+      "learning_rate": 8.504715024592132e-06,
+      "logits/chosen": -0.05482687056064606,
+      "logits/rejected": -0.003489813767373562,
+      "logps/chosen": -84.82199096679688,
+      "logps/rejected": -97.53748321533203,
+      "loss": 2.1753,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5944690704345703,
+      "rewards/margins": 0.6753374934196472,
+      "rewards/rejected": -3.269806385040283,
+      "step": 588
+    },
+    {
+      "epoch": 1.0690281562216166,
+      "grad_norm": 1.8869906663894653,
+      "learning_rate": 8.499928346984512e-06,
+      "logits/chosen": 0.13489654660224915,
+      "logits/rejected": 0.052815720438957214,
+      "logps/chosen": -90.98455047607422,
+      "logps/rejected": -91.64305877685547,
+      "loss": 2.3005,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.8705921173095703,
+      "rewards/margins": 0.555056095123291,
+      "rewards/rejected": -3.4256482124328613,
+      "step": 589
+    },
+    {
+      "epoch": 1.0708446866485013,
+      "grad_norm": 2.286853790283203,
+      "learning_rate": 8.495135552000657e-06,
+      "logits/chosen": -0.0365738607943058,
+      "logits/rejected": 0.09564587473869324,
+      "logps/chosen": -93.99796295166016,
+      "logps/rejected": -111.09819793701172,
+      "loss": 2.1574,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.484596014022827,
+      "rewards/margins": 0.840925395488739,
+      "rewards/rejected": -3.325521469116211,
+      "step": 590
+    },
+    {
+      "epoch": 1.072661217075386,
+      "grad_norm": 1.7640435695648193,
+      "learning_rate": 8.490336649413909e-06,
+      "logits/chosen": 0.06625102460384369,
+      "logits/rejected": 0.14394444227218628,
+      "logps/chosen": -74.65644836425781,
+      "logps/rejected": -92.41899871826172,
+      "loss": 2.0387,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.453733205795288,
+      "rewards/margins": 0.7862347364425659,
+      "rewards/rejected": -3.2399678230285645,
+      "step": 591
+    },
+    {
+      "epoch": 1.0744777475022707,
+      "grad_norm": 1.5737273693084717,
+      "learning_rate": 8.485531649010063e-06,
+      "logits/chosen": -0.0012233639135956764,
+      "logits/rejected": 0.021886199712753296,
+      "logps/chosen": -79.36224365234375,
+      "logps/rejected": -94.19145965576172,
+      "loss": 2.3886,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.6413750648498535,
+      "rewards/margins": 0.7800988554954529,
+      "rewards/rejected": -3.421473979949951,
+      "step": 592
+    },
+    {
+      "epoch": 1.0762942779291553,
+      "grad_norm": 2.4172728061676025,
+      "learning_rate": 8.480720560587356e-06,
+      "logits/chosen": 0.04091513529419899,
+      "logits/rejected": 0.16047289967536926,
+      "logps/chosen": -76.38783264160156,
+      "logps/rejected": -94.16845703125,
+      "loss": 2.2219,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.719104290008545,
+      "rewards/margins": 0.9459499716758728,
+      "rewards/rejected": -3.6650545597076416,
+      "step": 593
+    },
+    {
+      "epoch": 1.07811080835604,
+      "grad_norm": 1.6836284399032593,
+      "learning_rate": 8.475903393956434e-06,
+      "logits/chosen": 0.07094185054302216,
+      "logits/rejected": 0.09420361369848251,
+      "logps/chosen": -82.47905731201172,
+      "logps/rejected": -90.06558990478516,
+      "loss": 2.1268,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3819637298583984,
+      "rewards/margins": 0.5947781801223755,
+      "rewards/rejected": -2.9767420291900635,
+      "step": 594
+    },
+    {
+      "epoch": 1.0799273387829247,
+      "grad_norm": 1.808720588684082,
+      "learning_rate": 8.471080158940336e-06,
+      "logits/chosen": 0.010593242943286896,
+      "logits/rejected": 0.03635484725236893,
+      "logps/chosen": -82.94917297363281,
+      "logps/rejected": -94.52174377441406,
+      "loss": 1.9352,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.120471239089966,
+      "rewards/margins": 0.8991338014602661,
+      "rewards/rejected": -3.0196051597595215,
+      "step": 595
+    },
+    {
+      "epoch": 1.0817438692098094,
+      "grad_norm": 1.809515118598938,
+      "learning_rate": 8.46625086537448e-06,
+      "logits/chosen": 0.0202183797955513,
+      "logits/rejected": 0.1138802096247673,
+      "logps/chosen": -84.1740951538086,
+      "logps/rejected": -99.66446685791016,
+      "loss": 2.1384,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7415525913238525,
+      "rewards/margins": 0.8209500908851624,
+      "rewards/rejected": -3.562502861022949,
+      "step": 596
+    },
+    {
+      "epoch": 1.0835603996366938,
+      "grad_norm": 2.023786783218384,
+      "learning_rate": 8.46141552310664e-06,
+      "logits/chosen": 0.11318185180425644,
+      "logits/rejected": 0.1275079846382141,
+      "logps/chosen": -78.02458953857422,
+      "logps/rejected": -89.36206817626953,
+      "loss": 1.9695,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5616073608398438,
+      "rewards/margins": 0.7586690783500671,
+      "rewards/rejected": -3.3202764987945557,
+      "step": 597
+    },
+    {
+      "epoch": 1.0853769300635785,
+      "grad_norm": 1.7670155763626099,
+      "learning_rate": 8.456574141996913e-06,
+      "logits/chosen": 0.09054378420114517,
+      "logits/rejected": 0.17597083747386932,
+      "logps/chosen": -79.7802505493164,
+      "logps/rejected": -95.21873474121094,
+      "loss": 1.9499,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.336801767349243,
+      "rewards/margins": 0.7675365209579468,
+      "rewards/rejected": -3.1043381690979004,
+      "step": 598
+    },
+    {
+      "epoch": 1.0871934604904632,
+      "grad_norm": 2.033724308013916,
+      "learning_rate": 8.451726731917722e-06,
+      "logits/chosen": 0.08810828626155853,
+      "logits/rejected": 0.1388731151819229,
+      "logps/chosen": -75.62480163574219,
+      "logps/rejected": -90.9974365234375,
+      "loss": 1.993,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2493839263916016,
+      "rewards/margins": 0.9519067406654358,
+      "rewards/rejected": -3.2012906074523926,
+      "step": 599
+    },
+    {
+      "epoch": 1.0890099909173478,
+      "grad_norm": 2.0463502407073975,
+      "learning_rate": 8.446873302753783e-06,
+      "logits/chosen": -0.017319753766059875,
+      "logits/rejected": 0.026935823261737823,
+      "logps/chosen": -91.20410919189453,
+      "logps/rejected": -97.56501770019531,
+      "loss": 2.5042,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5798678398132324,
+      "rewards/margins": 0.46915650367736816,
+      "rewards/rejected": -3.0490241050720215,
+      "step": 600
+    },
+    {
+      "epoch": 1.0908265213442325,
+      "grad_norm": 1.944875717163086,
+      "learning_rate": 8.44201386440208e-06,
+      "logits/chosen": 0.1267091929912567,
+      "logits/rejected": 0.12764671444892883,
+      "logps/chosen": -82.59441375732422,
+      "logps/rejected": -91.89214324951172,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.495068311691284,
+      "rewards/margins": 0.6413819789886475,
+      "rewards/rejected": -3.1364500522613525,
+      "step": 601
+    },
+    {
+      "epoch": 1.0926430517711172,
+      "grad_norm": 1.8729236125946045,
+      "learning_rate": 8.437148426771852e-06,
+      "logits/chosen": 0.15159347653388977,
+      "logits/rejected": 0.16930025815963745,
+      "logps/chosen": -80.54000854492188,
+      "logps/rejected": -93.16973114013672,
+      "loss": 1.8605,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4500646591186523,
+      "rewards/margins": 0.9991447925567627,
+      "rewards/rejected": -3.449209690093994,
+      "step": 602
+    },
+    {
+      "epoch": 1.0944595821980019,
+      "grad_norm": 1.9284056425094604,
+      "learning_rate": 8.432276999784578e-06,
+      "logits/chosen": 0.16441625356674194,
+      "logits/rejected": 0.17125487327575684,
+      "logps/chosen": -83.31804656982422,
+      "logps/rejected": -91.39764404296875,
+      "loss": 1.9416,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.577606201171875,
+      "rewards/margins": 0.8178077936172485,
+      "rewards/rejected": -3.395413875579834,
+      "step": 603
+    },
+    {
+      "epoch": 1.0962761126248866,
+      "grad_norm": 2.0095536708831787,
+      "learning_rate": 8.427399593373941e-06,
+      "logits/chosen": 0.03275620937347412,
+      "logits/rejected": 0.033026132732629776,
+      "logps/chosen": -83.01813507080078,
+      "logps/rejected": -93.54053497314453,
+      "loss": 2.0358,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.488004684448242,
+      "rewards/margins": 0.6699528694152832,
+      "rewards/rejected": -3.1579573154449463,
+      "step": 604
+    },
+    {
+      "epoch": 1.0980926430517712,
+      "grad_norm": 1.8138859272003174,
+      "learning_rate": 8.422516217485826e-06,
+      "logits/chosen": 0.1295854151248932,
+      "logits/rejected": 0.15053214132785797,
+      "logps/chosen": -87.74354553222656,
+      "logps/rejected": -94.38423919677734,
+      "loss": 2.6379,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.538604259490967,
+      "rewards/margins": 0.44896796345710754,
+      "rewards/rejected": -2.987572193145752,
+      "step": 605
+    },
+    {
+      "epoch": 1.0999091734786557,
+      "grad_norm": 2.262779951095581,
+      "learning_rate": 8.417626882078287e-06,
+      "logits/chosen": 0.07311846315860748,
+      "logits/rejected": 0.09329289197921753,
+      "logps/chosen": -88.92317199707031,
+      "logps/rejected": -105.15555572509766,
+      "loss": 1.8705,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.468625783920288,
+      "rewards/margins": 1.0499695539474487,
+      "rewards/rejected": -3.5185956954956055,
+      "step": 606
+    },
+    {
+      "epoch": 1.1017257039055404,
+      "grad_norm": 1.7983455657958984,
+      "learning_rate": 8.412731597121527e-06,
+      "logits/chosen": 0.06970179080963135,
+      "logits/rejected": 0.170780211687088,
+      "logps/chosen": -84.99759674072266,
+      "logps/rejected": -91.03089904785156,
+      "loss": 2.3611,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.31594181060791,
+      "rewards/margins": 0.46637701988220215,
+      "rewards/rejected": -2.7823188304901123,
+      "step": 607
+    },
+    {
+      "epoch": 1.103542234332425,
+      "grad_norm": 1.8828606605529785,
+      "learning_rate": 8.407830372597884e-06,
+      "logits/chosen": 0.10343387722969055,
+      "logits/rejected": 0.15009143948554993,
+      "logps/chosen": -74.23157501220703,
+      "logps/rejected": -85.4384994506836,
+      "loss": 2.0535,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3184473514556885,
+      "rewards/margins": 0.7085882425308228,
+      "rewards/rejected": -3.027035713195801,
+      "step": 608
+    },
+    {
+      "epoch": 1.1053587647593097,
+      "grad_norm": 1.8570231199264526,
+      "learning_rate": 8.402923218501813e-06,
+      "logits/chosen": 0.0035794638097286224,
+      "logits/rejected": 0.044682763516902924,
+      "logps/chosen": -82.44414520263672,
+      "logps/rejected": -93.22355651855469,
+      "loss": 2.1751,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.584066867828369,
+      "rewards/margins": 0.7579058408737183,
+      "rewards/rejected": -3.3419723510742188,
+      "step": 609
+    },
+    {
+      "epoch": 1.1071752951861944,
+      "grad_norm": 1.9866671562194824,
+      "learning_rate": 8.39801014483985e-06,
+      "logits/chosen": -0.015922199934720993,
+      "logits/rejected": 0.10281078517436981,
+      "logps/chosen": -99.91399383544922,
+      "logps/rejected": -113.79450988769531,
+      "loss": 2.2522,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.595137119293213,
+      "rewards/margins": 0.7955703139305115,
+      "rewards/rejected": -3.390707492828369,
+      "step": 610
+    },
+    {
+      "epoch": 1.108991825613079,
+      "grad_norm": 2.148136854171753,
+      "learning_rate": 8.393091161630612e-06,
+      "logits/chosen": 0.04068015143275261,
+      "logits/rejected": 0.09121442586183548,
+      "logps/chosen": -85.49508666992188,
+      "logps/rejected": -97.27752685546875,
+      "loss": 1.9467,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.673704147338867,
+      "rewards/margins": 0.8281559348106384,
+      "rewards/rejected": -3.5018603801727295,
+      "step": 611
+    },
+    {
+      "epoch": 1.1108083560399638,
+      "grad_norm": 2.0157570838928223,
+      "learning_rate": 8.388166278904759e-06,
+      "logits/chosen": 0.03451048582792282,
+      "logits/rejected": 0.10373912751674652,
+      "logps/chosen": -79.24334716796875,
+      "logps/rejected": -91.64152526855469,
+      "loss": 2.041,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4238996505737305,
+      "rewards/margins": 0.7874577045440674,
+      "rewards/rejected": -3.211357593536377,
+      "step": 612
+    },
+    {
+      "epoch": 1.1126248864668482,
+      "grad_norm": 1.7337363958358765,
+      "learning_rate": 8.383235506704986e-06,
+      "logits/chosen": 0.06933741271495819,
+      "logits/rejected": 0.10624522715806961,
+      "logps/chosen": -77.62388610839844,
+      "logps/rejected": -91.96408081054688,
+      "loss": 2.0593,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.304389476776123,
+      "rewards/margins": 0.6717057228088379,
+      "rewards/rejected": -2.97609543800354,
+      "step": 613
+    },
+    {
+      "epoch": 1.1144414168937329,
+      "grad_norm": 3.6357574462890625,
+      "learning_rate": 8.378298855085996e-06,
+      "logits/chosen": 0.08138547092676163,
+      "logits/rejected": 0.0961306244134903,
+      "logps/chosen": -87.44918823242188,
+      "logps/rejected": -94.0655746459961,
+      "loss": 2.1543,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5593717098236084,
+      "rewards/margins": 0.6174351572990417,
+      "rewards/rejected": -3.176806688308716,
+      "step": 614
+    },
+    {
+      "epoch": 1.1162579473206176,
+      "grad_norm": 2.273991107940674,
+      "learning_rate": 8.373356334114484e-06,
+      "logits/chosen": 0.029560726135969162,
+      "logits/rejected": 0.11155828088521957,
+      "logps/chosen": -91.85681915283203,
+      "logps/rejected": -103.11241912841797,
+      "loss": 2.2401,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5434679985046387,
+      "rewards/margins": 0.7232552170753479,
+      "rewards/rejected": -3.266723394393921,
+      "step": 615
+    },
+    {
+      "epoch": 1.1180744777475022,
+      "grad_norm": 2.1391029357910156,
+      "learning_rate": 8.368407953869105e-06,
+      "logits/chosen": 0.05583259090781212,
+      "logits/rejected": 0.04287164285778999,
+      "logps/chosen": -92.72290802001953,
+      "logps/rejected": -99.79463195800781,
+      "loss": 1.8293,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.395390272140503,
+      "rewards/margins": 0.9053142666816711,
+      "rewards/rejected": -3.3007044792175293,
+      "step": 616
+    },
+    {
+      "epoch": 1.119891008174387,
+      "grad_norm": 1.6499502658843994,
+      "learning_rate": 8.363453724440471e-06,
+      "logits/chosen": 0.06658023595809937,
+      "logits/rejected": 0.04786865413188934,
+      "logps/chosen": -88.62269592285156,
+      "logps/rejected": -95.76766967773438,
+      "loss": 1.9608,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.522400379180908,
+      "rewards/margins": 0.8233134746551514,
+      "rewards/rejected": -3.3457136154174805,
+      "step": 617
+    },
+    {
+      "epoch": 1.1217075386012716,
+      "grad_norm": 2.018575668334961,
+      "learning_rate": 8.358493655931119e-06,
+      "logits/chosen": 0.06770970672369003,
+      "logits/rejected": 0.08037856221199036,
+      "logps/chosen": -82.23643493652344,
+      "logps/rejected": -97.71478271484375,
+      "loss": 1.8448,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.201833724975586,
+      "rewards/margins": 1.0529935359954834,
+      "rewards/rejected": -3.2548277378082275,
+      "step": 618
+    },
+    {
+      "epoch": 1.1235240690281563,
+      "grad_norm": 2.0033349990844727,
+      "learning_rate": 8.353527758455491e-06,
+      "logits/chosen": 0.11088447272777557,
+      "logits/rejected": 0.09635978937149048,
+      "logps/chosen": -81.90861511230469,
+      "logps/rejected": -92.48001098632812,
+      "loss": 2.2057,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.494230270385742,
+      "rewards/margins": 0.7774206399917603,
+      "rewards/rejected": -3.271650791168213,
+      "step": 619
+    },
+    {
+      "epoch": 1.125340599455041,
+      "grad_norm": 1.700210452079773,
+      "learning_rate": 8.348556042139918e-06,
+      "logits/chosen": 0.08965969830751419,
+      "logits/rejected": 0.06873422861099243,
+      "logps/chosen": -80.28114318847656,
+      "logps/rejected": -84.04791259765625,
+      "loss": 2.4366,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.773008108139038,
+      "rewards/margins": 0.5039690732955933,
+      "rewards/rejected": -3.276977300643921,
+      "step": 620
+    },
+    {
+      "epoch": 1.1271571298819256,
+      "grad_norm": 2.623145580291748,
+      "learning_rate": 8.343578517122594e-06,
+      "logits/chosen": 0.08407651633024216,
+      "logits/rejected": 0.09678006917238235,
+      "logps/chosen": -81.90699005126953,
+      "logps/rejected": -94.30622863769531,
+      "loss": 2.126,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4066429138183594,
+      "rewards/margins": 0.8733029365539551,
+      "rewards/rejected": -3.2799456119537354,
+      "step": 621
+    },
+    {
+      "epoch": 1.12897366030881,
+      "grad_norm": 1.8616830110549927,
+      "learning_rate": 8.338595193553559e-06,
+      "logits/chosen": 0.002603452652692795,
+      "logits/rejected": 0.13452833890914917,
+      "logps/chosen": -73.35641479492188,
+      "logps/rejected": -84.70191192626953,
+      "loss": 2.084,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3735039234161377,
+      "rewards/margins": 0.7769300937652588,
+      "rewards/rejected": -3.1504342555999756,
+      "step": 622
+    },
+    {
+      "epoch": 1.1307901907356948,
+      "grad_norm": 1.8286879062652588,
+      "learning_rate": 8.333606081594678e-06,
+      "logits/chosen": 0.06104264035820961,
+      "logits/rejected": 0.10199623554944992,
+      "logps/chosen": -76.39376831054688,
+      "logps/rejected": -88.08850860595703,
+      "loss": 1.7885,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.130338191986084,
+      "rewards/margins": 0.848843514919281,
+      "rewards/rejected": -2.9791817665100098,
+      "step": 623
+    },
+    {
+      "epoch": 1.1326067211625794,
+      "grad_norm": 1.4803669452667236,
+      "learning_rate": 8.32861119141962e-06,
+      "logits/chosen": 0.07387635856866837,
+      "logits/rejected": 0.11718127131462097,
+      "logps/chosen": -82.92483520507812,
+      "logps/rejected": -91.46949005126953,
+      "loss": 1.7624,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.291823148727417,
+      "rewards/margins": 0.9700915217399597,
+      "rewards/rejected": -3.2619142532348633,
+      "step": 624
+    },
+    {
+      "epoch": 1.134423251589464,
+      "grad_norm": 1.6750752925872803,
+      "learning_rate": 8.323610533213835e-06,
+      "logits/chosen": 0.10784655809402466,
+      "logits/rejected": 0.17897658050060272,
+      "logps/chosen": -75.01900482177734,
+      "logps/rejected": -88.1203384399414,
+      "loss": 2.1956,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4190964698791504,
+      "rewards/margins": 0.688554584980011,
+      "rewards/rejected": -3.1076512336730957,
+      "step": 625
+    },
+    {
+      "epoch": 1.1362397820163488,
+      "grad_norm": 1.8641157150268555,
+      "learning_rate": 8.318604117174533e-06,
+      "logits/chosen": 0.05499793961644173,
+      "logits/rejected": 0.0563310906291008,
+      "logps/chosen": -83.42897033691406,
+      "logps/rejected": -94.2099380493164,
+      "loss": 1.675,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2912588119506836,
+      "rewards/margins": 0.9781790375709534,
+      "rewards/rejected": -3.2694382667541504,
+      "step": 626
+    },
+    {
+      "epoch": 1.1380563124432335,
+      "grad_norm": 1.7625656127929688,
+      "learning_rate": 8.313591953510674e-06,
+      "logits/chosen": 0.0831976979970932,
+      "logits/rejected": 0.12448858469724655,
+      "logps/chosen": -80.08613586425781,
+      "logps/rejected": -85.6810302734375,
+      "loss": 2.3784,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.453763008117676,
+      "rewards/margins": 0.46461760997772217,
+      "rewards/rejected": -2.9183804988861084,
+      "step": 627
+    },
+    {
+      "epoch": 1.1398728428701181,
+      "grad_norm": 2.10945463180542,
+      "learning_rate": 8.308574052442928e-06,
+      "logits/chosen": 0.09317971020936966,
+      "logits/rejected": 0.12335637211799622,
+      "logps/chosen": -82.36027526855469,
+      "logps/rejected": -94.93681335449219,
+      "loss": 2.1623,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5503153800964355,
+      "rewards/margins": 0.6077253222465515,
+      "rewards/rejected": -3.158040761947632,
+      "step": 628
+    },
+    {
+      "epoch": 1.1416893732970028,
+      "grad_norm": 1.8368538618087769,
+      "learning_rate": 8.30355042420367e-06,
+      "logits/chosen": 0.08310627192258835,
+      "logits/rejected": 0.11056395620107651,
+      "logps/chosen": -81.20982360839844,
+      "logps/rejected": -92.69551086425781,
+      "loss": 1.8051,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2482149600982666,
+      "rewards/margins": 0.9621560573577881,
+      "rewards/rejected": -3.2103710174560547,
+      "step": 629
+    },
+    {
+      "epoch": 1.1435059037238875,
+      "grad_norm": 1.7961188554763794,
+      "learning_rate": 8.29852107903695e-06,
+      "logits/chosen": 0.13976231217384338,
+      "logits/rejected": 0.11740799248218536,
+      "logps/chosen": -70.01700592041016,
+      "logps/rejected": -76.1086654663086,
+      "loss": 2.3401,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.239321231842041,
+      "rewards/margins": 0.5643521547317505,
+      "rewards/rejected": -2.80367374420166,
+      "step": 630
+    },
+    {
+      "epoch": 1.145322434150772,
+      "grad_norm": 2.041840076446533,
+      "learning_rate": 8.293486027198483e-06,
+      "logits/chosen": 0.06489598006010056,
+      "logits/rejected": 0.11150355637073517,
+      "logps/chosen": -95.31941223144531,
+      "logps/rejected": -96.98028564453125,
+      "loss": 2.2151,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5194008350372314,
+      "rewards/margins": 0.5945936441421509,
+      "rewards/rejected": -3.113994836807251,
+      "step": 631
+    },
+    {
+      "epoch": 1.1471389645776566,
+      "grad_norm": 2.7394556999206543,
+      "learning_rate": 8.288445278955615e-06,
+      "logits/chosen": 0.10089049488306046,
+      "logits/rejected": 0.08729755133390427,
+      "logps/chosen": -95.64276123046875,
+      "logps/rejected": -99.41658020019531,
+      "loss": 2.6777,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.5987305641174316,
+      "rewards/margins": 0.4612637162208557,
+      "rewards/rejected": -3.0599937438964844,
+      "step": 632
+    },
+    {
+      "epoch": 1.1489554950045413,
+      "grad_norm": 2.024982213973999,
+      "learning_rate": 8.283398844587308e-06,
+      "logits/chosen": 0.023156292736530304,
+      "logits/rejected": 0.06789979338645935,
+      "logps/chosen": -89.90803527832031,
+      "logps/rejected": -96.06836700439453,
+      "loss": 1.9289,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.1415646076202393,
+      "rewards/margins": 0.7636609673500061,
+      "rewards/rejected": -2.9052255153656006,
+      "step": 633
+    },
+    {
+      "epoch": 1.150772025431426,
+      "grad_norm": 1.8750615119934082,
+      "learning_rate": 8.278346734384122e-06,
+      "logits/chosen": 0.02243964746594429,
+      "logits/rejected": 0.12660253047943115,
+      "logps/chosen": -71.34333038330078,
+      "logps/rejected": -86.67922973632812,
+      "loss": 1.8554,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.2015068531036377,
+      "rewards/margins": 0.9946616291999817,
+      "rewards/rejected": -3.1961684226989746,
+      "step": 634
+    },
+    {
+      "epoch": 1.1525885558583107,
+      "grad_norm": 1.9947502613067627,
+      "learning_rate": 8.27328895864819e-06,
+      "logits/chosen": 0.05387189984321594,
+      "logits/rejected": 0.08354412019252777,
+      "logps/chosen": -85.38788604736328,
+      "logps/rejected": -97.08065795898438,
+      "loss": 2.3917,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.595083236694336,
+      "rewards/margins": 0.5549375414848328,
+      "rewards/rejected": -3.1500213146209717,
+      "step": 635
+    },
+    {
+      "epoch": 1.1544050862851953,
+      "grad_norm": 2.145918607711792,
+      "learning_rate": 8.268225527693193e-06,
+      "logits/chosen": 0.10176774114370346,
+      "logits/rejected": 0.14395561814308167,
+      "logps/chosen": -76.02487182617188,
+      "logps/rejected": -86.05180358886719,
+      "loss": 1.8909,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.0001683235168457,
+      "rewards/margins": 0.8724240064620972,
+      "rewards/rejected": -2.8725922107696533,
+      "step": 636
+    },
+    {
+      "epoch": 1.15622161671208,
+      "grad_norm": 1.6830824613571167,
+      "learning_rate": 8.263156451844353e-06,
+      "logits/chosen": 0.10138621181249619,
+      "logits/rejected": 0.11966821551322937,
+      "logps/chosen": -76.56681060791016,
+      "logps/rejected": -88.1928939819336,
+      "loss": 1.9508,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.247589588165283,
+      "rewards/margins": 0.8286643028259277,
+      "rewards/rejected": -3.076253890991211,
+      "step": 637
+    },
+    {
+      "epoch": 1.1580381471389645,
+      "grad_norm": 2.0992536544799805,
+      "learning_rate": 8.258081741438396e-06,
+      "logits/chosen": 0.07541397213935852,
+      "logits/rejected": 0.09907414764165878,
+      "logps/chosen": -90.69495391845703,
+      "logps/rejected": -101.71548461914062,
+      "loss": 2.1933,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4336788654327393,
+      "rewards/margins": 0.7834519147872925,
+      "rewards/rejected": -3.217130422592163,
+      "step": 638
+    },
+    {
+      "epoch": 1.1598546775658491,
+      "grad_norm": 1.8066256046295166,
+      "learning_rate": 8.25300140682354e-06,
+      "logits/chosen": 0.14558832347393036,
+      "logits/rejected": 0.14358943700790405,
+      "logps/chosen": -82.73606872558594,
+      "logps/rejected": -86.27526092529297,
+      "loss": 2.0025,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.393693208694458,
+      "rewards/margins": 0.7131556272506714,
+      "rewards/rejected": -3.106848955154419,
+      "step": 639
+    },
+    {
+      "epoch": 1.1616712079927338,
+      "grad_norm": 2.3521695137023926,
+      "learning_rate": 8.247915458359473e-06,
+      "logits/chosen": 0.05236164107918739,
+      "logits/rejected": 0.1095740869641304,
+      "logps/chosen": -83.70420837402344,
+      "logps/rejected": -100.09117889404297,
+      "loss": 2.3348,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5044684410095215,
+      "rewards/margins": 0.7525762319564819,
+      "rewards/rejected": -3.257044553756714,
+      "step": 640
+    },
+    {
+      "epoch": 1.1634877384196185,
+      "grad_norm": 2.343412160873413,
+      "learning_rate": 8.242823906417329e-06,
+      "logits/chosen": -0.00503757456317544,
+      "logits/rejected": 0.06355556100606918,
+      "logps/chosen": -80.41883087158203,
+      "logps/rejected": -95.32416534423828,
+      "loss": 2.2004,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3209891319274902,
+      "rewards/margins": 0.745349109172821,
+      "rewards/rejected": -3.066338539123535,
+      "step": 641
+    },
+    {
+      "epoch": 1.1653042688465032,
+      "grad_norm": 1.9915282726287842,
+      "learning_rate": 8.237726761379668e-06,
+      "logits/chosen": 0.02087419107556343,
+      "logits/rejected": 0.08556030690670013,
+      "logps/chosen": -84.41456604003906,
+      "logps/rejected": -97.61527252197266,
+      "loss": 2.0052,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.385272741317749,
+      "rewards/margins": 0.7285323739051819,
+      "rewards/rejected": -3.113805055618286,
+      "step": 642
+    },
+    {
+      "epoch": 1.1671207992733879,
+      "grad_norm": 2.4018874168395996,
+      "learning_rate": 8.232624033640458e-06,
+      "logits/chosen": 0.0840807855129242,
+      "logits/rejected": 0.03736239671707153,
+      "logps/chosen": -83.6908950805664,
+      "logps/rejected": -88.65196228027344,
+      "loss": 2.5072,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.494255542755127,
+      "rewards/margins": 0.5595706701278687,
+      "rewards/rejected": -3.053826332092285,
+      "step": 643
+    },
+    {
+      "epoch": 1.1689373297002725,
+      "grad_norm": 2.3055484294891357,
+      "learning_rate": 8.227515733605049e-06,
+      "logits/chosen": 0.05399080738425255,
+      "logits/rejected": 0.13454601168632507,
+      "logps/chosen": -77.62064361572266,
+      "logps/rejected": -86.54924011230469,
+      "loss": 2.0759,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3616065979003906,
+      "rewards/margins": 0.7132005095481873,
+      "rewards/rejected": -3.0748066902160645,
+      "step": 644
+    },
+    {
+      "epoch": 1.1707538601271572,
+      "grad_norm": 1.989559292793274,
+      "learning_rate": 8.222401871690153e-06,
+      "logits/chosen": 0.03339620307087898,
+      "logits/rejected": 0.06584692001342773,
+      "logps/chosen": -74.14878845214844,
+      "logps/rejected": -89.0107421875,
+      "loss": 1.9992,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3560426235198975,
+      "rewards/margins": 0.7924355864524841,
+      "rewards/rejected": -3.1484780311584473,
+      "step": 645
+    },
+    {
+      "epoch": 1.1725703905540419,
+      "grad_norm": 1.7572104930877686,
+      "learning_rate": 8.217282458323825e-06,
+      "logits/chosen": 0.09702566266059875,
+      "logits/rejected": 0.06184859201312065,
+      "logps/chosen": -82.05592346191406,
+      "logps/rejected": -87.59920501708984,
+      "loss": 2.0353,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.121622085571289,
+      "rewards/margins": 0.707017183303833,
+      "rewards/rejected": -2.828639268875122,
+      "step": 646
+    },
+    {
+      "epoch": 1.1743869209809263,
+      "grad_norm": 1.8265388011932373,
+      "learning_rate": 8.21215750394544e-06,
+      "logits/chosen": 0.12195339798927307,
+      "logits/rejected": 0.09588825702667236,
+      "logps/chosen": -76.14807891845703,
+      "logps/rejected": -85.36613464355469,
+      "loss": 2.0684,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.2251272201538086,
+      "rewards/margins": 0.8967663049697876,
+      "rewards/rejected": -3.1218934059143066,
+      "step": 647
+    },
+    {
+      "epoch": 1.176203451407811,
+      "grad_norm": 1.9061174392700195,
+      "learning_rate": 8.207027019005675e-06,
+      "logits/chosen": 0.09651574492454529,
+      "logits/rejected": 0.10208628326654434,
+      "logps/chosen": -86.28733825683594,
+      "logps/rejected": -97.5046157836914,
+      "loss": 2.1319,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3506147861480713,
+      "rewards/margins": 0.7415947318077087,
+      "rewards/rejected": -3.0922091007232666,
+      "step": 648
+    },
+    {
+      "epoch": 1.1780199818346957,
+      "grad_norm": 1.999624252319336,
+      "learning_rate": 8.20189101396648e-06,
+      "logits/chosen": 0.08931821584701538,
+      "logits/rejected": 0.09364073723554611,
+      "logps/chosen": -84.32461547851562,
+      "logps/rejected": -96.70581817626953,
+      "loss": 1.7919,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4468250274658203,
+      "rewards/margins": 0.8684096932411194,
+      "rewards/rejected": -3.315234661102295,
+      "step": 649
+    },
+    {
+      "epoch": 1.1798365122615804,
+      "grad_norm": 1.5666590929031372,
+      "learning_rate": 8.196749499301062e-06,
+      "logits/chosen": 0.09945614635944366,
+      "logits/rejected": 0.12043865770101547,
+      "logps/chosen": -81.99935150146484,
+      "logps/rejected": -96.36174011230469,
+      "loss": 1.8932,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4319562911987305,
+      "rewards/margins": 0.8513790369033813,
+      "rewards/rejected": -3.2833354473114014,
+      "step": 650
+    },
+    {
+      "epoch": 1.181653042688465,
+      "grad_norm": 1.9403899908065796,
+      "learning_rate": 8.191602485493868e-06,
+      "logits/chosen": 0.12361271679401398,
+      "logits/rejected": 0.14710690081119537,
+      "logps/chosen": -82.11424255371094,
+      "logps/rejected": -89.37992858886719,
+      "loss": 2.0615,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.161297082901001,
+      "rewards/margins": 0.7220393419265747,
+      "rewards/rejected": -2.8833365440368652,
+      "step": 651
+    },
+    {
+      "epoch": 1.1834695731153497,
+      "grad_norm": 1.8558425903320312,
+      "learning_rate": 8.186449983040552e-06,
+      "logits/chosen": 0.07215605676174164,
+      "logits/rejected": 0.06390087306499481,
+      "logps/chosen": -81.64322662353516,
+      "logps/rejected": -85.39015197753906,
+      "loss": 2.0064,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.320831775665283,
+      "rewards/margins": 0.7350802421569824,
+      "rewards/rejected": -3.0559120178222656,
+      "step": 652
+    },
+    {
+      "epoch": 1.1852861035422344,
+      "grad_norm": 1.7956428527832031,
+      "learning_rate": 8.181292002447966e-06,
+      "logits/chosen": 0.11276388168334961,
+      "logits/rejected": 0.06517648696899414,
+      "logps/chosen": -76.14613342285156,
+      "logps/rejected": -82.24524688720703,
+      "loss": 2.0473,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2219183444976807,
+      "rewards/margins": 0.6231520175933838,
+      "rewards/rejected": -2.8450703620910645,
+      "step": 653
+    },
+    {
+      "epoch": 1.1871026339691189,
+      "grad_norm": 2.1648740768432617,
+      "learning_rate": 8.17612855423413e-06,
+      "logits/chosen": 0.0334198959171772,
+      "logits/rejected": 0.013921715319156647,
+      "logps/chosen": -83.02323150634766,
+      "logps/rejected": -92.95353698730469,
+      "loss": 1.9884,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.0769619941711426,
+      "rewards/margins": 0.851192057132721,
+      "rewards/rejected": -2.9281537532806396,
+      "step": 654
+    },
+    {
+      "epoch": 1.1889191643960035,
+      "grad_norm": 1.921157717704773,
+      "learning_rate": 8.170959648928214e-06,
+      "logits/chosen": 0.073361836373806,
+      "logits/rejected": 0.13112977147102356,
+      "logps/chosen": -83.41218566894531,
+      "logps/rejected": -92.03260803222656,
+      "loss": 2.0203,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7512502670288086,
+      "rewards/margins": 0.8279457092285156,
+      "rewards/rejected": -3.579195737838745,
+      "step": 655
+    },
+    {
+      "epoch": 1.1907356948228882,
+      "grad_norm": 2.4273247718811035,
+      "learning_rate": 8.165785297070516e-06,
+      "logits/chosen": 0.027622584253549576,
+      "logits/rejected": 0.08521658927202225,
+      "logps/chosen": -84.34210205078125,
+      "logps/rejected": -101.76324462890625,
+      "loss": 1.9023,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6731176376342773,
+      "rewards/margins": 1.004132866859436,
+      "rewards/rejected": -3.677250385284424,
+      "step": 656
+    },
+    {
+      "epoch": 1.192552225249773,
+      "grad_norm": 1.8262163400650024,
+      "learning_rate": 8.16060550921244e-06,
+      "logits/chosen": 0.03431132063269615,
+      "logits/rejected": 0.06454990059137344,
+      "logps/chosen": -76.59049987792969,
+      "logps/rejected": -97.41972351074219,
+      "loss": 2.0105,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3002851009368896,
+      "rewards/margins": 0.7388467788696289,
+      "rewards/rejected": -3.0391316413879395,
+      "step": 657
+    },
+    {
+      "epoch": 1.1943687556766576,
+      "grad_norm": 1.9906315803527832,
+      "learning_rate": 8.155420295916474e-06,
+      "logits/chosen": 0.08427983522415161,
+      "logits/rejected": 0.025464089587330818,
+      "logps/chosen": -84.34020233154297,
+      "logps/rejected": -92.31369018554688,
+      "loss": 1.9895,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3555400371551514,
+      "rewards/margins": 0.8297086954116821,
+      "rewards/rejected": -3.185249090194702,
+      "step": 658
+    },
+    {
+      "epoch": 1.1961852861035422,
+      "grad_norm": 1.9980812072753906,
+      "learning_rate": 8.150229667756172e-06,
+      "logits/chosen": 0.07556813955307007,
+      "logits/rejected": 0.07593435049057007,
+      "logps/chosen": -85.46858215332031,
+      "logps/rejected": -90.94758605957031,
+      "loss": 2.2687,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.679569959640503,
+      "rewards/margins": 0.5908665657043457,
+      "rewards/rejected": -3.2704365253448486,
+      "step": 659
+    },
+    {
+      "epoch": 1.198001816530427,
+      "grad_norm": 2.0758330821990967,
+      "learning_rate": 8.14503363531613e-06,
+      "logits/chosen": 0.11312856525182724,
+      "logits/rejected": 0.08777043223381042,
+      "logps/chosen": -80.32341003417969,
+      "logps/rejected": -84.89898681640625,
+      "loss": 2.1371,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5170207023620605,
+      "rewards/margins": 0.632743239402771,
+      "rewards/rejected": -3.149764060974121,
+      "step": 660
+    },
+    {
+      "epoch": 1.1998183469573116,
+      "grad_norm": 2.765526533126831,
+      "learning_rate": 8.139832209191961e-06,
+      "logits/chosen": 0.17475244402885437,
+      "logits/rejected": 0.13103297352790833,
+      "logps/chosen": -92.03308868408203,
+      "logps/rejected": -94.15422821044922,
+      "loss": 2.4151,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.3920867443084717,
+      "rewards/margins": 0.518372118473053,
+      "rewards/rejected": -2.910458564758301,
+      "step": 661
+    },
+    {
+      "epoch": 1.2016348773841963,
+      "grad_norm": 2.0190646648406982,
+      "learning_rate": 8.13462539999028e-06,
+      "logits/chosen": 0.040579523891210556,
+      "logits/rejected": 0.060216568410396576,
+      "logps/chosen": -96.32977294921875,
+      "logps/rejected": -105.37261962890625,
+      "loss": 2.1216,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5936977863311768,
+      "rewards/margins": 1.040100336074829,
+      "rewards/rejected": -3.6337978839874268,
+      "step": 662
+    },
+    {
+      "epoch": 1.2034514078110807,
+      "grad_norm": 2.53027606010437,
+      "learning_rate": 8.129413218328674e-06,
+      "logits/chosen": 0.07498883455991745,
+      "logits/rejected": 0.042916588485240936,
+      "logps/chosen": -83.17558288574219,
+      "logps/rejected": -89.55461883544922,
+      "loss": 2.1232,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6350390911102295,
+      "rewards/margins": 0.7082449793815613,
+      "rewards/rejected": -3.3432838916778564,
+      "step": 663
+    },
+    {
+      "epoch": 1.2052679382379654,
+      "grad_norm": 1.6480835676193237,
+      "learning_rate": 8.124195674835695e-06,
+      "logits/chosen": 0.1062261089682579,
+      "logits/rejected": 0.09598012268543243,
+      "logps/chosen": -84.5193099975586,
+      "logps/rejected": -95.05423736572266,
+      "loss": 1.7544,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.1563069820404053,
+      "rewards/margins": 0.8859087228775024,
+      "rewards/rejected": -3.042215585708618,
+      "step": 664
+    },
+    {
+      "epoch": 1.20708446866485,
+      "grad_norm": 2.126664161682129,
+      "learning_rate": 8.118972780150817e-06,
+      "logits/chosen": -0.01870904304087162,
+      "logits/rejected": 0.1347496211528778,
+      "logps/chosen": -79.35530853271484,
+      "logps/rejected": -105.41546630859375,
+      "loss": 2.0524,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6006765365600586,
+      "rewards/margins": 1.0105364322662354,
+      "rewards/rejected": -3.611213207244873,
+      "step": 665
+    },
+    {
+      "epoch": 1.2089009990917348,
+      "grad_norm": 2.117645263671875,
+      "learning_rate": 8.113744544924434e-06,
+      "logits/chosen": 0.01287349034100771,
+      "logits/rejected": 0.019422955811023712,
+      "logps/chosen": -105.86250305175781,
+      "logps/rejected": -117.42082977294922,
+      "loss": 1.8973,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.636453151702881,
+      "rewards/margins": 0.8021495342254639,
+      "rewards/rejected": -3.4386026859283447,
+      "step": 666
+    },
+    {
+      "epoch": 1.2107175295186194,
+      "grad_norm": 2.956285238265991,
+      "learning_rate": 8.108510979817828e-06,
+      "logits/chosen": 0.047061942517757416,
+      "logits/rejected": 0.11715231090784073,
+      "logps/chosen": -89.56261444091797,
+      "logps/rejected": -103.57279968261719,
+      "loss": 2.0627,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5317227840423584,
+      "rewards/margins": 0.8301056623458862,
+      "rewards/rejected": -3.361828327178955,
+      "step": 667
+    },
+    {
+      "epoch": 1.2125340599455041,
+      "grad_norm": 1.897623062133789,
+      "learning_rate": 8.10327209550315e-06,
+      "logits/chosen": 0.017857536673545837,
+      "logits/rejected": 0.12088865786790848,
+      "logps/chosen": -95.75634002685547,
+      "logps/rejected": -113.0296630859375,
+      "loss": 1.9091,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6507534980773926,
+      "rewards/margins": 0.963871955871582,
+      "rewards/rejected": -3.6146254539489746,
+      "step": 668
+    },
+    {
+      "epoch": 1.2143505903723888,
+      "grad_norm": 2.176003932952881,
+      "learning_rate": 8.098027902663396e-06,
+      "logits/chosen": -0.0014993082731962204,
+      "logits/rejected": 0.11425516754388809,
+      "logps/chosen": -79.22991943359375,
+      "logps/rejected": -96.39210510253906,
+      "loss": 1.8977,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.582876443862915,
+      "rewards/margins": 0.9855522513389587,
+      "rewards/rejected": -3.5684289932250977,
+      "step": 669
+    },
+    {
+      "epoch": 1.2161671207992735,
+      "grad_norm": 2.140666961669922,
+      "learning_rate": 8.092778411992388e-06,
+      "logits/chosen": 0.04144468903541565,
+      "logits/rejected": 0.0556509830057621,
+      "logps/chosen": -80.35415649414062,
+      "logps/rejected": -97.16475677490234,
+      "loss": 1.9182,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.66176176071167,
+      "rewards/margins": 0.9624839425086975,
+      "rewards/rejected": -3.6242458820343018,
+      "step": 670
+    },
+    {
+      "epoch": 1.2179836512261581,
+      "grad_norm": 1.937705636024475,
+      "learning_rate": 8.087523634194755e-06,
+      "logits/chosen": 0.03470912575721741,
+      "logits/rejected": 0.12844812870025635,
+      "logps/chosen": -81.73896026611328,
+      "logps/rejected": -98.39558410644531,
+      "loss": 2.032,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.406296730041504,
+      "rewards/margins": 0.9326379895210266,
+      "rewards/rejected": -3.3389346599578857,
+      "step": 671
+    },
+    {
+      "epoch": 1.2198001816530426,
+      "grad_norm": 1.8903594017028809,
+      "learning_rate": 8.082263579985898e-06,
+      "logits/chosen": 0.03509046137332916,
+      "logits/rejected": 0.0663415864109993,
+      "logps/chosen": -81.79936981201172,
+      "logps/rejected": -96.15756225585938,
+      "loss": 1.7671,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.2738354206085205,
+      "rewards/margins": 1.1403576135635376,
+      "rewards/rejected": -3.4141931533813477,
+      "step": 672
+    },
+    {
+      "epoch": 1.2216167120799273,
+      "grad_norm": 2.206265687942505,
+      "learning_rate": 8.076998260091989e-06,
+      "logits/chosen": 0.016251683235168457,
+      "logits/rejected": 0.07814528048038483,
+      "logps/chosen": -80.23258209228516,
+      "logps/rejected": -85.3895034790039,
+      "loss": 2.2589,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6450579166412354,
+      "rewards/margins": 0.5741185545921326,
+      "rewards/rejected": -3.2191762924194336,
+      "step": 673
+    },
+    {
+      "epoch": 1.223433242506812,
+      "grad_norm": 2.3217952251434326,
+      "learning_rate": 8.071727685249929e-06,
+      "logits/chosen": 0.15181973576545715,
+      "logits/rejected": 0.11617043614387512,
+      "logps/chosen": -85.15460205078125,
+      "logps/rejected": -83.82669830322266,
+      "loss": 2.2534,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.530156135559082,
+      "rewards/margins": 0.6635551452636719,
+      "rewards/rejected": -3.193711519241333,
+      "step": 674
+    },
+    {
+      "epoch": 1.2252497729336966,
+      "grad_norm": 2.011244297027588,
+      "learning_rate": 8.06645186620734e-06,
+      "logits/chosen": 0.02763158269226551,
+      "logits/rejected": 0.04441550746560097,
+      "logps/chosen": -81.14067840576172,
+      "logps/rejected": -95.907958984375,
+      "loss": 1.8979,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.693267583847046,
+      "rewards/margins": 0.9938480854034424,
+      "rewards/rejected": -3.6871156692504883,
+      "step": 675
+    },
+    {
+      "epoch": 1.2270663033605813,
+      "grad_norm": 2.3554768562316895,
+      "learning_rate": 8.061170813722533e-06,
+      "logits/chosen": 0.06379646062850952,
+      "logits/rejected": 0.052463918924331665,
+      "logps/chosen": -92.42412567138672,
+      "logps/rejected": -92.37240600585938,
+      "loss": 2.4093,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.7538669109344482,
+      "rewards/margins": 0.5145044326782227,
+      "rewards/rejected": -3.268371105194092,
+      "step": 676
+    },
+    {
+      "epoch": 1.228882833787466,
+      "grad_norm": 2.2725136280059814,
+      "learning_rate": 8.055884538564493e-06,
+      "logits/chosen": 0.028828933835029602,
+      "logits/rejected": 0.08082776516675949,
+      "logps/chosen": -76.20146179199219,
+      "logps/rejected": -94.82791900634766,
+      "loss": 1.7849,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.348304510116577,
+      "rewards/margins": 1.1067698001861572,
+      "rewards/rejected": -3.4550740718841553,
+      "step": 677
+    },
+    {
+      "epoch": 1.2306993642143507,
+      "grad_norm": 2.3565657138824463,
+      "learning_rate": 8.050593051512859e-06,
+      "logits/chosen": 0.09367989003658295,
+      "logits/rejected": 0.06132106855511665,
+      "logps/chosen": -84.85393524169922,
+      "logps/rejected": -91.87137603759766,
+      "loss": 2.162,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5945935249328613,
+      "rewards/margins": 0.7115421295166016,
+      "rewards/rejected": -3.306135416030884,
+      "step": 678
+    },
+    {
+      "epoch": 1.2325158946412351,
+      "grad_norm": 2.2214298248291016,
+      "learning_rate": 8.045296363357891e-06,
+      "logits/chosen": 0.015750454738736153,
+      "logits/rejected": -0.001254781149327755,
+      "logps/chosen": -94.99285125732422,
+      "logps/rejected": -107.42745971679688,
+      "loss": 2.0311,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.769968032836914,
+      "rewards/margins": 0.9211719632148743,
+      "rewards/rejected": -3.6911399364471436,
+      "step": 679
+    },
+    {
+      "epoch": 1.2343324250681198,
+      "grad_norm": 1.9377881288528442,
+      "learning_rate": 8.039994484900463e-06,
+      "logits/chosen": -0.011005287989974022,
+      "logits/rejected": 0.05877537280321121,
+      "logps/chosen": -84.62346649169922,
+      "logps/rejected": -99.59872436523438,
+      "loss": 1.8343,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.554427146911621,
+      "rewards/margins": 1.086029291152954,
+      "rewards/rejected": -3.640456199645996,
+      "step": 680
+    },
+    {
+      "epoch": 1.2361489554950045,
+      "grad_norm": 2.665781259536743,
+      "learning_rate": 8.034687426952023e-06,
+      "logits/chosen": 0.030874190852046013,
+      "logits/rejected": 0.09339022636413574,
+      "logps/chosen": -82.25083923339844,
+      "logps/rejected": -99.68885803222656,
+      "loss": 2.0966,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.847683906555176,
+      "rewards/margins": 0.8753347396850586,
+      "rewards/rejected": -3.7230184078216553,
+      "step": 681
+    },
+    {
+      "epoch": 1.2379654859218892,
+      "grad_norm": 1.8999613523483276,
+      "learning_rate": 8.029375200334588e-06,
+      "logits/chosen": 0.07361680269241333,
+      "logits/rejected": 0.08493717014789581,
+      "logps/chosen": -88.70954895019531,
+      "logps/rejected": -102.10923767089844,
+      "loss": 1.7236,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.518230676651001,
+      "rewards/margins": 1.0492362976074219,
+      "rewards/rejected": -3.567467212677002,
+      "step": 682
+    },
+    {
+      "epoch": 1.2397820163487738,
+      "grad_norm": 2.179049491882324,
+      "learning_rate": 8.024057815880716e-06,
+      "logits/chosen": 0.12260061502456665,
+      "logits/rejected": 0.09896722435951233,
+      "logps/chosen": -82.61366271972656,
+      "logps/rejected": -89.37825775146484,
+      "loss": 2.0931,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3436989784240723,
+      "rewards/margins": 0.7696498036384583,
+      "rewards/rejected": -3.1133487224578857,
+      "step": 683
+    },
+    {
+      "epoch": 1.2415985467756585,
+      "grad_norm": 1.8527759313583374,
+      "learning_rate": 8.018735284433475e-06,
+      "logits/chosen": 0.007166730239987373,
+      "logits/rejected": 0.04468049108982086,
+      "logps/chosen": -88.59947204589844,
+      "logps/rejected": -97.66645050048828,
+      "loss": 1.9353,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.669034957885742,
+      "rewards/margins": 0.8667396306991577,
+      "rewards/rejected": -3.5357747077941895,
+      "step": 684
+    },
+    {
+      "epoch": 1.2434150772025432,
+      "grad_norm": 2.013792037963867,
+      "learning_rate": 8.013407616846436e-06,
+      "logits/chosen": 0.05774620547890663,
+      "logits/rejected": 0.06977065652608871,
+      "logps/chosen": -83.88308715820312,
+      "logps/rejected": -93.36985778808594,
+      "loss": 1.9716,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.5362911224365234,
+      "rewards/margins": 0.9699288010597229,
+      "rewards/rejected": -3.5062198638916016,
+      "step": 685
+    },
+    {
+      "epoch": 1.2452316076294279,
+      "grad_norm": 1.9147852659225464,
+      "learning_rate": 8.008074823983642e-06,
+      "logits/chosen": 0.10071337223052979,
+      "logits/rejected": 0.09696929156780243,
+      "logps/chosen": -87.06709289550781,
+      "logps/rejected": -98.84783172607422,
+      "loss": 1.7989,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7123823165893555,
+      "rewards/margins": 0.9586774110794067,
+      "rewards/rejected": -3.6710598468780518,
+      "step": 686
+    },
+    {
+      "epoch": 1.2470481380563125,
+      "grad_norm": 3.8477859497070312,
+      "learning_rate": 8.002736916719582e-06,
+      "logits/chosen": 0.09164638817310333,
+      "logits/rejected": 0.14311164617538452,
+      "logps/chosen": -94.75381469726562,
+      "logps/rejected": -101.14093017578125,
+      "loss": 2.3677,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.7352569103240967,
+      "rewards/margins": 0.5952449440956116,
+      "rewards/rejected": -3.3305020332336426,
+      "step": 687
+    },
+    {
+      "epoch": 1.248864668483197,
+      "grad_norm": 2.1601197719573975,
+      "learning_rate": 7.997393905939183e-06,
+      "logits/chosen": 0.14653781056404114,
+      "logits/rejected": 0.217861607670784,
+      "logps/chosen": -78.05258178710938,
+      "logps/rejected": -87.37406921386719,
+      "loss": 2.2395,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4850776195526123,
+      "rewards/margins": 0.5597268342971802,
+      "rewards/rejected": -3.044804573059082,
+      "step": 688
+    },
+    {
+      "epoch": 1.2506811989100817,
+      "grad_norm": 2.1657280921936035,
+      "learning_rate": 7.992045802537772e-06,
+      "logits/chosen": 0.008183173835277557,
+      "logits/rejected": 0.07762319594621658,
+      "logps/chosen": -76.81200408935547,
+      "logps/rejected": -89.77823638916016,
+      "loss": 2.0968,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.395460844039917,
+      "rewards/margins": 0.7663928270339966,
+      "rewards/rejected": -3.161853790283203,
+      "step": 689
+    },
+    {
+      "epoch": 1.2524977293369663,
+      "grad_norm": 2.0229945182800293,
+      "learning_rate": 7.986692617421063e-06,
+      "logits/chosen": 0.049828238785266876,
+      "logits/rejected": 0.12920916080474854,
+      "logps/chosen": -89.49681854248047,
+      "logps/rejected": -94.05538177490234,
+      "loss": 2.3605,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5402889251708984,
+      "rewards/margins": 0.6041077375411987,
+      "rewards/rejected": -3.1443967819213867,
+      "step": 690
+    },
+    {
+      "epoch": 1.254314259763851,
+      "grad_norm": 2.1632978916168213,
+      "learning_rate": 7.981334361505131e-06,
+      "logits/chosen": 0.01283620297908783,
+      "logits/rejected": 0.06363464891910553,
+      "logps/chosen": -90.30138397216797,
+      "logps/rejected": -109.25923156738281,
+      "loss": 1.9464,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5346744060516357,
+      "rewards/margins": 1.0385441780090332,
+      "rewards/rejected": -3.573218584060669,
+      "step": 691
+    },
+    {
+      "epoch": 1.2561307901907357,
+      "grad_norm": 2.1562981605529785,
+      "learning_rate": 7.975971045716398e-06,
+      "logits/chosen": 0.12652680277824402,
+      "logits/rejected": 0.15553821623325348,
+      "logps/chosen": -74.45803833007812,
+      "logps/rejected": -90.87694549560547,
+      "loss": 1.8599,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.415904998779297,
+      "rewards/margins": 0.9220768809318542,
+      "rewards/rejected": -3.337982177734375,
+      "step": 692
+    },
+    {
+      "epoch": 1.2579473206176204,
+      "grad_norm": 2.007781982421875,
+      "learning_rate": 7.970602680991594e-06,
+      "logits/chosen": 0.05228351801633835,
+      "logits/rejected": 0.12716761231422424,
+      "logps/chosen": -96.6414794921875,
+      "logps/rejected": -115.6935043334961,
+      "loss": 2.0777,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5941665172576904,
+      "rewards/margins": 1.015256643295288,
+      "rewards/rejected": -3.6094233989715576,
+      "step": 693
+    },
+    {
+      "epoch": 1.259763851044505,
+      "grad_norm": 2.2393176555633545,
+      "learning_rate": 7.965229278277749e-06,
+      "logits/chosen": 0.11106812953948975,
+      "logits/rejected": 0.10329913347959518,
+      "logps/chosen": -91.51345825195312,
+      "logps/rejected": -95.78245544433594,
+      "loss": 2.1322,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6126160621643066,
+      "rewards/margins": 0.8137804865837097,
+      "rewards/rejected": -3.426396608352661,
+      "step": 694
+    },
+    {
+      "epoch": 1.2615803814713895,
+      "grad_norm": 1.9000614881515503,
+      "learning_rate": 7.959850848532172e-06,
+      "logits/chosen": 0.053980953991413116,
+      "logits/rejected": 0.1285300850868225,
+      "logps/chosen": -80.65451049804688,
+      "logps/rejected": -96.31999206542969,
+      "loss": 1.8157,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.449226140975952,
+      "rewards/margins": 0.9224156141281128,
+      "rewards/rejected": -3.3716418743133545,
+      "step": 695
+    },
+    {
+      "epoch": 1.2633969118982744,
+      "grad_norm": 2.1616997718811035,
+      "learning_rate": 7.95446740272241e-06,
+      "logits/chosen": 0.0943533331155777,
+      "logits/rejected": 0.09738673269748688,
+      "logps/chosen": -86.31460571289062,
+      "logps/rejected": -95.93961334228516,
+      "loss": 2.3068,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.4675066471099854,
+      "rewards/margins": 0.5155179500579834,
+      "rewards/rejected": -2.9830245971679688,
+      "step": 696
+    },
+    {
+      "epoch": 1.2652134423251589,
+      "grad_norm": 2.040947675704956,
+      "learning_rate": 7.949078951826254e-06,
+      "logits/chosen": 0.05092187598347664,
+      "logits/rejected": 0.10424528270959854,
+      "logps/chosen": -78.41385650634766,
+      "logps/rejected": -96.27998352050781,
+      "loss": 1.9729,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.392333507537842,
+      "rewards/margins": 0.8541865348815918,
+      "rewards/rejected": -3.246520519256592,
+      "step": 697
+    },
+    {
+      "epoch": 1.2670299727520435,
+      "grad_norm": 2.123605251312256,
+      "learning_rate": 7.943685506831689e-06,
+      "logits/chosen": 0.1472083032131195,
+      "logits/rejected": 0.07182697206735611,
+      "logps/chosen": -87.48722839355469,
+      "logps/rejected": -89.31311798095703,
+      "loss": 2.1536,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.34977388381958,
+      "rewards/margins": 0.6212242245674133,
+      "rewards/rejected": -2.9709980487823486,
+      "step": 698
+    },
+    {
+      "epoch": 1.2688465031789282,
+      "grad_norm": 2.11053729057312,
+      "learning_rate": 7.938287078736889e-06,
+      "logits/chosen": 0.08341722190380096,
+      "logits/rejected": 0.12655527889728546,
+      "logps/chosen": -79.16002655029297,
+      "logps/rejected": -91.57137298583984,
+      "loss": 2.0396,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4659664630889893,
+      "rewards/margins": 0.7776908874511719,
+      "rewards/rejected": -3.2436575889587402,
+      "step": 699
+    },
+    {
+      "epoch": 1.270663033605813,
+      "grad_norm": 1.9555423259735107,
+      "learning_rate": 7.932883678550191e-06,
+      "logits/chosen": -0.06148137152194977,
+      "logits/rejected": 0.055927492678165436,
+      "logps/chosen": -81.51953125,
+      "logps/rejected": -95.884521484375,
+      "loss": 1.9223,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5550904273986816,
+      "rewards/margins": 0.8093082904815674,
+      "rewards/rejected": -3.364398956298828,
+      "step": 700
+    },
+    {
+      "epoch": 1.2724795640326976,
+      "grad_norm": 2.031662702560425,
+      "learning_rate": 7.927475317290065e-06,
+      "logits/chosen": 0.04235261678695679,
+      "logits/rejected": 0.1255102902650833,
+      "logps/chosen": -74.28742218017578,
+      "logps/rejected": -86.79241180419922,
+      "loss": 1.9669,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.138885498046875,
+      "rewards/margins": 0.8455573916435242,
+      "rewards/rejected": -2.984442949295044,
+      "step": 701
+    },
+    {
+      "epoch": 1.2742960944595823,
+      "grad_norm": 2.131220817565918,
+      "learning_rate": 7.922062005985105e-06,
+      "logits/chosen": 0.15667293965816498,
+      "logits/rejected": 0.12293773889541626,
+      "logps/chosen": -76.34903717041016,
+      "logps/rejected": -80.57623291015625,
+      "loss": 2.2422,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.408144474029541,
+      "rewards/margins": 0.5897243618965149,
+      "rewards/rejected": -2.997868776321411,
+      "step": 702
+    },
+    {
+      "epoch": 1.276112624886467,
+      "grad_norm": 1.9162285327911377,
+      "learning_rate": 7.916643755673994e-06,
+      "logits/chosen": -0.00397343747317791,
+      "logits/rejected": -0.01795889064669609,
+      "logps/chosen": -72.12644958496094,
+      "logps/rejected": -80.48601531982422,
+      "loss": 2.286,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.1793017387390137,
+      "rewards/margins": 0.6844039559364319,
+      "rewards/rejected": -2.863705635070801,
+      "step": 703
+    },
+    {
+      "epoch": 1.2779291553133514,
+      "grad_norm": 2.877826690673828,
+      "learning_rate": 7.911220577405485e-06,
+      "logits/chosen": 0.031911637634038925,
+      "logits/rejected": -0.02112415060400963,
+      "logps/chosen": -92.17039489746094,
+      "logps/rejected": -100.05145263671875,
+      "loss": 2.2664,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.5374324321746826,
+      "rewards/margins": 0.6277021765708923,
+      "rewards/rejected": -3.1651346683502197,
+      "step": 704
+    },
+    {
+      "epoch": 1.2797456857402363,
+      "grad_norm": 2.437596082687378,
+      "learning_rate": 7.905792482238387e-06,
+      "logits/chosen": 0.05161967873573303,
+      "logits/rejected": 0.07132750749588013,
+      "logps/chosen": -83.6578140258789,
+      "logps/rejected": -91.72476959228516,
+      "loss": 2.2835,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.62272310256958,
+      "rewards/margins": 0.7025635242462158,
+      "rewards/rejected": -3.325286865234375,
+      "step": 705
+    },
+    {
+      "epoch": 1.2815622161671207,
+      "grad_norm": 1.9355360269546509,
+      "learning_rate": 7.900359481241531e-06,
+      "logits/chosen": 0.11531795561313629,
+      "logits/rejected": 0.1957155019044876,
+      "logps/chosen": -70.70565795898438,
+      "logps/rejected": -83.15331268310547,
+      "loss": 1.9752,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.568023681640625,
+      "rewards/margins": 0.748910129070282,
+      "rewards/rejected": -3.3169333934783936,
+      "step": 706
+    },
+    {
+      "epoch": 1.2833787465940054,
+      "grad_norm": 1.9978153705596924,
+      "learning_rate": 7.89492158549375e-06,
+      "logits/chosen": 0.04371386021375656,
+      "logits/rejected": 0.07748796790838242,
+      "logps/chosen": -80.5776138305664,
+      "logps/rejected": -90.11798095703125,
+      "loss": 2.0422,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.1579325199127197,
+      "rewards/margins": 0.726677656173706,
+      "rewards/rejected": -2.8846099376678467,
+      "step": 707
+    },
+    {
+      "epoch": 1.28519527702089,
+      "grad_norm": 1.982696771621704,
+      "learning_rate": 7.889478806083859e-06,
+      "logits/chosen": 0.11542633175849915,
+      "logits/rejected": 0.09008399397134781,
+      "logps/chosen": -80.96235656738281,
+      "logps/rejected": -84.60002899169922,
+      "loss": 2.1817,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.0073299407958984,
+      "rewards/margins": 0.626833438873291,
+      "rewards/rejected": -2.6341633796691895,
+      "step": 708
+    },
+    {
+      "epoch": 1.2870118074477748,
+      "grad_norm": 2.2399492263793945,
+      "learning_rate": 7.884031154110636e-06,
+      "logits/chosen": 0.13016283512115479,
+      "logits/rejected": 0.10277393460273743,
+      "logps/chosen": -76.04215240478516,
+      "logps/rejected": -82.029296875,
+      "loss": 2.1021,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.4932732582092285,
+      "rewards/margins": 0.7273231744766235,
+      "rewards/rejected": -3.2205960750579834,
+      "step": 709
+    },
+    {
+      "epoch": 1.2888283378746594,
+      "grad_norm": 1.7780169248580933,
+      "learning_rate": 7.87857864068279e-06,
+      "logits/chosen": 0.015022790059447289,
+      "logits/rejected": 0.08707302063703537,
+      "logps/chosen": -85.66944122314453,
+      "logps/rejected": -101.31831359863281,
+      "loss": 1.8779,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4976251125335693,
+      "rewards/margins": 0.9009003639221191,
+      "rewards/rejected": -3.3985257148742676,
+      "step": 710
+    },
+    {
+      "epoch": 1.2906448683015441,
+      "grad_norm": 2.2210092544555664,
+      "learning_rate": 7.873121276918947e-06,
+      "logits/chosen": 0.02679639309644699,
+      "logits/rejected": 0.00411232328042388,
+      "logps/chosen": -87.88114929199219,
+      "logps/rejected": -98.84779357910156,
+      "loss": 2.2478,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.56654953956604,
+      "rewards/margins": 0.7424182295799255,
+      "rewards/rejected": -3.3089675903320312,
+      "step": 711
+    },
+    {
+      "epoch": 1.2924613987284288,
+      "grad_norm": 2.3803563117980957,
+      "learning_rate": 7.867659073947618e-06,
+      "logits/chosen": 0.12869741022586823,
+      "logits/rejected": 0.17313969135284424,
+      "logps/chosen": -78.2127685546875,
+      "logps/rejected": -89.54530334472656,
+      "loss": 2.2313,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.2719531059265137,
+      "rewards/margins": 0.6112346649169922,
+      "rewards/rejected": -2.883187770843506,
+      "step": 712
+    },
+    {
+      "epoch": 1.2942779291553133,
+      "grad_norm": 1.8979357481002808,
+      "learning_rate": 7.862192042907191e-06,
+      "logits/chosen": -0.002289058640599251,
+      "logits/rejected": 0.01941063068807125,
+      "logps/chosen": -79.63894653320312,
+      "logps/rejected": -99.8336410522461,
+      "loss": 1.8824,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.396660566329956,
+      "rewards/margins": 1.1458237171173096,
+      "rewards/rejected": -3.5424838066101074,
+      "step": 713
+    },
+    {
+      "epoch": 1.296094459582198,
+      "grad_norm": 2.069608449935913,
+      "learning_rate": 7.856720194945892e-06,
+      "logits/chosen": 0.061066634953022,
+      "logits/rejected": 0.0689641609787941,
+      "logps/chosen": -78.00001525878906,
+      "logps/rejected": -87.9764404296875,
+      "loss": 2.0765,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3631246089935303,
+      "rewards/margins": 0.8202179670333862,
+      "rewards/rejected": -3.183342695236206,
+      "step": 714
+    },
+    {
+      "epoch": 1.2979109900090826,
+      "grad_norm": 2.080536365509033,
+      "learning_rate": 7.85124354122177e-06,
+      "logits/chosen": 0.055848486721515656,
+      "logits/rejected": 0.09090499579906464,
+      "logps/chosen": -86.07003784179688,
+      "logps/rejected": -95.0962905883789,
+      "loss": 2.0101,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5897908210754395,
+      "rewards/margins": 0.6455370187759399,
+      "rewards/rejected": -3.23532772064209,
+      "step": 715
+    },
+    {
+      "epoch": 1.2997275204359673,
+      "grad_norm": 1.9409488439559937,
+      "learning_rate": 7.845762092902678e-06,
+      "logits/chosen": 0.17071956396102905,
+      "logits/rejected": 0.11957748234272003,
+      "logps/chosen": -85.08267211914062,
+      "logps/rejected": -100.11673736572266,
+      "loss": 1.9544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2637064456939697,
+      "rewards/margins": 0.9579347968101501,
+      "rewards/rejected": -3.2216408252716064,
+      "step": 716
+    },
+    {
+      "epoch": 1.301544050862852,
+      "grad_norm": 2.3988044261932373,
+      "learning_rate": 7.840275861166243e-06,
+      "logits/chosen": 0.07587475329637527,
+      "logits/rejected": 0.018805254250764847,
+      "logps/chosen": -92.8540267944336,
+      "logps/rejected": -97.47180938720703,
+      "loss": 2.3578,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.429079294204712,
+      "rewards/margins": 0.5412003993988037,
+      "rewards/rejected": -2.9702794551849365,
+      "step": 717
+    },
+    {
+      "epoch": 1.3033605812897366,
+      "grad_norm": 2.509941577911377,
+      "learning_rate": 7.834784857199846e-06,
+      "logits/chosen": 0.0847601592540741,
+      "logits/rejected": 0.039042674005031586,
+      "logps/chosen": -80.2916030883789,
+      "logps/rejected": -81.85697174072266,
+      "loss": 2.0909,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4719648361206055,
+      "rewards/margins": 0.6977176666259766,
+      "rewards/rejected": -3.169682264328003,
+      "step": 718
+    },
+    {
+      "epoch": 1.3051771117166213,
+      "grad_norm": 2.192031145095825,
+      "learning_rate": 7.8292890922006e-06,
+      "logits/chosen": 0.048360832035541534,
+      "logits/rejected": 0.10979942232370377,
+      "logps/chosen": -72.04978942871094,
+      "logps/rejected": -85.74111938476562,
+      "loss": 2.1544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.367837905883789,
+      "rewards/margins": 0.6129733920097351,
+      "rewards/rejected": -2.98081111907959,
+      "step": 719
+    },
+    {
+      "epoch": 1.3069936421435058,
+      "grad_norm": 2.33526873588562,
+      "learning_rate": 7.82378857737533e-06,
+      "logits/chosen": 0.09924662113189697,
+      "logits/rejected": 0.10467047989368439,
+      "logps/chosen": -85.21525573730469,
+      "logps/rejected": -94.86468505859375,
+      "loss": 2.0783,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.466061592102051,
+      "rewards/margins": 0.7927428483963013,
+      "rewards/rejected": -3.2588045597076416,
+      "step": 720
+    },
+    {
+      "epoch": 1.3088101725703907,
+      "grad_norm": 2.381256341934204,
+      "learning_rate": 7.818283323940539e-06,
+      "logits/chosen": 0.1228807270526886,
+      "logits/rejected": 0.1338506042957306,
+      "logps/chosen": -90.72122192382812,
+      "logps/rejected": -96.53125,
+      "loss": 2.0252,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4308130741119385,
+      "rewards/margins": 0.6035470366477966,
+      "rewards/rejected": -3.03436017036438,
+      "step": 721
+    },
+    {
+      "epoch": 1.3106267029972751,
+      "grad_norm": 1.942866325378418,
+      "learning_rate": 7.8127733431224e-06,
+      "logits/chosen": 0.05967065691947937,
+      "logits/rejected": 0.08946052193641663,
+      "logps/chosen": -78.91156768798828,
+      "logps/rejected": -90.6188735961914,
+      "loss": 1.9435,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.333373546600342,
+      "rewards/margins": 0.7353364825248718,
+      "rewards/rejected": -3.0687103271484375,
+      "step": 722
+    },
+    {
+      "epoch": 1.3124432334241598,
+      "grad_norm": 2.068608283996582,
+      "learning_rate": 7.80725864615672e-06,
+      "logits/chosen": 0.01783425733447075,
+      "logits/rejected": 0.06930528581142426,
+      "logps/chosen": -81.48194885253906,
+      "logps/rejected": -96.17157745361328,
+      "loss": 2.0554,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.423696756362915,
+      "rewards/margins": 0.8400081396102905,
+      "rewards/rejected": -3.263705253601074,
+      "step": 723
+    },
+    {
+      "epoch": 1.3142597638510445,
+      "grad_norm": 2.349059820175171,
+      "learning_rate": 7.801739244288929e-06,
+      "logits/chosen": 0.08266536891460419,
+      "logits/rejected": 0.09768007695674896,
+      "logps/chosen": -82.81443786621094,
+      "logps/rejected": -92.47273254394531,
+      "loss": 1.8288,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.347062587738037,
+      "rewards/margins": 0.839061975479126,
+      "rewards/rejected": -3.186124563217163,
+      "step": 724
+    },
+    {
+      "epoch": 1.3160762942779292,
+      "grad_norm": 2.1312410831451416,
+      "learning_rate": 7.79621514877405e-06,
+      "logits/chosen": 0.05873649939894676,
+      "logits/rejected": 0.10713300108909607,
+      "logps/chosen": -99.603759765625,
+      "logps/rejected": -110.6041259765625,
+      "loss": 1.8875,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4636247158050537,
+      "rewards/margins": 0.8619670271873474,
+      "rewards/rejected": -3.325591564178467,
+      "step": 725
+    },
+    {
+      "epoch": 1.3178928247048138,
+      "grad_norm": 2.137871742248535,
+      "learning_rate": 7.79068637087667e-06,
+      "logits/chosen": 0.07839182764291763,
+      "logits/rejected": 0.1225314810872078,
+      "logps/chosen": -82.73197937011719,
+      "logps/rejected": -101.63593292236328,
+      "loss": 1.7664,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.491330862045288,
+      "rewards/margins": 1.1549298763275146,
+      "rewards/rejected": -3.646261215209961,
+      "step": 726
+    },
+    {
+      "epoch": 1.3197093551316985,
+      "grad_norm": 2.030689001083374,
+      "learning_rate": 7.785152921870939e-06,
+      "logits/chosen": 0.034951481968164444,
+      "logits/rejected": 0.06388817727565765,
+      "logps/chosen": -85.28425598144531,
+      "logps/rejected": -101.43257904052734,
+      "loss": 2.3856,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6969833374023438,
+      "rewards/margins": 0.727889895439148,
+      "rewards/rejected": -3.4248733520507812,
+      "step": 727
+    },
+    {
+      "epoch": 1.3215258855585832,
+      "grad_norm": 2.65952205657959,
+      "learning_rate": 7.779614813040513e-06,
+      "logits/chosen": 0.1679934561252594,
+      "logits/rejected": 0.18826396763324738,
+      "logps/chosen": -80.64529418945312,
+      "logps/rejected": -87.35059356689453,
+      "loss": 1.9723,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.259064197540283,
+      "rewards/margins": 0.7135547995567322,
+      "rewards/rejected": -2.97261905670166,
+      "step": 728
+    },
+    {
+      "epoch": 1.3233424159854676,
+      "grad_norm": 2.1841068267822266,
+      "learning_rate": 7.774072055678565e-06,
+      "logits/chosen": 0.11853618919849396,
+      "logits/rejected": 0.1412927508354187,
+      "logps/chosen": -80.26649475097656,
+      "logps/rejected": -88.19939422607422,
+      "loss": 2.1344,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5820248126983643,
+      "rewards/margins": 0.8093788623809814,
+      "rewards/rejected": -3.3914034366607666,
+      "step": 729
+    },
+    {
+      "epoch": 1.3251589464123525,
+      "grad_norm": 2.463083505630493,
+      "learning_rate": 7.768524661087744e-06,
+      "logits/chosen": 0.05979528650641441,
+      "logits/rejected": 0.10566453635692596,
+      "logps/chosen": -88.6331787109375,
+      "logps/rejected": -97.57271575927734,
+      "loss": 2.1545,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4209208488464355,
+      "rewards/margins": 0.7680116295814514,
+      "rewards/rejected": -3.188931941986084,
+      "step": 730
+    },
+    {
+      "epoch": 1.326975476839237,
+      "grad_norm": 2.3693671226501465,
+      "learning_rate": 7.76297264058015e-06,
+      "logits/chosen": 0.05617160722613335,
+      "logits/rejected": 0.017447492107748985,
+      "logps/chosen": -91.15104675292969,
+      "logps/rejected": -91.32392883300781,
+      "loss": 2.0652,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.469024181365967,
+      "rewards/margins": 0.7454814910888672,
+      "rewards/rejected": -3.214505672454834,
+      "step": 731
+    },
+    {
+      "epoch": 1.3287920072661217,
+      "grad_norm": 2.0553715229034424,
+      "learning_rate": 7.757416005477323e-06,
+      "logits/chosen": 0.10786914825439453,
+      "logits/rejected": 0.1281609684228897,
+      "logps/chosen": -83.35604095458984,
+      "logps/rejected": -91.0313949584961,
+      "loss": 2.0932,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.75126314163208,
+      "rewards/margins": 0.7679738402366638,
+      "rewards/rejected": -3.5192370414733887,
+      "step": 732
+    },
+    {
+      "epoch": 1.3306085376930064,
+      "grad_norm": 2.6259214878082275,
+      "learning_rate": 7.751854767110205e-06,
+      "logits/chosen": 0.025147411972284317,
+      "logits/rejected": 0.11643064022064209,
+      "logps/chosen": -74.00305938720703,
+      "logps/rejected": -99.70087432861328,
+      "loss": 1.9757,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.617929220199585,
+      "rewards/margins": 1.131290078163147,
+      "rewards/rejected": -3.7492191791534424,
+      "step": 733
+    },
+    {
+      "epoch": 1.332425068119891,
+      "grad_norm": 2.21140193939209,
+      "learning_rate": 7.746288936819134e-06,
+      "logits/chosen": 0.036435484886169434,
+      "logits/rejected": 0.09736932814121246,
+      "logps/chosen": -79.6103515625,
+      "logps/rejected": -97.5123519897461,
+      "loss": 1.999,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.716804027557373,
+      "rewards/margins": 0.9280580282211304,
+      "rewards/rejected": -3.644861936569214,
+      "step": 734
+    },
+    {
+      "epoch": 1.3342415985467757,
+      "grad_norm": 1.8063753843307495,
+      "learning_rate": 7.740718525953804e-06,
+      "logits/chosen": 0.012957969680428505,
+      "logits/rejected": 0.06407644599676132,
+      "logps/chosen": -86.7850112915039,
+      "logps/rejected": -104.12468719482422,
+      "loss": 1.8674,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5402910709381104,
+      "rewards/margins": 0.9172405004501343,
+      "rewards/rejected": -3.457531690597534,
+      "step": 735
+    },
+    {
+      "epoch": 1.3360581289736604,
+      "grad_norm": 2.384291410446167,
+      "learning_rate": 7.735143545873252e-06,
+      "logits/chosen": 0.15138015151023865,
+      "logits/rejected": 0.1674022376537323,
+      "logps/chosen": -78.28397369384766,
+      "logps/rejected": -88.92375183105469,
+      "loss": 1.8622,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6077139377593994,
+      "rewards/margins": 0.9488463401794434,
+      "rewards/rejected": -3.5565602779388428,
+      "step": 736
+    },
+    {
+      "epoch": 1.337874659400545,
+      "grad_norm": 2.1717827320098877,
+      "learning_rate": 7.729564007945834e-06,
+      "logits/chosen": 0.07187292724847794,
+      "logits/rejected": 0.12309285253286362,
+      "logps/chosen": -87.5180892944336,
+      "logps/rejected": -93.48638153076172,
+      "loss": 2.2686,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4702401161193848,
+      "rewards/margins": 0.6287782192230225,
+      "rewards/rejected": -3.099018096923828,
+      "step": 737
+    },
+    {
+      "epoch": 1.3396911898274295,
+      "grad_norm": 2.6149754524230957,
+      "learning_rate": 7.723979923549201e-06,
+      "logits/chosen": 0.10500997304916382,
+      "logits/rejected": 0.13640286028385162,
+      "logps/chosen": -76.4052963256836,
+      "logps/rejected": -84.28958892822266,
+      "loss": 2.4316,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.9436187744140625,
+      "rewards/margins": 0.5192191004753113,
+      "rewards/rejected": -3.4628381729125977,
+      "step": 738
+    },
+    {
+      "epoch": 1.3415077202543142,
+      "grad_norm": 2.2604503631591797,
+      "learning_rate": 7.718391304070274e-06,
+      "logits/chosen": 0.019504152238368988,
+      "logits/rejected": 0.0677918866276741,
+      "logps/chosen": -96.57980346679688,
+      "logps/rejected": -111.43043518066406,
+      "loss": 2.1929,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.63236141204834,
+      "rewards/margins": 0.8397969007492065,
+      "rewards/rejected": -3.4721579551696777,
+      "step": 739
+    },
+    {
+      "epoch": 1.3433242506811989,
+      "grad_norm": 2.827580451965332,
+      "learning_rate": 7.712798160905218e-06,
+      "logits/chosen": 0.028927450999617577,
+      "logits/rejected": 0.060569021850824356,
+      "logps/chosen": -85.33325958251953,
+      "logps/rejected": -87.87812042236328,
+      "loss": 2.2098,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.669008255004883,
+      "rewards/margins": 0.6498023271560669,
+      "rewards/rejected": -3.31881046295166,
+      "step": 740
+    },
+    {
+      "epoch": 1.3451407811080835,
+      "grad_norm": 2.375645160675049,
+      "learning_rate": 7.707200505459427e-06,
+      "logits/chosen": 0.04526009410619736,
+      "logits/rejected": 0.0799727588891983,
+      "logps/chosen": -89.62974548339844,
+      "logps/rejected": -99.64259338378906,
+      "loss": 2.2352,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.709292411804199,
+      "rewards/margins": 0.6063066720962524,
+      "rewards/rejected": -3.3155994415283203,
+      "step": 741
+    },
+    {
+      "epoch": 1.3469573115349682,
+      "grad_norm": 2.4972751140594482,
+      "learning_rate": 7.701598349147496e-06,
+      "logits/chosen": -0.020723650231957436,
+      "logits/rejected": 0.04764425382018089,
+      "logps/chosen": -85.17987060546875,
+      "logps/rejected": -98.80146789550781,
+      "loss": 2.0034,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.656331777572632,
+      "rewards/margins": 0.9565580487251282,
+      "rewards/rejected": -3.6128900051116943,
+      "step": 742
+    },
+    {
+      "epoch": 1.348773841961853,
+      "grad_norm": 2.23154616355896,
+      "learning_rate": 7.695991703393196e-06,
+      "logits/chosen": 0.062117062509059906,
+      "logits/rejected": 0.018935926258563995,
+      "logps/chosen": -99.47203063964844,
+      "logps/rejected": -104.25543212890625,
+      "loss": 2.3257,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.538417339324951,
+      "rewards/margins": 0.5829829573631287,
+      "rewards/rejected": -3.1214003562927246,
+      "step": 743
+    },
+    {
+      "epoch": 1.3505903723887376,
+      "grad_norm": 2.1097710132598877,
+      "learning_rate": 7.690380579629454e-06,
+      "logits/chosen": 0.07178185135126114,
+      "logits/rejected": 0.13060711324214935,
+      "logps/chosen": -82.14419555664062,
+      "logps/rejected": -94.78469848632812,
+      "loss": 1.9158,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.5844056606292725,
+      "rewards/margins": 0.95278000831604,
+      "rewards/rejected": -3.5371859073638916,
+      "step": 744
+    },
+    {
+      "epoch": 1.352406902815622,
+      "grad_norm": 2.6826353073120117,
+      "learning_rate": 7.684764989298329e-06,
+      "logits/chosen": 0.10977941751480103,
+      "logits/rejected": 0.11871352046728134,
+      "logps/chosen": -87.78453826904297,
+      "logps/rejected": -88.22404479980469,
+      "loss": 2.405,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.647420644760132,
+      "rewards/margins": 0.5518589019775391,
+      "rewards/rejected": -3.199279546737671,
+      "step": 745
+    },
+    {
+      "epoch": 1.354223433242507,
+      "grad_norm": 2.235769510269165,
+      "learning_rate": 7.679144943850986e-06,
+      "logits/chosen": 0.05031013488769531,
+      "logits/rejected": 0.01593434251844883,
+      "logps/chosen": -94.69499206542969,
+      "logps/rejected": -99.25582885742188,
+      "loss": 2.4038,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.653748035430908,
+      "rewards/margins": 0.7265360355377197,
+      "rewards/rejected": -3.380283832550049,
+      "step": 746
+    },
+    {
+      "epoch": 1.3560399636693914,
+      "grad_norm": 2.4417803287506104,
+      "learning_rate": 7.673520454747681e-06,
+      "logits/chosen": 0.04040508344769478,
+      "logits/rejected": 0.07281962782144547,
+      "logps/chosen": -77.71315002441406,
+      "logps/rejected": -94.82330322265625,
+      "loss": 1.6934,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4112367630004883,
+      "rewards/margins": 1.1407581567764282,
+      "rewards/rejected": -3.551994800567627,
+      "step": 747
+    },
+    {
+      "epoch": 1.357856494096276,
+      "grad_norm": 1.7434958219528198,
+      "learning_rate": 7.667891533457718e-06,
+      "logits/chosen": 0.1453830599784851,
+      "logits/rejected": 0.10198508203029633,
+      "logps/chosen": -79.32072448730469,
+      "logps/rejected": -88.78571319580078,
+      "loss": 2.0308,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6976542472839355,
+      "rewards/margins": 0.8518580198287964,
+      "rewards/rejected": -3.5495121479034424,
+      "step": 748
+    },
+    {
+      "epoch": 1.3596730245231607,
+      "grad_norm": 2.364915609359741,
+      "learning_rate": 7.662258191459456e-06,
+      "logits/chosen": 0.07202528417110443,
+      "logits/rejected": 0.07270742952823639,
+      "logps/chosen": -82.5189208984375,
+      "logps/rejected": -87.38330841064453,
+      "loss": 2.2639,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.583577871322632,
+      "rewards/margins": 0.5575169324874878,
+      "rewards/rejected": -3.141094923019409,
+      "step": 749
+    },
+    {
+      "epoch": 1.3614895549500454,
+      "grad_norm": 1.9622827768325806,
+      "learning_rate": 7.656620440240257e-06,
+      "logits/chosen": 0.0625351071357727,
+      "logits/rejected": 0.1307908296585083,
+      "logps/chosen": -84.89385223388672,
+      "logps/rejected": -92.00534057617188,
+      "loss": 2.0835,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.679396629333496,
+      "rewards/margins": 0.7283416986465454,
+      "rewards/rejected": -3.407738208770752,
+      "step": 750
+    },
+    {
+      "epoch": 1.36330608537693,
+      "grad_norm": 3.2049853801727295,
+      "learning_rate": 7.650978291296475e-06,
+      "logits/chosen": 0.06182331219315529,
+      "logits/rejected": 0.08293959498405457,
+      "logps/chosen": -77.00846099853516,
+      "logps/rejected": -85.75076293945312,
+      "loss": 2.1026,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4312846660614014,
+      "rewards/margins": 0.7030216455459595,
+      "rewards/rejected": -3.134305953979492,
+      "step": 751
+    },
+    {
+      "epoch": 1.3651226158038148,
+      "grad_norm": 2.065100908279419,
+      "learning_rate": 7.645331756133438e-06,
+      "logits/chosen": 0.12972712516784668,
+      "logits/rejected": 0.12206097692251205,
+      "logps/chosen": -88.62772369384766,
+      "logps/rejected": -98.35625457763672,
+      "loss": 1.8117,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.59391450881958,
+      "rewards/margins": 0.8645298480987549,
+      "rewards/rejected": -3.458444833755493,
+      "step": 752
+    },
+    {
+      "epoch": 1.3669391462306995,
+      "grad_norm": 2.078524112701416,
+      "learning_rate": 7.639680846265411e-06,
+      "logits/chosen": 0.09852910041809082,
+      "logits/rejected": 0.12210299074649811,
+      "logps/chosen": -81.41858673095703,
+      "logps/rejected": -92.6720199584961,
+      "loss": 1.9148,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4295332431793213,
+      "rewards/margins": 0.8047889471054077,
+      "rewards/rejected": -3.2343220710754395,
+      "step": 753
+    },
+    {
+      "epoch": 1.368755676657584,
+      "grad_norm": 1.991837739944458,
+      "learning_rate": 7.634025573215585e-06,
+      "logits/chosen": -0.02604215405881405,
+      "logits/rejected": 0.048631928861141205,
+      "logps/chosen": -95.89967346191406,
+      "logps/rejected": -109.60630798339844,
+      "loss": 1.9067,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.491032361984253,
+      "rewards/margins": 1.0193102359771729,
+      "rewards/rejected": -3.5103423595428467,
+      "step": 754
+    },
+    {
+      "epoch": 1.3705722070844686,
+      "grad_norm": 2.560757875442505,
+      "learning_rate": 7.628365948516047e-06,
+      "logits/chosen": 0.12844571471214294,
+      "logits/rejected": 0.11145608872175217,
+      "logps/chosen": -81.85939025878906,
+      "logps/rejected": -92.21177673339844,
+      "loss": 2.195,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6527202129364014,
+      "rewards/margins": 0.7286384105682373,
+      "rewards/rejected": -3.3813586235046387,
+      "step": 755
+    },
+    {
+      "epoch": 1.3723887375113533,
+      "grad_norm": 2.5189437866210938,
+      "learning_rate": 7.622701983707756e-06,
+      "logits/chosen": 0.049444377422332764,
+      "logits/rejected": 0.07628104835748672,
+      "logps/chosen": -80.98538970947266,
+      "logps/rejected": -89.72940063476562,
+      "loss": 2.1411,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7696704864501953,
+      "rewards/margins": 0.7333941459655762,
+      "rewards/rejected": -3.5030641555786133,
+      "step": 756
+    },
+    {
+      "epoch": 1.374205267938238,
+      "grad_norm": 2.5741689205169678,
+      "learning_rate": 7.617033690340523e-06,
+      "logits/chosen": 0.02790415659546852,
+      "logits/rejected": 0.0068018268793821335,
+      "logps/chosen": -83.76559448242188,
+      "logps/rejected": -91.61640167236328,
+      "loss": 2.203,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.734015941619873,
+      "rewards/margins": 0.6714599132537842,
+      "rewards/rejected": -3.405475616455078,
+      "step": 757
+    },
+    {
+      "epoch": 1.3760217983651226,
+      "grad_norm": 1.7192715406417847,
+      "learning_rate": 7.611361079972984e-06,
+      "logits/chosen": 0.1048116460442543,
+      "logits/rejected": 0.13254055380821228,
+      "logps/chosen": -67.2003402709961,
+      "logps/rejected": -91.35997009277344,
+      "loss": 1.5709,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.3325514793395996,
+      "rewards/margins": 1.2846351861953735,
+      "rewards/rejected": -3.617187023162842,
+      "step": 758
+    },
+    {
+      "epoch": 1.3778383287920073,
+      "grad_norm": 2.5107131004333496,
+      "learning_rate": 7.605684164172581e-06,
+      "logits/chosen": 0.04970414191484451,
+      "logits/rejected": 0.06727412343025208,
+      "logps/chosen": -78.22857666015625,
+      "logps/rejected": -91.56071472167969,
+      "loss": 2.2132,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.5088820457458496,
+      "rewards/margins": 0.8137573003768921,
+      "rewards/rejected": -3.322639226913452,
+      "step": 759
+    },
+    {
+      "epoch": 1.379654859218892,
+      "grad_norm": 2.327336311340332,
+      "learning_rate": 7.600002954515532e-06,
+      "logits/chosen": 0.09795573353767395,
+      "logits/rejected": 0.04695986211299896,
+      "logps/chosen": -85.50299835205078,
+      "logps/rejected": -92.48783111572266,
+      "loss": 1.9836,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.55703067779541,
+      "rewards/margins": 0.7251678109169006,
+      "rewards/rejected": -3.282198429107666,
+      "step": 760
+    },
+    {
+      "epoch": 1.3814713896457766,
+      "grad_norm": 2.2488017082214355,
+      "learning_rate": 7.5943174625868146e-06,
+      "logits/chosen": 0.00528271310031414,
+      "logits/rejected": 0.05780588090419769,
+      "logps/chosen": -90.95349884033203,
+      "logps/rejected": -98.3586654663086,
+      "loss": 2.4998,
+      "rewards/accuracies": 0.640625,
+      "rewards/chosen": -2.74594783782959,
+      "rewards/margins": 0.4672420024871826,
+      "rewards/rejected": -3.2131900787353516,
+      "step": 761
+    },
+    {
+      "epoch": 1.3832879200726613,
+      "grad_norm": 2.4573490619659424,
+      "learning_rate": 7.588627699980134e-06,
+      "logits/chosen": 0.023196734488010406,
+      "logits/rejected": 0.07495579123497009,
+      "logps/chosen": -77.51380157470703,
+      "logps/rejected": -95.95264434814453,
+      "loss": 1.8657,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.519115447998047,
+      "rewards/margins": 1.0751680135726929,
+      "rewards/rejected": -3.59428334236145,
+      "step": 762
+    },
+    {
+      "epoch": 1.3851044504995458,
+      "grad_norm": 2.718414783477783,
+      "learning_rate": 7.582933678297911e-06,
+      "logits/chosen": -0.006191038526594639,
+      "logits/rejected": 0.08378598839044571,
+      "logps/chosen": -85.86323547363281,
+      "logps/rejected": -98.59344482421875,
+      "loss": 2.3292,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.69970703125,
+      "rewards/margins": 0.7494994401931763,
+      "rewards/rejected": -3.4492063522338867,
+      "step": 763
+    },
+    {
+      "epoch": 1.3869209809264305,
+      "grad_norm": 3.296757936477661,
+      "learning_rate": 7.577235409151242e-06,
+      "logits/chosen": 0.03207925334572792,
+      "logits/rejected": 0.11028114706277847,
+      "logps/chosen": -77.22196960449219,
+      "logps/rejected": -87.56900024414062,
+      "loss": 2.5069,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.9215140342712402,
+      "rewards/margins": 0.6305817365646362,
+      "rewards/rejected": -3.552095890045166,
+      "step": 764
+    },
+    {
+      "epoch": 1.3887375113533151,
+      "grad_norm": 2.6371073722839355,
+      "learning_rate": 7.5715329041598955e-06,
+      "logits/chosen": 0.06983280181884766,
+      "logits/rejected": 0.0031841248273849487,
+      "logps/chosen": -93.26805877685547,
+      "logps/rejected": -100.43709564208984,
+      "loss": 2.1158,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.602299928665161,
+      "rewards/margins": 0.8267786502838135,
+      "rewards/rejected": -3.4290785789489746,
+      "step": 765
+    },
+    {
+      "epoch": 1.3905540417801998,
+      "grad_norm": 2.557405471801758,
+      "learning_rate": 7.565826174952267e-06,
+      "logits/chosen": 0.0866529792547226,
+      "logits/rejected": 0.11220981180667877,
+      "logps/chosen": -81.10444641113281,
+      "logps/rejected": -89.5517807006836,
+      "loss": 2.0158,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.469362258911133,
+      "rewards/margins": 0.6984151601791382,
+      "rewards/rejected": -3.1677775382995605,
+      "step": 766
+    },
+    {
+      "epoch": 1.3923705722070845,
+      "grad_norm": 2.013252019882202,
+      "learning_rate": 7.560115233165375e-06,
+      "logits/chosen": 0.10826343297958374,
+      "logits/rejected": 0.09555835276842117,
+      "logps/chosen": -90.81183624267578,
+      "logps/rejected": -99.46356201171875,
+      "loss": 1.9386,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6625754833221436,
+      "rewards/margins": 0.9187412858009338,
+      "rewards/rejected": -3.5813164710998535,
+      "step": 767
+    },
+    {
+      "epoch": 1.3941871026339692,
+      "grad_norm": 2.1441266536712646,
+      "learning_rate": 7.554400090444822e-06,
+      "logits/chosen": 0.04958781227469444,
+      "logits/rejected": 0.10418076813220978,
+      "logps/chosen": -73.46155548095703,
+      "logps/rejected": -83.24677276611328,
+      "loss": 1.9933,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.4550204277038574,
+      "rewards/margins": 0.7676887512207031,
+      "rewards/rejected": -3.2227089405059814,
+      "step": 768
+    },
+    {
+      "epoch": 1.3960036330608538,
+      "grad_norm": 2.401198387145996,
+      "learning_rate": 7.54868075844478e-06,
+      "logits/chosen": 0.10376621782779694,
+      "logits/rejected": 0.1655123233795166,
+      "logps/chosen": -80.3701171875,
+      "logps/rejected": -95.8973159790039,
+      "loss": 1.784,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2211532592773438,
+      "rewards/margins": 1.070261836051941,
+      "rewards/rejected": -3.291415214538574,
+      "step": 769
+    },
+    {
+      "epoch": 1.3978201634877383,
+      "grad_norm": 1.93374502658844,
+      "learning_rate": 7.5429572488279615e-06,
+      "logits/chosen": 0.081766776740551,
+      "logits/rejected": 0.1179564967751503,
+      "logps/chosen": -73.5924301147461,
+      "logps/rejected": -85.84019470214844,
+      "loss": 2.2637,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.3363659381866455,
+      "rewards/margins": 0.6945525407791138,
+      "rewards/rejected": -3.030918598175049,
+      "step": 770
+    },
+    {
+      "epoch": 1.3996366939146232,
+      "grad_norm": 2.4809932708740234,
+      "learning_rate": 7.5372295732656e-06,
+      "logits/chosen": 0.11731807887554169,
+      "logits/rejected": 0.05732544884085655,
+      "logps/chosen": -89.25398254394531,
+      "logps/rejected": -90.94881439208984,
+      "loss": 2.412,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.435483694076538,
+      "rewards/margins": 0.5316743850708008,
+      "rewards/rejected": -2.967158317565918,
+      "step": 771
+    },
+    {
+      "epoch": 1.4014532243415077,
+      "grad_norm": 2.622253894805908,
+      "learning_rate": 7.531497743437424e-06,
+      "logits/chosen": -0.02307087928056717,
+      "logits/rejected": 0.034644536674022675,
+      "logps/chosen": -79.53402709960938,
+      "logps/rejected": -95.38764953613281,
+      "loss": 2.1841,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5010921955108643,
+      "rewards/margins": 0.7504414916038513,
+      "rewards/rejected": -3.2515335083007812,
+      "step": 772
+    },
+    {
+      "epoch": 1.4032697547683923,
+      "grad_norm": 2.341421365737915,
+      "learning_rate": 7.525761771031632e-06,
+      "logits/chosen": 0.09784074872732162,
+      "logits/rejected": 0.1302230954170227,
+      "logps/chosen": -78.1878433227539,
+      "logps/rejected": -88.39629364013672,
+      "loss": 2.064,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.63604736328125,
+      "rewards/margins": 0.6544671654701233,
+      "rewards/rejected": -3.2905144691467285,
+      "step": 773
+    },
+    {
+      "epoch": 1.405086285195277,
+      "grad_norm": 2.4410409927368164,
+      "learning_rate": 7.520021667744869e-06,
+      "logits/chosen": 0.03553074598312378,
+      "logits/rejected": 0.024142932146787643,
+      "logps/chosen": -92.7217025756836,
+      "logps/rejected": -99.62353515625,
+      "loss": 2.077,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5202903747558594,
+      "rewards/margins": 0.9078298211097717,
+      "rewards/rejected": -3.4281198978424072,
+      "step": 774
+    },
+    {
+      "epoch": 1.4069028156221617,
+      "grad_norm": 2.1428725719451904,
+      "learning_rate": 7.514277445282206e-06,
+      "logits/chosen": 0.12562477588653564,
+      "logits/rejected": 0.11905661970376968,
+      "logps/chosen": -85.4616470336914,
+      "logps/rejected": -101.02755737304688,
+      "loss": 1.8298,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.236478328704834,
+      "rewards/margins": 0.9414302110671997,
+      "rewards/rejected": -3.177908420562744,
+      "step": 775
+    },
+    {
+      "epoch": 1.4087193460490464,
+      "grad_norm": 2.2605643272399902,
+      "learning_rate": 7.508529115357115e-06,
+      "logits/chosen": 0.08080364018678665,
+      "logits/rejected": 0.14329738914966583,
+      "logps/chosen": -86.23857116699219,
+      "logps/rejected": -94.77734375,
+      "loss": 2.1025,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3885529041290283,
+      "rewards/margins": 0.8040337562561035,
+      "rewards/rejected": -3.1925864219665527,
+      "step": 776
+    },
+    {
+      "epoch": 1.410535876475931,
+      "grad_norm": 2.1767449378967285,
+      "learning_rate": 7.502776689691439e-06,
+      "logits/chosen": 0.061047837138175964,
+      "logits/rejected": 0.049486950039863586,
+      "logps/chosen": -79.15450286865234,
+      "logps/rejected": -89.98368072509766,
+      "loss": 2.032,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.129742383956909,
+      "rewards/margins": 0.7515033483505249,
+      "rewards/rejected": -2.8812458515167236,
+      "step": 777
+    },
+    {
+      "epoch": 1.4123524069028157,
+      "grad_norm": 2.2512662410736084,
+      "learning_rate": 7.4970201800153765e-06,
+      "logits/chosen": 0.08744536340236664,
+      "logits/rejected": 0.10895106941461563,
+      "logps/chosen": -68.41326141357422,
+      "logps/rejected": -82.72618103027344,
+      "loss": 2.0954,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.360013008117676,
+      "rewards/margins": 0.856154203414917,
+      "rewards/rejected": -3.2161672115325928,
+      "step": 778
+    },
+    {
+      "epoch": 1.4141689373297002,
+      "grad_norm": 2.1528515815734863,
+      "learning_rate": 7.491259598067452e-06,
+      "logits/chosen": 0.10324890911579132,
+      "logits/rejected": 0.017273597419261932,
+      "logps/chosen": -101.93807983398438,
+      "logps/rejected": -110.01669311523438,
+      "loss": 1.8012,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2577784061431885,
+      "rewards/margins": 0.8829382658004761,
+      "rewards/rejected": -3.140716552734375,
+      "step": 779
+    },
+    {
+      "epoch": 1.4159854677565848,
+      "grad_norm": 2.24306321144104,
+      "learning_rate": 7.485494955594495e-06,
+      "logits/chosen": 0.0638672262430191,
+      "logits/rejected": 0.08847285807132721,
+      "logps/chosen": -83.1998291015625,
+      "logps/rejected": -87.86129760742188,
+      "loss": 2.215,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.975947380065918,
+      "rewards/margins": 0.5703001022338867,
+      "rewards/rejected": -2.546247720718384,
+      "step": 780
+    },
+    {
+      "epoch": 1.4178019981834695,
+      "grad_norm": 2.103116035461426,
+      "learning_rate": 7.479726264351619e-06,
+      "logits/chosen": 0.09146749973297119,
+      "logits/rejected": 0.12321461737155914,
+      "logps/chosen": -81.99695587158203,
+      "logps/rejected": -89.16120910644531,
+      "loss": 1.9022,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.159066677093506,
+      "rewards/margins": 0.8253484964370728,
+      "rewards/rejected": -2.984415054321289,
+      "step": 781
+    },
+    {
+      "epoch": 1.4196185286103542,
+      "grad_norm": 2.1531925201416016,
+      "learning_rate": 7.473953536102185e-06,
+      "logits/chosen": 0.1091499999165535,
+      "logits/rejected": 0.13049989938735962,
+      "logps/chosen": -81.74616241455078,
+      "logps/rejected": -94.27207946777344,
+      "loss": 1.9222,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.445281982421875,
+      "rewards/margins": 0.8645297288894653,
+      "rewards/rejected": -3.30981183052063,
+      "step": 782
+    },
+    {
+      "epoch": 1.4214350590372389,
+      "grad_norm": 2.489908218383789,
+      "learning_rate": 7.468176782617793e-06,
+      "logits/chosen": -0.04253482073545456,
+      "logits/rejected": 0.10589614510536194,
+      "logps/chosen": -83.19894409179688,
+      "logps/rejected": -99.28282928466797,
+      "loss": 1.7606,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2363641262054443,
+      "rewards/margins": 1.0474493503570557,
+      "rewards/rejected": -3.2838134765625,
+      "step": 783
+    },
+    {
+      "epoch": 1.4232515894641236,
+      "grad_norm": 2.4771764278411865,
+      "learning_rate": 7.462396015678249e-06,
+      "logits/chosen": 0.02688070759177208,
+      "logits/rejected": 0.05978942662477493,
+      "logps/chosen": -75.48037719726562,
+      "logps/rejected": -87.0803451538086,
+      "loss": 1.892,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3419995307922363,
+      "rewards/margins": 0.8640164732933044,
+      "rewards/rejected": -3.2060160636901855,
+      "step": 784
+    },
+    {
+      "epoch": 1.4250681198910082,
+      "grad_norm": 2.3507192134857178,
+      "learning_rate": 7.456611247071544e-06,
+      "logits/chosen": 0.13474830985069275,
+      "logits/rejected": 0.08549933135509491,
+      "logps/chosen": -78.67109680175781,
+      "logps/rejected": -80.96075439453125,
+      "loss": 2.1411,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.477113723754883,
+      "rewards/margins": 0.652525007724762,
+      "rewards/rejected": -3.129638671875,
+      "step": 785
+    },
+    {
+      "epoch": 1.4268846503178927,
+      "grad_norm": 2.009307861328125,
+      "learning_rate": 7.450822488593829e-06,
+      "logits/chosen": 0.04329455643892288,
+      "logits/rejected": 0.05247782543301582,
+      "logps/chosen": -89.56121063232422,
+      "logps/rejected": -99.5736083984375,
+      "loss": 1.8014,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.2546510696411133,
+      "rewards/margins": 0.9092133045196533,
+      "rewards/rejected": -3.1638646125793457,
+      "step": 786
+    },
+    {
+      "epoch": 1.4287011807447776,
+      "grad_norm": 2.3687586784362793,
+      "learning_rate": 7.445029752049387e-06,
+      "logits/chosen": 0.04104599729180336,
+      "logits/rejected": 0.08001314103603363,
+      "logps/chosen": -77.9354248046875,
+      "logps/rejected": -92.67184448242188,
+      "loss": 1.7508,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.2342448234558105,
+      "rewards/margins": 0.9252514839172363,
+      "rewards/rejected": -3.159496307373047,
+      "step": 787
+    },
+    {
+      "epoch": 1.430517711171662,
+      "grad_norm": 1.8127901554107666,
+      "learning_rate": 7.439233049250622e-06,
+      "logits/chosen": 0.07227849960327148,
+      "logits/rejected": 0.0907898098230362,
+      "logps/chosen": -71.49815368652344,
+      "logps/rejected": -82.32070922851562,
+      "loss": 1.8207,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.33886456489563,
+      "rewards/margins": 0.906674325466156,
+      "rewards/rejected": -3.2455389499664307,
+      "step": 788
+    },
+    {
+      "epoch": 1.4323342415985467,
+      "grad_norm": 2.0310990810394287,
+      "learning_rate": 7.433432392018013e-06,
+      "logits/chosen": 0.07254546135663986,
+      "logits/rejected": 0.08750150352716446,
+      "logps/chosen": -87.65977478027344,
+      "logps/rejected": -105.3295669555664,
+      "loss": 1.781,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.4844467639923096,
+      "rewards/margins": 0.8295308351516724,
+      "rewards/rejected": -3.3139777183532715,
+      "step": 789
+    },
+    {
+      "epoch": 1.4341507720254314,
+      "grad_norm": 2.093871831893921,
+      "learning_rate": 7.427627792180116e-06,
+      "logits/chosen": 0.09880155324935913,
+      "logits/rejected": 0.15118218958377838,
+      "logps/chosen": -75.81531524658203,
+      "logps/rejected": -88.76158142089844,
+      "loss": 2.2499,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6412038803100586,
+      "rewards/margins": 0.7693830132484436,
+      "rewards/rejected": -3.4105868339538574,
+      "step": 790
+    },
+    {
+      "epoch": 1.435967302452316,
+      "grad_norm": 2.393105983734131,
+      "learning_rate": 7.4218192615735185e-06,
+      "logits/chosen": 0.08802710473537445,
+      "logits/rejected": 0.04610452428460121,
+      "logps/chosen": -99.03189849853516,
+      "logps/rejected": -103.37187194824219,
+      "loss": 1.8155,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3743648529052734,
+      "rewards/margins": 0.8601830005645752,
+      "rewards/rejected": -3.2345480918884277,
+      "step": 791
+    },
+    {
+      "epoch": 1.4377838328792008,
+      "grad_norm": 2.148087978363037,
+      "learning_rate": 7.416006812042827e-06,
+      "logits/chosen": 0.05752583593130112,
+      "logits/rejected": 0.034685466438531876,
+      "logps/chosen": -89.9745101928711,
+      "logps/rejected": -97.91671752929688,
+      "loss": 1.9892,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.483428716659546,
+      "rewards/margins": 0.7583959102630615,
+      "rewards/rejected": -3.2418243885040283,
+      "step": 792
+    },
+    {
+      "epoch": 1.4396003633060854,
+      "grad_norm": 2.605093479156494,
+      "learning_rate": 7.410190455440638e-06,
+      "logits/chosen": 0.07939667254686356,
+      "logits/rejected": 0.05258120596408844,
+      "logps/chosen": -81.38958740234375,
+      "logps/rejected": -92.0197982788086,
+      "loss": 1.8609,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4127330780029297,
+      "rewards/margins": 0.8712365627288818,
+      "rewards/rejected": -3.2839698791503906,
+      "step": 793
+    },
+    {
+      "epoch": 1.44141689373297,
+      "grad_norm": 2.3762154579162598,
+      "learning_rate": 7.4043702036275136e-06,
+      "logits/chosen": 0.04353059083223343,
+      "logits/rejected": 0.08619395643472672,
+      "logps/chosen": -83.58119201660156,
+      "logps/rejected": -94.14823150634766,
+      "loss": 2.4321,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": -2.5087859630584717,
+      "rewards/margins": 0.4951530396938324,
+      "rewards/rejected": -3.003939151763916,
+      "step": 794
+    },
+    {
+      "epoch": 1.4432334241598546,
+      "grad_norm": 2.350267171859741,
+      "learning_rate": 7.398546068471961e-06,
+      "logits/chosen": 0.18240118026733398,
+      "logits/rejected": 0.14964665472507477,
+      "logps/chosen": -79.20641326904297,
+      "logps/rejected": -86.13677978515625,
+      "loss": 2.1999,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4007363319396973,
+      "rewards/margins": 0.6775516867637634,
+      "rewards/rejected": -3.0782880783081055,
+      "step": 795
+    },
+    {
+      "epoch": 1.4450499545867395,
+      "grad_norm": 2.183232545852661,
+      "learning_rate": 7.392718061850407e-06,
+      "logits/chosen": 0.06714559346437454,
+      "logits/rejected": 0.046004436910152435,
+      "logps/chosen": -95.4114990234375,
+      "logps/rejected": -101.2634048461914,
+      "loss": 1.9802,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4638829231262207,
+      "rewards/margins": 0.7256942987442017,
+      "rewards/rejected": -3.189577579498291,
+      "step": 796
+    },
+    {
+      "epoch": 1.446866485013624,
+      "grad_norm": 2.2636830806732178,
+      "learning_rate": 7.3868861956471725e-06,
+      "logits/chosen": 0.025870440527796745,
+      "logits/rejected": 0.05159129202365875,
+      "logps/chosen": -80.84700775146484,
+      "logps/rejected": -88.84320831298828,
+      "loss": 1.7361,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.451986312866211,
+      "rewards/margins": 0.9137101173400879,
+      "rewards/rejected": -3.3656961917877197,
+      "step": 797
+    },
+    {
+      "epoch": 1.4486830154405086,
+      "grad_norm": 2.3395354747772217,
+      "learning_rate": 7.381050481754447e-06,
+      "logits/chosen": 0.06643021106719971,
+      "logits/rejected": 0.1107374057173729,
+      "logps/chosen": -88.24702453613281,
+      "logps/rejected": -97.00286865234375,
+      "loss": 2.3738,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.6666600704193115,
+      "rewards/margins": 0.5626463294029236,
+      "rewards/rejected": -3.22930645942688,
+      "step": 798
+    },
+    {
+      "epoch": 1.4504995458673933,
+      "grad_norm": 2.936711549758911,
+      "learning_rate": 7.375210932072265e-06,
+      "logits/chosen": 0.04520021751523018,
+      "logits/rejected": 0.06952523440122604,
+      "logps/chosen": -82.35380554199219,
+      "logps/rejected": -89.27215576171875,
+      "loss": 2.4222,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5209689140319824,
+      "rewards/margins": 0.5571924448013306,
+      "rewards/rejected": -3.0781612396240234,
+      "step": 799
+    },
+    {
+      "epoch": 1.452316076294278,
+      "grad_norm": 2.342952013015747,
+      "learning_rate": 7.36936755850849e-06,
+      "logits/chosen": 0.016199974343180656,
+      "logits/rejected": 0.07721687853336334,
+      "logps/chosen": -83.52679443359375,
+      "logps/rejected": -95.04873657226562,
+      "loss": 1.8844,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.487191677093506,
+      "rewards/margins": 0.9364355206489563,
+      "rewards/rejected": -3.4236273765563965,
+      "step": 800
+    },
+    {
+      "epoch": 1.4541326067211626,
+      "grad_norm": 2.611804246902466,
+      "learning_rate": 7.363520372978774e-06,
+      "logits/chosen": 0.0884896069765091,
+      "logits/rejected": 0.12684395909309387,
+      "logps/chosen": -73.71459197998047,
+      "logps/rejected": -78.7441635131836,
+      "loss": 2.5897,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -2.3413586616516113,
+      "rewards/margins": 0.3767651617527008,
+      "rewards/rejected": -2.7181236743927,
+      "step": 801
+    },
+    {
+      "epoch": 1.4559491371480473,
+      "grad_norm": 2.2446863651275635,
+      "learning_rate": 7.357669387406548e-06,
+      "logits/chosen": 0.04801754653453827,
+      "logits/rejected": 0.14114932715892792,
+      "logps/chosen": -72.67373657226562,
+      "logps/rejected": -88.53012084960938,
+      "loss": 1.8134,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4397833347320557,
+      "rewards/margins": 0.9794785976409912,
+      "rewards/rejected": -3.419261932373047,
+      "step": 802
+    },
+    {
+      "epoch": 1.457765667574932,
+      "grad_norm": 2.2070155143737793,
+      "learning_rate": 7.351814613722991e-06,
+      "logits/chosen": 0.09963800758123398,
+      "logits/rejected": 0.09044703096151352,
+      "logps/chosen": -93.37407684326172,
+      "logps/rejected": -101.1263656616211,
+      "loss": 2.0345,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4976069927215576,
+      "rewards/margins": 0.7321794033050537,
+      "rewards/rejected": -3.2297863960266113,
+      "step": 803
+    },
+    {
+      "epoch": 1.4595821980018164,
+      "grad_norm": 2.339240789413452,
+      "learning_rate": 7.3459560638670035e-06,
+      "logits/chosen": 0.06331755220890045,
+      "logits/rejected": 0.09320802986621857,
+      "logps/chosen": -84.83074951171875,
+      "logps/rejected": -90.84583282470703,
+      "loss": 1.7825,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.2360713481903076,
+      "rewards/margins": 0.8683581352233887,
+      "rewards/rejected": -3.104429244995117,
+      "step": 804
+    },
+    {
+      "epoch": 1.461398728428701,
+      "grad_norm": 2.2252860069274902,
+      "learning_rate": 7.3400937497851906e-06,
+      "logits/chosen": 0.06259419769048691,
+      "logits/rejected": 0.05668449029326439,
+      "logps/chosen": -76.97444915771484,
+      "logps/rejected": -95.10333251953125,
+      "loss": 1.9808,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.475806713104248,
+      "rewards/margins": 0.9982489943504333,
+      "rewards/rejected": -3.4740560054779053,
+      "step": 805
+    },
+    {
+      "epoch": 1.4632152588555858,
+      "grad_norm": 2.056358814239502,
+      "learning_rate": 7.334227683431832e-06,
+      "logits/chosen": 0.0795949399471283,
+      "logits/rejected": 0.11871220171451569,
+      "logps/chosen": -82.24653625488281,
+      "logps/rejected": -104.25836944580078,
+      "loss": 1.8753,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.397820234298706,
+      "rewards/margins": 1.0821467638015747,
+      "rewards/rejected": -3.479966878890991,
+      "step": 806
+    },
+    {
+      "epoch": 1.4650317892824705,
+      "grad_norm": 2.7462871074676514,
+      "learning_rate": 7.328357876768858e-06,
+      "logits/chosen": 0.08270560950040817,
+      "logits/rejected": 0.11420280486345291,
+      "logps/chosen": -76.36333465576172,
+      "logps/rejected": -86.0474853515625,
+      "loss": 2.3324,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.573794364929199,
+      "rewards/margins": 0.7034318447113037,
+      "rewards/rejected": -3.277225971221924,
+      "step": 807
+    },
+    {
+      "epoch": 1.4668483197093551,
+      "grad_norm": 2.5223159790039062,
+      "learning_rate": 7.322484341765824e-06,
+      "logits/chosen": 0.17336739599704742,
+      "logits/rejected": 0.15756890177726746,
+      "logps/chosen": -82.3572998046875,
+      "logps/rejected": -94.081787109375,
+      "loss": 2.0898,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3327016830444336,
+      "rewards/margins": 0.8286820650100708,
+      "rewards/rejected": -3.161383867263794,
+      "step": 808
+    },
+    {
+      "epoch": 1.4686648501362398,
+      "grad_norm": 2.4014031887054443,
+      "learning_rate": 7.316607090399894e-06,
+      "logits/chosen": 0.0774412751197815,
+      "logits/rejected": 0.016062507405877113,
+      "logps/chosen": -79.8453598022461,
+      "logps/rejected": -80.36003875732422,
+      "loss": 2.0743,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6673779487609863,
+      "rewards/margins": 0.6888259649276733,
+      "rewards/rejected": -3.356204032897949,
+      "step": 809
+    },
+    {
+      "epoch": 1.4704813805631245,
+      "grad_norm": 2.3294990062713623,
+      "learning_rate": 7.310726134655807e-06,
+      "logits/chosen": 0.03737390413880348,
+      "logits/rejected": 0.1001349613070488,
+      "logps/chosen": -81.04705810546875,
+      "logps/rejected": -96.8980712890625,
+      "loss": 1.7474,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.27017879486084,
+      "rewards/margins": 1.0432392358779907,
+      "rewards/rejected": -3.313417673110962,
+      "step": 810
+    },
+    {
+      "epoch": 1.472297910990009,
+      "grad_norm": 2.287733793258667,
+      "learning_rate": 7.30484148652585e-06,
+      "logits/chosen": 0.0469367690384388,
+      "logits/rejected": 0.06735274195671082,
+      "logps/chosen": -80.7060546875,
+      "logps/rejected": -93.6329345703125,
+      "loss": 2.0516,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.463186740875244,
+      "rewards/margins": 0.7543611526489258,
+      "rewards/rejected": -3.21754789352417,
+      "step": 811
+    },
+    {
+      "epoch": 1.4741144414168939,
+      "grad_norm": 3.1462066173553467,
+      "learning_rate": 7.298953158009852e-06,
+      "logits/chosen": 0.09930360317230225,
+      "logits/rejected": 0.13193194568157196,
+      "logps/chosen": -82.47021484375,
+      "logps/rejected": -94.59288024902344,
+      "loss": 1.9513,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.323047161102295,
+      "rewards/margins": 0.8602778911590576,
+      "rewards/rejected": -3.1833250522613525,
+      "step": 812
+    },
+    {
+      "epoch": 1.4759309718437783,
+      "grad_norm": 2.2134501934051514,
+      "learning_rate": 7.293061161115138e-06,
+      "logits/chosen": 0.04882989823818207,
+      "logits/rejected": 0.07393565028905869,
+      "logps/chosen": -69.4496841430664,
+      "logps/rejected": -79.9421157836914,
+      "loss": 1.9357,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.233206272125244,
+      "rewards/margins": 0.7889159321784973,
+      "rewards/rejected": -3.0221221446990967,
+      "step": 813
+    },
+    {
+      "epoch": 1.477747502270663,
+      "grad_norm": 2.3149659633636475,
+      "learning_rate": 7.287165507856513e-06,
+      "logits/chosen": 0.05155353993177414,
+      "logits/rejected": 0.036280907690525055,
+      "logps/chosen": -87.41694641113281,
+      "logps/rejected": -97.0438461303711,
+      "loss": 1.9108,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.27425217628479,
+      "rewards/margins": 0.8441011309623718,
+      "rewards/rejected": -3.1183536052703857,
+      "step": 814
+    },
+    {
+      "epoch": 1.4795640326975477,
+      "grad_norm": 2.4460105895996094,
+      "learning_rate": 7.281266210256244e-06,
+      "logits/chosen": 0.06415215134620667,
+      "logits/rejected": 0.08172871172428131,
+      "logps/chosen": -94.13727569580078,
+      "logps/rejected": -108.59651184082031,
+      "loss": 1.9764,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.476318120956421,
+      "rewards/margins": 0.8187392354011536,
+      "rewards/rejected": -3.2950570583343506,
+      "step": 815
+    },
+    {
+      "epoch": 1.4813805631244323,
+      "grad_norm": 2.5813546180725098,
+      "learning_rate": 7.275363280344023e-06,
+      "logits/chosen": 0.10900135338306427,
+      "logits/rejected": 0.06012497842311859,
+      "logps/chosen": -86.26004028320312,
+      "logps/rejected": -92.2495346069336,
+      "loss": 2.1791,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4611434936523438,
+      "rewards/margins": 0.6075427532196045,
+      "rewards/rejected": -3.0686862468719482,
+      "step": 816
+    },
+    {
+      "epoch": 1.483197093551317,
+      "grad_norm": 2.3455967903137207,
+      "learning_rate": 7.269456730156954e-06,
+      "logits/chosen": 0.08659522235393524,
+      "logits/rejected": 0.1016351729631424,
+      "logps/chosen": -75.75826263427734,
+      "logps/rejected": -85.91856384277344,
+      "loss": 2.0077,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4886691570281982,
+      "rewards/margins": 0.9004117846488953,
+      "rewards/rejected": -3.3890810012817383,
+      "step": 817
+    },
+    {
+      "epoch": 1.4850136239782017,
+      "grad_norm": 2.612858533859253,
+      "learning_rate": 7.26354657173952e-06,
+      "logits/chosen": 0.08592739701271057,
+      "logits/rejected": 0.15154355764389038,
+      "logps/chosen": -86.36177062988281,
+      "logps/rejected": -95.97001647949219,
+      "loss": 2.1468,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6705503463745117,
+      "rewards/margins": 0.7056388854980469,
+      "rewards/rejected": -3.376189708709717,
+      "step": 818
+    },
+    {
+      "epoch": 1.4868301544050864,
+      "grad_norm": 2.39699387550354,
+      "learning_rate": 7.257632817143566e-06,
+      "logits/chosen": 0.03663618862628937,
+      "logits/rejected": 0.06003079563379288,
+      "logps/chosen": -73.82808685302734,
+      "logps/rejected": -86.11414337158203,
+      "loss": 2.0372,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.251981258392334,
+      "rewards/margins": 0.8644815683364868,
+      "rewards/rejected": -3.1164627075195312,
+      "step": 819
+    },
+    {
+      "epoch": 1.4886466848319708,
+      "grad_norm": 2.1887619495391846,
+      "learning_rate": 7.251715478428263e-06,
+      "logits/chosen": -0.026794860139489174,
+      "logits/rejected": 0.11405564099550247,
+      "logps/chosen": -96.14276123046875,
+      "logps/rejected": -116.06658172607422,
+      "loss": 1.786,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.527817964553833,
+      "rewards/margins": 1.0386499166488647,
+      "rewards/rejected": -3.5664680004119873,
+      "step": 820
+    },
+    {
+      "epoch": 1.4904632152588557,
+      "grad_norm": 4.235897064208984,
+      "learning_rate": 7.2457945676600975e-06,
+      "logits/chosen": 0.10090241581201553,
+      "logits/rejected": 0.10542615503072739,
+      "logps/chosen": -78.00686645507812,
+      "logps/rejected": -86.51849365234375,
+      "loss": 2.1653,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.6344261169433594,
+      "rewards/margins": 0.6512963771820068,
+      "rewards/rejected": -3.285722494125366,
+      "step": 821
+    },
+    {
+      "epoch": 1.4922797456857402,
+      "grad_norm": 2.189833402633667,
+      "learning_rate": 7.239870096912835e-06,
+      "logits/chosen": 0.0326794758439064,
+      "logits/rejected": 0.0896943062543869,
+      "logps/chosen": -89.783447265625,
+      "logps/rejected": -102.53228759765625,
+      "loss": 2.0422,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5607566833496094,
+      "rewards/margins": 1.009847640991211,
+      "rewards/rejected": -3.5706043243408203,
+      "step": 822
+    },
+    {
+      "epoch": 1.4940962761126249,
+      "grad_norm": 2.7656984329223633,
+      "learning_rate": 7.233942078267507e-06,
+      "logits/chosen": 0.03892774134874344,
+      "logits/rejected": 0.10934557020664215,
+      "logps/chosen": -87.70173645019531,
+      "logps/rejected": -98.54985046386719,
+      "loss": 1.9344,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.796427011489868,
+      "rewards/margins": 0.7776911854743958,
+      "rewards/rejected": -3.574118137359619,
+      "step": 823
+    },
+    {
+      "epoch": 1.4959128065395095,
+      "grad_norm": 2.1388044357299805,
+      "learning_rate": 7.228010523812372e-06,
+      "logits/chosen": 0.04929535835981369,
+      "logits/rejected": 0.05947191268205643,
+      "logps/chosen": -100.03392791748047,
+      "logps/rejected": -106.81358337402344,
+      "loss": 2.2769,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.775858163833618,
+      "rewards/margins": 0.8662734031677246,
+      "rewards/rejected": -3.6421315670013428,
+      "step": 824
+    },
+    {
+      "epoch": 1.4977293369663942,
+      "grad_norm": 2.8936967849731445,
+      "learning_rate": 7.222075445642904e-06,
+      "logits/chosen": 0.05339030176401138,
+      "logits/rejected": 0.09673048555850983,
+      "logps/chosen": -91.19949340820312,
+      "logps/rejected": -98.978515625,
+      "loss": 2.0686,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.4693431854248047,
+      "rewards/margins": 0.6974996328353882,
+      "rewards/rejected": -3.1668429374694824,
+      "step": 825
+    },
+    {
+      "epoch": 1.4995458673932789,
+      "grad_norm": 2.069953203201294,
+      "learning_rate": 7.216136855861763e-06,
+      "logits/chosen": 0.07213248312473297,
+      "logits/rejected": 0.14326205849647522,
+      "logps/chosen": -82.97685241699219,
+      "logps/rejected": -95.03224182128906,
+      "loss": 1.4687,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": -2.4275951385498047,
+      "rewards/margins": 1.1413557529449463,
+      "rewards/rejected": -3.568950653076172,
+      "step": 826
+    },
+    {
+      "epoch": 1.5013623978201633,
+      "grad_norm": 2.1628522872924805,
+      "learning_rate": 7.2101947665787605e-06,
+      "logits/chosen": 0.03247791528701782,
+      "logits/rejected": 0.13113634288311005,
+      "logps/chosen": -83.87806701660156,
+      "logps/rejected": -96.8788833618164,
+      "loss": 1.9276,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.425245761871338,
+      "rewards/margins": 0.938535213470459,
+      "rewards/rejected": -3.363780975341797,
+      "step": 827
+    },
+    {
+      "epoch": 1.5031789282470482,
+      "grad_norm": 2.1415798664093018,
+      "learning_rate": 7.20424918991086e-06,
+      "logits/chosen": 0.03731315955519676,
+      "logits/rejected": 0.038553059101104736,
+      "logps/chosen": -83.6593017578125,
+      "logps/rejected": -93.7667007446289,
+      "loss": 1.8234,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.623645305633545,
+      "rewards/margins": 1.0124547481536865,
+      "rewards/rejected": -3.6361002922058105,
+      "step": 828
+    },
+    {
+      "epoch": 1.5049954586739327,
+      "grad_norm": 2.155017375946045,
+      "learning_rate": 7.198300137982121e-06,
+      "logits/chosen": 0.05089723318815231,
+      "logits/rejected": 0.0881565734744072,
+      "logps/chosen": -87.8398666381836,
+      "logps/rejected": -99.52027130126953,
+      "loss": 1.8965,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.6992077827453613,
+      "rewards/margins": 0.9189518690109253,
+      "rewards/rejected": -3.618159770965576,
+      "step": 829
+    },
+    {
+      "epoch": 1.5068119891008176,
+      "grad_norm": 2.2586467266082764,
+      "learning_rate": 7.1923476229237e-06,
+      "logits/chosen": 0.032588109374046326,
+      "logits/rejected": 0.041385456919670105,
+      "logps/chosen": -88.36325073242188,
+      "logps/rejected": -97.6312026977539,
+      "loss": 1.8165,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.743462562561035,
+      "rewards/margins": 0.9551791548728943,
+      "rewards/rejected": -3.698641300201416,
+      "step": 830
+    },
+    {
+      "epoch": 1.508628519527702,
+      "grad_norm": 2.1857669353485107,
+      "learning_rate": 7.186391656873809e-06,
+      "logits/chosen": 0.09781689196825027,
+      "logits/rejected": 0.15005499124526978,
+      "logps/chosen": -85.13274383544922,
+      "logps/rejected": -98.7428207397461,
+      "loss": 1.8323,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.4565935134887695,
+      "rewards/margins": 0.9222261309623718,
+      "rewards/rejected": -3.378819704055786,
+      "step": 831
+    },
+    {
+      "epoch": 1.5104450499545867,
+      "grad_norm": 2.0807337760925293,
+      "learning_rate": 7.180432251977702e-06,
+      "logits/chosen": 0.0358676053583622,
+      "logits/rejected": 0.07221710681915283,
+      "logps/chosen": -82.19658660888672,
+      "logps/rejected": -93.883056640625,
+      "loss": 2.0109,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4384050369262695,
+      "rewards/margins": 0.8013818860054016,
+      "rewards/rejected": -3.2397866249084473,
+      "step": 832
+    },
+    {
+      "epoch": 1.5122615803814714,
+      "grad_norm": 2.5140273571014404,
+      "learning_rate": 7.1744694203876405e-06,
+      "logits/chosen": 0.057662852108478546,
+      "logits/rejected": 0.07171612232923508,
+      "logps/chosen": -83.77603912353516,
+      "logps/rejected": -94.08065795898438,
+      "loss": 1.8424,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.402648687362671,
+      "rewards/margins": 0.952576756477356,
+      "rewards/rejected": -3.3552253246307373,
+      "step": 833
+    },
+    {
+      "epoch": 1.514078110808356,
+      "grad_norm": 2.1008050441741943,
+      "learning_rate": 7.168503174262881e-06,
+      "logits/chosen": 0.0810992568731308,
+      "logits/rejected": 0.13944530487060547,
+      "logps/chosen": -77.3492660522461,
+      "logps/rejected": -90.80574035644531,
+      "loss": 1.8164,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.4869396686553955,
+      "rewards/margins": 1.0017035007476807,
+      "rewards/rejected": -3.488643169403076,
+      "step": 834
+    },
+    {
+      "epoch": 1.5158946412352408,
+      "grad_norm": 2.228290319442749,
+      "learning_rate": 7.162533525769633e-06,
+      "logits/chosen": -0.05040167272090912,
+      "logits/rejected": 0.09420110285282135,
+      "logps/chosen": -84.03011322021484,
+      "logps/rejected": -93.4085464477539,
+      "loss": 2.0479,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6713035106658936,
+      "rewards/margins": 0.8382018804550171,
+      "rewards/rejected": -3.509505033493042,
+      "step": 835
+    },
+    {
+      "epoch": 1.5177111716621252,
+      "grad_norm": 2.6196677684783936,
+      "learning_rate": 7.156560487081052e-06,
+      "logits/chosen": 0.06445930898189545,
+      "logits/rejected": 0.09075027704238892,
+      "logps/chosen": -81.3261489868164,
+      "logps/rejected": -97.19158935546875,
+      "loss": 1.83,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7067627906799316,
+      "rewards/margins": 1.0383878946304321,
+      "rewards/rejected": -3.745150566101074,
+      "step": 836
+    },
+    {
+      "epoch": 1.5195277020890101,
+      "grad_norm": 2.3788716793060303,
+      "learning_rate": 7.150584070377202e-06,
+      "logits/chosen": 0.023997552692890167,
+      "logits/rejected": 0.084604412317276,
+      "logps/chosen": -86.31644439697266,
+      "logps/rejected": -103.74675750732422,
+      "loss": 1.9901,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.7836720943450928,
+      "rewards/margins": 0.9425742626190186,
+      "rewards/rejected": -3.7262461185455322,
+      "step": 837
+    },
+    {
+      "epoch": 1.5213442325158946,
+      "grad_norm": 2.6834628582000732,
+      "learning_rate": 7.144604287845041e-06,
+      "logits/chosen": 0.015507448464632034,
+      "logits/rejected": 0.07179054617881775,
+      "logps/chosen": -79.04088592529297,
+      "logps/rejected": -95.68939208984375,
+      "loss": 1.7565,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.706390142440796,
+      "rewards/margins": 1.0823572874069214,
+      "rewards/rejected": -3.7887473106384277,
+      "step": 838
+    },
+    {
+      "epoch": 1.5231607629427792,
+      "grad_norm": 2.56843638420105,
+      "learning_rate": 7.1386211516783835e-06,
+      "logits/chosen": -0.017521001398563385,
+      "logits/rejected": -0.061632271856069565,
+      "logps/chosen": -80.90260314941406,
+      "logps/rejected": -93.43807220458984,
+      "loss": 1.8708,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.495241165161133,
+      "rewards/margins": 1.119477391242981,
+      "rewards/rejected": -3.614718437194824,
+      "step": 839
+    },
+    {
+      "epoch": 1.524977293369664,
+      "grad_norm": 2.1130528450012207,
+      "learning_rate": 7.132634674077884e-06,
+      "logits/chosen": 0.0070226918905973434,
+      "logits/rejected": 0.12307396531105042,
+      "logps/chosen": -80.71980285644531,
+      "logps/rejected": -101.35388946533203,
+      "loss": 1.8225,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7102115154266357,
+      "rewards/margins": 1.0462895631790161,
+      "rewards/rejected": -3.7565011978149414,
+      "step": 840
+    },
+    {
+      "epoch": 1.5267938237965486,
+      "grad_norm": 2.270901918411255,
+      "learning_rate": 7.12664486725102e-06,
+      "logits/chosen": 0.09991536289453506,
+      "logits/rejected": 0.11764054000377655,
+      "logps/chosen": -84.9771499633789,
+      "logps/rejected": -100.33434295654297,
+      "loss": 2.2036,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7242026329040527,
+      "rewards/margins": 0.846297562122345,
+      "rewards/rejected": -3.570500373840332,
+      "step": 841
+    },
+    {
+      "epoch": 1.5286103542234333,
+      "grad_norm": 2.743037462234497,
+      "learning_rate": 7.120651743412043e-06,
+      "logits/chosen": 0.025931421667337418,
+      "logits/rejected": -0.005241448059678078,
+      "logps/chosen": -85.05397033691406,
+      "logps/rejected": -92.75206756591797,
+      "loss": 2.0745,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6730573177337646,
+      "rewards/margins": 1.065488338470459,
+      "rewards/rejected": -3.7385458946228027,
+      "step": 842
+    },
+    {
+      "epoch": 1.5304268846503177,
+      "grad_norm": 2.7490670680999756,
+      "learning_rate": 7.114655314781979e-06,
+      "logits/chosen": 0.07127739489078522,
+      "logits/rejected": 0.04277154803276062,
+      "logps/chosen": -83.83795166015625,
+      "logps/rejected": -84.39205169677734,
+      "loss": 2.9054,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7217230796813965,
+      "rewards/margins": 0.3551695644855499,
+      "rewards/rejected": -3.076892852783203,
+      "step": 843
+    },
+    {
+      "epoch": 1.5322434150772026,
+      "grad_norm": 2.204655170440674,
+      "learning_rate": 7.10865559358859e-06,
+      "logits/chosen": 0.034612249583005905,
+      "logits/rejected": 0.07071523368358612,
+      "logps/chosen": -77.57713317871094,
+      "logps/rejected": -88.81067657470703,
+      "loss": 2.0337,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.794020175933838,
+      "rewards/margins": 0.7705952525138855,
+      "rewards/rejected": -3.564615488052368,
+      "step": 844
+    },
+    {
+      "epoch": 1.534059945504087,
+      "grad_norm": 2.373863935470581,
+      "learning_rate": 7.102652592066354e-06,
+      "logits/chosen": 0.1263967752456665,
+      "logits/rejected": 0.10894232988357544,
+      "logps/chosen": -78.27210998535156,
+      "logps/rejected": -84.76968383789062,
+      "loss": 2.0159,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7315664291381836,
+      "rewards/margins": 0.8217055201530457,
+      "rewards/rejected": -3.553272247314453,
+      "step": 845
+    },
+    {
+      "epoch": 1.535876475930972,
+      "grad_norm": 2.7803144454956055,
+      "learning_rate": 7.096646322456433e-06,
+      "logits/chosen": 0.12285368889570236,
+      "logits/rejected": 0.13744983077049255,
+      "logps/chosen": -82.08031463623047,
+      "logps/rejected": -88.31858825683594,
+      "loss": 2.2342,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5975961685180664,
+      "rewards/margins": 0.7215418219566345,
+      "rewards/rejected": -3.3191380500793457,
+      "step": 846
+    },
+    {
+      "epoch": 1.5376930063578564,
+      "grad_norm": 2.3556156158447266,
+      "learning_rate": 7.090636797006658e-06,
+      "logits/chosen": 0.10716073215007782,
+      "logits/rejected": 0.146171435713768,
+      "logps/chosen": -90.718994140625,
+      "logps/rejected": -105.76301574707031,
+      "loss": 1.9024,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.929590940475464,
+      "rewards/margins": 0.8752865195274353,
+      "rewards/rejected": -3.804877281188965,
+      "step": 847
+    },
+    {
+      "epoch": 1.5395095367847411,
+      "grad_norm": 2.5544545650482178,
+      "learning_rate": 7.0846240279714984e-06,
+      "logits/chosen": 0.0810776948928833,
+      "logits/rejected": 0.0847952738404274,
+      "logps/chosen": -85.46900939941406,
+      "logps/rejected": -91.54446411132812,
+      "loss": 1.8777,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6484267711639404,
+      "rewards/margins": 0.8561803102493286,
+      "rewards/rejected": -3.5046072006225586,
+      "step": 848
+    },
+    {
+      "epoch": 1.5413260672116258,
+      "grad_norm": 2.667482852935791,
+      "learning_rate": 7.078608027612035e-06,
+      "logits/chosen": 0.05214187875390053,
+      "logits/rejected": 0.009745856747031212,
+      "logps/chosen": -83.0794677734375,
+      "logps/rejected": -96.57014465332031,
+      "loss": 2.0238,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.473628282546997,
+      "rewards/margins": 1.0539703369140625,
+      "rewards/rejected": -3.5275986194610596,
+      "step": 849
+    },
+    {
+      "epoch": 1.5431425976385105,
+      "grad_norm": 2.492583751678467,
+      "learning_rate": 7.072588808195944e-06,
+      "logits/chosen": 0.10039053112268448,
+      "logits/rejected": 0.05160602927207947,
+      "logps/chosen": -83.8985366821289,
+      "logps/rejected": -87.81637573242188,
+      "loss": 2.2203,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.873837471008301,
+      "rewards/margins": 0.702816367149353,
+      "rewards/rejected": -3.5766537189483643,
+      "step": 850
+    },
+    {
+      "epoch": 1.5449591280653951,
+      "grad_norm": 2.280651807785034,
+      "learning_rate": 7.066566381997457e-06,
+      "logits/chosen": 0.08965125679969788,
+      "logits/rejected": 0.08822537213563919,
+      "logps/chosen": -90.04351043701172,
+      "logps/rejected": -101.63990783691406,
+      "loss": 1.7834,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.407850742340088,
+      "rewards/margins": 1.1601994037628174,
+      "rewards/rejected": -3.5680501461029053,
+      "step": 851
+    },
+    {
+      "epoch": 1.5467756584922796,
+      "grad_norm": 1.9988317489624023,
+      "learning_rate": 7.060540761297353e-06,
+      "logits/chosen": 0.0349108949303627,
+      "logits/rejected": 0.12820343673229218,
+      "logps/chosen": -80.97975158691406,
+      "logps/rejected": -99.21882629394531,
+      "loss": 1.7877,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5994269847869873,
+      "rewards/margins": 0.9676470756530762,
+      "rewards/rejected": -3.5670742988586426,
+      "step": 852
+    },
+    {
+      "epoch": 1.5485921889191645,
+      "grad_norm": 2.590162992477417,
+      "learning_rate": 7.05451195838292e-06,
+      "logits/chosen": 0.10045134276151657,
+      "logits/rejected": 0.14448747038841248,
+      "logps/chosen": -77.86140441894531,
+      "logps/rejected": -89.10052490234375,
+      "loss": 2.1282,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7747833728790283,
+      "rewards/margins": 0.7405270934104919,
+      "rewards/rejected": -3.515310764312744,
+      "step": 853
+    },
+    {
+      "epoch": 1.550408719346049,
+      "grad_norm": 2.6667673587799072,
+      "learning_rate": 7.048479985547941e-06,
+      "logits/chosen": 0.02231951802968979,
+      "logits/rejected": 0.12828901410102844,
+      "logps/chosen": -78.2810287475586,
+      "logps/rejected": -97.22885131835938,
+      "loss": 1.8773,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4572958946228027,
+      "rewards/margins": 0.9423343539237976,
+      "rewards/rejected": -3.399630069732666,
+      "step": 854
+    },
+    {
+      "epoch": 1.5522252497729339,
+      "grad_norm": 1.7987407445907593,
+      "learning_rate": 7.042444855092653e-06,
+      "logits/chosen": 0.061210960149765015,
+      "logits/rejected": 0.08236300945281982,
+      "logps/chosen": -79.21247100830078,
+      "logps/rejected": -94.42364501953125,
+      "loss": 2.144,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.486496686935425,
+      "rewards/margins": 0.8338861465454102,
+      "rewards/rejected": -3.320382833480835,
+      "step": 855
+    },
+    {
+      "epoch": 1.5540417801998183,
+      "grad_norm": 2.5392510890960693,
+      "learning_rate": 7.0364065793237406e-06,
+      "logits/chosen": 0.034715939313173294,
+      "logits/rejected": -0.004529901314526796,
+      "logps/chosen": -87.7265853881836,
+      "logps/rejected": -99.10587310791016,
+      "loss": 1.8283,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.661682605743408,
+      "rewards/margins": 0.9480312466621399,
+      "rewards/rejected": -3.609713554382324,
+      "step": 856
+    },
+    {
+      "epoch": 1.555858310626703,
+      "grad_norm": 2.442492961883545,
+      "learning_rate": 7.0303651705542985e-06,
+      "logits/chosen": 0.0268821120262146,
+      "logits/rejected": 0.09976281225681305,
+      "logps/chosen": -91.73252868652344,
+      "logps/rejected": -100.85424041748047,
+      "loss": 2.1762,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.6233739852905273,
+      "rewards/margins": 0.893970787525177,
+      "rewards/rejected": -3.5173447132110596,
+      "step": 857
+    },
+    {
+      "epoch": 1.5576748410535877,
+      "grad_norm": 2.857959270477295,
+      "learning_rate": 7.024320641103812e-06,
+      "logits/chosen": 0.15326820313930511,
+      "logits/rejected": 0.1491074562072754,
+      "logps/chosen": -82.46217346191406,
+      "logps/rejected": -87.15203857421875,
+      "loss": 2.4176,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.4776577949523926,
+      "rewards/margins": 0.4634871482849121,
+      "rewards/rejected": -2.9411449432373047,
+      "step": 858
+    },
+    {
+      "epoch": 1.5594913714804723,
+      "grad_norm": 2.5892186164855957,
+      "learning_rate": 7.018273003298129e-06,
+      "logits/chosen": 0.04699310287833214,
+      "logits/rejected": 0.033118702471256256,
+      "logps/chosen": -79.92018127441406,
+      "logps/rejected": -89.115966796875,
+      "loss": 2.0407,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5493407249450684,
+      "rewards/margins": 0.8943923115730286,
+      "rewards/rejected": -3.443732738494873,
+      "step": 859
+    },
+    {
+      "epoch": 1.561307901907357,
+      "grad_norm": 2.2549710273742676,
+      "learning_rate": 7.0122222694694356e-06,
+      "logits/chosen": 0.12253312766551971,
+      "logits/rejected": 0.1437494158744812,
+      "logps/chosen": -75.97108459472656,
+      "logps/rejected": -90.68475341796875,
+      "loss": 1.7125,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.593079090118408,
+      "rewards/margins": 1.1156821250915527,
+      "rewards/rejected": -3.70876145362854,
+      "step": 860
+    },
+    {
+      "epoch": 1.5631244323342415,
+      "grad_norm": 2.22299861907959,
+      "learning_rate": 7.006168451956233e-06,
+      "logits/chosen": 0.11537674069404602,
+      "logits/rejected": 0.045418016612529755,
+      "logps/chosen": -86.62315368652344,
+      "logps/rejected": -87.4183120727539,
+      "loss": 2.0229,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.457634687423706,
+      "rewards/margins": 0.7262483835220337,
+      "rewards/rejected": -3.183882713317871,
+      "step": 861
+    },
+    {
+      "epoch": 1.5649409627611264,
+      "grad_norm": 2.734768867492676,
+      "learning_rate": 7.0001115631033065e-06,
+      "logits/chosen": 0.04321238026022911,
+      "logits/rejected": 0.04483964294195175,
+      "logps/chosen": -82.62739562988281,
+      "logps/rejected": -94.7513656616211,
+      "loss": 1.9925,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.520638942718506,
+      "rewards/margins": 0.8393934965133667,
+      "rewards/rejected": -3.360032320022583,
+      "step": 862
+    },
+    {
+      "epoch": 1.5667574931880108,
+      "grad_norm": 2.2604191303253174,
+      "learning_rate": 6.994051615261711e-06,
+      "logits/chosen": 0.0408606193959713,
+      "logits/rejected": 0.047267988324165344,
+      "logps/chosen": -80.66863250732422,
+      "logps/rejected": -95.75201416015625,
+      "loss": 1.8486,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.533386707305908,
+      "rewards/margins": 1.029523491859436,
+      "rewards/rejected": -3.5629100799560547,
+      "step": 863
+    },
+    {
+      "epoch": 1.5685740236148955,
+      "grad_norm": 2.274548292160034,
+      "learning_rate": 6.987988620788731e-06,
+      "logits/chosen": 0.010744954459369183,
+      "logits/rejected": 0.052073340862989426,
+      "logps/chosen": -82.55726623535156,
+      "logps/rejected": -93.57171630859375,
+      "loss": 1.9414,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.571608543395996,
+      "rewards/margins": 0.828125,
+      "rewards/rejected": -3.399733781814575,
+      "step": 864
+    },
+    {
+      "epoch": 1.5703905540417802,
+      "grad_norm": 2.2155966758728027,
+      "learning_rate": 6.98192259204787e-06,
+      "logits/chosen": 0.082489974796772,
+      "logits/rejected": 0.11105959117412567,
+      "logps/chosen": -72.24876403808594,
+      "logps/rejected": -88.81566619873047,
+      "loss": 1.7377,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.400843381881714,
+      "rewards/margins": 1.03752601146698,
+      "rewards/rejected": -3.4383695125579834,
+      "step": 865
+    },
+    {
+      "epoch": 1.5722070844686649,
+      "grad_norm": 2.4514036178588867,
+      "learning_rate": 6.9758535414088166e-06,
+      "logits/chosen": 0.029958020895719528,
+      "logits/rejected": 0.0316297747194767,
+      "logps/chosen": -88.95033264160156,
+      "logps/rejected": -92.07289123535156,
+      "loss": 2.3553,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.583791732788086,
+      "rewards/margins": 0.5526307225227356,
+      "rewards/rejected": -3.1364223957061768,
+      "step": 866
+    },
+    {
+      "epoch": 1.5740236148955495,
+      "grad_norm": 2.701751470565796,
+      "learning_rate": 6.969781481247423e-06,
+      "logits/chosen": 0.08873719722032547,
+      "logits/rejected": 0.05975634232163429,
+      "logps/chosen": -83.68199157714844,
+      "logps/rejected": -91.66279602050781,
+      "loss": 2.0972,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.5379390716552734,
+      "rewards/margins": 0.720723569393158,
+      "rewards/rejected": -3.258662462234497,
+      "step": 867
+    },
+    {
+      "epoch": 1.575840145322434,
+      "grad_norm": 2.887376308441162,
+      "learning_rate": 6.963706423945674e-06,
+      "logits/chosen": 0.05533679947257042,
+      "logits/rejected": 0.02152944728732109,
+      "logps/chosen": -81.51123046875,
+      "logps/rejected": -95.27003479003906,
+      "loss": 2.0153,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.5463404655456543,
+      "rewards/margins": 0.968324601650238,
+      "rewards/rejected": -3.514665126800537,
+      "step": 868
+    },
+    {
+      "epoch": 1.577656675749319,
+      "grad_norm": 2.442810535430908,
+      "learning_rate": 6.957628381891674e-06,
+      "logits/chosen": -0.04294189438223839,
+      "logits/rejected": 0.03631982207298279,
+      "logps/chosen": -77.20703125,
+      "logps/rejected": -93.48060607910156,
+      "loss": 2.1137,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.508554220199585,
+      "rewards/margins": 0.8056586384773254,
+      "rewards/rejected": -3.3142127990722656,
+      "step": 869
+    },
+    {
+      "epoch": 1.5794732061762033,
+      "grad_norm": 2.456723928451538,
+      "learning_rate": 6.951547367479604e-06,
+      "logits/chosen": 0.025873543694615364,
+      "logits/rejected": -0.04093042016029358,
+      "logps/chosen": -86.80729675292969,
+      "logps/rejected": -96.05107116699219,
+      "loss": 2.0419,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.3767457008361816,
+      "rewards/margins": 0.8237012028694153,
+      "rewards/rejected": -3.200446844100952,
+      "step": 870
+    },
+    {
+      "epoch": 1.5812897366030882,
+      "grad_norm": 2.332174062728882,
+      "learning_rate": 6.945463393109712e-06,
+      "logits/chosen": 0.006868576630949974,
+      "logits/rejected": 0.09155347943305969,
+      "logps/chosen": -85.2925796508789,
+      "logps/rejected": -104.44573211669922,
+      "loss": 2.0281,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.5827579498291016,
+      "rewards/margins": 0.9421724081039429,
+      "rewards/rejected": -3.524930477142334,
+      "step": 871
+    },
+    {
+      "epoch": 1.5831062670299727,
+      "grad_norm": 3.058819055557251,
+      "learning_rate": 6.939376471188285e-06,
+      "logits/chosen": 0.07147755473852158,
+      "logits/rejected": 0.03247075155377388,
+      "logps/chosen": -85.51266479492188,
+      "logps/rejected": -96.28234100341797,
+      "loss": 2.2254,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.8757948875427246,
+      "rewards/margins": 0.7729579210281372,
+      "rewards/rejected": -3.6487529277801514,
+      "step": 872
+    },
+    {
+      "epoch": 1.5849227974568574,
+      "grad_norm": 2.271951198577881,
+      "learning_rate": 6.933286614127613e-06,
+      "logits/chosen": 0.04744710773229599,
+      "logits/rejected": 0.06620746105909348,
+      "logps/chosen": -75.55758666992188,
+      "logps/rejected": -77.83596801757812,
+      "loss": 2.0876,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.806255340576172,
+      "rewards/margins": 0.6999154686927795,
+      "rewards/rejected": -3.5061709880828857,
+      "step": 873
+    },
+    {
+      "epoch": 1.586739327883742,
+      "grad_norm": 2.9459800720214844,
+      "learning_rate": 6.927193834345976e-06,
+      "logits/chosen": 0.025282222777605057,
+      "logits/rejected": 0.08691692352294922,
+      "logps/chosen": -88.87342834472656,
+      "logps/rejected": -101.53143310546875,
+      "loss": 2.0501,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.600255012512207,
+      "rewards/margins": 0.8346633911132812,
+      "rewards/rejected": -3.4349184036254883,
+      "step": 874
+    },
+    {
+      "epoch": 1.5885558583106267,
+      "grad_norm": 2.279984474182129,
+      "learning_rate": 6.921098144267613e-06,
+      "logits/chosen": 0.027605077251791954,
+      "logits/rejected": 0.09924539178609848,
+      "logps/chosen": -85.24947357177734,
+      "logps/rejected": -97.9180679321289,
+      "loss": 1.9226,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.842658281326294,
+      "rewards/margins": 0.8581206202507019,
+      "rewards/rejected": -3.7007789611816406,
+      "step": 875
+    },
+    {
+      "epoch": 1.5903723887375114,
+      "grad_norm": 3.2044014930725098,
+      "learning_rate": 6.914999556322697e-06,
+      "logits/chosen": 0.06477774679660797,
+      "logits/rejected": 0.03072739765048027,
+      "logps/chosen": -96.84561157226562,
+      "logps/rejected": -98.88067626953125,
+      "loss": 2.3308,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.960343360900879,
+      "rewards/margins": 0.545425295829773,
+      "rewards/rejected": -3.5057687759399414,
+      "step": 876
+    },
+    {
+      "epoch": 1.5921889191643959,
+      "grad_norm": 2.679797410964966,
+      "learning_rate": 6.90889808294731e-06,
+      "logits/chosen": 0.04393957555294037,
+      "logits/rejected": 0.02098749577999115,
+      "logps/chosen": -88.12010192871094,
+      "logps/rejected": -96.62691497802734,
+      "loss": 2.297,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.7419230937957764,
+      "rewards/margins": 0.6746903657913208,
+      "rewards/rejected": -3.4166133403778076,
+      "step": 877
+    },
+    {
+      "epoch": 1.5940054495912808,
+      "grad_norm": 2.7355191707611084,
+      "learning_rate": 6.9027937365834206e-06,
+      "logits/chosen": 0.07665219157934189,
+      "logits/rejected": 0.10063200443983078,
+      "logps/chosen": -86.96102905273438,
+      "logps/rejected": -101.6490478515625,
+      "loss": 1.9587,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6776695251464844,
+      "rewards/margins": 0.838853120803833,
+      "rewards/rejected": -3.5165226459503174,
+      "step": 878
+    },
+    {
+      "epoch": 1.5958219800181652,
+      "grad_norm": 3.1400301456451416,
+      "learning_rate": 6.896686529678853e-06,
+      "logits/chosen": 0.10836174339056015,
+      "logits/rejected": 0.14626868069171906,
+      "logps/chosen": -85.75300598144531,
+      "logps/rejected": -101.42611694335938,
+      "loss": 2.0544,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.7651474475860596,
+      "rewards/margins": 0.8746659159660339,
+      "rewards/rejected": -3.63981294631958,
+      "step": 879
+    },
+    {
+      "epoch": 1.5976385104450501,
+      "grad_norm": 2.9733288288116455,
+      "learning_rate": 6.890576474687264e-06,
+      "logits/chosen": 0.05398109182715416,
+      "logits/rejected": 0.011274321004748344,
+      "logps/chosen": -91.39241790771484,
+      "logps/rejected": -98.18811798095703,
+      "loss": 2.1884,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.658442497253418,
+      "rewards/margins": 0.6335260272026062,
+      "rewards/rejected": -3.291968584060669,
+      "step": 880
+    },
+    {
+      "epoch": 1.5994550408719346,
+      "grad_norm": 2.4757211208343506,
+      "learning_rate": 6.884463584068121e-06,
+      "logits/chosen": 0.10966772586107254,
+      "logits/rejected": 0.0948777049779892,
+      "logps/chosen": -84.34974670410156,
+      "logps/rejected": -92.52738952636719,
+      "loss": 1.8273,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.648416519165039,
+      "rewards/margins": 0.8553692102432251,
+      "rewards/rejected": -3.5037853717803955,
+      "step": 881
+    },
+    {
+      "epoch": 1.6012715712988193,
+      "grad_norm": 2.846766710281372,
+      "learning_rate": 6.878347870286673e-06,
+      "logits/chosen": 0.03845957666635513,
+      "logits/rejected": 0.027499085292220116,
+      "logps/chosen": -77.8682861328125,
+      "logps/rejected": -86.57948303222656,
+      "loss": 1.9555,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -2.8318474292755127,
+      "rewards/margins": 0.8005434274673462,
+      "rewards/rejected": -3.6323909759521484,
+      "step": 882
+    },
+    {
+      "epoch": 1.603088101725704,
+      "grad_norm": 2.80184268951416,
+      "learning_rate": 6.872229345813924e-06,
+      "logits/chosen": 0.08280351012945175,
+      "logits/rejected": 0.11747594177722931,
+      "logps/chosen": -80.64393615722656,
+      "logps/rejected": -94.31782531738281,
+      "loss": 1.7596,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.782545804977417,
+      "rewards/margins": 0.9048705101013184,
+      "rewards/rejected": -3.6874163150787354,
+      "step": 883
+    },
+    {
+      "epoch": 1.6049046321525886,
+      "grad_norm": 2.607837200164795,
+      "learning_rate": 6.8661080231266104e-06,
+      "logits/chosen": 0.11254586279392242,
+      "logits/rejected": 0.06956940144300461,
+      "logps/chosen": -84.59523010253906,
+      "logps/rejected": -91.33222198486328,
+      "loss": 2.176,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.859100818634033,
+      "rewards/margins": 0.6652751564979553,
+      "rewards/rejected": -3.524376153945923,
+      "step": 884
+    },
+    {
+      "epoch": 1.6067211625794733,
+      "grad_norm": 3.3895885944366455,
+      "learning_rate": 6.859983914707177e-06,
+      "logits/chosen": 0.11012329161167145,
+      "logits/rejected": 0.1130778044462204,
+      "logps/chosen": -92.46974182128906,
+      "logps/rejected": -103.50257110595703,
+      "loss": 2.0558,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -3.044379234313965,
+      "rewards/margins": 0.958195686340332,
+      "rewards/rejected": -4.002574920654297,
+      "step": 885
+    },
+    {
+      "epoch": 1.6085376930063577,
+      "grad_norm": 2.8371028900146484,
+      "learning_rate": 6.853857033043745e-06,
+      "logits/chosen": 0.05022600293159485,
+      "logits/rejected": 0.0587327815592289,
+      "logps/chosen": -87.489013671875,
+      "logps/rejected": -105.74163055419922,
+      "loss": 2.1291,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.5108165740966797,
+      "rewards/margins": 0.822872519493103,
+      "rewards/rejected": -3.333688735961914,
+      "step": 886
+    },
+    {
+      "epoch": 1.6103542234332426,
+      "grad_norm": 2.347508668899536,
+      "learning_rate": 6.847727390630093e-06,
+      "logits/chosen": 0.1272752583026886,
+      "logits/rejected": 0.12413996458053589,
+      "logps/chosen": -87.31372833251953,
+      "logps/rejected": -103.10004425048828,
+      "loss": 1.9655,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.733205795288086,
+      "rewards/margins": 1.1384268999099731,
+      "rewards/rejected": -3.8716328144073486,
+      "step": 887
+    },
+    {
+      "epoch": 1.612170753860127,
+      "grad_norm": 3.185014247894287,
+      "learning_rate": 6.84159499996563e-06,
+      "logits/chosen": 0.11554019898176193,
+      "logits/rejected": 0.1505330353975296,
+      "logps/chosen": -83.4138412475586,
+      "logps/rejected": -97.09869384765625,
+      "loss": 1.9864,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.906578540802002,
+      "rewards/margins": 1.0419787168502808,
+      "rewards/rejected": -3.948557138442993,
+      "step": 888
+    },
+    {
+      "epoch": 1.6139872842870118,
+      "grad_norm": 2.601205348968506,
+      "learning_rate": 6.83545987355537e-06,
+      "logits/chosen": 0.10080607980489731,
+      "logits/rejected": 0.06827899813652039,
+      "logps/chosen": -81.38069915771484,
+      "logps/rejected": -92.65945434570312,
+      "loss": 1.7634,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.6402342319488525,
+      "rewards/margins": 1.1383386850357056,
+      "rewards/rejected": -3.7785730361938477,
+      "step": 889
+    },
+    {
+      "epoch": 1.6158038147138964,
+      "grad_norm": 2.91386079788208,
+      "learning_rate": 6.829322023909901e-06,
+      "logits/chosen": 0.06657513976097107,
+      "logits/rejected": 0.05152777582406998,
+      "logps/chosen": -90.99147033691406,
+      "logps/rejected": -93.03759765625,
+      "loss": 2.3534,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.659510374069214,
+      "rewards/margins": 0.5215647220611572,
+      "rewards/rejected": -3.181075096130371,
+      "step": 890
+    },
+    {
+      "epoch": 1.6176203451407811,
+      "grad_norm": 2.6477596759796143,
+      "learning_rate": 6.823181463545368e-06,
+      "logits/chosen": 0.01778372749686241,
+      "logits/rejected": 0.06530646234750748,
+      "logps/chosen": -85.58574676513672,
+      "logps/rejected": -101.24583435058594,
+      "loss": 1.9453,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.543752670288086,
+      "rewards/margins": 0.894145667552948,
+      "rewards/rejected": -3.4378981590270996,
+      "step": 891
+    },
+    {
+      "epoch": 1.6194368755676658,
+      "grad_norm": 2.449430465698242,
+      "learning_rate": 6.817038204983444e-06,
+      "logits/chosen": 0.10009394586086273,
+      "logits/rejected": 0.11922462284564972,
+      "logps/chosen": -75.22452545166016,
+      "logps/rejected": -89.5363998413086,
+      "loss": 1.9703,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.5327696800231934,
+      "rewards/margins": 1.006116271018982,
+      "rewards/rejected": -3.538886070251465,
+      "step": 892
+    },
+    {
+      "epoch": 1.6212534059945503,
+      "grad_norm": 2.757439374923706,
+      "learning_rate": 6.810892260751302e-06,
+      "logits/chosen": 0.054223865270614624,
+      "logits/rejected": 0.11705522984266281,
+      "logps/chosen": -84.47832489013672,
+      "logps/rejected": -107.8006591796875,
+      "loss": 1.9296,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.757310152053833,
+      "rewards/margins": 1.0707166194915771,
+      "rewards/rejected": -3.828026294708252,
+      "step": 893
+    },
+    {
+      "epoch": 1.6230699364214352,
+      "grad_norm": 2.5267086029052734,
+      "learning_rate": 6.804743643381592e-06,
+      "logits/chosen": 0.07622219622135162,
+      "logits/rejected": 0.05764157325029373,
+      "logps/chosen": -81.3788070678711,
+      "logps/rejected": -87.3583984375,
+      "loss": 1.857,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6934287548065186,
+      "rewards/margins": 0.9804441928863525,
+      "rewards/rejected": -3.673872947692871,
+      "step": 894
+    },
+    {
+      "epoch": 1.6248864668483196,
+      "grad_norm": 2.3378968238830566,
+      "learning_rate": 6.798592365412415e-06,
+      "logits/chosen": 0.08304903656244278,
+      "logits/rejected": 0.10661397874355316,
+      "logps/chosen": -84.71305847167969,
+      "logps/rejected": -100.10174560546875,
+      "loss": 1.8795,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.6393752098083496,
+      "rewards/margins": 0.9716106057167053,
+      "rewards/rejected": -3.610985517501831,
+      "step": 895
+    },
+    {
+      "epoch": 1.6267029972752045,
+      "grad_norm": 3.1945290565490723,
+      "learning_rate": 6.792438439387299e-06,
+      "logits/chosen": 0.09845674782991409,
+      "logits/rejected": 0.037844423204660416,
+      "logps/chosen": -87.18885803222656,
+      "logps/rejected": -91.7021484375,
+      "loss": 2.4462,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.7558887004852295,
+      "rewards/margins": 0.6585051417350769,
+      "rewards/rejected": -3.414393663406372,
+      "step": 896
+    },
+    {
+      "epoch": 1.628519527702089,
+      "grad_norm": 2.678049087524414,
+      "learning_rate": 6.7862818778551695e-06,
+      "logits/chosen": -0.059538260102272034,
+      "logits/rejected": 0.028212856501340866,
+      "logps/chosen": -91.52423858642578,
+      "logps/rejected": -110.80790710449219,
+      "loss": 2.0741,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.6971120834350586,
+      "rewards/margins": 0.9168665409088135,
+      "rewards/rejected": -3.6139791011810303,
+      "step": 897
+    },
+    {
+      "epoch": 1.6303360581289736,
+      "grad_norm": 2.703550338745117,
+      "learning_rate": 6.780122693370329e-06,
+      "logits/chosen": -0.007212355732917786,
+      "logits/rejected": 0.08302198350429535,
+      "logps/chosen": -78.15343475341797,
+      "logps/rejected": -88.86624908447266,
+      "loss": 2.1919,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.428589344024658,
+      "rewards/margins": 0.6562613844871521,
+      "rewards/rejected": -3.084850549697876,
+      "step": 898
+    },
+    {
+      "epoch": 1.6321525885558583,
+      "grad_norm": 2.5438737869262695,
+      "learning_rate": 6.773960898492428e-06,
+      "logits/chosen": 0.10631975531578064,
+      "logits/rejected": 0.1630961000919342,
+      "logps/chosen": -79.30477142333984,
+      "logps/rejected": -94.29690551757812,
+      "loss": 1.8887,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.7491891384124756,
+      "rewards/margins": 0.8967292308807373,
+      "rewards/rejected": -3.645918369293213,
+      "step": 899
+    },
+    {
+      "epoch": 1.633969118982743,
+      "grad_norm": 2.6740238666534424,
+      "learning_rate": 6.767796505786435e-06,
+      "logits/chosen": 0.04583762586116791,
+      "logits/rejected": 0.05214417725801468,
+      "logps/chosen": -90.60303497314453,
+      "logps/rejected": -99.45308685302734,
+      "loss": 2.2367,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.576535224914551,
+      "rewards/margins": 0.6341556310653687,
+      "rewards/rejected": -3.210691213607788,
+      "step": 900
+    },
+    {
+      "epoch": 1.6357856494096277,
+      "grad_norm": 2.8195323944091797,
+      "learning_rate": 6.761629527822622e-06,
+      "logits/chosen": 0.11805088073015213,
+      "logits/rejected": 0.03617120161652565,
+      "logps/chosen": -79.40000915527344,
+      "logps/rejected": -79.39904022216797,
+      "loss": 2.0995,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.64365816116333,
+      "rewards/margins": 0.7394881248474121,
+      "rewards/rejected": -3.383146047592163,
+      "step": 901
+    },
+    {
+      "epoch": 1.6376021798365121,
+      "grad_norm": 2.666821241378784,
+      "learning_rate": 6.755459977176533e-06,
+      "logits/chosen": 0.017997367307543755,
+      "logits/rejected": 0.09890137612819672,
+      "logps/chosen": -89.78903198242188,
+      "logps/rejected": -111.10968017578125,
+      "loss": 1.8919,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.503241539001465,
+      "rewards/margins": 1.0098021030426025,
+      "rewards/rejected": -3.5130434036254883,
+      "step": 902
+    },
+    {
+      "epoch": 1.639418710263397,
+      "grad_norm": 2.3226935863494873,
+      "learning_rate": 6.749287866428953e-06,
+      "logits/chosen": 0.09748942404985428,
+      "logits/rejected": 0.09682448208332062,
+      "logps/chosen": -82.52100372314453,
+      "logps/rejected": -91.55313873291016,
+      "loss": 1.7682,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5952441692352295,
+      "rewards/margins": 1.0089478492736816,
+      "rewards/rejected": -3.604191780090332,
+      "step": 903
+    },
+    {
+      "epoch": 1.6412352406902815,
+      "grad_norm": 2.479614019393921,
+      "learning_rate": 6.743113208165893e-06,
+      "logits/chosen": 0.029578503221273422,
+      "logits/rejected": 0.07657214999198914,
+      "logps/chosen": -100.56556701660156,
+      "logps/rejected": -102.63167572021484,
+      "loss": 2.5709,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": -2.717003345489502,
+      "rewards/margins": 0.48937955498695374,
+      "rewards/rejected": -3.206382989883423,
+      "step": 904
+    },
+    {
+      "epoch": 1.6430517711171662,
+      "grad_norm": 3.3280487060546875,
+      "learning_rate": 6.736936014978552e-06,
+      "logits/chosen": 0.04310007020831108,
+      "logits/rejected": 0.05533391237258911,
+      "logps/chosen": -102.93861389160156,
+      "logps/rejected": -110.60729217529297,
+      "loss": 1.9292,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3910446166992188,
+      "rewards/margins": 0.9039056897163391,
+      "rewards/rejected": -3.294950485229492,
+      "step": 905
+    },
+    {
+      "epoch": 1.6448683015440508,
+      "grad_norm": 2.6267921924591064,
+      "learning_rate": 6.730756299463308e-06,
+      "logits/chosen": 0.06594766676425934,
+      "logits/rejected": 0.11423090845346451,
+      "logps/chosen": -79.65184020996094,
+      "logps/rejected": -90.49372863769531,
+      "loss": 2.2204,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.69381046295166,
+      "rewards/margins": 0.8501594066619873,
+      "rewards/rejected": -3.5439698696136475,
+      "step": 906
+    },
+    {
+      "epoch": 1.6466848319709355,
+      "grad_norm": 2.4378793239593506,
+      "learning_rate": 6.724574074221673e-06,
+      "logits/chosen": 0.043015994131565094,
+      "logits/rejected": 0.14672963321208954,
+      "logps/chosen": -78.83383178710938,
+      "logps/rejected": -93.22097778320312,
+      "loss": 1.9456,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.480109691619873,
+      "rewards/margins": 0.8949618339538574,
+      "rewards/rejected": -3.3750715255737305,
+      "step": 907
+    },
+    {
+      "epoch": 1.6485013623978202,
+      "grad_norm": 2.400392770767212,
+      "learning_rate": 6.718389351860285e-06,
+      "logits/chosen": 0.08273115754127502,
+      "logits/rejected": 0.06503652036190033,
+      "logps/chosen": -87.24859619140625,
+      "logps/rejected": -95.38743591308594,
+      "loss": 1.7985,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.356613874435425,
+      "rewards/margins": 0.8771872520446777,
+      "rewards/rejected": -3.2338013648986816,
+      "step": 908
+    },
+    {
+      "epoch": 1.6503178928247049,
+      "grad_norm": 2.42429780960083,
+      "learning_rate": 6.712202144990863e-06,
+      "logits/chosen": 0.10674826800823212,
+      "logits/rejected": 0.08516664803028107,
+      "logps/chosen": -89.37960815429688,
+      "logps/rejected": -99.21661376953125,
+      "loss": 2.1353,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6561362743377686,
+      "rewards/margins": 0.7742940187454224,
+      "rewards/rejected": -3.4304304122924805,
+      "step": 909
+    },
+    {
+      "epoch": 1.6521344232515895,
+      "grad_norm": 2.6496243476867676,
+      "learning_rate": 6.7060124662302066e-06,
+      "logits/chosen": 0.14033779501914978,
+      "logits/rejected": 0.07252150774002075,
+      "logps/chosen": -80.96002197265625,
+      "logps/rejected": -82.9183120727539,
+      "loss": 2.0659,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.427546739578247,
+      "rewards/margins": 0.8414397835731506,
+      "rewards/rejected": -3.268986463546753,
+      "step": 910
+    },
+    {
+      "epoch": 1.653950953678474,
+      "grad_norm": 3.0638275146484375,
+      "learning_rate": 6.699820328200143e-06,
+      "logits/chosen": 0.058655206114053726,
+      "logits/rejected": 0.08714447915554047,
+      "logps/chosen": -88.06983947753906,
+      "logps/rejected": -95.36041259765625,
+      "loss": 2.0403,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.2383832931518555,
+      "rewards/margins": 0.9142959117889404,
+      "rewards/rejected": -3.152679443359375,
+      "step": 911
+    },
+    {
+      "epoch": 1.655767484105359,
+      "grad_norm": 2.2842650413513184,
+      "learning_rate": 6.69362574352752e-06,
+      "logits/chosen": -0.011715320870280266,
+      "logits/rejected": 0.03974226117134094,
+      "logps/chosen": -84.77818298339844,
+      "logps/rejected": -94.44766235351562,
+      "loss": 2.091,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4050097465515137,
+      "rewards/margins": 0.7936041951179504,
+      "rewards/rejected": -3.1986136436462402,
+      "step": 912
+    },
+    {
+      "epoch": 1.6575840145322434,
+      "grad_norm": 3.1531357765197754,
+      "learning_rate": 6.6874287248441794e-06,
+      "logits/chosen": 0.08675999939441681,
+      "logits/rejected": 0.06470594555139542,
+      "logps/chosen": -80.276123046875,
+      "logps/rejected": -90.92185974121094,
+      "loss": 2.0593,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.28584361076355,
+      "rewards/margins": 0.7852586507797241,
+      "rewards/rejected": -3.0711023807525635,
+      "step": 913
+    },
+    {
+      "epoch": 1.659400544959128,
+      "grad_norm": 2.7664525508880615,
+      "learning_rate": 6.681229284786915e-06,
+      "logits/chosen": 0.05306999385356903,
+      "logits/rejected": 0.051360100507736206,
+      "logps/chosen": -83.74992370605469,
+      "logps/rejected": -90.31805419921875,
+      "loss": 2.2115,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.623281955718994,
+      "rewards/margins": 0.6698784232139587,
+      "rewards/rejected": -3.2931606769561768,
+      "step": 914
+    },
+    {
+      "epoch": 1.6612170753860127,
+      "grad_norm": 2.890864610671997,
+      "learning_rate": 6.675027435997469e-06,
+      "logits/chosen": 0.04681715369224548,
+      "logits/rejected": 0.1523803025484085,
+      "logps/chosen": -74.13650512695312,
+      "logps/rejected": -93.63848876953125,
+      "loss": 1.988,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.540015697479248,
+      "rewards/margins": 0.9194749593734741,
+      "rewards/rejected": -3.4594907760620117,
+      "step": 915
+    },
+    {
+      "epoch": 1.6630336058128974,
+      "grad_norm": 2.2408931255340576,
+      "learning_rate": 6.668823191122487e-06,
+      "logits/chosen": 0.08750308305025101,
+      "logits/rejected": 0.06304562091827393,
+      "logps/chosen": -87.24713897705078,
+      "logps/rejected": -96.94601440429688,
+      "loss": 1.9141,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.300774097442627,
+      "rewards/margins": 0.8336561322212219,
+      "rewards/rejected": -3.134430170059204,
+      "step": 916
+    },
+    {
+      "epoch": 1.664850136239782,
+      "grad_norm": 2.584038019180298,
+      "learning_rate": 6.662616562813512e-06,
+      "logits/chosen": 0.0785432755947113,
+      "logits/rejected": 0.05843789130449295,
+      "logps/chosen": -88.5465087890625,
+      "logps/rejected": -96.19537353515625,
+      "loss": 1.7811,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.419790267944336,
+      "rewards/margins": 0.844237744808197,
+      "rewards/rejected": -3.2640278339385986,
+      "step": 917
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 2.4870214462280273,
+      "learning_rate": 6.6564075637269295e-06,
+      "logits/chosen": 0.06065249443054199,
+      "logits/rejected": 0.0761883556842804,
+      "logps/chosen": -76.00924682617188,
+      "logps/rejected": -85.09017944335938,
+      "loss": 1.9802,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3578128814697266,
+      "rewards/margins": 0.8296225070953369,
+      "rewards/rejected": -3.1874351501464844,
+      "step": 918
+    },
+    {
+      "epoch": 1.6684831970935514,
+      "grad_norm": 2.0682199001312256,
+      "learning_rate": 6.650196206523977e-06,
+      "logits/chosen": 0.0461968369781971,
+      "logits/rejected": 0.08261042088270187,
+      "logps/chosen": -72.26417541503906,
+      "logps/rejected": -77.92495727539062,
+      "loss": 2.0028,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.1353201866149902,
+      "rewards/margins": 0.6552493572235107,
+      "rewards/rejected": -2.790569305419922,
+      "step": 919
+    },
+    {
+      "epoch": 1.6702997275204359,
+      "grad_norm": 2.482517719268799,
+      "learning_rate": 6.643982503870693e-06,
+      "logits/chosen": 0.12246346473693848,
+      "logits/rejected": 0.03192145377397537,
+      "logps/chosen": -89.27822875976562,
+      "logps/rejected": -89.58540344238281,
+      "loss": 2.0059,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.4135119915008545,
+      "rewards/margins": 0.6876509785652161,
+      "rewards/rejected": -3.101162910461426,
+      "step": 920
+    },
+    {
+      "epoch": 1.6721162579473208,
+      "grad_norm": 2.84334659576416,
+      "learning_rate": 6.637766468437894e-06,
+      "logits/chosen": 0.027422966435551643,
+      "logits/rejected": 0.019883258268237114,
+      "logps/chosen": -77.09085845947266,
+      "logps/rejected": -88.27078247070312,
+      "loss": 1.9826,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.446176052093506,
+      "rewards/margins": 0.8014341592788696,
+      "rewards/rejected": -3.247610092163086,
+      "step": 921
+    },
+    {
+      "epoch": 1.6739327883742052,
+      "grad_norm": 2.386929512023926,
+      "learning_rate": 6.631548112901164e-06,
+      "logits/chosen": 0.025946754962205887,
+      "logits/rejected": 0.06377127766609192,
+      "logps/chosen": -82.12971496582031,
+      "logps/rejected": -97.94608306884766,
+      "loss": 1.9747,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.573225736618042,
+      "rewards/margins": 0.7901925444602966,
+      "rewards/rejected": -3.3634183406829834,
+      "step": 922
+    },
+    {
+      "epoch": 1.67574931880109,
+      "grad_norm": 2.3462514877319336,
+      "learning_rate": 6.6253274499408095e-06,
+      "logits/chosen": -0.009377829730510712,
+      "logits/rejected": 0.0917067900300026,
+      "logps/chosen": -83.40922546386719,
+      "logps/rejected": -103.09521484375,
+      "loss": 1.7486,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.368823766708374,
+      "rewards/margins": 0.9489405751228333,
+      "rewards/rejected": -3.3177640438079834,
+      "step": 923
+    },
+    {
+      "epoch": 1.6775658492279746,
+      "grad_norm": 2.3609023094177246,
+      "learning_rate": 6.619104492241848e-06,
+      "logits/chosen": 0.013379395008087158,
+      "logits/rejected": 0.1470850110054016,
+      "logps/chosen": -75.76339721679688,
+      "logps/rejected": -95.58364868164062,
+      "loss": 1.7618,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.3512778282165527,
+      "rewards/margins": 1.085105061531067,
+      "rewards/rejected": -3.4363832473754883,
+      "step": 924
+    },
+    {
+      "epoch": 1.6793823796548593,
+      "grad_norm": 2.4780282974243164,
+      "learning_rate": 6.61287925249397e-06,
+      "logits/chosen": 0.03466716408729553,
+      "logits/rejected": 0.06602051854133606,
+      "logps/chosen": -85.4227294921875,
+      "logps/rejected": -94.31620025634766,
+      "loss": 2.1665,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6465556621551514,
+      "rewards/margins": 0.6559208035469055,
+      "rewards/rejected": -3.302476644515991,
+      "step": 925
+    },
+    {
+      "epoch": 1.681198910081744,
+      "grad_norm": 4.518651485443115,
+      "learning_rate": 6.606651743391528e-06,
+      "logits/chosen": 0.00949503481388092,
+      "logits/rejected": 0.06442085653543472,
+      "logps/chosen": -88.59455108642578,
+      "logps/rejected": -105.29449462890625,
+      "loss": 2.0174,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.2852275371551514,
+      "rewards/margins": 0.9270148873329163,
+      "rewards/rejected": -3.2122421264648438,
+      "step": 926
+    },
+    {
+      "epoch": 1.6830154405086284,
+      "grad_norm": 2.9527838230133057,
+      "learning_rate": 6.6004219776334925e-06,
+      "logits/chosen": 0.06499281525611877,
+      "logits/rejected": 0.011767326854169369,
+      "logps/chosen": -87.16426849365234,
+      "logps/rejected": -91.7620620727539,
+      "loss": 2.0636,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.3328614234924316,
+      "rewards/margins": 0.7728868126869202,
+      "rewards/rejected": -3.105748176574707,
+      "step": 927
+    },
+    {
+      "epoch": 1.6848319709355133,
+      "grad_norm": 2.7613911628723145,
+      "learning_rate": 6.594189967923445e-06,
+      "logits/chosen": 0.10220185667276382,
+      "logits/rejected": 0.11646111309528351,
+      "logps/chosen": -88.42827606201172,
+      "logps/rejected": -100.41016387939453,
+      "loss": 1.8203,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": -2.3946800231933594,
+      "rewards/margins": 0.9652897119522095,
+      "rewards/rejected": -3.3599696159362793,
+      "step": 928
+    },
+    {
+      "epoch": 1.6866485013623977,
+      "grad_norm": 2.674398899078369,
+      "learning_rate": 6.587955726969533e-06,
+      "logits/chosen": 0.07171659916639328,
+      "logits/rejected": 0.04297471046447754,
+      "logps/chosen": -90.65692138671875,
+      "logps/rejected": -93.96014404296875,
+      "loss": 2.5874,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.538684606552124,
+      "rewards/margins": 0.4615382254123688,
+      "rewards/rejected": -3.00022292137146,
+      "step": 929
+    },
+    {
+      "epoch": 1.6884650317892824,
+      "grad_norm": 2.5353660583496094,
+      "learning_rate": 6.581719267484466e-06,
+      "logits/chosen": 0.13961178064346313,
+      "logits/rejected": 0.13023585081100464,
+      "logps/chosen": -77.078369140625,
+      "logps/rejected": -86.81641387939453,
+      "loss": 1.7334,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.3547840118408203,
+      "rewards/margins": 0.9528897404670715,
+      "rewards/rejected": -3.307673692703247,
+      "step": 930
+    },
+    {
+      "epoch": 1.690281562216167,
+      "grad_norm": 2.8513107299804688,
+      "learning_rate": 6.575480602185468e-06,
+      "logits/chosen": 0.12496423721313477,
+      "logits/rejected": 0.14027045667171478,
+      "logps/chosen": -84.40544891357422,
+      "logps/rejected": -98.02970123291016,
+      "loss": 2.2827,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.501647472381592,
+      "rewards/margins": 0.8144313097000122,
+      "rewards/rejected": -3.3160786628723145,
+      "step": 931
+    },
+    {
+      "epoch": 1.6920980926430518,
+      "grad_norm": 2.7415926456451416,
+      "learning_rate": 6.56923974379426e-06,
+      "logits/chosen": 0.032566964626312256,
+      "logits/rejected": 0.09472635388374329,
+      "logps/chosen": -91.36210632324219,
+      "logps/rejected": -107.93646240234375,
+      "loss": 2.1175,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.199713945388794,
+      "rewards/margins": 0.8857415914535522,
+      "rewards/rejected": -3.0854556560516357,
+      "step": 932
+    },
+    {
+      "epoch": 1.6939146230699365,
+      "grad_norm": 2.986032009124756,
+      "learning_rate": 6.562996705037046e-06,
+      "logits/chosen": 0.028138715773820877,
+      "logits/rejected": 0.08858595788478851,
+      "logps/chosen": -78.46945190429688,
+      "logps/rejected": -92.57295989990234,
+      "loss": 2.384,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.4276278018951416,
+      "rewards/margins": 0.7274990677833557,
+      "rewards/rejected": -3.1551268100738525,
+      "step": 933
+    },
+    {
+      "epoch": 1.695731153496821,
+      "grad_norm": 2.294461488723755,
+      "learning_rate": 6.556751498644461e-06,
+      "logits/chosen": 0.1189623773097992,
+      "logits/rejected": 0.11807019263505936,
+      "logps/chosen": -72.875244140625,
+      "logps/rejected": -77.55158996582031,
+      "loss": 2.0037,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.3735506534576416,
+      "rewards/margins": 0.6975762844085693,
+      "rewards/rejected": -3.07112717628479,
+      "step": 934
+    },
+    {
+      "epoch": 1.6975476839237058,
+      "grad_norm": 2.706350803375244,
+      "learning_rate": 6.550504137351576e-06,
+      "logits/chosen": 0.03257827088236809,
+      "logits/rejected": 0.06685741245746613,
+      "logps/chosen": -93.078369140625,
+      "logps/rejected": -100.50369262695312,
+      "loss": 2.3693,
+      "rewards/accuracies": 0.671875,
+      "rewards/chosen": -2.5124664306640625,
+      "rewards/margins": 0.7416088581085205,
+      "rewards/rejected": -3.254075050354004,
+      "step": 935
+    },
+    {
+      "epoch": 1.6993642143505903,
+      "grad_norm": 2.6172664165496826,
+      "learning_rate": 6.544254633897843e-06,
+      "logits/chosen": 0.10361647605895996,
+      "logits/rejected": 0.1024157926440239,
+      "logps/chosen": -83.01170349121094,
+      "logps/rejected": -94.26844787597656,
+      "loss": 2.02,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -2.4206433296203613,
+      "rewards/margins": 0.8451172113418579,
+      "rewards/rejected": -3.265760660171509,
+      "step": 936
+    },
+    {
+      "epoch": 1.7011807447774752,
+      "grad_norm": 2.4271178245544434,
+      "learning_rate": 6.538003001027089e-06,
+      "logits/chosen": 0.0540219321846962,
+      "logits/rejected": 0.1337059736251831,
+      "logps/chosen": -82.78097534179688,
+      "logps/rejected": -102.0522689819336,
+      "loss": 1.8375,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.637333631515503,
+      "rewards/margins": 1.0766263008117676,
+      "rewards/rejected": -3.7139596939086914,
+      "step": 937
+    },
+    {
+      "epoch": 1.7029972752043596,
+      "grad_norm": 3.2782678604125977,
+      "learning_rate": 6.531749251487482e-06,
+      "logits/chosen": 0.09090890735387802,
+      "logits/rejected": 0.08031464368104935,
+      "logps/chosen": -77.13628387451172,
+      "logps/rejected": -79.11683654785156,
+      "loss": 2.6753,
+      "rewards/accuracies": 0.609375,
+      "rewards/chosen": -2.3901565074920654,
+      "rewards/margins": 0.39642852544784546,
+      "rewards/rejected": -2.7865848541259766,
+      "step": 938
+    },
+    {
+      "epoch": 1.7048138056312443,
+      "grad_norm": 3.1739110946655273,
+      "learning_rate": 6.525493398031508e-06,
+      "logits/chosen": 0.08761344850063324,
+      "logits/rejected": 0.1552104502916336,
+      "logps/chosen": -83.78360748291016,
+      "logps/rejected": -97.63893127441406,
+      "loss": 2.2164,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -2.603111743927002,
+      "rewards/margins": 0.6446429491043091,
+      "rewards/rejected": -3.2477540969848633,
+      "step": 939
+    },
+    {
+      "epoch": 1.706630336058129,
+      "grad_norm": 3.4325766563415527,
+      "learning_rate": 6.5192354534159406e-06,
+      "logits/chosen": 0.018010685220360756,
+      "logits/rejected": 0.07210525125265121,
+      "logps/chosen": -96.70736694335938,
+      "logps/rejected": -112.33457946777344,
+      "loss": 1.7358,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.5734646320343018,
+      "rewards/margins": 1.2066644430160522,
+      "rewards/rejected": -3.7801291942596436,
+      "step": 940
+    },
+    {
+      "epoch": 1.7084468664850136,
+      "grad_norm": 2.4167909622192383,
+      "learning_rate": 6.5129754304018165e-06,
+      "logits/chosen": 0.04140740633010864,
+      "logits/rejected": 0.08365748077630997,
+      "logps/chosen": -76.82650756835938,
+      "logps/rejected": -84.57887268066406,
+      "loss": 2.3819,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.239987373352051,
+      "rewards/margins": 0.5318999290466309,
+      "rewards/rejected": -2.7718873023986816,
+      "step": 941
+    },
+    {
+      "epoch": 1.7102633969118983,
+      "grad_norm": 2.5350871086120605,
+      "learning_rate": 6.506713341754415e-06,
+      "logits/chosen": 0.021460117772221565,
+      "logits/rejected": 0.10231097042560577,
+      "logps/chosen": -79.95002746582031,
+      "logps/rejected": -96.06900024414062,
+      "loss": 1.7958,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.4041337966918945,
+      "rewards/margins": 1.1107407808303833,
+      "rewards/rejected": -3.5148746967315674,
+      "step": 942
+    },
+    {
+      "epoch": 1.7120799273387828,
+      "grad_norm": 2.602478504180908,
+      "learning_rate": 6.500449200243224e-06,
+      "logits/chosen": 0.08819068968296051,
+      "logits/rejected": 0.10114292055368423,
+      "logps/chosen": -72.37012481689453,
+      "logps/rejected": -82.5709228515625,
+      "loss": 2.1943,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.402527332305908,
+      "rewards/margins": 0.6120179295539856,
+      "rewards/rejected": -3.014545202255249,
+      "step": 943
+    },
+    {
+      "epoch": 1.7138964577656677,
+      "grad_norm": 2.405487537384033,
+      "learning_rate": 6.494183018641919e-06,
+      "logits/chosen": 0.020005354657769203,
+      "logits/rejected": 0.07407501339912415,
+      "logps/chosen": -73.03540802001953,
+      "logps/rejected": -84.96491241455078,
+      "loss": 1.9642,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6638689041137695,
+      "rewards/margins": 0.8980711102485657,
+      "rewards/rejected": -3.5619401931762695,
+      "step": 944
+    },
+    {
+      "epoch": 1.7157129881925521,
+      "grad_norm": 2.8340280055999756,
+      "learning_rate": 6.487914809728335e-06,
+      "logits/chosen": 0.09565074741840363,
+      "logits/rejected": 0.13740883767604828,
+      "logps/chosen": -90.7655258178711,
+      "logps/rejected": -105.01213073730469,
+      "loss": 2.102,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.418067216873169,
+      "rewards/margins": 0.7691794633865356,
+      "rewards/rejected": -3.187246799468994,
+      "step": 945
+    },
+    {
+      "epoch": 1.717529518619437,
+      "grad_norm": 2.5804646015167236,
+      "learning_rate": 6.481644586284443e-06,
+      "logits/chosen": 0.06889508664608002,
+      "logits/rejected": 0.0715349018573761,
+      "logps/chosen": -90.50798034667969,
+      "logps/rejected": -96.70076751708984,
+      "loss": 1.8887,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -2.6765780448913574,
+      "rewards/margins": 0.9487190246582031,
+      "rewards/rejected": -3.6252973079681396,
+      "step": 946
+    },
+    {
+      "epoch": 1.7193460490463215,
+      "grad_norm": 2.947023868560791,
+      "learning_rate": 6.475372361096316e-06,
+      "logits/chosen": 0.057566914707422256,
+      "logits/rejected": 0.12709926068782806,
+      "logps/chosen": -77.96834564208984,
+      "logps/rejected": -92.29097747802734,
+      "loss": 2.1631,
+      "rewards/accuracies": 0.703125,
+      "rewards/chosen": -2.5991628170013428,
+      "rewards/margins": 0.8114217519760132,
+      "rewards/rejected": -3.4105844497680664,
+      "step": 947
+    },
+    {
+      "epoch": 1.7211625794732062,
+      "grad_norm": 2.337834596633911,
+      "learning_rate": 6.469098146954119e-06,
+      "logits/chosen": 0.08289645612239838,
+      "logits/rejected": 0.11055514216423035,
+      "logps/chosen": -85.81585693359375,
+      "logps/rejected": -98.47114562988281,
+      "loss": 1.6936,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.407276153564453,
+      "rewards/margins": 0.9422338008880615,
+      "rewards/rejected": -3.3495099544525146,
+      "step": 948
+    },
+    {
+      "epoch": 1.7229791099000908,
+      "grad_norm": 2.2546920776367188,
+      "learning_rate": 6.462821956652064e-06,
+      "logits/chosen": 0.035123877227306366,
+      "logits/rejected": 0.07462943345308304,
+      "logps/chosen": -76.06634521484375,
+      "logps/rejected": -84.29540252685547,
+      "loss": 2.1523,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.3520922660827637,
+      "rewards/margins": 0.698824405670166,
+      "rewards/rejected": -3.050916910171509,
+      "step": 949
+    },
+    {
+      "epoch": 1.7247956403269755,
+      "grad_norm": 2.8522515296936035,
+      "learning_rate": 6.456543802988395e-06,
+      "logits/chosen": 0.07937800884246826,
+      "logits/rejected": 0.09652063250541687,
+      "logps/chosen": -83.96992492675781,
+      "logps/rejected": -97.54905700683594,
+      "loss": 1.9428,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": -2.6706998348236084,
+      "rewards/margins": 0.8279770612716675,
+      "rewards/rejected": -3.4986767768859863,
+      "step": 950
+    },
+    {
+      "epoch": 1.7266121707538602,
+      "grad_norm": 2.6932287216186523,
+      "learning_rate": 6.450263698765363e-06,
+      "logits/chosen": 0.07110100239515305,
+      "logits/rejected": 0.09598114341497421,
+      "logps/chosen": -84.68408966064453,
+      "logps/rejected": -96.74703979492188,
+      "loss": 1.7506,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -2.7609381675720215,
+      "rewards/margins": 0.9440479278564453,
+      "rewards/rejected": -3.7049858570098877,
+      "step": 951
+    },
+    {
+      "epoch": 1.7284287011807447,
+      "grad_norm": 2.2383601665496826,
+      "learning_rate": 6.4439816567891925e-06,
+      "logits/chosen": 0.03344205766916275,
+      "logits/rejected": 0.060385581105947495,
+      "logps/chosen": -90.16828155517578,
+      "logps/rejected": -102.0325927734375,
+      "loss": 1.6958,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": -2.6442179679870605,
+      "rewards/margins": 1.103801965713501,
+      "rewards/rejected": -3.7480199337005615,
+      "step": 952
+    },
+    {
+      "epoch": 1.7302452316076296,
+      "grad_norm": 2.574528455734253,
+      "learning_rate": 6.43769768987006e-06,
+      "logits/chosen": 0.10802514851093292,
+      "logits/rejected": 0.06120527908205986,
+      "logps/chosen": -90.94403076171875,
+      "logps/rejected": -96.3043441772461,
+      "loss": 1.9554,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.774078845977783,
+      "rewards/margins": 0.967216968536377,
+      "rewards/rejected": -3.74129581451416,
+      "step": 953
+    },
+    {
+      "epoch": 1.732061762034514,
+      "grad_norm": 2.6094908714294434,
+      "learning_rate": 6.431411810822069e-06,
+      "logits/chosen": 0.013344529084861279,
+      "logits/rejected": 0.06106187403202057,
+      "logps/chosen": -82.14124298095703,
+      "logps/rejected": -97.54022979736328,
+      "loss": 2.1501,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.814952850341797,
+      "rewards/margins": 0.8146507143974304,
+      "rewards/rejected": -3.629603624343872,
+      "step": 954
+    },
+    {
+      "epoch": 1.7338782924613987,
+      "grad_norm": 2.762138605117798,
+      "learning_rate": 6.425124032463223e-06,
+      "logits/chosen": 0.03482529893517494,
+      "logits/rejected": 0.03158363327383995,
+      "logps/chosen": -84.45268249511719,
+      "logps/rejected": -91.9566650390625,
+      "loss": 2.1222,
+      "rewards/accuracies": 0.734375,
+      "rewards/chosen": -2.973424196243286,
+      "rewards/margins": 0.6969504952430725,
+      "rewards/rejected": -3.670374631881714,
+      "step": 955
+    },
+    {
+      "epoch": 1.7356948228882834,
+      "grad_norm": 2.802934169769287,
+      "learning_rate": 6.418834367615395e-06,
+      "logits/chosen": 0.09849456697702408,
+      "logits/rejected": 0.07233191281557083,
+      "logps/chosen": -85.41230010986328,
+      "logps/rejected": -92.97038269042969,
+      "loss": 2.1891,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -2.4898338317871094,
+      "rewards/margins": 0.8130300045013428,
+      "rewards/rejected": -3.3028640747070312,
+      "step": 956
+    },
+    {
+      "epoch": 1.737511353315168,
+      "grad_norm": 2.2555596828460693,
+      "learning_rate": 6.4125428291043064e-06,
+      "logits/chosen": -0.0023345965892076492,
+      "logits/rejected": 0.08920536190271378,
+      "logps/chosen": -79.65885925292969,
+      "logps/rejected": -86.8980712890625,
+      "loss": 2.0766,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.5075061321258545,
+      "rewards/margins": 0.6272554993629456,
+      "rewards/rejected": -3.1347618103027344,
+      "step": 957
+    },
+    {
+      "epoch": 1.7393278837420527,
+      "grad_norm": 3.097147226333618,
+      "learning_rate": 6.4062494297595025e-06,
+      "logits/chosen": 0.024910222738981247,
+      "logits/rejected": 0.07345539331436157,
+      "logps/chosen": -97.62705993652344,
+      "logps/rejected": -112.78961944580078,
+      "loss": 2.0315,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": -2.446077823638916,
+      "rewards/margins": 0.7760088443756104,
+      "rewards/rejected": -3.2220866680145264,
+      "step": 958
+    },
+    {
+      "epoch": 1.7411444141689372,
+      "grad_norm": 2.4640140533447266,
+      "learning_rate": 6.399954182414317e-06,
+      "logits/chosen": 0.002200646325945854,
+      "logits/rejected": 0.054388273507356644,
+      "logps/chosen": -84.92884063720703,
+      "logps/rejected": -102.98699188232422,
+      "loss": 1.9766,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -2.6910696029663086,
+      "rewards/margins": 0.8245669603347778,
+      "rewards/rejected": -3.515636444091797,
+      "step": 959
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2200,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 137,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.505440174871675e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-959/training_args.bin b/checkpoint-959/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f3be2c1b8e6fda2e41962f8c44df9def5640a19f
--- /dev/null
+++ b/checkpoint-959/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fb07809f6de56695691d7f9f9a747370262f2b5db2dfb93064fc44f1fcbe2a
+size 7736
diff --git a/checkpoint-959/zero_to_fp32.py b/checkpoint-959/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69ecd9acb5a235ffbf927091051106d902b3d39
--- /dev/null
+++ b/checkpoint-959/zero_to_fp32.py
@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/special_tokens_map.json b/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..712a88f318397d0e8cc49e023063cacfbdacdf0d
--- /dev/null
+++ b/special_tokens_map.json
@@ -0,0 +1,39 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|finetune_right_pad_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/tokenizer.json b/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2
--- /dev/null
+++ b/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/tokenizer_config.json b/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e1adcc0c23bace7262ee80c9597c65c5cf6eb99b
--- /dev/null
+++ b/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|eot_id|>",
+    "<|eom_id|>"
+  ],
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 2048,
+  "pad_token": "<|finetune_right_pad_id|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/train_results.json b/train_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..f79cbd52b02b920694c001d8856154bc13f2f095
--- /dev/null
+++ b/train_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 3.0,
+    "total_flos": 216736934658048.0,
+    "train_loss": 0.47654047750291373,
+    "train_runtime": 2544.5527,
+    "train_samples_per_second": 0.519,
+    "train_steps_per_second": 0.033
+}
\ No newline at end of file
diff --git a/trainer_log.jsonl b/trainer_log.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..05bac272f02fcdbe5e65c0605ee283416c7ed893
--- /dev/null
+++ b/trainer_log.jsonl
@@ -0,0 +1,1199 @@
+{"current_steps": 1, "total_steps": 2200, "loss": 2.7726, "accuracy": 0.0, "lr": 9.99999541185824e-06, "epoch": 0.0018165304268846503, "percentage": 0.05, "elapsed_time": "0:01:53", "remaining_time": "2 days, 21:09:11"}
+{"current_steps": 2, "total_steps": 2200, "loss": 2.7516, "accuracy": 0.578125, "lr": 9.999981647442309e-06, "epoch": 0.0036330608537693005, "percentage": 0.09, "elapsed_time": "0:04:01", "remaining_time": "3 days, 1:44:38"}
+{"current_steps": 3, "total_steps": 2200, "loss": 2.7902, "accuracy": 0.4375, "lr": 9.999958706780277e-06, "epoch": 0.005449591280653951, "percentage": 0.14, "elapsed_time": "0:05:43", "remaining_time": "2 days, 21:54:48"}
+{"current_steps": 4, "total_steps": 2200, "loss": 2.7537, "accuracy": 0.546875, "lr": 9.999926589918927e-06, "epoch": 0.007266121707538601, "percentage": 0.18, "elapsed_time": "0:07:38", "remaining_time": "2 days, 21:57:39"}
+{"current_steps": 5, "total_steps": 2200, "loss": 2.79, "accuracy": 0.453125, "lr": 9.999885296923748e-06, "epoch": 0.009082652134423252, "percentage": 0.23, "elapsed_time": "0:09:46", "remaining_time": "2 days, 23:32:12"}
+{"current_steps": 6, "total_steps": 2200, "loss": 2.7732, "accuracy": 0.5, "lr": 9.999834827878942e-06, "epoch": 0.010899182561307902, "percentage": 0.27, "elapsed_time": "0:11:45", "remaining_time": "2 days, 23:42:05"}
+{"current_steps": 7, "total_steps": 2200, "loss": 2.7716, "accuracy": 0.578125, "lr": 9.999775182887429e-06, "epoch": 0.012715712988192553, "percentage": 0.32, "elapsed_time": "0:14:05", "remaining_time": "3 days, 1:37:06"}
+{"current_steps": 8, "total_steps": 2200, "loss": 2.7544, "accuracy": 0.59375, "lr": 9.999706362070833e-06, "epoch": 0.014532243415077202, "percentage": 0.36, "elapsed_time": "0:16:10", "remaining_time": "3 days, 1:51:00"}
+{"current_steps": 9, "total_steps": 2200, "loss": 2.7688, "accuracy": 0.515625, "lr": 9.99962836556949e-06, "epoch": 0.01634877384196185, "percentage": 0.41, "elapsed_time": "0:18:04", "remaining_time": "3 days, 1:19:24"}
+{"current_steps": 10, "total_steps": 2200, "loss": 2.7827, "accuracy": 0.46875, "lr": 9.99954119354245e-06, "epoch": 0.018165304268846504, "percentage": 0.45, "elapsed_time": "0:19:56", "remaining_time": "3 days, 0:48:51"}
+{"current_steps": 11, "total_steps": 2200, "loss": 2.7565, "accuracy": 0.625, "lr": 9.999444846167473e-06, "epoch": 0.019981834695731154, "percentage": 0.5, "elapsed_time": "0:21:48", "remaining_time": "3 days, 0:20:50"}
+{"current_steps": 12, "total_steps": 2200, "loss": 2.7732, "accuracy": 0.5, "lr": 9.999339323641027e-06, "epoch": 0.021798365122615803, "percentage": 0.55, "elapsed_time": "0:23:35", "remaining_time": "2 days, 23:40:39"}
+{"current_steps": 13, "total_steps": 2200, "loss": 2.7693, "accuracy": 0.578125, "lr": 9.99922462617829e-06, "epoch": 0.023614895549500452, "percentage": 0.59, "elapsed_time": "0:25:37", "remaining_time": "2 days, 23:49:37"}
+{"current_steps": 14, "total_steps": 2200, "loss": 2.7715, "accuracy": 0.578125, "lr": 9.999100754013152e-06, "epoch": 0.025431425976385105, "percentage": 0.64, "elapsed_time": "0:27:50", "remaining_time": "3 days, 0:26:22"}
+{"current_steps": 15, "total_steps": 2200, "loss": 2.7405, "accuracy": 0.734375, "lr": 9.998967707398207e-06, "epoch": 0.027247956403269755, "percentage": 0.68, "elapsed_time": "0:29:37", "remaining_time": "2 days, 23:55:14"}
+{"current_steps": 16, "total_steps": 2200, "loss": 2.7628, "accuracy": 0.484375, "lr": 9.998825486604765e-06, "epoch": 0.029064486830154404, "percentage": 0.73, "elapsed_time": "0:31:40", "remaining_time": "3 days, 0:04:25"}
+{"current_steps": 17, "total_steps": 2200, "loss": 2.7568, "accuracy": 0.578125, "lr": 9.998674091922833e-06, "epoch": 0.030881017257039057, "percentage": 0.77, "elapsed_time": "0:33:21", "remaining_time": "2 days, 23:24:02"}
+{"current_steps": 18, "total_steps": 2200, "loss": 2.7888, "accuracy": 0.515625, "lr": 9.998513523661136e-06, "epoch": 0.0326975476839237, "percentage": 0.82, "elapsed_time": "0:35:07", "remaining_time": "2 days, 22:57:27"}
+{"current_steps": 19, "total_steps": 2200, "loss": 2.7652, "accuracy": 0.578125, "lr": 9.998343782147098e-06, "epoch": 0.03451407811080836, "percentage": 0.86, "elapsed_time": "0:36:53", "remaining_time": "2 days, 22:34:02"}
+{"current_steps": 20, "total_steps": 2200, "loss": 2.7607, "accuracy": 0.5625, "lr": 9.998164867726853e-06, "epoch": 0.03633060853769301, "percentage": 0.91, "elapsed_time": "0:38:49", "remaining_time": "2 days, 22:32:16"}
+{"current_steps": 21, "total_steps": 2200, "loss": 2.7752, "accuracy": 0.46875, "lr": 9.997976780765237e-06, "epoch": 0.03814713896457766, "percentage": 0.95, "elapsed_time": "0:41:06", "remaining_time": "2 days, 23:05:18"}
+{"current_steps": 22, "total_steps": 2200, "loss": 2.7661, "accuracy": 0.53125, "lr": 9.997779521645793e-06, "epoch": 0.03996366939146231, "percentage": 1.0, "elapsed_time": "0:42:48", "remaining_time": "2 days, 22:38:33"}
+{"current_steps": 23, "total_steps": 2200, "loss": 2.7526, "accuracy": 0.59375, "lr": 9.997573090770766e-06, "epoch": 0.04178019981834696, "percentage": 1.05, "elapsed_time": "0:45:00", "remaining_time": "2 days, 22:59:31"}
+{"current_steps": 24, "total_steps": 2200, "loss": 2.7534, "accuracy": 0.65625, "lr": 9.997357488561107e-06, "epoch": 0.043596730245231606, "percentage": 1.09, "elapsed_time": "0:47:02", "remaining_time": "2 days, 23:04:58"}
+{"current_steps": 25, "total_steps": 2200, "loss": 2.7602, "accuracy": 0.5625, "lr": 9.997132715456464e-06, "epoch": 0.045413260672116255, "percentage": 1.14, "elapsed_time": "0:49:03", "remaining_time": "2 days, 23:08:42"}
+{"current_steps": 26, "total_steps": 2200, "loss": 2.762, "accuracy": 0.578125, "lr": 9.99689877191519e-06, "epoch": 0.047229791099000905, "percentage": 1.18, "elapsed_time": "0:51:12", "remaining_time": "2 days, 23:21:34"}
+{"current_steps": 27, "total_steps": 2200, "loss": 2.7519, "accuracy": 0.5, "lr": 9.996655658414331e-06, "epoch": 0.04904632152588556, "percentage": 1.23, "elapsed_time": "0:53:19", "remaining_time": "2 days, 23:32:12"}
+{"current_steps": 28, "total_steps": 2200, "loss": 2.743, "accuracy": 0.5625, "lr": 9.996403375449647e-06, "epoch": 0.05086285195277021, "percentage": 1.27, "elapsed_time": "0:55:28", "remaining_time": "2 days, 23:43:40"}
+{"current_steps": 29, "total_steps": 2200, "loss": 2.7641, "accuracy": 0.546875, "lr": 9.996141923535582e-06, "epoch": 0.05267938237965486, "percentage": 1.32, "elapsed_time": "0:57:18", "remaining_time": "2 days, 23:30:05"}
+{"current_steps": 30, "total_steps": 2200, "loss": 2.7519, "accuracy": 0.5625, "lr": 9.995871303205282e-06, "epoch": 0.05449591280653951, "percentage": 1.36, "elapsed_time": "0:59:04", "remaining_time": "2 days, 23:13:24"}
+{"current_steps": 31, "total_steps": 2200, "loss": 2.7253, "accuracy": 0.65625, "lr": 9.995591515010589e-06, "epoch": 0.05631244323342416, "percentage": 1.41, "elapsed_time": "1:01:05", "remaining_time": "2 days, 23:14:38"}
+{"current_steps": 32, "total_steps": 2200, "loss": 2.7571, "accuracy": 0.515625, "lr": 9.99530255952204e-06, "epoch": 0.05812897366030881, "percentage": 1.45, "elapsed_time": "1:03:15", "remaining_time": "2 days, 23:25:24"}
+{"current_steps": 33, "total_steps": 2200, "loss": 2.7627, "accuracy": 0.53125, "lr": 9.995004437328866e-06, "epoch": 0.05994550408719346, "percentage": 1.5, "elapsed_time": "1:05:10", "remaining_time": "2 days, 23:19:44"}
+{"current_steps": 34, "total_steps": 2200, "loss": 2.7437, "accuracy": 0.53125, "lr": 9.994697149038989e-06, "epoch": 0.061762034514078114, "percentage": 1.55, "elapsed_time": "1:07:18", "remaining_time": "2 days, 23:27:45"}
+{"current_steps": 35, "total_steps": 2200, "loss": 2.7221, "accuracy": 0.609375, "lr": 9.994380695279025e-06, "epoch": 0.06357856494096276, "percentage": 1.59, "elapsed_time": "1:09:18", "remaining_time": "2 days, 23:26:45"}
+{"current_steps": 36, "total_steps": 2200, "loss": 2.7179, "accuracy": 0.640625, "lr": 9.994055076694276e-06, "epoch": 0.0653950953678474, "percentage": 1.64, "elapsed_time": "1:11:23", "remaining_time": "2 days, 23:31:46"}
+{"current_steps": 37, "total_steps": 2200, "loss": 2.7505, "accuracy": 0.5625, "lr": 9.993720293948739e-06, "epoch": 0.06721162579473206, "percentage": 1.68, "elapsed_time": "1:13:41", "remaining_time": "2 days, 23:48:24"}
+{"current_steps": 38, "total_steps": 2200, "loss": 2.7007, "accuracy": 0.6875, "lr": 9.993376347725091e-06, "epoch": 0.06902815622161672, "percentage": 1.73, "elapsed_time": "1:16:02", "remaining_time": "3 days, 0:06:25"}
+{"current_steps": 39, "total_steps": 2200, "loss": 2.7406, "accuracy": 0.59375, "lr": 9.993023238724696e-06, "epoch": 0.07084468664850137, "percentage": 1.77, "elapsed_time": "1:18:23", "remaining_time": "3 days, 0:24:02"}
+{"current_steps": 40, "total_steps": 2200, "loss": 2.7422, "accuracy": 0.578125, "lr": 9.99266096766761e-06, "epoch": 0.07266121707538602, "percentage": 1.82, "elapsed_time": "1:20:39", "remaining_time": "3 days, 0:35:13"}
+{"current_steps": 41, "total_steps": 2200, "loss": 2.7505, "accuracy": 0.5625, "lr": 9.992289535292565e-06, "epoch": 0.07447774750227067, "percentage": 1.86, "elapsed_time": "1:22:23", "remaining_time": "3 days, 0:18:35"}
+{"current_steps": 42, "total_steps": 2200, "loss": 2.7855, "accuracy": 0.515625, "lr": 9.991908942356977e-06, "epoch": 0.07629427792915532, "percentage": 1.91, "elapsed_time": "1:24:25", "remaining_time": "3 days, 0:17:26"}
+{"current_steps": 43, "total_steps": 2200, "loss": 2.7203, "accuracy": 0.6875, "lr": 9.991519189636937e-06, "epoch": 0.07811080835603997, "percentage": 1.95, "elapsed_time": "1:26:22", "remaining_time": "3 days, 0:12:25"}
+{"current_steps": 44, "total_steps": 2200, "loss": 2.7302, "accuracy": 0.65625, "lr": 9.991120277927224e-06, "epoch": 0.07992733878292461, "percentage": 2.0, "elapsed_time": "1:28:17", "remaining_time": "3 days, 0:06:15"}
+{"current_steps": 45, "total_steps": 2200, "loss": 2.7288, "accuracy": 0.640625, "lr": 9.990712208041284e-06, "epoch": 0.08174386920980926, "percentage": 2.05, "elapsed_time": "1:30:31", "remaining_time": "3 days, 0:15:21"}
+{"current_steps": 46, "total_steps": 2200, "loss": 2.7003, "accuracy": 0.625, "lr": 9.990294980811244e-06, "epoch": 0.08356039963669391, "percentage": 2.09, "elapsed_time": "1:32:32", "remaining_time": "3 days, 0:13:20"}
+{"current_steps": 47, "total_steps": 2200, "loss": 2.6636, "accuracy": 0.671875, "lr": 9.989868597087907e-06, "epoch": 0.08537693006357856, "percentage": 2.14, "elapsed_time": "1:34:37", "remaining_time": "3 days, 0:14:47"}
+{"current_steps": 48, "total_steps": 2200, "loss": 2.6658, "accuracy": 0.625, "lr": 9.989433057740738e-06, "epoch": 0.08719346049046321, "percentage": 2.18, "elapsed_time": "1:36:34", "remaining_time": "3 days, 0:09:49"}
+{"current_steps": 49, "total_steps": 2200, "loss": 2.6777, "accuracy": 0.609375, "lr": 9.98898836365788e-06, "epoch": 0.08900999091734786, "percentage": 2.23, "elapsed_time": "1:38:35", "remaining_time": "3 days, 0:08:09"}
+{"current_steps": 50, "total_steps": 2200, "loss": 2.68, "accuracy": 0.578125, "lr": 9.988534515746141e-06, "epoch": 0.09082652134423251, "percentage": 2.27, "elapsed_time": "1:40:39", "remaining_time": "3 days, 0:08:38"}
+{"current_steps": 51, "total_steps": 2200, "loss": 2.6104, "accuracy": 0.640625, "lr": 9.988071514930998e-06, "epoch": 0.09264305177111716, "percentage": 2.32, "elapsed_time": "1:42:25", "remaining_time": "2 days, 23:55:37"}
+{"current_steps": 52, "total_steps": 2200, "loss": 2.6207, "accuracy": 0.640625, "lr": 9.987599362156587e-06, "epoch": 0.09445958219800181, "percentage": 2.36, "elapsed_time": "1:44:17", "remaining_time": "2 days, 23:47:46"}
+{"current_steps": 53, "total_steps": 2200, "loss": 2.7093, "accuracy": 0.5, "lr": 9.987118058385712e-06, "epoch": 0.09627611262488647, "percentage": 2.41, "elapsed_time": "1:46:19", "remaining_time": "2 days, 23:47:17"}
+{"current_steps": 54, "total_steps": 2200, "loss": 2.7725, "accuracy": 0.53125, "lr": 9.986627604599835e-06, "epoch": 0.09809264305177112, "percentage": 2.45, "elapsed_time": "1:48:09", "remaining_time": "2 days, 23:38:18"}
+{"current_steps": 55, "total_steps": 2200, "loss": 2.6826, "accuracy": 0.53125, "lr": 9.986128001799077e-06, "epoch": 0.09990917347865577, "percentage": 2.5, "elapsed_time": "1:50:23", "remaining_time": "2 days, 23:45:04"}
+{"current_steps": 56, "total_steps": 2200, "loss": 2.664, "accuracy": 0.671875, "lr": 9.985619251002214e-06, "epoch": 0.10172570390554042, "percentage": 2.55, "elapsed_time": "1:52:19", "remaining_time": "2 days, 23:40:31"}
+{"current_steps": 57, "total_steps": 2200, "loss": 2.6165, "accuracy": 0.703125, "lr": 9.985101353246676e-06, "epoch": 0.10354223433242507, "percentage": 2.59, "elapsed_time": "1:54:24", "remaining_time": "2 days, 23:41:35"}
+{"current_steps": 58, "total_steps": 2200, "loss": 2.551, "accuracy": 0.734375, "lr": 9.98457430958855e-06, "epoch": 0.10535876475930972, "percentage": 2.64, "elapsed_time": "1:56:19", "remaining_time": "2 days, 23:36:02"}
+{"current_steps": 59, "total_steps": 2200, "loss": 2.6345, "accuracy": 0.671875, "lr": 9.984038121102569e-06, "epoch": 0.10717529518619437, "percentage": 2.68, "elapsed_time": "1:58:08", "remaining_time": "2 days, 23:27:15"}
+{"current_steps": 60, "total_steps": 2200, "loss": 2.7099, "accuracy": 0.546875, "lr": 9.983492788882113e-06, "epoch": 0.10899182561307902, "percentage": 2.73, "elapsed_time": "1:59:44", "remaining_time": "2 days, 23:10:32"}
+{"current_steps": 61, "total_steps": 2200, "loss": 2.6373, "accuracy": 0.65625, "lr": 9.98293831403921e-06, "epoch": 0.11080835603996367, "percentage": 2.77, "elapsed_time": "2:01:43", "remaining_time": "2 days, 23:08:09"}
+{"current_steps": 62, "total_steps": 2200, "loss": 2.5465, "accuracy": 0.625, "lr": 9.982374697704532e-06, "epoch": 0.11262488646684832, "percentage": 2.82, "elapsed_time": "2:03:53", "remaining_time": "2 days, 23:12:16"}
+{"current_steps": 63, "total_steps": 2200, "loss": 2.5876, "accuracy": 0.6875, "lr": 9.981801941027388e-06, "epoch": 0.11444141689373297, "percentage": 2.86, "elapsed_time": "2:05:48", "remaining_time": "2 days, 23:07:40"}
+{"current_steps": 64, "total_steps": 2200, "loss": 2.7096, "accuracy": 0.609375, "lr": 9.981220045175731e-06, "epoch": 0.11625794732061762, "percentage": 2.91, "elapsed_time": "2:07:48", "remaining_time": "2 days, 23:05:48"}
+{"current_steps": 65, "total_steps": 2200, "loss": 2.5447, "accuracy": 0.59375, "lr": 9.980629011336149e-06, "epoch": 0.11807447774750227, "percentage": 2.95, "elapsed_time": "2:09:46", "remaining_time": "2 days, 23:02:52"}
+{"current_steps": 66, "total_steps": 2200, "loss": 2.6013, "accuracy": 0.640625, "lr": 9.98002884071386e-06, "epoch": 0.11989100817438691, "percentage": 3.0, "elapsed_time": "2:12:05", "remaining_time": "2 days, 23:11:04"}
+{"current_steps": 67, "total_steps": 2200, "loss": 2.6008, "accuracy": 0.640625, "lr": 9.97941953453272e-06, "epoch": 0.12170753860127158, "percentage": 3.05, "elapsed_time": "2:14:01", "remaining_time": "2 days, 23:06:34"}
+{"current_steps": 68, "total_steps": 2200, "loss": 2.6714, "accuracy": 0.640625, "lr": 9.978801094035207e-06, "epoch": 0.12352406902815623, "percentage": 3.09, "elapsed_time": "2:16:02", "remaining_time": "2 days, 23:05:13"}
+{"current_steps": 69, "total_steps": 2200, "loss": 2.6884, "accuracy": 0.515625, "lr": 9.978173520482429e-06, "epoch": 0.12534059945504086, "percentage": 3.14, "elapsed_time": "2:17:53", "remaining_time": "2 days, 22:58:49"}
+{"current_steps": 70, "total_steps": 2200, "loss": 2.5226, "accuracy": 0.640625, "lr": 9.97753681515412e-06, "epoch": 0.1271571298819255, "percentage": 3.18, "elapsed_time": "2:19:56", "remaining_time": "2 days, 22:58:25"}
+{"current_steps": 71, "total_steps": 2200, "loss": 2.6945, "accuracy": 0.546875, "lr": 9.976890979348628e-06, "epoch": 0.12897366030881016, "percentage": 3.23, "elapsed_time": "2:21:52", "remaining_time": "2 days, 22:54:03"}
+{"current_steps": 72, "total_steps": 2200, "loss": 2.7689, "accuracy": 0.546875, "lr": 9.976236014382934e-06, "epoch": 0.1307901907356948, "percentage": 3.27, "elapsed_time": "2:23:43", "remaining_time": "2 days, 22:47:53"}
+{"current_steps": 73, "total_steps": 2200, "loss": 2.5859, "accuracy": 0.59375, "lr": 9.975571921592618e-06, "epoch": 0.13260672116257946, "percentage": 3.32, "elapsed_time": "2:25:59", "remaining_time": "2 days, 22:53:41"}
+{"current_steps": 74, "total_steps": 2200, "loss": 2.4062, "accuracy": 0.671875, "lr": 9.97489870233188e-06, "epoch": 0.1344232515894641, "percentage": 3.36, "elapsed_time": "2:27:44", "remaining_time": "2 days, 22:44:48"}
+{"current_steps": 75, "total_steps": 2200, "loss": 2.8556, "accuracy": 0.5625, "lr": 9.974216357973539e-06, "epoch": 0.1362397820163488, "percentage": 3.41, "elapsed_time": "2:29:56", "remaining_time": "2 days, 22:48:07"}
+{"current_steps": 76, "total_steps": 2200, "loss": 2.5576, "accuracy": 0.671875, "lr": 9.973524889909007e-06, "epoch": 0.13805631244323344, "percentage": 3.45, "elapsed_time": "2:32:00", "remaining_time": "2 days, 22:48:10"}
+{"current_steps": 77, "total_steps": 2200, "loss": 2.7845, "accuracy": 0.453125, "lr": 9.97282429954831e-06, "epoch": 0.13987284287011809, "percentage": 3.5, "elapsed_time": "2:34:03", "remaining_time": "2 days, 22:47:48"}
+{"current_steps": 78, "total_steps": 2200, "loss": 2.5834, "accuracy": 0.6875, "lr": 9.972114588320073e-06, "epoch": 0.14168937329700274, "percentage": 3.55, "elapsed_time": "2:36:26", "remaining_time": "2 days, 22:56:06"}
+{"current_steps": 79, "total_steps": 2200, "loss": 2.4955, "accuracy": 0.671875, "lr": 9.97139575767152e-06, "epoch": 0.14350590372388738, "percentage": 3.59, "elapsed_time": "2:38:22", "remaining_time": "2 days, 22:51:51"}
+{"current_steps": 80, "total_steps": 2200, "loss": 2.5763, "accuracy": 0.6875, "lr": 9.970667809068476e-06, "epoch": 0.14532243415077203, "percentage": 3.64, "elapsed_time": "2:40:48", "remaining_time": "2 days, 23:01:19"}
+{"current_steps": 81, "total_steps": 2200, "loss": 2.6829, "accuracy": 0.578125, "lr": 9.969930743995351e-06, "epoch": 0.14713896457765668, "percentage": 3.68, "elapsed_time": "2:42:37", "remaining_time": "2 days, 22:54:22"}
+{"current_steps": 82, "total_steps": 2200, "loss": 2.6277, "accuracy": 0.609375, "lr": 9.969184563955152e-06, "epoch": 0.14895549500454133, "percentage": 3.73, "elapsed_time": "2:44:58", "remaining_time": "2 days, 23:01:01"}
+{"current_steps": 83, "total_steps": 2200, "loss": 2.4821, "accuracy": 0.59375, "lr": 9.968429270469467e-06, "epoch": 0.15077202543142598, "percentage": 3.77, "elapsed_time": "2:46:55", "remaining_time": "2 days, 22:57:33"}
+{"current_steps": 84, "total_steps": 2200, "loss": 2.6289, "accuracy": 0.609375, "lr": 9.967664865078472e-06, "epoch": 0.15258855585831063, "percentage": 3.82, "elapsed_time": "2:49:05", "remaining_time": "2 days, 22:59:39"}
+{"current_steps": 85, "total_steps": 2200, "loss": 2.7702, "accuracy": 0.59375, "lr": 9.966891349340922e-06, "epoch": 0.15440508628519528, "percentage": 3.86, "elapsed_time": "2:51:04", "remaining_time": "2 days, 22:56:36"}
+{"current_steps": 86, "total_steps": 2200, "loss": 2.5889, "accuracy": 0.65625, "lr": 9.966108724834151e-06, "epoch": 0.15622161671207993, "percentage": 3.91, "elapsed_time": "2:53:07", "remaining_time": "2 days, 22:55:31"}
+{"current_steps": 87, "total_steps": 2200, "loss": 2.5522, "accuracy": 0.6875, "lr": 9.965316993154069e-06, "epoch": 0.15803814713896458, "percentage": 3.95, "elapsed_time": "2:55:15", "remaining_time": "2 days, 22:56:37"}
+{"current_steps": 88, "total_steps": 2200, "loss": 2.7432, "accuracy": 0.53125, "lr": 9.964516155915152e-06, "epoch": 0.15985467756584923, "percentage": 4.0, "elapsed_time": "2:57:44", "remaining_time": "2 days, 23:05:41"}
+{"current_steps": 89, "total_steps": 2200, "loss": 2.5472, "accuracy": 0.640625, "lr": 9.963706214750446e-06, "epoch": 0.16167120799273388, "percentage": 4.05, "elapsed_time": "2:59:55", "remaining_time": "2 days, 23:07:32"}
+{"current_steps": 90, "total_steps": 2200, "loss": 2.5339, "accuracy": 0.5625, "lr": 9.962887171311563e-06, "epoch": 0.16348773841961853, "percentage": 4.09, "elapsed_time": "3:01:37", "remaining_time": "2 days, 22:58:08"}
+{"current_steps": 91, "total_steps": 2200, "loss": 2.5165, "accuracy": 0.640625, "lr": 9.962059027268676e-06, "epoch": 0.16530426884650318, "percentage": 4.14, "elapsed_time": "3:03:35", "remaining_time": "2 days, 22:54:50"}
+{"current_steps": 92, "total_steps": 2200, "loss": 2.4254, "accuracy": 0.71875, "lr": 9.961221784310514e-06, "epoch": 0.16712079927338783, "percentage": 4.18, "elapsed_time": "3:05:31", "remaining_time": "2 days, 22:50:46"}
+{"current_steps": 93, "total_steps": 2200, "loss": 2.7543, "accuracy": 0.578125, "lr": 9.96037544414436e-06, "epoch": 0.16893732970027248, "percentage": 4.23, "elapsed_time": "3:07:51", "remaining_time": "2 days, 22:56:08"}
+{"current_steps": 94, "total_steps": 2200, "loss": 2.7134, "accuracy": 0.625, "lr": 9.959520008496054e-06, "epoch": 0.17075386012715713, "percentage": 4.27, "elapsed_time": "3:09:24", "remaining_time": "2 days, 22:43:29"}
+{"current_steps": 95, "total_steps": 2200, "loss": 2.5031, "accuracy": 0.6875, "lr": 9.95865547910997e-06, "epoch": 0.17257039055404177, "percentage": 4.32, "elapsed_time": "3:11:17", "remaining_time": "2 days, 22:38:38"}
+{"current_steps": 96, "total_steps": 2200, "loss": 2.7239, "accuracy": 0.578125, "lr": 9.957781857749043e-06, "epoch": 0.17438692098092642, "percentage": 4.36, "elapsed_time": "3:13:02", "remaining_time": "2 days, 22:30:47"}
+{"current_steps": 97, "total_steps": 2200, "loss": 2.7076, "accuracy": 0.6875, "lr": 9.956899146194732e-06, "epoch": 0.17620345140781107, "percentage": 4.41, "elapsed_time": "3:14:59", "remaining_time": "2 days, 22:27:26"}
+{"current_steps": 98, "total_steps": 2200, "loss": 2.4996, "accuracy": 0.671875, "lr": 9.956007346247043e-06, "epoch": 0.17801998183469572, "percentage": 4.45, "elapsed_time": "3:16:39", "remaining_time": "2 days, 22:18:17"}
+{"current_steps": 99, "total_steps": 2200, "loss": 2.5677, "accuracy": 0.65625, "lr": 9.95510645972451e-06, "epoch": 0.17983651226158037, "percentage": 4.5, "elapsed_time": "3:18:24", "remaining_time": "2 days, 22:10:43"}
+{"current_steps": 100, "total_steps": 2200, "loss": 2.7065, "accuracy": 0.5625, "lr": 9.954196488464198e-06, "epoch": 0.18165304268846502, "percentage": 4.55, "elapsed_time": "3:20:11", "remaining_time": "2 days, 22:03:59"}
+{"current_steps": 101, "total_steps": 2200, "loss": 2.5346, "accuracy": 0.671875, "lr": 9.953277434321696e-06, "epoch": 0.18346957311534967, "percentage": 4.59, "elapsed_time": "3:22:36", "remaining_time": "2 days, 22:10:38"}
+{"current_steps": 102, "total_steps": 2200, "loss": 2.4325, "accuracy": 0.6875, "lr": 9.952349299171117e-06, "epoch": 0.18528610354223432, "percentage": 4.64, "elapsed_time": "3:24:51", "remaining_time": "2 days, 22:13:29"}
+{"current_steps": 103, "total_steps": 2200, "loss": 2.3004, "accuracy": 0.671875, "lr": 9.95141208490509e-06, "epoch": 0.18710263396911897, "percentage": 4.68, "elapsed_time": "3:26:38", "remaining_time": "2 days, 22:07:13"}
+{"current_steps": 104, "total_steps": 2200, "loss": 2.5388, "accuracy": 0.625, "lr": 9.950465793434759e-06, "epoch": 0.18891916439600362, "percentage": 4.73, "elapsed_time": "3:28:49", "remaining_time": "2 days, 22:08:28"}
+{"current_steps": 105, "total_steps": 2200, "loss": 2.6443, "accuracy": 0.546875, "lr": 9.949510426689773e-06, "epoch": 0.1907356948228883, "percentage": 4.77, "elapsed_time": "3:30:40", "remaining_time": "2 days, 22:03:21"}
+{"current_steps": 106, "total_steps": 2200, "loss": 2.8494, "accuracy": 0.609375, "lr": 9.948545986618295e-06, "epoch": 0.19255222524977295, "percentage": 4.82, "elapsed_time": "3:32:18", "remaining_time": "2 days, 21:53:57"}
+{"current_steps": 107, "total_steps": 2200, "loss": 2.46, "accuracy": 0.6875, "lr": 9.947572475186984e-06, "epoch": 0.1943687556766576, "percentage": 4.86, "elapsed_time": "3:34:16", "remaining_time": "2 days, 21:51:25"}
+{"current_steps": 108, "total_steps": 2200, "loss": 2.6472, "accuracy": 0.609375, "lr": 9.946589894381002e-06, "epoch": 0.19618528610354224, "percentage": 4.91, "elapsed_time": "3:36:33", "remaining_time": "2 days, 21:54:45"}
+{"current_steps": 109, "total_steps": 2200, "loss": 2.7037, "accuracy": 0.5, "lr": 9.945598246204e-06, "epoch": 0.1980018165304269, "percentage": 4.95, "elapsed_time": "3:38:26", "remaining_time": "2 days, 21:50:28"}
+{"current_steps": 110, "total_steps": 2200, "loss": 2.6224, "accuracy": 0.5625, "lr": 9.94459753267812e-06, "epoch": 0.19981834695731154, "percentage": 5.0, "elapsed_time": "3:40:36", "remaining_time": "2 days, 21:51:37"}
+{"current_steps": 111, "total_steps": 2200, "loss": 2.6102, "accuracy": 0.640625, "lr": 9.943587755843996e-06, "epoch": 0.2016348773841962, "percentage": 5.05, "elapsed_time": "3:42:17", "remaining_time": "2 days, 21:43:30"}
+{"current_steps": 112, "total_steps": 2200, "loss": 2.4897, "accuracy": 0.6875, "lr": 9.942568917760733e-06, "epoch": 0.20345140781108084, "percentage": 5.09, "elapsed_time": "3:44:02", "remaining_time": "2 days, 21:36:48"}
+{"current_steps": 113, "total_steps": 2200, "loss": 2.5037, "accuracy": 0.703125, "lr": 9.941541020505924e-06, "epoch": 0.2052679382379655, "percentage": 5.14, "elapsed_time": "3:45:59", "remaining_time": "2 days, 21:33:52"}
+{"current_steps": 114, "total_steps": 2200, "loss": 2.6328, "accuracy": 0.5, "lr": 9.940504066175626e-06, "epoch": 0.20708446866485014, "percentage": 5.18, "elapsed_time": "3:47:55", "remaining_time": "2 days, 21:30:38"}
+{"current_steps": 115, "total_steps": 2200, "loss": 2.5752, "accuracy": 0.625, "lr": 9.939458056884375e-06, "epoch": 0.2089009990917348, "percentage": 5.23, "elapsed_time": "3:49:41", "remaining_time": "2 days, 21:24:16"}
+{"current_steps": 116, "total_steps": 2200, "loss": 2.6792, "accuracy": 0.625, "lr": 9.938402994765163e-06, "epoch": 0.21071752951861944, "percentage": 5.27, "elapsed_time": "3:51:45", "remaining_time": "2 days, 21:23:46"}
+{"current_steps": 117, "total_steps": 2200, "loss": 2.6029, "accuracy": 0.65625, "lr": 9.937338881969444e-06, "epoch": 0.2125340599455041, "percentage": 5.32, "elapsed_time": "3:53:52", "remaining_time": "2 days, 21:23:52"}
+{"current_steps": 118, "total_steps": 2200, "loss": 2.4687, "accuracy": 0.578125, "lr": 9.93626572066713e-06, "epoch": 0.21435059037238874, "percentage": 5.36, "elapsed_time": "3:55:49", "remaining_time": "2 days, 21:20:57"}
+{"current_steps": 119, "total_steps": 2200, "loss": 2.6288, "accuracy": 0.5625, "lr": 9.935183513046585e-06, "epoch": 0.2161671207992734, "percentage": 5.41, "elapsed_time": "3:58:11", "remaining_time": "2 days, 21:25:22"}
+{"current_steps": 120, "total_steps": 2200, "loss": 2.4658, "accuracy": 0.65625, "lr": 9.93409226131462e-06, "epoch": 0.21798365122615804, "percentage": 5.45, "elapsed_time": "4:00:31", "remaining_time": "2 days, 21:29:02"}
+{"current_steps": 121, "total_steps": 2200, "loss": 2.579, "accuracy": 0.6875, "lr": 9.932991967696484e-06, "epoch": 0.2198001816530427, "percentage": 5.5, "elapsed_time": "4:02:22", "remaining_time": "2 days, 21:24:26"}
+{"current_steps": 122, "total_steps": 2200, "loss": 2.6368, "accuracy": 0.609375, "lr": 9.931882634435868e-06, "epoch": 0.22161671207992734, "percentage": 5.55, "elapsed_time": "4:04:33", "remaining_time": "2 days, 21:25:22"}
+{"current_steps": 123, "total_steps": 2200, "loss": 2.4417, "accuracy": 0.703125, "lr": 9.930764263794898e-06, "epoch": 0.22343324250681199, "percentage": 5.59, "elapsed_time": "4:06:43", "remaining_time": "2 days, 21:26:06"}
+{"current_steps": 124, "total_steps": 2200, "loss": 2.6023, "accuracy": 0.6875, "lr": 9.929636858054124e-06, "epoch": 0.22524977293369663, "percentage": 5.64, "elapsed_time": "4:08:47", "remaining_time": "2 days, 21:25:16"}
+{"current_steps": 125, "total_steps": 2200, "loss": 2.4589, "accuracy": 0.71875, "lr": 9.928500419512522e-06, "epoch": 0.22706630336058128, "percentage": 5.68, "elapsed_time": "4:11:08", "remaining_time": "2 days, 21:28:50"}
+{"current_steps": 126, "total_steps": 2200, "loss": 2.4839, "accuracy": 0.6875, "lr": 9.927354950487491e-06, "epoch": 0.22888283378746593, "percentage": 5.73, "elapsed_time": "4:13:07", "remaining_time": "2 days, 21:26:38"}
+{"current_steps": 127, "total_steps": 2200, "loss": 2.5198, "accuracy": 0.625, "lr": 9.926200453314842e-06, "epoch": 0.23069936421435058, "percentage": 5.77, "elapsed_time": "4:15:11", "remaining_time": "2 days, 21:25:21"}
+{"current_steps": 128, "total_steps": 2200, "loss": 2.3547, "accuracy": 0.6875, "lr": 9.925036930348793e-06, "epoch": 0.23251589464123523, "percentage": 5.82, "elapsed_time": "4:17:09", "remaining_time": "2 days, 21:22:44"}
+{"current_steps": 129, "total_steps": 2200, "loss": 2.631, "accuracy": 0.609375, "lr": 9.92386438396197e-06, "epoch": 0.23433242506811988, "percentage": 5.86, "elapsed_time": "4:19:08", "remaining_time": "2 days, 21:20:20"}
+{"current_steps": 130, "total_steps": 2200, "loss": 2.4497, "accuracy": 0.65625, "lr": 9.9226828165454e-06, "epoch": 0.23614895549500453, "percentage": 5.91, "elapsed_time": "4:21:16", "remaining_time": "2 days, 21:20:22"}
+{"current_steps": 131, "total_steps": 2200, "loss": 2.4612, "accuracy": 0.734375, "lr": 9.921492230508507e-06, "epoch": 0.23796548592188918, "percentage": 5.95, "elapsed_time": "4:23:00", "remaining_time": "2 days, 21:14:02"}
+{"current_steps": 132, "total_steps": 2200, "loss": 2.4164, "accuracy": 0.65625, "lr": 9.9202926282791e-06, "epoch": 0.23978201634877383, "percentage": 6.0, "elapsed_time": "4:25:14", "remaining_time": "2 days, 21:15:33"}
+{"current_steps": 133, "total_steps": 2200, "loss": 2.7243, "accuracy": 0.609375, "lr": 9.919084012303378e-06, "epoch": 0.24159854677565848, "percentage": 6.05, "elapsed_time": "4:27:41", "remaining_time": "2 days, 21:20:11"}
+{"current_steps": 134, "total_steps": 2200, "loss": 2.8086, "accuracy": 0.609375, "lr": 9.917866385045918e-06, "epoch": 0.24341507720254316, "percentage": 6.09, "elapsed_time": "4:29:39", "remaining_time": "2 days, 21:17:32"}
+{"current_steps": 135, "total_steps": 2200, "loss": 2.4774, "accuracy": 0.65625, "lr": 9.916639748989677e-06, "epoch": 0.2452316076294278, "percentage": 6.14, "elapsed_time": "4:31:54", "remaining_time": "2 days, 21:19:08"}
+{"current_steps": 136, "total_steps": 2200, "loss": 2.4399, "accuracy": 0.65625, "lr": 9.915404106635979e-06, "epoch": 0.24704813805631246, "percentage": 6.18, "elapsed_time": "4:33:50", "remaining_time": "2 days, 21:15:51"}
+{"current_steps": 137, "total_steps": 2200, "loss": 2.5281, "accuracy": 0.65625, "lr": 9.914159460504512e-06, "epoch": 0.2488646684831971, "percentage": 6.23, "elapsed_time": "4:35:47", "remaining_time": "2 days, 21:12:56"}
+{"current_steps": 138, "total_steps": 2200, "loss": 2.5718, "accuracy": 0.671875, "lr": 9.912905813133325e-06, "epoch": 0.2506811989100817, "percentage": 6.27, "elapsed_time": "4:38:49", "remaining_time": "2 days, 21:26:14"}
+{"current_steps": 139, "total_steps": 2200, "loss": 2.5334, "accuracy": 0.59375, "lr": 9.911643167078827e-06, "epoch": 0.2524977293369664, "percentage": 6.32, "elapsed_time": "4:40:56", "remaining_time": "2 days, 21:25:32"}
+{"current_steps": 140, "total_steps": 2200, "loss": 2.454, "accuracy": 0.625, "lr": 9.91037152491577e-06, "epoch": 0.254314259763851, "percentage": 6.36, "elapsed_time": "4:43:15", "remaining_time": "2 days, 21:28:03"}
+{"current_steps": 141, "total_steps": 2200, "loss": 2.4511, "accuracy": 0.640625, "lr": 9.909090889237257e-06, "epoch": 0.2561307901907357, "percentage": 6.41, "elapsed_time": "4:44:59", "remaining_time": "2 days, 21:21:44"}
+{"current_steps": 142, "total_steps": 2200, "loss": 2.4476, "accuracy": 0.65625, "lr": 9.907801262654725e-06, "epoch": 0.2579473206176203, "percentage": 6.45, "elapsed_time": "4:47:04", "remaining_time": "2 days, 21:20:40"}
+{"current_steps": 143, "total_steps": 2200, "loss": 2.5982, "accuracy": 0.515625, "lr": 9.906502647797946e-06, "epoch": 0.259763851044505, "percentage": 6.5, "elapsed_time": "4:48:56", "remaining_time": "2 days, 21:16:24"}
+{"current_steps": 144, "total_steps": 2200, "loss": 2.8244, "accuracy": 0.53125, "lr": 9.905195047315024e-06, "epoch": 0.2615803814713896, "percentage": 6.55, "elapsed_time": "4:51:22", "remaining_time": "2 days, 21:20:15"}
+{"current_steps": 145, "total_steps": 2200, "loss": 2.5832, "accuracy": 0.625, "lr": 9.903878463872384e-06, "epoch": 0.2633969118982743, "percentage": 6.59, "elapsed_time": "4:53:21", "remaining_time": "2 days, 21:17:29"}
+{"current_steps": 146, "total_steps": 2200, "loss": 2.3888, "accuracy": 0.65625, "lr": 9.902552900154769e-06, "epoch": 0.2652134423251589, "percentage": 6.64, "elapsed_time": "4:55:17", "remaining_time": "2 days, 21:14:16"}
+{"current_steps": 147, "total_steps": 2200, "loss": 2.5421, "accuracy": 0.609375, "lr": 9.90121835886523e-06, "epoch": 0.2670299727520436, "percentage": 6.68, "elapsed_time": "4:57:34", "remaining_time": "2 days, 21:15:49"}
+{"current_steps": 148, "total_steps": 2200, "loss": 2.5082, "accuracy": 0.6875, "lr": 9.899874842725136e-06, "epoch": 0.2688465031789282, "percentage": 6.73, "elapsed_time": "4:59:27", "remaining_time": "2 days, 21:11:56"}
+{"current_steps": 149, "total_steps": 2200, "loss": 2.4807, "accuracy": 0.65625, "lr": 9.898522354474144e-06, "epoch": 0.2706630336058129, "percentage": 6.77, "elapsed_time": "5:01:20", "remaining_time": "2 days, 21:07:56"}
+{"current_steps": 150, "total_steps": 2200, "loss": 2.6237, "accuracy": 0.65625, "lr": 9.897160896870217e-06, "epoch": 0.2724795640326976, "percentage": 6.82, "elapsed_time": "5:03:10", "remaining_time": "2 days, 21:03:22"}
+{"current_steps": 151, "total_steps": 2200, "loss": 2.7214, "accuracy": 0.59375, "lr": 9.895790472689605e-06, "epoch": 0.2742960944595822, "percentage": 6.86, "elapsed_time": "5:04:59", "remaining_time": "2 days, 20:58:35"}
+{"current_steps": 152, "total_steps": 2200, "loss": 2.5094, "accuracy": 0.671875, "lr": 9.894411084726837e-06, "epoch": 0.2761126248864669, "percentage": 6.91, "elapsed_time": "5:06:58", "remaining_time": "2 days, 20:56:00"}
+{"current_steps": 153, "total_steps": 2200, "loss": 2.4789, "accuracy": 0.5625, "lr": 9.893022735794728e-06, "epoch": 0.2779291553133515, "percentage": 6.95, "elapsed_time": "5:09:00", "remaining_time": "2 days, 20:54:13"}
+{"current_steps": 154, "total_steps": 2200, "loss": 2.2804, "accuracy": 0.75, "lr": 9.891625428724365e-06, "epoch": 0.27974568574023617, "percentage": 7.0, "elapsed_time": "5:10:56", "remaining_time": "2 days, 20:51:00"}
+{"current_steps": 155, "total_steps": 2200, "loss": 2.5929, "accuracy": 0.65625, "lr": 9.890219166365097e-06, "epoch": 0.2815622161671208, "percentage": 7.05, "elapsed_time": "5:13:20", "remaining_time": "2 days, 20:53:59"}
+{"current_steps": 156, "total_steps": 2200, "loss": 2.3491, "accuracy": 0.734375, "lr": 9.888803951584537e-06, "epoch": 0.28337874659400547, "percentage": 7.09, "elapsed_time": "5:15:36", "remaining_time": "2 days, 20:55:18"}
+{"current_steps": 157, "total_steps": 2200, "loss": 2.6179, "accuracy": 0.609375, "lr": 9.887379787268558e-06, "epoch": 0.2851952770208901, "percentage": 7.14, "elapsed_time": "5:17:43", "remaining_time": "2 days, 20:54:34"}
+{"current_steps": 158, "total_steps": 2200, "loss": 2.5262, "accuracy": 0.625, "lr": 9.885946676321279e-06, "epoch": 0.28701180744777477, "percentage": 7.18, "elapsed_time": "5:19:42", "remaining_time": "2 days, 20:51:57"}
+{"current_steps": 159, "total_steps": 2200, "loss": 2.3984, "accuracy": 0.671875, "lr": 9.884504621665059e-06, "epoch": 0.2888283378746594, "percentage": 7.23, "elapsed_time": "5:21:44", "remaining_time": "2 days, 20:50:06"}
+{"current_steps": 160, "total_steps": 2200, "loss": 3.0326, "accuracy": 0.484375, "lr": 9.883053626240503e-06, "epoch": 0.29064486830154407, "percentage": 7.27, "elapsed_time": "5:24:05", "remaining_time": "2 days, 20:52:15"}
+{"current_steps": 161, "total_steps": 2200, "loss": 3.1665, "accuracy": 0.546875, "lr": 9.881593693006438e-06, "epoch": 0.2924613987284287, "percentage": 7.32, "elapsed_time": "5:26:06", "remaining_time": "2 days, 20:50:00"}
+{"current_steps": 162, "total_steps": 2200, "loss": 2.8382, "accuracy": 0.578125, "lr": 9.880124824939927e-06, "epoch": 0.29427792915531337, "percentage": 7.36, "elapsed_time": "5:27:56", "remaining_time": "2 days, 20:45:39"}
+{"current_steps": 163, "total_steps": 2200, "loss": 2.3183, "accuracy": 0.703125, "lr": 9.878647025036245e-06, "epoch": 0.296094459582198, "percentage": 7.41, "elapsed_time": "5:30:20", "remaining_time": "2 days, 20:48:10"}
+{"current_steps": 164, "total_steps": 2200, "loss": 2.3323, "accuracy": 0.75, "lr": 9.877160296308886e-06, "epoch": 0.29791099000908267, "percentage": 7.45, "elapsed_time": "5:32:10", "remaining_time": "2 days, 20:43:54"}
+{"current_steps": 165, "total_steps": 2200, "loss": 2.4058, "accuracy": 0.65625, "lr": 9.875664641789545e-06, "epoch": 0.2997275204359673, "percentage": 7.5, "elapsed_time": "5:34:03", "remaining_time": "2 days, 20:40:01"}
+{"current_steps": 166, "total_steps": 2200, "loss": 2.2759, "accuracy": 0.703125, "lr": 9.874160064528124e-06, "epoch": 0.30154405086285196, "percentage": 7.55, "elapsed_time": "5:35:37", "remaining_time": "2 days, 20:32:25"}
+{"current_steps": 167, "total_steps": 2200, "loss": 2.368, "accuracy": 0.65625, "lr": 9.872646567592719e-06, "epoch": 0.3033605812897366, "percentage": 7.59, "elapsed_time": "5:37:43", "remaining_time": "2 days, 20:31:24"}
+{"current_steps": 168, "total_steps": 2200, "loss": 2.5178, "accuracy": 0.640625, "lr": 9.871124154069613e-06, "epoch": 0.30517711171662126, "percentage": 7.64, "elapsed_time": "5:39:39", "remaining_time": "2 days, 20:28:13"}
+{"current_steps": 169, "total_steps": 2200, "loss": 2.7762, "accuracy": 0.578125, "lr": 9.86959282706327e-06, "epoch": 0.3069936421435059, "percentage": 7.68, "elapsed_time": "5:41:59", "remaining_time": "2 days, 20:29:59"}
+{"current_steps": 170, "total_steps": 2200, "loss": 2.419, "accuracy": 0.59375, "lr": 9.868052589696337e-06, "epoch": 0.30881017257039056, "percentage": 7.73, "elapsed_time": "5:43:50", "remaining_time": "2 days, 20:25:46"}
+{"current_steps": 171, "total_steps": 2200, "loss": 2.3475, "accuracy": 0.71875, "lr": 9.866503445109621e-06, "epoch": 0.3106267029972752, "percentage": 7.77, "elapsed_time": "5:45:34", "remaining_time": "2 days, 20:20:20"}
+{"current_steps": 172, "total_steps": 2200, "loss": 2.6719, "accuracy": 0.546875, "lr": 9.864945396462101e-06, "epoch": 0.31244323342415986, "percentage": 7.82, "elapsed_time": "5:47:59", "remaining_time": "2 days, 20:23:01"}
+{"current_steps": 173, "total_steps": 2200, "loss": 2.4931, "accuracy": 0.671875, "lr": 9.86337844693091e-06, "epoch": 0.3142597638510445, "percentage": 7.86, "elapsed_time": "5:49:47", "remaining_time": "2 days, 20:18:23"}
+{"current_steps": 174, "total_steps": 2200, "loss": 2.4972, "accuracy": 0.625, "lr": 9.861802599711329e-06, "epoch": 0.31607629427792916, "percentage": 7.91, "elapsed_time": "5:51:50", "remaining_time": "2 days, 20:16:40"}
+{"current_steps": 175, "total_steps": 2200, "loss": 2.4456, "accuracy": 0.6875, "lr": 9.860217858016783e-06, "epoch": 0.3178928247048138, "percentage": 7.95, "elapsed_time": "5:53:52", "remaining_time": "2 days, 20:14:53"}
+{"current_steps": 176, "total_steps": 2200, "loss": 2.5003, "accuracy": 0.65625, "lr": 9.858624225078841e-06, "epoch": 0.31970935513169846, "percentage": 8.0, "elapsed_time": "5:56:11", "remaining_time": "2 days, 20:16:12"}
+{"current_steps": 177, "total_steps": 2200, "loss": 2.5589, "accuracy": 0.640625, "lr": 9.857021704147195e-06, "epoch": 0.3215258855585831, "percentage": 8.05, "elapsed_time": "5:58:21", "remaining_time": "2 days, 20:15:54"}
+{"current_steps": 178, "total_steps": 2200, "loss": 2.4595, "accuracy": 0.625, "lr": 9.855410298489663e-06, "epoch": 0.32334241598546776, "percentage": 8.09, "elapsed_time": "6:00:32", "remaining_time": "2 days, 20:15:36"}
+{"current_steps": 179, "total_steps": 2200, "loss": 2.2606, "accuracy": 0.65625, "lr": 9.853790011392186e-06, "epoch": 0.32515894641235243, "percentage": 8.14, "elapsed_time": "6:02:37", "remaining_time": "2 days, 20:14:15"}
+{"current_steps": 180, "total_steps": 2200, "loss": 2.4039, "accuracy": 0.671875, "lr": 9.852160846158808e-06, "epoch": 0.32697547683923706, "percentage": 8.18, "elapsed_time": "6:04:37", "remaining_time": "2 days, 20:11:50"}
+{"current_steps": 181, "total_steps": 2200, "loss": 2.5716, "accuracy": 0.546875, "lr": 9.850522806111681e-06, "epoch": 0.32879200726612173, "percentage": 8.23, "elapsed_time": "6:06:41", "remaining_time": "2 days, 20:10:24"}
+{"current_steps": 182, "total_steps": 2200, "loss": 2.2623, "accuracy": 0.703125, "lr": 9.848875894591055e-06, "epoch": 0.33060853769300635, "percentage": 8.27, "elapsed_time": "6:08:35", "remaining_time": "2 days, 20:06:59"}
+{"current_steps": 183, "total_steps": 2200, "loss": 2.348, "accuracy": 0.671875, "lr": 9.847220114955269e-06, "epoch": 0.33242506811989103, "percentage": 8.32, "elapsed_time": "6:10:28", "remaining_time": "2 days, 20:03:15"}
+{"current_steps": 184, "total_steps": 2200, "loss": 2.4969, "accuracy": 0.5625, "lr": 9.845555470580746e-06, "epoch": 0.33424159854677565, "percentage": 8.36, "elapsed_time": "6:12:27", "remaining_time": "2 days, 20:00:50"}
+{"current_steps": 185, "total_steps": 2200, "loss": 2.5595, "accuracy": 0.671875, "lr": 9.843881964861985e-06, "epoch": 0.33605812897366033, "percentage": 8.41, "elapsed_time": "6:14:24", "remaining_time": "2 days, 19:58:04"}
+{"current_steps": 186, "total_steps": 2200, "loss": 2.3544, "accuracy": 0.734375, "lr": 9.842199601211556e-06, "epoch": 0.33787465940054495, "percentage": 8.45, "elapsed_time": "6:16:17", "remaining_time": "2 days, 19:54:30"}
+{"current_steps": 187, "total_steps": 2200, "loss": 2.5754, "accuracy": 0.65625, "lr": 9.840508383060092e-06, "epoch": 0.33969118982742963, "percentage": 8.5, "elapsed_time": "6:18:06", "remaining_time": "2 days, 19:50:18"}
+{"current_steps": 188, "total_steps": 2200, "loss": 2.6229, "accuracy": 0.640625, "lr": 9.838808313856281e-06, "epoch": 0.34150772025431425, "percentage": 8.55, "elapsed_time": "6:20:34", "remaining_time": "2 days, 19:53:00"}
+{"current_steps": 189, "total_steps": 2200, "loss": 2.7017, "accuracy": 0.6875, "lr": 9.83709939706686e-06, "epoch": 0.34332425068119893, "percentage": 8.59, "elapsed_time": "6:22:21", "remaining_time": "2 days, 19:48:18"}
+{"current_steps": 190, "total_steps": 2200, "loss": 2.5936, "accuracy": 0.5625, "lr": 9.835381636176604e-06, "epoch": 0.34514078110808355, "percentage": 8.64, "elapsed_time": "6:24:28", "remaining_time": "2 days, 19:47:21"}
+{"current_steps": 191, "total_steps": 2200, "loss": 2.6442, "accuracy": 0.609375, "lr": 9.833655034688336e-06, "epoch": 0.3469573115349682, "percentage": 8.68, "elapsed_time": "6:26:26", "remaining_time": "2 days, 19:44:38"}
+{"current_steps": 192, "total_steps": 2200, "loss": 2.3599, "accuracy": 0.65625, "lr": 9.831919596122888e-06, "epoch": 0.34877384196185285, "percentage": 8.73, "elapsed_time": "6:28:30", "remaining_time": "2 days, 19:43:09"}
+{"current_steps": 193, "total_steps": 2200, "loss": 2.4369, "accuracy": 0.671875, "lr": 9.830175324019125e-06, "epoch": 0.3505903723887375, "percentage": 8.77, "elapsed_time": "6:30:36", "remaining_time": "2 days, 19:41:51"}
+{"current_steps": 194, "total_steps": 2200, "loss": 2.3612, "accuracy": 0.671875, "lr": 9.828422221933924e-06, "epoch": 0.35240690281562215, "percentage": 8.82, "elapsed_time": "6:32:43", "remaining_time": "2 days, 19:40:53"}
+{"current_steps": 195, "total_steps": 2200, "loss": 2.2288, "accuracy": 0.703125, "lr": 9.826660293442158e-06, "epoch": 0.3542234332425068, "percentage": 8.86, "elapsed_time": "6:34:39", "remaining_time": "2 days, 19:37:49"}
+{"current_steps": 196, "total_steps": 2200, "loss": 2.4671, "accuracy": 0.640625, "lr": 9.824889542136714e-06, "epoch": 0.35603996366939145, "percentage": 8.91, "elapsed_time": "6:37:00", "remaining_time": "2 days, 19:39:12"}
+{"current_steps": 197, "total_steps": 2200, "loss": 2.3705, "accuracy": 0.703125, "lr": 9.823109971628459e-06, "epoch": 0.3578564940962761, "percentage": 8.95, "elapsed_time": "6:39:31", "remaining_time": "2 days, 19:42:14"}
+{"current_steps": 198, "total_steps": 2200, "loss": 2.2915, "accuracy": 0.75, "lr": 9.821321585546244e-06, "epoch": 0.35967302452316074, "percentage": 9.0, "elapsed_time": "6:41:30", "remaining_time": "2 days, 19:39:40"}
+{"current_steps": 199, "total_steps": 2200, "loss": 2.3, "accuracy": 0.765625, "lr": 9.819524387536905e-06, "epoch": 0.3614895549500454, "percentage": 9.05, "elapsed_time": "6:43:54", "remaining_time": "2 days, 19:41:27"}
+{"current_steps": 200, "total_steps": 2200, "loss": 2.2872, "accuracy": 0.71875, "lr": 9.81771838126524e-06, "epoch": 0.36330608537693004, "percentage": 9.09, "elapsed_time": "6:46:05", "remaining_time": "2 days, 19:40:51"}
+{"current_steps": 201, "total_steps": 2200, "loss": 2.4279, "accuracy": 0.65625, "lr": 9.815903570414006e-06, "epoch": 0.3651226158038147, "percentage": 9.14, "elapsed_time": "6:48:02", "remaining_time": "2 days, 19:38:08"}
+{"current_steps": 202, "total_steps": 2200, "loss": 2.4603, "accuracy": 0.65625, "lr": 9.814079958683925e-06, "epoch": 0.36693914623069934, "percentage": 9.18, "elapsed_time": "6:49:54", "remaining_time": "2 days, 19:34:27"}
+{"current_steps": 203, "total_steps": 2200, "loss": 2.8134, "accuracy": 0.625, "lr": 9.812247549793656e-06, "epoch": 0.368755676657584, "percentage": 9.23, "elapsed_time": "6:52:04", "remaining_time": "2 days, 19:33:42"}
+{"current_steps": 204, "total_steps": 2200, "loss": 2.4092, "accuracy": 0.75, "lr": 9.810406347479798e-06, "epoch": 0.37057220708446864, "percentage": 9.27, "elapsed_time": "6:54:09", "remaining_time": "2 days, 19:32:10"}
+{"current_steps": 205, "total_steps": 2200, "loss": 2.3507, "accuracy": 0.671875, "lr": 9.808556355496885e-06, "epoch": 0.3723887375113533, "percentage": 9.32, "elapsed_time": "6:56:38", "remaining_time": "2 days, 19:34:34"}
+{"current_steps": 206, "total_steps": 2200, "loss": 2.1888, "accuracy": 0.734375, "lr": 9.806697577617371e-06, "epoch": 0.37420526793823794, "percentage": 9.36, "elapsed_time": "6:58:51", "remaining_time": "2 days, 19:34:19"}
+{"current_steps": 207, "total_steps": 2200, "loss": 2.7312, "accuracy": 0.609375, "lr": 9.804830017631631e-06, "epoch": 0.3760217983651226, "percentage": 9.41, "elapsed_time": "7:00:55", "remaining_time": "2 days, 19:32:42"}
+{"current_steps": 208, "total_steps": 2200, "loss": 2.7324, "accuracy": 0.625, "lr": 9.802953679347943e-06, "epoch": 0.37783832879200724, "percentage": 9.45, "elapsed_time": "7:02:35", "remaining_time": "2 days, 19:27:07"}
+{"current_steps": 209, "total_steps": 2200, "loss": 2.5237, "accuracy": 0.609375, "lr": 9.801068566592486e-06, "epoch": 0.3796548592188919, "percentage": 9.5, "elapsed_time": "7:05:05", "remaining_time": "2 days, 19:29:37"}
+{"current_steps": 210, "total_steps": 2200, "loss": 2.5179, "accuracy": 0.6875, "lr": 9.799174683209336e-06, "epoch": 0.3814713896457766, "percentage": 9.55, "elapsed_time": "7:07:04", "remaining_time": "2 days, 19:27:06"}
+{"current_steps": 211, "total_steps": 2200, "loss": 2.3892, "accuracy": 0.65625, "lr": 9.79727203306045e-06, "epoch": 0.3832879200726612, "percentage": 9.59, "elapsed_time": "7:09:23", "remaining_time": "2 days, 19:27:44"}
+{"current_steps": 212, "total_steps": 2200, "loss": 2.3246, "accuracy": 0.671875, "lr": 9.79536062002566e-06, "epoch": 0.3851044504995459, "percentage": 9.64, "elapsed_time": "7:11:28", "remaining_time": "2 days, 19:26:02"}
+{"current_steps": 213, "total_steps": 2200, "loss": 2.4585, "accuracy": 0.796875, "lr": 9.793440448002676e-06, "epoch": 0.3869209809264305, "percentage": 9.68, "elapsed_time": "7:13:37", "remaining_time": "2 days, 19:25:12"}
+{"current_steps": 214, "total_steps": 2200, "loss": 2.3441, "accuracy": 0.671875, "lr": 9.791511520907056e-06, "epoch": 0.3887375113533152, "percentage": 9.73, "elapsed_time": "7:15:48", "remaining_time": "2 days, 19:24:29"}
+{"current_steps": 215, "total_steps": 2200, "loss": 2.2192, "accuracy": 0.703125, "lr": 9.789573842672223e-06, "epoch": 0.3905540417801998, "percentage": 9.77, "elapsed_time": "7:17:46", "remaining_time": "2 days, 19:21:49"}
+{"current_steps": 216, "total_steps": 2200, "loss": 1.9498, "accuracy": 0.765625, "lr": 9.787627417249441e-06, "epoch": 0.3923705722070845, "percentage": 9.82, "elapsed_time": "7:20:15", "remaining_time": "2 days, 19:23:54"}
+{"current_steps": 217, "total_steps": 2200, "loss": 2.2455, "accuracy": 0.734375, "lr": 9.785672248607807e-06, "epoch": 0.3941871026339691, "percentage": 9.86, "elapsed_time": "7:22:16", "remaining_time": "2 days, 19:21:41"}
+{"current_steps": 218, "total_steps": 2200, "loss": 2.6526, "accuracy": 0.625, "lr": 9.78370834073425e-06, "epoch": 0.3960036330608538, "percentage": 9.91, "elapsed_time": "7:24:15", "remaining_time": "2 days, 19:19:04"}
+{"current_steps": 219, "total_steps": 2200, "loss": 2.3689, "accuracy": 0.671875, "lr": 9.781735697633526e-06, "epoch": 0.3978201634877384, "percentage": 9.95, "elapsed_time": "7:26:00", "remaining_time": "2 days, 19:14:26"}
+{"current_steps": 220, "total_steps": 2200, "loss": 3.2204, "accuracy": 0.578125, "lr": 9.779754323328192e-06, "epoch": 0.3996366939146231, "percentage": 10.0, "elapsed_time": "7:27:39", "remaining_time": "2 days, 19:08:54"}
+{"current_steps": 221, "total_steps": 2200, "loss": 2.3358, "accuracy": 0.71875, "lr": 9.777764221858616e-06, "epoch": 0.4014532243415077, "percentage": 10.05, "elapsed_time": "7:29:28", "remaining_time": "2 days, 19:04:54"}
+{"current_steps": 222, "total_steps": 2200, "loss": 2.4037, "accuracy": 0.75, "lr": 9.775765397282963e-06, "epoch": 0.4032697547683924, "percentage": 10.09, "elapsed_time": "7:31:20", "remaining_time": "2 days, 19:01:23"}
+{"current_steps": 223, "total_steps": 2200, "loss": 2.6149, "accuracy": 0.703125, "lr": 9.773757853677182e-06, "epoch": 0.405086285195277, "percentage": 10.14, "elapsed_time": "7:33:28", "remaining_time": "2 days, 19:00:12"}
+{"current_steps": 224, "total_steps": 2200, "loss": 2.3215, "accuracy": 0.703125, "lr": 9.771741595135009e-06, "epoch": 0.4069028156221617, "percentage": 10.18, "elapsed_time": "7:35:43", "remaining_time": "2 days, 19:00:05"}
+{"current_steps": 225, "total_steps": 2200, "loss": 2.5366, "accuracy": 0.65625, "lr": 9.769716625767939e-06, "epoch": 0.4087193460490463, "percentage": 10.23, "elapsed_time": "7:37:41", "remaining_time": "2 days, 18:57:28"}
+{"current_steps": 226, "total_steps": 2200, "loss": 2.3992, "accuracy": 0.609375, "lr": 9.767682949705243e-06, "epoch": 0.410535876475931, "percentage": 10.27, "elapsed_time": "7:39:27", "remaining_time": "2 days, 18:53:06"}
+{"current_steps": 227, "total_steps": 2200, "loss": 2.488, "accuracy": 0.640625, "lr": 9.765640571093938e-06, "epoch": 0.4123524069028156, "percentage": 10.32, "elapsed_time": "7:41:17", "remaining_time": "2 days, 18:49:22"}
+{"current_steps": 228, "total_steps": 2200, "loss": 2.466, "accuracy": 0.71875, "lr": 9.76358949409879e-06, "epoch": 0.4141689373297003, "percentage": 10.36, "elapsed_time": "7:43:27", "remaining_time": "2 days, 18:48:27"}
+{"current_steps": 229, "total_steps": 2200, "loss": 2.282, "accuracy": 0.765625, "lr": 9.7615297229023e-06, "epoch": 0.4159854677565849, "percentage": 10.41, "elapsed_time": "7:45:10", "remaining_time": "2 days, 18:43:49"}
+{"current_steps": 230, "total_steps": 2200, "loss": 2.1345, "accuracy": 0.703125, "lr": 9.759461261704705e-06, "epoch": 0.4178019981834696, "percentage": 10.45, "elapsed_time": "7:47:10", "remaining_time": "2 days, 18:41:28"}
+{"current_steps": 231, "total_steps": 2200, "loss": 2.3606, "accuracy": 0.65625, "lr": 9.757384114723954e-06, "epoch": 0.4196185286103542, "percentage": 10.5, "elapsed_time": "7:49:18", "remaining_time": "2 days, 18:40:13"}
+{"current_steps": 232, "total_steps": 2200, "loss": 2.4171, "accuracy": 0.671875, "lr": 9.755298286195712e-06, "epoch": 0.4214350590372389, "percentage": 10.55, "elapsed_time": "7:51:42", "remaining_time": "2 days, 18:41:20"}
+{"current_steps": 233, "total_steps": 2200, "loss": 2.7548, "accuracy": 0.703125, "lr": 9.753203780373348e-06, "epoch": 0.4232515894641235, "percentage": 10.59, "elapsed_time": "7:54:00", "remaining_time": "2 days, 18:41:36"}
+{"current_steps": 234, "total_steps": 2200, "loss": 2.4369, "accuracy": 0.75, "lr": 9.751100601527922e-06, "epoch": 0.4250681198910082, "percentage": 10.64, "elapsed_time": "7:55:58", "remaining_time": "2 days, 18:39:03"}
+{"current_steps": 235, "total_steps": 2200, "loss": 2.3437, "accuracy": 0.71875, "lr": 9.748988753948183e-06, "epoch": 0.4268846503178928, "percentage": 10.68, "elapsed_time": "7:58:27", "remaining_time": "2 days, 18:40:42"}
+{"current_steps": 236, "total_steps": 2200, "loss": 2.6303, "accuracy": 0.625, "lr": 9.746868241940554e-06, "epoch": 0.4287011807447775, "percentage": 10.73, "elapsed_time": "8:00:25", "remaining_time": "2 days, 18:38:03"}
+{"current_steps": 237, "total_steps": 2200, "loss": 2.3038, "accuracy": 0.671875, "lr": 9.744739069829132e-06, "epoch": 0.4305177111716621, "percentage": 10.77, "elapsed_time": "8:02:04", "remaining_time": "2 days, 18:32:53"}
+{"current_steps": 238, "total_steps": 2200, "loss": 2.2981, "accuracy": 0.734375, "lr": 9.742601241955666e-06, "epoch": 0.4323342415985468, "percentage": 10.82, "elapsed_time": "8:04:35", "remaining_time": "2 days, 18:34:46"}
+{"current_steps": 239, "total_steps": 2200, "loss": 2.0885, "accuracy": 0.75, "lr": 9.740454762679562e-06, "epoch": 0.43415077202543145, "percentage": 10.86, "elapsed_time": "8:06:35", "remaining_time": "2 days, 18:32:26"}
+{"current_steps": 240, "total_steps": 2200, "loss": 2.3774, "accuracy": 0.75, "lr": 9.738299636377863e-06, "epoch": 0.4359673024523161, "percentage": 10.91, "elapsed_time": "8:08:35", "remaining_time": "2 days, 18:30:07"}
+{"current_steps": 241, "total_steps": 2200, "loss": 1.9408, "accuracy": 0.75, "lr": 9.736135867445246e-06, "epoch": 0.43778383287920075, "percentage": 10.95, "elapsed_time": "8:10:26", "remaining_time": "2 days, 18:26:39"}
+{"current_steps": 242, "total_steps": 2200, "loss": 2.4779, "accuracy": 0.625, "lr": 9.733963460294016e-06, "epoch": 0.4396003633060854, "percentage": 11.0, "elapsed_time": "8:12:26", "remaining_time": "2 days, 18:24:21"}
+{"current_steps": 243, "total_steps": 2200, "loss": 2.4426, "accuracy": 0.640625, "lr": 9.731782419354087e-06, "epoch": 0.44141689373297005, "percentage": 11.05, "elapsed_time": "8:14:25", "remaining_time": "2 days, 18:21:54"}
+{"current_steps": 244, "total_steps": 2200, "loss": 2.4137, "accuracy": 0.65625, "lr": 9.729592749072981e-06, "epoch": 0.44323342415985467, "percentage": 11.09, "elapsed_time": "8:16:50", "remaining_time": "2 days, 18:22:56"}
+{"current_steps": 245, "total_steps": 2200, "loss": 2.3143, "accuracy": 0.75, "lr": 9.727394453915817e-06, "epoch": 0.44504995458673935, "percentage": 11.14, "elapsed_time": "8:18:39", "remaining_time": "2 days, 18:19:01"}
+{"current_steps": 246, "total_steps": 2200, "loss": 2.3278, "accuracy": 0.609375, "lr": 9.725187538365304e-06, "epoch": 0.44686648501362397, "percentage": 11.18, "elapsed_time": "8:20:50", "remaining_time": "2 days, 18:18:17"}
+{"current_steps": 247, "total_steps": 2200, "loss": 2.3939, "accuracy": 0.6875, "lr": 9.722972006921725e-06, "epoch": 0.44868301544050865, "percentage": 11.23, "elapsed_time": "8:22:49", "remaining_time": "2 days, 18:15:44"}
+{"current_steps": 248, "total_steps": 2200, "loss": 2.4935, "accuracy": 0.6875, "lr": 9.720747864102935e-06, "epoch": 0.45049954586739327, "percentage": 11.27, "elapsed_time": "8:24:51", "remaining_time": "2 days, 18:13:44"}
+{"current_steps": 249, "total_steps": 2200, "loss": 2.2988, "accuracy": 0.640625, "lr": 9.718515114444347e-06, "epoch": 0.45231607629427795, "percentage": 11.32, "elapsed_time": "8:26:21", "remaining_time": "2 days, 18:07:31"}
+{"current_steps": 250, "total_steps": 2200, "loss": 2.4975, "accuracy": 0.671875, "lr": 9.716273762498929e-06, "epoch": 0.45413260672116257, "percentage": 11.36, "elapsed_time": "8:28:04", "remaining_time": "2 days, 18:03:02"}
+{"current_steps": 251, "total_steps": 2200, "loss": 2.627, "accuracy": 0.625, "lr": 9.714023812837185e-06, "epoch": 0.45594913714804725, "percentage": 11.41, "elapsed_time": "8:29:51", "remaining_time": "2 days, 17:58:57"}
+{"current_steps": 252, "total_steps": 2200, "loss": 2.1641, "accuracy": 0.71875, "lr": 9.711765270047155e-06, "epoch": 0.45776566757493187, "percentage": 11.45, "elapsed_time": "8:32:08", "remaining_time": "2 days, 17:58:55"}
+{"current_steps": 253, "total_steps": 2200, "loss": 2.58, "accuracy": 0.546875, "lr": 9.709498138734405e-06, "epoch": 0.45958219800181654, "percentage": 11.5, "elapsed_time": "8:34:21", "remaining_time": "2 days, 17:58:15"}
+{"current_steps": 254, "total_steps": 2200, "loss": 2.7899, "accuracy": 0.515625, "lr": 9.707222423522004e-06, "epoch": 0.46139872842870117, "percentage": 11.55, "elapsed_time": "8:36:31", "remaining_time": "2 days, 17:57:20"}
+{"current_steps": 255, "total_steps": 2200, "loss": 2.1796, "accuracy": 0.75, "lr": 9.704938129050535e-06, "epoch": 0.46321525885558584, "percentage": 11.59, "elapsed_time": "8:38:48", "remaining_time": "2 days, 17:57:11"}
+{"current_steps": 256, "total_steps": 2200, "loss": 2.2113, "accuracy": 0.703125, "lr": 9.702645259978072e-06, "epoch": 0.46503178928247046, "percentage": 11.64, "elapsed_time": "8:40:56", "remaining_time": "2 days, 17:55:57"}
+{"current_steps": 257, "total_steps": 2200, "loss": 2.3793, "accuracy": 0.65625, "lr": 9.700343820980172e-06, "epoch": 0.46684831970935514, "percentage": 11.68, "elapsed_time": "8:43:00", "remaining_time": "2 days, 17:54:04"}
+{"current_steps": 258, "total_steps": 2200, "loss": 2.0527, "accuracy": 0.84375, "lr": 9.698033816749874e-06, "epoch": 0.46866485013623976, "percentage": 11.73, "elapsed_time": "8:45:03", "remaining_time": "2 days, 17:52:13"}
+{"current_steps": 259, "total_steps": 2200, "loss": 1.8381, "accuracy": 0.859375, "lr": 9.695715251997676e-06, "epoch": 0.47048138056312444, "percentage": 11.77, "elapsed_time": "8:47:11", "remaining_time": "2 days, 17:50:50"}
+{"current_steps": 260, "total_steps": 2200, "loss": 2.2716, "accuracy": 0.71875, "lr": 9.693388131451536e-06, "epoch": 0.47229791099000906, "percentage": 11.82, "elapsed_time": "8:48:49", "remaining_time": "2 days, 17:45:53"}
+{"current_steps": 261, "total_steps": 2200, "loss": 2.3255, "accuracy": 0.703125, "lr": 9.691052459856858e-06, "epoch": 0.47411444141689374, "percentage": 11.86, "elapsed_time": "8:51:06", "remaining_time": "2 days, 17:45:39"}
+{"current_steps": 262, "total_steps": 2200, "loss": 2.5126, "accuracy": 0.640625, "lr": 9.688708241976484e-06, "epoch": 0.47593097184377836, "percentage": 11.91, "elapsed_time": "8:53:17", "remaining_time": "2 days, 17:44:43"}
+{"current_steps": 263, "total_steps": 2200, "loss": 2.0243, "accuracy": 0.84375, "lr": 9.686355482590679e-06, "epoch": 0.47774750227066304, "percentage": 11.95, "elapsed_time": "8:55:35", "remaining_time": "2 days, 17:44:37"}
+{"current_steps": 264, "total_steps": 2200, "loss": 2.6106, "accuracy": 0.8125, "lr": 9.683994186497132e-06, "epoch": 0.47956403269754766, "percentage": 12.0, "elapsed_time": "8:57:37", "remaining_time": "2 days, 17:42:36"}
+{"current_steps": 265, "total_steps": 2200, "loss": 2.1509, "accuracy": 0.703125, "lr": 9.681624358510936e-06, "epoch": 0.48138056312443234, "percentage": 12.05, "elapsed_time": "8:59:27", "remaining_time": "2 days, 17:39:02"}
+{"current_steps": 266, "total_steps": 2200, "loss": 2.8119, "accuracy": 0.625, "lr": 9.679246003464585e-06, "epoch": 0.48319709355131696, "percentage": 12.09, "elapsed_time": "9:01:19", "remaining_time": "2 days, 17:35:45"}
+{"current_steps": 267, "total_steps": 2200, "loss": 2.2386, "accuracy": 0.703125, "lr": 9.676859126207957e-06, "epoch": 0.48501362397820164, "percentage": 12.14, "elapsed_time": "9:03:06", "remaining_time": "2 days, 17:31:56"}
+{"current_steps": 268, "total_steps": 2200, "loss": 2.8564, "accuracy": 0.625, "lr": 9.674463731608309e-06, "epoch": 0.4868301544050863, "percentage": 12.18, "elapsed_time": "9:04:56", "remaining_time": "2 days, 17:28:28"}
+{"current_steps": 269, "total_steps": 2200, "loss": 2.1801, "accuracy": 0.75, "lr": 9.672059824550268e-06, "epoch": 0.48864668483197093, "percentage": 12.23, "elapsed_time": "9:07:10", "remaining_time": "2 days, 17:27:52"}
+{"current_steps": 270, "total_steps": 2200, "loss": 2.342, "accuracy": 0.671875, "lr": 9.669647409935822e-06, "epoch": 0.4904632152588556, "percentage": 12.27, "elapsed_time": "9:09:11", "remaining_time": "2 days, 17:25:43"}
+{"current_steps": 271, "total_steps": 2200, "loss": 2.2264, "accuracy": 0.609375, "lr": 9.667226492684302e-06, "epoch": 0.49227974568574023, "percentage": 12.32, "elapsed_time": "9:11:28", "remaining_time": "2 days, 17:25:27"}
+{"current_steps": 272, "total_steps": 2200, "loss": 2.4723, "accuracy": 0.734375, "lr": 9.66479707773238e-06, "epoch": 0.4940962761126249, "percentage": 12.36, "elapsed_time": "9:13:39", "remaining_time": "2 days, 17:24:24"}
+{"current_steps": 273, "total_steps": 2200, "loss": 2.5565, "accuracy": 0.640625, "lr": 9.662359170034058e-06, "epoch": 0.49591280653950953, "percentage": 12.41, "elapsed_time": "9:15:25", "remaining_time": "2 days, 17:20:30"}
+{"current_steps": 274, "total_steps": 2200, "loss": 2.1908, "accuracy": 0.703125, "lr": 9.659912774560654e-06, "epoch": 0.4977293369663942, "percentage": 12.45, "elapsed_time": "9:17:22", "remaining_time": "2 days, 17:17:56"}
+{"current_steps": 275, "total_steps": 2200, "loss": 3.0939, "accuracy": 0.71875, "lr": 9.65745789630079e-06, "epoch": 0.49954586739327883, "percentage": 12.5, "elapsed_time": "9:20:29", "remaining_time": "2 days, 17:23:27"}
+{"current_steps": 276, "total_steps": 2200, "loss": 2.7161, "accuracy": 0.53125, "lr": 9.654994540260396e-06, "epoch": 0.5013623978201635, "percentage": 12.55, "elapsed_time": "9:22:44", "remaining_time": "2 days, 17:22:52"}
+{"current_steps": 277, "total_steps": 2200, "loss": 2.5366, "accuracy": 0.65625, "lr": 9.65252271146268e-06, "epoch": 0.5031789282470481, "percentage": 12.59, "elapsed_time": "9:24:36", "remaining_time": "2 days, 17:19:38"}
+{"current_steps": 278, "total_steps": 2200, "loss": 2.6085, "accuracy": 0.640625, "lr": 9.650042414948133e-06, "epoch": 0.5049954586739328, "percentage": 12.64, "elapsed_time": "9:26:38", "remaining_time": "2 days, 17:17:36"}
+{"current_steps": 279, "total_steps": 2200, "loss": 2.3156, "accuracy": 0.671875, "lr": 9.64755365577451e-06, "epoch": 0.5068119891008175, "percentage": 12.68, "elapsed_time": "9:28:31", "remaining_time": "2 days, 17:14:26"}
+{"current_steps": 280, "total_steps": 2200, "loss": 2.8085, "accuracy": 0.609375, "lr": 9.645056439016827e-06, "epoch": 0.508628519527702, "percentage": 12.73, "elapsed_time": "9:30:42", "remaining_time": "2 days, 17:13:22"}
+{"current_steps": 281, "total_steps": 2200, "loss": 2.2314, "accuracy": 0.75, "lr": 9.642550769767342e-06, "epoch": 0.5104450499545867, "percentage": 12.77, "elapsed_time": "9:33:08", "remaining_time": "2 days, 17:14:03"}
+{"current_steps": 282, "total_steps": 2200, "loss": 2.4359, "accuracy": 0.765625, "lr": 9.640036653135548e-06, "epoch": 0.5122615803814714, "percentage": 12.82, "elapsed_time": "9:34:45", "remaining_time": "2 days, 17:09:11"}
+{"current_steps": 283, "total_steps": 2200, "loss": 2.2478, "accuracy": 0.65625, "lr": 9.637514094248172e-06, "epoch": 0.5140781108083561, "percentage": 12.86, "elapsed_time": "9:36:35", "remaining_time": "2 days, 17:05:47"}
+{"current_steps": 284, "total_steps": 2200, "loss": 2.1538, "accuracy": 0.765625, "lr": 9.634983098249146e-06, "epoch": 0.5158946412352406, "percentage": 12.91, "elapsed_time": "9:38:11", "remaining_time": "2 days, 17:00:45"}
+{"current_steps": 285, "total_steps": 2200, "loss": 2.2988, "accuracy": 0.65625, "lr": 9.632443670299616e-06, "epoch": 0.5177111716621253, "percentage": 12.95, "elapsed_time": "9:40:42", "remaining_time": "2 days, 17:01:58"}
+{"current_steps": 286, "total_steps": 2200, "loss": 2.2831, "accuracy": 0.65625, "lr": 9.629895815577915e-06, "epoch": 0.51952770208901, "percentage": 13.0, "elapsed_time": "9:43:06", "remaining_time": "2 days, 17:02:19"}
+{"current_steps": 287, "total_steps": 2200, "loss": 2.5101, "accuracy": 0.71875, "lr": 9.627339539279564e-06, "epoch": 0.5213442325158947, "percentage": 13.05, "elapsed_time": "9:45:13", "remaining_time": "2 days, 17:00:49"}
+{"current_steps": 288, "total_steps": 2200, "loss": 2.5236, "accuracy": 0.703125, "lr": 9.624774846617254e-06, "epoch": 0.5231607629427792, "percentage": 13.09, "elapsed_time": "9:47:03", "remaining_time": "2 days, 16:57:22"}
+{"current_steps": 289, "total_steps": 2200, "loss": 2.3891, "accuracy": 0.703125, "lr": 9.622201742820839e-06, "epoch": 0.5249772933696639, "percentage": 13.14, "elapsed_time": "9:48:51", "remaining_time": "2 days, 16:53:44"}
+{"current_steps": 290, "total_steps": 2200, "loss": 2.4569, "accuracy": 0.71875, "lr": 9.619620233137328e-06, "epoch": 0.5267938237965486, "percentage": 13.18, "elapsed_time": "9:50:54", "remaining_time": "2 days, 16:51:47"}
+{"current_steps": 291, "total_steps": 2200, "loss": 2.4021, "accuracy": 0.671875, "lr": 9.617030322830868e-06, "epoch": 0.5286103542234333, "percentage": 13.23, "elapsed_time": "9:52:38", "remaining_time": "2 days, 16:47:50"}
+{"current_steps": 292, "total_steps": 2200, "loss": 2.3636, "accuracy": 0.640625, "lr": 9.614432017182736e-06, "epoch": 0.5304268846503178, "percentage": 13.27, "elapsed_time": "9:55:07", "remaining_time": "2 days, 16:48:39"}
+{"current_steps": 293, "total_steps": 2200, "loss": 2.343, "accuracy": 0.671875, "lr": 9.611825321491331e-06, "epoch": 0.5322434150772025, "percentage": 13.32, "elapsed_time": "9:57:30", "remaining_time": "2 days, 16:48:57"}
+{"current_steps": 294, "total_steps": 2200, "loss": 2.1163, "accuracy": 0.671875, "lr": 9.609210241072158e-06, "epoch": 0.5340599455040872, "percentage": 13.36, "elapsed_time": "9:59:33", "remaining_time": "2 days, 16:46:56"}
+{"current_steps": 295, "total_steps": 2200, "loss": 2.5075, "accuracy": 0.65625, "lr": 9.606586781257822e-06, "epoch": 0.5358764759309719, "percentage": 13.41, "elapsed_time": "10:01:37", "remaining_time": "2 days, 16:45:05"}
+{"current_steps": 296, "total_steps": 2200, "loss": 2.1959, "accuracy": 0.828125, "lr": 9.603954947398016e-06, "epoch": 0.5376930063578564, "percentage": 13.45, "elapsed_time": "10:03:06", "remaining_time": "2 days, 16:39:24"}
+{"current_steps": 297, "total_steps": 2200, "loss": 2.3538, "accuracy": 0.6875, "lr": 9.601314744859504e-06, "epoch": 0.5395095367847411, "percentage": 13.5, "elapsed_time": "10:05:04", "remaining_time": "2 days, 16:36:59"}
+{"current_steps": 298, "total_steps": 2200, "loss": 2.4531, "accuracy": 0.578125, "lr": 9.598666179026123e-06, "epoch": 0.5413260672116258, "percentage": 13.55, "elapsed_time": "10:06:46", "remaining_time": "2 days, 16:32:43"}
+{"current_steps": 299, "total_steps": 2200, "loss": 2.8549, "accuracy": 0.609375, "lr": 9.596009255298755e-06, "epoch": 0.5431425976385105, "percentage": 13.59, "elapsed_time": "10:09:11", "remaining_time": "2 days, 16:33:09"}
+{"current_steps": 300, "total_steps": 2200, "loss": 2.4644, "accuracy": 0.65625, "lr": 9.593343979095334e-06, "epoch": 0.5449591280653951, "percentage": 13.64, "elapsed_time": "10:11:08", "remaining_time": "2 days, 16:30:31"}
+{"current_steps": 301, "total_steps": 2200, "loss": 2.4159, "accuracy": 0.625, "lr": 9.590670355850819e-06, "epoch": 0.5467756584922797, "percentage": 13.68, "elapsed_time": "10:13:02", "remaining_time": "2 days, 16:27:38"}
+{"current_steps": 302, "total_steps": 2200, "loss": 2.5826, "accuracy": 0.640625, "lr": 9.587988391017198e-06, "epoch": 0.5485921889191644, "percentage": 13.73, "elapsed_time": "10:15:09", "remaining_time": "2 days, 16:26:10"}
+{"current_steps": 303, "total_steps": 2200, "loss": 2.2064, "accuracy": 0.75, "lr": 9.585298090063459e-06, "epoch": 0.5504087193460491, "percentage": 13.77, "elapsed_time": "10:16:48", "remaining_time": "2 days, 16:21:38"}
+{"current_steps": 304, "total_steps": 2200, "loss": 2.3792, "accuracy": 0.734375, "lr": 9.582599458475598e-06, "epoch": 0.5522252497729337, "percentage": 13.82, "elapsed_time": "10:18:39", "remaining_time": "2 days, 16:18:28"}
+{"current_steps": 305, "total_steps": 2200, "loss": 2.2105, "accuracy": 0.75, "lr": 9.579892501756593e-06, "epoch": 0.5540417801998183, "percentage": 13.86, "elapsed_time": "10:20:51", "remaining_time": "2 days, 16:17:29"}
+{"current_steps": 306, "total_steps": 2200, "loss": 2.1687, "accuracy": 0.75, "lr": 9.5771772254264e-06, "epoch": 0.555858310626703, "percentage": 13.91, "elapsed_time": "10:23:08", "remaining_time": "2 days, 16:16:56"}
+{"current_steps": 307, "total_steps": 2200, "loss": 2.2116, "accuracy": 0.765625, "lr": 9.57445363502194e-06, "epoch": 0.5576748410535877, "percentage": 13.95, "elapsed_time": "10:24:54", "remaining_time": "2 days, 16:13:13"}
+{"current_steps": 308, "total_steps": 2200, "loss": 1.9942, "accuracy": 0.765625, "lr": 9.571721736097089e-06, "epoch": 0.5594913714804723, "percentage": 14.0, "elapsed_time": "10:27:12", "remaining_time": "2 days, 16:12:51"}
+{"current_steps": 309, "total_steps": 2200, "loss": 2.4018, "accuracy": 0.65625, "lr": 9.568981534222664e-06, "epoch": 0.5613079019073569, "percentage": 14.05, "elapsed_time": "10:29:02", "remaining_time": "2 days, 16:09:35"}
+{"current_steps": 310, "total_steps": 2200, "loss": 2.6286, "accuracy": 0.71875, "lr": 9.566233034986413e-06, "epoch": 0.5631244323342416, "percentage": 14.09, "elapsed_time": "10:30:54", "remaining_time": "2 days, 16:06:31"}
+{"current_steps": 311, "total_steps": 2200, "loss": 2.4426, "accuracy": 0.71875, "lr": 9.563476243993008e-06, "epoch": 0.5649409627611263, "percentage": 14.14, "elapsed_time": "10:33:07", "remaining_time": "2 days, 16:05:32"}
+{"current_steps": 312, "total_steps": 2200, "loss": 2.578, "accuracy": 0.671875, "lr": 9.56071116686402e-06, "epoch": 0.5667574931880109, "percentage": 14.18, "elapsed_time": "10:35:06", "remaining_time": "2 days, 16:03:10"}
+{"current_steps": 313, "total_steps": 2200, "loss": 2.3854, "accuracy": 0.75, "lr": 9.557937809237927e-06, "epoch": 0.5685740236148955, "percentage": 14.23, "elapsed_time": "10:37:16", "remaining_time": "2 days, 16:01:59"}
+{"current_steps": 314, "total_steps": 2200, "loss": 2.6795, "accuracy": 0.59375, "lr": 9.555156176770087e-06, "epoch": 0.5703905540417802, "percentage": 14.27, "elapsed_time": "10:39:24", "remaining_time": "2 days, 16:00:28"}
+{"current_steps": 315, "total_steps": 2200, "loss": 2.6181, "accuracy": 0.625, "lr": 9.552366275132733e-06, "epoch": 0.5722070844686649, "percentage": 14.32, "elapsed_time": "10:41:43", "remaining_time": "2 days, 16:00:08"}
+{"current_steps": 316, "total_steps": 2200, "loss": 2.5849, "accuracy": 0.65625, "lr": 9.54956811001496e-06, "epoch": 0.5740236148955495, "percentage": 14.36, "elapsed_time": "10:44:02", "remaining_time": "2 days, 15:59:49"}
+{"current_steps": 317, "total_steps": 2200, "loss": 2.3208, "accuracy": 0.640625, "lr": 9.546761687122715e-06, "epoch": 0.5758401453224341, "percentage": 14.41, "elapsed_time": "10:45:57", "remaining_time": "2 days, 15:57:03"}
+{"current_steps": 318, "total_steps": 2200, "loss": 2.2106, "accuracy": 0.71875, "lr": 9.54394701217878e-06, "epoch": 0.5776566757493188, "percentage": 14.45, "elapsed_time": "10:48:25", "remaining_time": "2 days, 15:57:32"}
+{"current_steps": 319, "total_steps": 2200, "loss": 2.6435, "accuracy": 0.609375, "lr": 9.541124090922771e-06, "epoch": 0.5794732061762035, "percentage": 14.5, "elapsed_time": "10:50:32", "remaining_time": "2 days, 15:55:55"}
+{"current_steps": 320, "total_steps": 2200, "loss": 2.2137, "accuracy": 0.6875, "lr": 9.538292929111114e-06, "epoch": 0.5812897366030881, "percentage": 14.55, "elapsed_time": "10:52:32", "remaining_time": "2 days, 15:53:42"}
+{"current_steps": 321, "total_steps": 2200, "loss": 2.6036, "accuracy": 0.6875, "lr": 9.535453532517039e-06, "epoch": 0.5831062670299727, "percentage": 14.59, "elapsed_time": "10:54:59", "remaining_time": "2 days, 15:54:04"}
+{"current_steps": 322, "total_steps": 2200, "loss": 2.4565, "accuracy": 0.75, "lr": 9.532605906930575e-06, "epoch": 0.5849227974568574, "percentage": 14.64, "elapsed_time": "10:56:43", "remaining_time": "2 days, 15:50:13"}
+{"current_steps": 323, "total_steps": 2200, "loss": 2.2729, "accuracy": 0.734375, "lr": 9.529750058158522e-06, "epoch": 0.5867393278837421, "percentage": 14.68, "elapsed_time": "10:58:37", "remaining_time": "2 days, 15:47:19"}
+{"current_steps": 324, "total_steps": 2200, "loss": 2.2075, "accuracy": 0.703125, "lr": 9.526885992024453e-06, "epoch": 0.5885558583106267, "percentage": 14.73, "elapsed_time": "11:00:48", "remaining_time": "2 days, 15:46:11"}
+{"current_steps": 325, "total_steps": 2200, "loss": 2.5473, "accuracy": 0.640625, "lr": 9.524013714368702e-06, "epoch": 0.5903723887375113, "percentage": 14.77, "elapsed_time": "11:02:41", "remaining_time": "2 days, 15:43:11"}
+{"current_steps": 326, "total_steps": 2200, "loss": 2.4866, "accuracy": 0.703125, "lr": 9.521133231048338e-06, "epoch": 0.592188919164396, "percentage": 14.82, "elapsed_time": "11:04:51", "remaining_time": "2 days, 15:41:52"}
+{"current_steps": 327, "total_steps": 2200, "loss": 2.2602, "accuracy": 0.6875, "lr": 9.51824454793717e-06, "epoch": 0.5940054495912807, "percentage": 14.86, "elapsed_time": "11:07:19", "remaining_time": "2 days, 15:42:16"}
+{"current_steps": 328, "total_steps": 2200, "loss": 2.3946, "accuracy": 0.640625, "lr": 9.515347670925728e-06, "epoch": 0.5958219800181653, "percentage": 14.91, "elapsed_time": "11:09:36", "remaining_time": "2 days, 15:41:42"}
+{"current_steps": 329, "total_steps": 2200, "loss": 2.5085, "accuracy": 0.640625, "lr": 9.512442605921245e-06, "epoch": 0.59763851044505, "percentage": 14.95, "elapsed_time": "11:11:17", "remaining_time": "2 days, 15:37:37"}
+{"current_steps": 330, "total_steps": 2200, "loss": 2.4419, "accuracy": 0.625, "lr": 9.509529358847655e-06, "epoch": 0.5994550408719346, "percentage": 15.0, "elapsed_time": "11:13:22", "remaining_time": "2 days, 15:35:47"}
+{"current_steps": 331, "total_steps": 2200, "loss": 2.4273, "accuracy": 0.65625, "lr": 9.506607935645579e-06, "epoch": 0.6012715712988193, "percentage": 15.05, "elapsed_time": "11:15:21", "remaining_time": "2 days, 15:33:24"}
+{"current_steps": 332, "total_steps": 2200, "loss": 2.6941, "accuracy": 0.640625, "lr": 9.503678342272306e-06, "epoch": 0.6030881017257039, "percentage": 15.09, "elapsed_time": "11:17:33", "remaining_time": "2 days, 15:32:19"}
+{"current_steps": 333, "total_steps": 2200, "loss": 2.0931, "accuracy": 0.765625, "lr": 9.500740584701785e-06, "epoch": 0.6049046321525886, "percentage": 15.14, "elapsed_time": "11:19:37", "remaining_time": "2 days, 15:30:25"}
+{"current_steps": 334, "total_steps": 2200, "loss": 2.4006, "accuracy": 0.71875, "lr": 9.497794668924617e-06, "epoch": 0.6067211625794732, "percentage": 15.18, "elapsed_time": "11:21:27", "remaining_time": "2 days, 15:27:09"}
+{"current_steps": 335, "total_steps": 2200, "loss": 2.3301, "accuracy": 0.71875, "lr": 9.494840600948038e-06, "epoch": 0.6085376930063578, "percentage": 15.23, "elapsed_time": "11:23:24", "remaining_time": "2 days, 15:24:37"}
+{"current_steps": 336, "total_steps": 2200, "loss": 2.5593, "accuracy": 0.640625, "lr": 9.491878386795906e-06, "epoch": 0.6103542234332425, "percentage": 15.27, "elapsed_time": "11:25:09", "remaining_time": "2 days, 15:21:02"}
+{"current_steps": 337, "total_steps": 2200, "loss": 2.8282, "accuracy": 0.59375, "lr": 9.488908032508691e-06, "epoch": 0.6121707538601272, "percentage": 15.32, "elapsed_time": "11:27:33", "remaining_time": "2 days, 15:20:58"}
+{"current_steps": 338, "total_steps": 2200, "loss": 2.4029, "accuracy": 0.65625, "lr": 9.485929544143462e-06, "epoch": 0.6139872842870118, "percentage": 15.36, "elapsed_time": "11:29:37", "remaining_time": "2 days, 15:19:05"}
+{"current_steps": 339, "total_steps": 2200, "loss": 2.2269, "accuracy": 0.6875, "lr": 9.482942927773876e-06, "epoch": 0.6158038147138964, "percentage": 15.41, "elapsed_time": "11:31:49", "remaining_time": "2 days, 15:17:52"}
+{"current_steps": 340, "total_steps": 2200, "loss": 2.3328, "accuracy": 0.65625, "lr": 9.479948189490164e-06, "epoch": 0.6176203451407811, "percentage": 15.45, "elapsed_time": "11:33:53", "remaining_time": "2 days, 15:16:02"}
+{"current_steps": 341, "total_steps": 2200, "loss": 2.4269, "accuracy": 0.71875, "lr": 9.476945335399122e-06, "epoch": 0.6194368755676658, "percentage": 15.5, "elapsed_time": "11:36:03", "remaining_time": "2 days, 15:14:37"}
+{"current_steps": 342, "total_steps": 2200, "loss": 2.4709, "accuracy": 0.71875, "lr": 9.473934371624087e-06, "epoch": 0.6212534059945504, "percentage": 15.55, "elapsed_time": "11:37:57", "remaining_time": "2 days, 15:11:52"}
+{"current_steps": 343, "total_steps": 2200, "loss": 2.2678, "accuracy": 0.6875, "lr": 9.47091530430494e-06, "epoch": 0.623069936421435, "percentage": 15.59, "elapsed_time": "11:40:06", "remaining_time": "2 days, 15:10:20"}
+{"current_steps": 344, "total_steps": 2200, "loss": 2.7867, "accuracy": 0.6875, "lr": 9.467888139598086e-06, "epoch": 0.6248864668483197, "percentage": 15.64, "elapsed_time": "11:42:21", "remaining_time": "2 days, 15:09:28"}
+{"current_steps": 345, "total_steps": 2200, "loss": 2.2777, "accuracy": 0.6875, "lr": 9.464852883676441e-06, "epoch": 0.6267029972752044, "percentage": 15.68, "elapsed_time": "11:44:22", "remaining_time": "2 days, 15:07:19"}
+{"current_steps": 346, "total_steps": 2200, "loss": 2.0438, "accuracy": 0.75, "lr": 9.461809542729421e-06, "epoch": 0.628519527702089, "percentage": 15.73, "elapsed_time": "11:46:48", "remaining_time": "2 days, 15:07:21"}
+{"current_steps": 347, "total_steps": 2200, "loss": 2.3299, "accuracy": 0.6875, "lr": 9.458758122962926e-06, "epoch": 0.6303360581289736, "percentage": 15.77, "elapsed_time": "11:49:03", "remaining_time": "2 days, 15:06:25"}
+{"current_steps": 348, "total_steps": 2200, "loss": 2.3439, "accuracy": 0.671875, "lr": 9.455698630599332e-06, "epoch": 0.6321525885558583, "percentage": 15.82, "elapsed_time": "11:51:13", "remaining_time": "2 days, 15:05:00"}
+{"current_steps": 349, "total_steps": 2200, "loss": 2.1232, "accuracy": 0.796875, "lr": 9.452631071877478e-06, "epoch": 0.633969118982743, "percentage": 15.86, "elapsed_time": "11:53:08", "remaining_time": "2 days, 15:02:19"}
+{"current_steps": 350, "total_steps": 2200, "loss": 2.9175, "accuracy": 0.53125, "lr": 9.449555453052652e-06, "epoch": 0.6357856494096276, "percentage": 15.91, "elapsed_time": "11:54:45", "remaining_time": "2 days, 14:57:59"}
+{"current_steps": 351, "total_steps": 2200, "loss": 2.3833, "accuracy": 0.6875, "lr": 9.446471780396573e-06, "epoch": 0.6376021798365122, "percentage": 15.95, "elapsed_time": "11:56:38", "remaining_time": "2 days, 14:55:05"}
+{"current_steps": 352, "total_steps": 2200, "loss": 2.5558, "accuracy": 0.65625, "lr": 9.443380060197387e-06, "epoch": 0.6394187102633969, "percentage": 16.0, "elapsed_time": "11:58:21", "remaining_time": "2 days, 14:51:21"}
+{"current_steps": 353, "total_steps": 2200, "loss": 2.6547, "accuracy": 0.59375, "lr": 9.440280298759653e-06, "epoch": 0.6412352406902816, "percentage": 16.05, "elapsed_time": "12:00:44", "remaining_time": "2 days, 14:51:07"}
+{"current_steps": 354, "total_steps": 2200, "loss": 2.3809, "accuracy": 0.609375, "lr": 9.437172502404318e-06, "epoch": 0.6430517711171662, "percentage": 16.09, "elapsed_time": "12:02:52", "remaining_time": "2 days, 14:49:32"}
+{"current_steps": 355, "total_steps": 2200, "loss": 2.006, "accuracy": 0.796875, "lr": 9.434056677468726e-06, "epoch": 0.6448683015440508, "percentage": 16.14, "elapsed_time": "12:04:59", "remaining_time": "2 days, 14:47:54"}
+{"current_steps": 356, "total_steps": 2200, "loss": 2.4388, "accuracy": 0.71875, "lr": 9.430932830306587e-06, "epoch": 0.6466848319709355, "percentage": 16.18, "elapsed_time": "12:06:54", "remaining_time": "2 days, 14:45:11"}
+{"current_steps": 357, "total_steps": 2200, "loss": 2.706, "accuracy": 0.5625, "lr": 9.427800967287963e-06, "epoch": 0.6485013623978202, "percentage": 16.23, "elapsed_time": "12:08:47", "remaining_time": "2 days, 14:42:22"}
+{"current_steps": 358, "total_steps": 2200, "loss": 2.3605, "accuracy": 0.65625, "lr": 9.424661094799273e-06, "epoch": 0.6503178928247049, "percentage": 16.27, "elapsed_time": "12:10:24", "remaining_time": "2 days, 14:38:08"}
+{"current_steps": 359, "total_steps": 2200, "loss": 1.7844, "accuracy": 0.828125, "lr": 9.421513219243262e-06, "epoch": 0.6521344232515894, "percentage": 16.32, "elapsed_time": "12:12:26", "remaining_time": "2 days, 14:36:03"}
+{"current_steps": 360, "total_steps": 2200, "loss": 2.4976, "accuracy": 0.671875, "lr": 9.418357347038999e-06, "epoch": 0.6539509536784741, "percentage": 16.36, "elapsed_time": "12:14:32", "remaining_time": "2 days, 14:34:19"}
+{"current_steps": 361, "total_steps": 2200, "loss": 2.1398, "accuracy": 0.703125, "lr": 9.415193484621852e-06, "epoch": 0.6557674841053588, "percentage": 16.41, "elapsed_time": "12:16:27", "remaining_time": "2 days, 14:31:38"}
+{"current_steps": 362, "total_steps": 2200, "loss": 2.2361, "accuracy": 0.75, "lr": 9.412021638443491e-06, "epoch": 0.6575840145322435, "percentage": 16.45, "elapsed_time": "12:18:29", "remaining_time": "2 days, 14:29:37"}
+{"current_steps": 363, "total_steps": 2200, "loss": 2.4009, "accuracy": 0.703125, "lr": 9.408841814971862e-06, "epoch": 0.659400544959128, "percentage": 16.5, "elapsed_time": "12:20:35", "remaining_time": "2 days, 14:27:49"}
+{"current_steps": 364, "total_steps": 2200, "loss": 2.141, "accuracy": 0.671875, "lr": 9.405654020691178e-06, "epoch": 0.6612170753860127, "percentage": 16.55, "elapsed_time": "12:22:52", "remaining_time": "2 days, 14:27:03"}
+{"current_steps": 365, "total_steps": 2200, "loss": 2.0511, "accuracy": 0.8125, "lr": 9.402458262101906e-06, "epoch": 0.6630336058128974, "percentage": 16.59, "elapsed_time": "12:25:03", "remaining_time": "2 days, 14:25:44"}
+{"current_steps": 366, "total_steps": 2200, "loss": 2.1756, "accuracy": 0.6875, "lr": 9.399254545720757e-06, "epoch": 0.6648501362397821, "percentage": 16.64, "elapsed_time": "12:27:13", "remaining_time": "2 days, 14:24:19"}
+{"current_steps": 367, "total_steps": 2200, "loss": 2.4018, "accuracy": 0.65625, "lr": 9.396042878080661e-06, "epoch": 0.6666666666666666, "percentage": 16.68, "elapsed_time": "12:29:08", "remaining_time": "2 days, 14:21:36"}
+{"current_steps": 368, "total_steps": 2200, "loss": 2.4262, "accuracy": 0.671875, "lr": 9.392823265730775e-06, "epoch": 0.6684831970935513, "percentage": 16.73, "elapsed_time": "12:30:50", "remaining_time": "2 days, 14:17:50"}
+{"current_steps": 369, "total_steps": 2200, "loss": 2.2511, "accuracy": 0.71875, "lr": 9.389595715236446e-06, "epoch": 0.670299727520436, "percentage": 16.77, "elapsed_time": "12:32:44", "remaining_time": "2 days, 14:15:07"}
+{"current_steps": 370, "total_steps": 2200, "loss": 2.3858, "accuracy": 0.65625, "lr": 9.386360233179206e-06, "epoch": 0.6721162579473207, "percentage": 16.82, "elapsed_time": "12:34:59", "remaining_time": "2 days, 14:14:06"}
+{"current_steps": 371, "total_steps": 2200, "loss": 2.6967, "accuracy": 0.71875, "lr": 9.383116826156775e-06, "epoch": 0.6739327883742052, "percentage": 16.86, "elapsed_time": "12:36:51", "remaining_time": "2 days, 14:11:14"}
+{"current_steps": 372, "total_steps": 2200, "loss": 2.4608, "accuracy": 0.671875, "lr": 9.37986550078302e-06, "epoch": 0.6757493188010899, "percentage": 16.91, "elapsed_time": "12:39:05", "remaining_time": "2 days, 14:10:09"}
+{"current_steps": 373, "total_steps": 2200, "loss": 2.4701, "accuracy": 0.625, "lr": 9.376606263687959e-06, "epoch": 0.6775658492279746, "percentage": 16.95, "elapsed_time": "12:41:09", "remaining_time": "2 days, 14:08:13"}
+{"current_steps": 374, "total_steps": 2200, "loss": 2.1297, "accuracy": 0.640625, "lr": 9.373339121517748e-06, "epoch": 0.6793823796548593, "percentage": 17.0, "elapsed_time": "12:43:16", "remaining_time": "2 days, 14:06:35"}
+{"current_steps": 375, "total_steps": 2200, "loss": 2.4704, "accuracy": 0.625, "lr": 9.370064080934654e-06, "epoch": 0.6811989100817438, "percentage": 17.05, "elapsed_time": "12:44:47", "remaining_time": "2 days, 14:01:58"}
+{"current_steps": 376, "total_steps": 2200, "loss": 2.2813, "accuracy": 0.671875, "lr": 9.366781148617056e-06, "epoch": 0.6830154405086285, "percentage": 17.09, "elapsed_time": "12:46:43", "remaining_time": "2 days, 13:59:27"}
+{"current_steps": 377, "total_steps": 2200, "loss": 2.1442, "accuracy": 0.75, "lr": 9.363490331259426e-06, "epoch": 0.6848319709355132, "percentage": 17.14, "elapsed_time": "12:48:46", "remaining_time": "2 days, 13:57:25"}
+{"current_steps": 378, "total_steps": 2200, "loss": 2.423, "accuracy": 0.6875, "lr": 9.360191635572313e-06, "epoch": 0.6866485013623979, "percentage": 17.18, "elapsed_time": "12:50:57", "remaining_time": "2 days, 13:56:07"}
+{"current_steps": 379, "total_steps": 2200, "loss": 2.4801, "accuracy": 0.625, "lr": 9.356885068282334e-06, "epoch": 0.6884650317892824, "percentage": 17.23, "elapsed_time": "12:53:07", "remaining_time": "2 days, 13:54:40"}
+{"current_steps": 380, "total_steps": 2200, "loss": 2.2711, "accuracy": 0.703125, "lr": 9.353570636132151e-06, "epoch": 0.6902815622161671, "percentage": 17.27, "elapsed_time": "12:55:16", "remaining_time": "2 days, 13:53:10"}
+{"current_steps": 381, "total_steps": 2200, "loss": 2.3863, "accuracy": 0.75, "lr": 9.350248345880471e-06, "epoch": 0.6920980926430518, "percentage": 17.32, "elapsed_time": "12:57:09", "remaining_time": "2 days, 13:50:23"}
+{"current_steps": 382, "total_steps": 2200, "loss": 2.5434, "accuracy": 0.640625, "lr": 9.346918204302022e-06, "epoch": 0.6939146230699365, "percentage": 17.36, "elapsed_time": "12:59:17", "remaining_time": "2 days, 13:48:45"}
+{"current_steps": 383, "total_steps": 2200, "loss": 2.1699, "accuracy": 0.71875, "lr": 9.343580218187544e-06, "epoch": 0.695731153496821, "percentage": 17.41, "elapsed_time": "13:01:29", "remaining_time": "2 days, 13:47:28"}
+{"current_steps": 384, "total_steps": 2200, "loss": 2.3138, "accuracy": 0.734375, "lr": 9.340234394343768e-06, "epoch": 0.6975476839237057, "percentage": 17.45, "elapsed_time": "13:03:18", "remaining_time": "2 days, 13:44:24"}
+{"current_steps": 385, "total_steps": 2200, "loss": 2.4864, "accuracy": 0.703125, "lr": 9.336880739593415e-06, "epoch": 0.6993642143505904, "percentage": 17.5, "elapsed_time": "13:05:38", "remaining_time": "2 days, 13:43:42"}
+{"current_steps": 386, "total_steps": 2200, "loss": 2.6878, "accuracy": 0.578125, "lr": 9.33351926077517e-06, "epoch": 0.701180744777475, "percentage": 17.55, "elapsed_time": "13:07:39", "remaining_time": "2 days, 13:41:33"}
+{"current_steps": 387, "total_steps": 2200, "loss": 2.343, "accuracy": 0.796875, "lr": 9.330149964743674e-06, "epoch": 0.7029972752043597, "percentage": 17.59, "elapsed_time": "13:09:17", "remaining_time": "2 days, 13:37:35"}
+{"current_steps": 388, "total_steps": 2200, "loss": 2.3036, "accuracy": 0.6875, "lr": 9.326772858369506e-06, "epoch": 0.7048138056312443, "percentage": 17.64, "elapsed_time": "13:11:11", "remaining_time": "2 days, 13:34:55"}
+{"current_steps": 389, "total_steps": 2200, "loss": 2.241, "accuracy": 0.75, "lr": 9.323387948539176e-06, "epoch": 0.706630336058129, "percentage": 17.68, "elapsed_time": "13:12:41", "remaining_time": "2 days, 13:30:22"}
+{"current_steps": 390, "total_steps": 2200, "loss": 2.284, "accuracy": 0.71875, "lr": 9.319995242155102e-06, "epoch": 0.7084468664850136, "percentage": 17.73, "elapsed_time": "13:15:00", "remaining_time": "2 days, 13:29:40"}
+{"current_steps": 391, "total_steps": 2200, "loss": 2.7281, "accuracy": 0.625, "lr": 9.316594746135608e-06, "epoch": 0.7102633969118983, "percentage": 17.77, "elapsed_time": "13:17:22", "remaining_time": "2 days, 13:29:06"}
+{"current_steps": 392, "total_steps": 2200, "loss": 2.5247, "accuracy": 0.59375, "lr": 9.313186467414892e-06, "epoch": 0.7120799273387829, "percentage": 17.82, "elapsed_time": "13:19:06", "remaining_time": "2 days, 13:25:42"}
+{"current_steps": 393, "total_steps": 2200, "loss": 2.5284, "accuracy": 0.609375, "lr": 9.30977041294303e-06, "epoch": 0.7138964577656676, "percentage": 17.86, "elapsed_time": "13:20:58", "remaining_time": "2 days, 13:22:49"}
+{"current_steps": 394, "total_steps": 2200, "loss": 2.3394, "accuracy": 0.703125, "lr": 9.306346589685956e-06, "epoch": 0.7157129881925522, "percentage": 17.91, "elapsed_time": "13:23:14", "remaining_time": "2 days, 13:21:52"}
+{"current_steps": 395, "total_steps": 2200, "loss": 2.3875, "accuracy": 0.6875, "lr": 9.302915004625435e-06, "epoch": 0.7175295186194369, "percentage": 17.95, "elapsed_time": "13:25:31", "remaining_time": "2 days, 13:20:58"}
+{"current_steps": 396, "total_steps": 2200, "loss": 2.7001, "accuracy": 0.65625, "lr": 9.29947566475907e-06, "epoch": 0.7193460490463215, "percentage": 18.0, "elapsed_time": "13:27:42", "remaining_time": "2 days, 13:19:35"}
+{"current_steps": 397, "total_steps": 2200, "loss": 2.347, "accuracy": 0.640625, "lr": 9.296028577100271e-06, "epoch": 0.7211625794732062, "percentage": 18.05, "elapsed_time": "13:29:34", "remaining_time": "2 days, 13:16:45"}
+{"current_steps": 398, "total_steps": 2200, "loss": 2.2229, "accuracy": 0.78125, "lr": 9.292573748678254e-06, "epoch": 0.7229791099000908, "percentage": 18.09, "elapsed_time": "13:31:58", "remaining_time": "2 days, 13:16:18"}
+{"current_steps": 399, "total_steps": 2200, "loss": 2.2271, "accuracy": 0.703125, "lr": 9.289111186538013e-06, "epoch": 0.7247956403269755, "percentage": 18.14, "elapsed_time": "13:34:02", "remaining_time": "2 days, 13:14:24"}
+{"current_steps": 400, "total_steps": 2200, "loss": 2.3595, "accuracy": 0.75, "lr": 9.285640897740316e-06, "epoch": 0.7266121707538601, "percentage": 18.18, "elapsed_time": "13:36:01", "remaining_time": "2 days, 13:12:05"}
+{"current_steps": 401, "total_steps": 2200, "loss": 2.3002, "accuracy": 0.71875, "lr": 9.282162889361686e-06, "epoch": 0.7284287011807448, "percentage": 18.23, "elapsed_time": "13:37:54", "remaining_time": "2 days, 13:09:23"}
+{"current_steps": 402, "total_steps": 2200, "loss": 2.1543, "accuracy": 0.75, "lr": 9.278677168494388e-06, "epoch": 0.7302452316076294, "percentage": 18.27, "elapsed_time": "13:39:45", "remaining_time": "2 days, 13:06:28"}
+{"current_steps": 403, "total_steps": 2200, "loss": 2.3315, "accuracy": 0.6875, "lr": 9.275183742246412e-06, "epoch": 0.7320617620345141, "percentage": 18.32, "elapsed_time": "13:42:01", "remaining_time": "2 days, 13:05:28"}
+{"current_steps": 404, "total_steps": 2200, "loss": 2.3528, "accuracy": 0.6875, "lr": 9.271682617741466e-06, "epoch": 0.7338782924613987, "percentage": 18.36, "elapsed_time": "13:44:08", "remaining_time": "2 days, 13:03:45"}
+{"current_steps": 405, "total_steps": 2200, "loss": 2.1505, "accuracy": 0.65625, "lr": 9.268173802118949e-06, "epoch": 0.7356948228882834, "percentage": 18.41, "elapsed_time": "13:45:59", "remaining_time": "2 days, 13:00:53"}
+{"current_steps": 406, "total_steps": 2200, "loss": 2.1873, "accuracy": 0.703125, "lr": 9.264657302533947e-06, "epoch": 0.737511353315168, "percentage": 18.45, "elapsed_time": "13:48:05", "remaining_time": "2 days, 12:59:07"}
+{"current_steps": 407, "total_steps": 2200, "loss": 2.3587, "accuracy": 0.703125, "lr": 9.261133126157218e-06, "epoch": 0.7393278837420527, "percentage": 18.5, "elapsed_time": "13:50:14", "remaining_time": "2 days, 12:57:33"}
+{"current_steps": 408, "total_steps": 2200, "loss": 2.444, "accuracy": 0.6875, "lr": 9.257601280175167e-06, "epoch": 0.7411444141689373, "percentage": 18.55, "elapsed_time": "13:51:55", "remaining_time": "2 days, 12:53:54"}
+{"current_steps": 409, "total_steps": 2200, "loss": 2.5324, "accuracy": 0.640625, "lr": 9.254061771789847e-06, "epoch": 0.742960944595822, "percentage": 18.59, "elapsed_time": "13:54:10", "remaining_time": "2 days, 12:52:48"}
+{"current_steps": 410, "total_steps": 2200, "loss": 2.1895, "accuracy": 0.640625, "lr": 9.25051460821893e-06, "epoch": 0.7447774750227066, "percentage": 18.64, "elapsed_time": "13:56:12", "remaining_time": "2 days, 12:50:46"}
+{"current_steps": 411, "total_steps": 2200, "loss": 2.46, "accuracy": 0.671875, "lr": 9.2469597966957e-06, "epoch": 0.7465940054495913, "percentage": 18.68, "elapsed_time": "13:58:18", "remaining_time": "2 days, 12:48:58"}
+{"current_steps": 412, "total_steps": 2200, "loss": 2.2929, "accuracy": 0.75, "lr": 9.243397344469037e-06, "epoch": 0.7484105358764759, "percentage": 18.73, "elapsed_time": "14:00:53", "remaining_time": "2 days, 12:49:18"}
+{"current_steps": 413, "total_steps": 2200, "loss": 1.8794, "accuracy": 0.8125, "lr": 9.239827258803402e-06, "epoch": 0.7502270663033606, "percentage": 18.77, "elapsed_time": "14:02:51", "remaining_time": "2 days, 12:46:58"}
+{"current_steps": 414, "total_steps": 2200, "loss": 2.5646, "accuracy": 0.6875, "lr": 9.23624954697882e-06, "epoch": 0.7520435967302452, "percentage": 18.82, "elapsed_time": "14:05:11", "remaining_time": "2 days, 12:46:07"}
+{"current_steps": 415, "total_steps": 2200, "loss": 2.2284, "accuracy": 0.765625, "lr": 9.232664216290868e-06, "epoch": 0.7538601271571299, "percentage": 18.86, "elapsed_time": "14:07:16", "remaining_time": "2 days, 12:44:16"}
+{"current_steps": 416, "total_steps": 2200, "loss": 2.2211, "accuracy": 0.6875, "lr": 9.229071274050663e-06, "epoch": 0.7556766575840145, "percentage": 18.91, "elapsed_time": "14:09:01", "remaining_time": "2 days, 12:41:02"}
+{"current_steps": 417, "total_steps": 2200, "loss": 2.4446, "accuracy": 0.59375, "lr": 9.225470727584835e-06, "epoch": 0.7574931880108992, "percentage": 18.95, "elapsed_time": "14:11:24", "remaining_time": "2 days, 12:40:25"}
+{"current_steps": 418, "total_steps": 2200, "loss": 2.7171, "accuracy": 0.59375, "lr": 9.221862584235527e-06, "epoch": 0.7593097184377838, "percentage": 19.0, "elapsed_time": "14:13:21", "remaining_time": "2 days, 12:37:59"}
+{"current_steps": 419, "total_steps": 2200, "loss": 2.4292, "accuracy": 0.59375, "lr": 9.218246851360374e-06, "epoch": 0.7611262488646685, "percentage": 19.05, "elapsed_time": "14:15:18", "remaining_time": "2 days, 12:35:36"}
+{"current_steps": 420, "total_steps": 2200, "loss": 2.5042, "accuracy": 0.609375, "lr": 9.214623536332483e-06, "epoch": 0.7629427792915532, "percentage": 19.09, "elapsed_time": "14:17:40", "remaining_time": "2 days, 12:34:54"}
+{"current_steps": 421, "total_steps": 2200, "loss": 2.2064, "accuracy": 0.703125, "lr": 9.210992646540425e-06, "epoch": 0.7647593097184378, "percentage": 19.14, "elapsed_time": "14:19:35", "remaining_time": "2 days, 12:32:22"}
+{"current_steps": 422, "total_steps": 2200, "loss": 2.2389, "accuracy": 0.6875, "lr": 9.207354189388214e-06, "epoch": 0.7665758401453224, "percentage": 19.18, "elapsed_time": "14:21:32", "remaining_time": "2 days, 12:29:54"}
+{"current_steps": 423, "total_steps": 2200, "loss": 2.2853, "accuracy": 0.71875, "lr": 9.203708172295299e-06, "epoch": 0.7683923705722071, "percentage": 19.23, "elapsed_time": "14:23:28", "remaining_time": "2 days, 12:27:25"}
+{"current_steps": 424, "total_steps": 2200, "loss": 2.3022, "accuracy": 0.6875, "lr": 9.200054602696544e-06, "epoch": 0.7702089009990918, "percentage": 19.27, "elapsed_time": "14:25:45", "remaining_time": "2 days, 12:26:24"}
+{"current_steps": 425, "total_steps": 2200, "loss": 2.5216, "accuracy": 0.671875, "lr": 9.196393488042213e-06, "epoch": 0.7720254314259763, "percentage": 19.32, "elapsed_time": "14:27:59", "remaining_time": "2 days, 12:25:09"}
+{"current_steps": 426, "total_steps": 2200, "loss": 2.0472, "accuracy": 0.734375, "lr": 9.192724835797956e-06, "epoch": 0.773841961852861, "percentage": 19.36, "elapsed_time": "14:29:53", "remaining_time": "2 days, 12:22:29"}
+{"current_steps": 427, "total_steps": 2200, "loss": 2.2618, "accuracy": 0.75, "lr": 9.189048653444796e-06, "epoch": 0.7756584922797457, "percentage": 19.41, "elapsed_time": "14:31:56", "remaining_time": "2 days, 12:20:29"}
+{"current_steps": 428, "total_steps": 2200, "loss": 2.4923, "accuracy": 0.703125, "lr": 9.185364948479109e-06, "epoch": 0.7774750227066304, "percentage": 19.45, "elapsed_time": "14:34:15", "remaining_time": "2 days, 12:19:35"}
+{"current_steps": 429, "total_steps": 2200, "loss": 2.5924, "accuracy": 0.671875, "lr": 9.181673728412605e-06, "epoch": 0.779291553133515, "percentage": 19.5, "elapsed_time": "14:36:16", "remaining_time": "2 days, 12:17:27"}
+{"current_steps": 430, "total_steps": 2200, "loss": 2.6208, "accuracy": 0.5625, "lr": 9.17797500077233e-06, "epoch": 0.7811080835603996, "percentage": 19.55, "elapsed_time": "14:38:27", "remaining_time": "2 days, 12:15:57"}
+{"current_steps": 431, "total_steps": 2200, "loss": 2.4676, "accuracy": 0.609375, "lr": 9.174268773100636e-06, "epoch": 0.7829246139872843, "percentage": 19.59, "elapsed_time": "14:40:04", "remaining_time": "2 days, 12:12:11"}
+{"current_steps": 432, "total_steps": 2200, "loss": 2.3366, "accuracy": 0.671875, "lr": 9.170555052955158e-06, "epoch": 0.784741144414169, "percentage": 19.64, "elapsed_time": "14:42:17", "remaining_time": "2 days, 12:10:51"}
+{"current_steps": 433, "total_steps": 2200, "loss": 2.2501, "accuracy": 0.703125, "lr": 9.166833847908825e-06, "epoch": 0.7865576748410535, "percentage": 19.68, "elapsed_time": "14:44:10", "remaining_time": "2 days, 12:08:08"}
+{"current_steps": 434, "total_steps": 2200, "loss": 2.5214, "accuracy": 0.703125, "lr": 9.163105165549819e-06, "epoch": 0.7883742052679382, "percentage": 19.73, "elapsed_time": "14:46:29", "remaining_time": "2 days, 12:07:14"}
+{"current_steps": 435, "total_steps": 2200, "loss": 2.053, "accuracy": 0.75, "lr": 9.159369013481574e-06, "epoch": 0.7901907356948229, "percentage": 19.77, "elapsed_time": "14:48:20", "remaining_time": "2 days, 12:04:26"}
+{"current_steps": 436, "total_steps": 2200, "loss": 2.4289, "accuracy": 0.703125, "lr": 9.155625399322754e-06, "epoch": 0.7920072661217076, "percentage": 19.82, "elapsed_time": "14:50:09", "remaining_time": "2 days, 12:01:27"}
+{"current_steps": 437, "total_steps": 2200, "loss": 2.0608, "accuracy": 0.734375, "lr": 9.15187433070724e-06, "epoch": 0.7938237965485921, "percentage": 19.86, "elapsed_time": "14:52:02", "remaining_time": "2 days, 11:58:46"}
+{"current_steps": 438, "total_steps": 2200, "loss": 2.338, "accuracy": 0.703125, "lr": 9.148115815284113e-06, "epoch": 0.7956403269754768, "percentage": 19.91, "elapsed_time": "14:53:57", "remaining_time": "2 days, 11:56:16"}
+{"current_steps": 439, "total_steps": 2200, "loss": 2.3485, "accuracy": 0.671875, "lr": 9.144349860717643e-06, "epoch": 0.7974568574023615, "percentage": 19.95, "elapsed_time": "14:56:06", "remaining_time": "2 days, 11:54:37"}
+{"current_steps": 440, "total_steps": 2200, "loss": 2.6513, "accuracy": 0.65625, "lr": 9.140576474687263e-06, "epoch": 0.7992733878292462, "percentage": 20.0, "elapsed_time": "14:58:02", "remaining_time": "2 days, 11:52:10"}
+{"current_steps": 441, "total_steps": 2200, "loss": 2.3956, "accuracy": 0.625, "lr": 9.13679566488757e-06, "epoch": 0.8010899182561307, "percentage": 20.05, "elapsed_time": "15:00:17", "remaining_time": "2 days, 11:50:56"}
+{"current_steps": 442, "total_steps": 2200, "loss": 2.243, "accuracy": 0.671875, "lr": 9.133007439028288e-06, "epoch": 0.8029064486830154, "percentage": 20.09, "elapsed_time": "15:02:44", "remaining_time": "2 days, 11:50:34"}
+{"current_steps": 443, "total_steps": 2200, "loss": 2.2574, "accuracy": 0.671875, "lr": 9.129211804834271e-06, "epoch": 0.8047229791099001, "percentage": 20.14, "elapsed_time": "15:04:41", "remaining_time": "2 days, 11:48:07"}
+{"current_steps": 444, "total_steps": 2200, "loss": 2.6959, "accuracy": 0.609375, "lr": 9.12540877004548e-06, "epoch": 0.8065395095367848, "percentage": 20.18, "elapsed_time": "15:06:29", "remaining_time": "2 days, 11:45:06"}
+{"current_steps": 445, "total_steps": 2200, "loss": 2.6036, "accuracy": 0.59375, "lr": 9.12159834241696e-06, "epoch": 0.8083560399636693, "percentage": 20.23, "elapsed_time": "15:08:18", "remaining_time": "2 days, 11:42:12"}
+{"current_steps": 446, "total_steps": 2200, "loss": 2.2157, "accuracy": 0.765625, "lr": 9.117780529718843e-06, "epoch": 0.810172570390554, "percentage": 20.27, "elapsed_time": "15:10:28", "remaining_time": "2 days, 11:40:40"}
+{"current_steps": 447, "total_steps": 2200, "loss": 2.6346, "accuracy": 0.734375, "lr": 9.113955339736309e-06, "epoch": 0.8119891008174387, "percentage": 20.32, "elapsed_time": "15:12:49", "remaining_time": "2 days, 11:39:47"}
+{"current_steps": 448, "total_steps": 2200, "loss": 2.3339, "accuracy": 0.703125, "lr": 9.11012278026959e-06, "epoch": 0.8138056312443234, "percentage": 20.36, "elapsed_time": "15:15:08", "remaining_time": "2 days, 11:38:51"}
+{"current_steps": 449, "total_steps": 2200, "loss": 2.8324, "accuracy": 0.609375, "lr": 9.106282859133936e-06, "epoch": 0.815622161671208, "percentage": 20.41, "elapsed_time": "15:17:13", "remaining_time": "2 days, 11:36:59"}
+{"current_steps": 450, "total_steps": 2200, "loss": 2.5562, "accuracy": 0.609375, "lr": 9.102435584159623e-06, "epoch": 0.8174386920980926, "percentage": 20.45, "elapsed_time": "15:19:02", "remaining_time": "2 days, 11:34:03"}
+{"current_steps": 451, "total_steps": 2200, "loss": 2.1826, "accuracy": 0.78125, "lr": 9.098580963191908e-06, "epoch": 0.8192552225249773, "percentage": 20.5, "elapsed_time": "15:21:17", "remaining_time": "2 days, 11:32:50"}
+{"current_steps": 452, "total_steps": 2200, "loss": 2.2843, "accuracy": 0.78125, "lr": 9.094719004091039e-06, "epoch": 0.821071752951862, "percentage": 20.55, "elapsed_time": "15:23:06", "remaining_time": "2 days, 11:29:55"}
+{"current_steps": 453, "total_steps": 2200, "loss": 2.3556, "accuracy": 0.640625, "lr": 9.090849714732217e-06, "epoch": 0.8228882833787466, "percentage": 20.59, "elapsed_time": "15:25:06", "remaining_time": "2 days, 11:27:39"}
+{"current_steps": 454, "total_steps": 2200, "loss": 2.2768, "accuracy": 0.671875, "lr": 9.086973103005602e-06, "epoch": 0.8247048138056312, "percentage": 20.64, "elapsed_time": "15:26:41", "remaining_time": "2 days, 11:23:51"}
+{"current_steps": 455, "total_steps": 2200, "loss": 2.3881, "accuracy": 0.703125, "lr": 9.08308917681628e-06, "epoch": 0.8265213442325159, "percentage": 20.68, "elapsed_time": "15:28:47", "remaining_time": "2 days, 11:22:05"}
+{"current_steps": 456, "total_steps": 2200, "loss": 2.8503, "accuracy": 0.640625, "lr": 9.07919794408425e-06, "epoch": 0.8283378746594006, "percentage": 20.73, "elapsed_time": "15:31:03", "remaining_time": "2 days, 11:20:51"}
+{"current_steps": 457, "total_steps": 2200, "loss": 2.1639, "accuracy": 0.703125, "lr": 9.075299412744417e-06, "epoch": 0.8301544050862852, "percentage": 20.77, "elapsed_time": "15:32:52", "remaining_time": "2 days, 11:17:58"}
+{"current_steps": 458, "total_steps": 2200, "loss": 2.4048, "accuracy": 0.609375, "lr": 9.07139359074656e-06, "epoch": 0.8319709355131698, "percentage": 20.82, "elapsed_time": "15:34:26", "remaining_time": "2 days, 11:14:10"}
+{"current_steps": 459, "total_steps": 2200, "loss": 2.8725, "accuracy": 0.546875, "lr": 9.067480486055333e-06, "epoch": 0.8337874659400545, "percentage": 20.86, "elapsed_time": "15:36:11", "remaining_time": "2 days, 11:11:00"}
+{"current_steps": 460, "total_steps": 2200, "loss": 2.2922, "accuracy": 0.734375, "lr": 9.063560106650238e-06, "epoch": 0.8356039963669392, "percentage": 20.91, "elapsed_time": "15:37:48", "remaining_time": "2 days, 11:07:20"}
+{"current_steps": 461, "total_steps": 2200, "loss": 2.3288, "accuracy": 0.640625, "lr": 9.059632460525613e-06, "epoch": 0.8374205267938238, "percentage": 20.95, "elapsed_time": "15:40:13", "remaining_time": "2 days, 11:06:44"}
+{"current_steps": 462, "total_steps": 2200, "loss": 2.3694, "accuracy": 0.6875, "lr": 9.055697555690607e-06, "epoch": 0.8392370572207084, "percentage": 21.0, "elapsed_time": "15:42:11", "remaining_time": "2 days, 11:04:25"}
+{"current_steps": 463, "total_steps": 2200, "loss": 2.5167, "accuracy": 0.5625, "lr": 9.051755400169182e-06, "epoch": 0.8410535876475931, "percentage": 21.05, "elapsed_time": "15:43:56", "remaining_time": "2 days, 11:01:17"}
+{"current_steps": 464, "total_steps": 2200, "loss": 2.4938, "accuracy": 0.71875, "lr": 9.047806002000075e-06, "epoch": 0.8428701180744778, "percentage": 21.09, "elapsed_time": "15:45:45", "remaining_time": "2 days, 10:58:26"}
+{"current_steps": 465, "total_steps": 2200, "loss": 2.5881, "accuracy": 0.671875, "lr": 9.043849369236799e-06, "epoch": 0.8446866485013624, "percentage": 21.14, "elapsed_time": "15:47:49", "remaining_time": "2 days, 10:56:29"}
+{"current_steps": 466, "total_steps": 2200, "loss": 2.3914, "accuracy": 0.6875, "lr": 9.039885509947616e-06, "epoch": 0.846503178928247, "percentage": 21.18, "elapsed_time": "15:49:40", "remaining_time": "2 days, 10:53:47"}
+{"current_steps": 467, "total_steps": 2200, "loss": 2.4194, "accuracy": 0.71875, "lr": 9.035914432215527e-06, "epoch": 0.8483197093551317, "percentage": 21.23, "elapsed_time": "15:52:02", "remaining_time": "2 days, 10:52:58"}
+{"current_steps": 468, "total_steps": 2200, "loss": 2.7262, "accuracy": 0.625, "lr": 9.031936144138247e-06, "epoch": 0.8501362397820164, "percentage": 21.27, "elapsed_time": "15:54:23", "remaining_time": "2 days, 10:52:02"}
+{"current_steps": 469, "total_steps": 2200, "loss": 2.7323, "accuracy": 0.703125, "lr": 9.027950653828202e-06, "epoch": 0.851952770208901, "percentage": 21.32, "elapsed_time": "15:56:12", "remaining_time": "2 days, 10:49:13"}
+{"current_steps": 470, "total_steps": 2200, "loss": 2.457, "accuracy": 0.625, "lr": 9.0239579694125e-06, "epoch": 0.8537693006357856, "percentage": 21.36, "elapsed_time": "15:58:05", "remaining_time": "2 days, 10:46:35"}
+{"current_steps": 471, "total_steps": 2200, "loss": 2.2809, "accuracy": 0.6875, "lr": 9.019958099032919e-06, "epoch": 0.8555858310626703, "percentage": 21.41, "elapsed_time": "16:00:02", "remaining_time": "2 days, 10:44:14"}
+{"current_steps": 472, "total_steps": 2200, "loss": 1.9536, "accuracy": 0.734375, "lr": 9.015951050845891e-06, "epoch": 0.857402361489555, "percentage": 21.45, "elapsed_time": "16:02:09", "remaining_time": "2 days, 10:42:27"}
+{"current_steps": 473, "total_steps": 2200, "loss": 2.5658, "accuracy": 0.703125, "lr": 9.011936833022485e-06, "epoch": 0.8592188919164396, "percentage": 21.5, "elapsed_time": "16:04:10", "remaining_time": "2 days, 10:40:21"}
+{"current_steps": 474, "total_steps": 2200, "loss": 2.1854, "accuracy": 0.796875, "lr": 9.00791545374839e-06, "epoch": 0.8610354223433242, "percentage": 21.55, "elapsed_time": "16:06:19", "remaining_time": "2 days, 10:38:43"}
+{"current_steps": 475, "total_steps": 2200, "loss": 2.1995, "accuracy": 0.75, "lr": 9.0038869212239e-06, "epoch": 0.8628519527702089, "percentage": 21.59, "elapsed_time": "16:08:24", "remaining_time": "2 days, 10:36:49"}
+{"current_steps": 476, "total_steps": 2200, "loss": 2.7347, "accuracy": 0.5625, "lr": 8.99985124366389e-06, "epoch": 0.8646684831970936, "percentage": 21.64, "elapsed_time": "16:10:32", "remaining_time": "2 days, 10:35:11"}
+{"current_steps": 477, "total_steps": 2200, "loss": 2.4449, "accuracy": 0.65625, "lr": 8.995808429297815e-06, "epoch": 0.8664850136239782, "percentage": 21.68, "elapsed_time": "16:12:26", "remaining_time": "2 days, 10:32:35"}
+{"current_steps": 478, "total_steps": 2200, "loss": 2.3072, "accuracy": 0.640625, "lr": 8.991758486369675e-06, "epoch": 0.8683015440508629, "percentage": 21.73, "elapsed_time": "16:14:53", "remaining_time": "2 days, 10:32:03"}
+{"current_steps": 479, "total_steps": 2200, "loss": 2.2815, "accuracy": 0.6875, "lr": 8.987701423138007e-06, "epoch": 0.8701180744777475, "percentage": 21.77, "elapsed_time": "16:16:36", "remaining_time": "2 days, 10:28:53"}
+{"current_steps": 480, "total_steps": 2200, "loss": 2.09, "accuracy": 0.734375, "lr": 8.983637247875872e-06, "epoch": 0.8719346049046321, "percentage": 21.82, "elapsed_time": "16:18:40", "remaining_time": "2 days, 10:26:55"}
+{"current_steps": 481, "total_steps": 2200, "loss": 2.4566, "accuracy": 0.578125, "lr": 8.979565968870831e-06, "epoch": 0.8737511353315168, "percentage": 21.86, "elapsed_time": "16:20:44", "remaining_time": "2 days, 10:24:58"}
+{"current_steps": 482, "total_steps": 2200, "loss": 2.2736, "accuracy": 0.703125, "lr": 8.975487594424927e-06, "epoch": 0.8755676657584015, "percentage": 21.91, "elapsed_time": "16:22:50", "remaining_time": "2 days, 10:23:09"}
+{"current_steps": 483, "total_steps": 2200, "loss": 2.3534, "accuracy": 0.671875, "lr": 8.971402132854677e-06, "epoch": 0.8773841961852861, "percentage": 21.95, "elapsed_time": "16:25:02", "remaining_time": "2 days, 10:21:40"}
+{"current_steps": 484, "total_steps": 2200, "loss": 2.5016, "accuracy": 0.65625, "lr": 8.967309592491052e-06, "epoch": 0.8792007266121707, "percentage": 22.0, "elapsed_time": "16:26:52", "remaining_time": "2 days, 10:18:55"}
+{"current_steps": 485, "total_steps": 2200, "loss": 1.9177, "accuracy": 0.78125, "lr": 8.963209981679451e-06, "epoch": 0.8810172570390554, "percentage": 22.05, "elapsed_time": "16:29:18", "remaining_time": "2 days, 10:18:15"}
+{"current_steps": 486, "total_steps": 2200, "loss": 2.1547, "accuracy": 0.765625, "lr": 8.959103308779696e-06, "epoch": 0.8828337874659401, "percentage": 22.09, "elapsed_time": "16:31:01", "remaining_time": "2 days, 10:15:04"}
+{"current_steps": 487, "total_steps": 2200, "loss": 2.2043, "accuracy": 0.6875, "lr": 8.954989582166009e-06, "epoch": 0.8846503178928247, "percentage": 22.14, "elapsed_time": "16:33:06", "remaining_time": "2 days, 10:13:12"}
+{"current_steps": 488, "total_steps": 2200, "loss": 2.0246, "accuracy": 0.734375, "lr": 8.95086881022699e-06, "epoch": 0.8864668483197093, "percentage": 22.18, "elapsed_time": "16:35:06", "remaining_time": "2 days, 10:11:00"}
+{"current_steps": 489, "total_steps": 2200, "loss": 2.5066, "accuracy": 0.671875, "lr": 8.946741001365616e-06, "epoch": 0.888283378746594, "percentage": 22.23, "elapsed_time": "16:36:59", "remaining_time": "2 days, 10:08:25"}
+{"current_steps": 490, "total_steps": 2200, "loss": 2.2655, "accuracy": 0.71875, "lr": 8.942606163999205e-06, "epoch": 0.8900999091734787, "percentage": 22.27, "elapsed_time": "16:38:43", "remaining_time": "2 days, 10:05:21"}
+{"current_steps": 491, "total_steps": 2200, "loss": 2.1676, "accuracy": 0.71875, "lr": 8.938464306559412e-06, "epoch": 0.8919164396003633, "percentage": 22.32, "elapsed_time": "16:40:57", "remaining_time": "2 days, 10:03:57"}
+{"current_steps": 492, "total_steps": 2200, "loss": 2.3435, "accuracy": 0.734375, "lr": 8.934315437492203e-06, "epoch": 0.8937329700272479, "percentage": 22.36, "elapsed_time": "16:43:10", "remaining_time": "2 days, 10:02:34"}
+{"current_steps": 493, "total_steps": 2200, "loss": 1.8051, "accuracy": 0.78125, "lr": 8.930159565257846e-06, "epoch": 0.8955495004541326, "percentage": 22.41, "elapsed_time": "16:45:13", "remaining_time": "2 days, 10:00:32"}
+{"current_steps": 494, "total_steps": 2200, "loss": 2.5853, "accuracy": 0.640625, "lr": 8.925996698330887e-06, "epoch": 0.8973660308810173, "percentage": 22.45, "elapsed_time": "16:47:25", "remaining_time": "2 days, 9:59:03"}
+{"current_steps": 495, "total_steps": 2200, "loss": 2.1072, "accuracy": 0.71875, "lr": 8.92182684520014e-06, "epoch": 0.8991825613079019, "percentage": 22.5, "elapsed_time": "16:49:26", "remaining_time": "2 days, 9:56:58"}
+{"current_steps": 496, "total_steps": 2200, "loss": 2.3116, "accuracy": 0.71875, "lr": 8.917650014368658e-06, "epoch": 0.9009990917347865, "percentage": 22.55, "elapsed_time": "16:51:13", "remaining_time": "2 days, 9:54:03"}
+{"current_steps": 497, "total_steps": 2200, "loss": 2.3541, "accuracy": 0.75, "lr": 8.913466214353728e-06, "epoch": 0.9028156221616712, "percentage": 22.59, "elapsed_time": "16:53:21", "remaining_time": "2 days, 9:52:20"}
+{"current_steps": 498, "total_steps": 2200, "loss": 2.0436, "accuracy": 0.78125, "lr": 8.909275453686845e-06, "epoch": 0.9046321525885559, "percentage": 22.64, "elapsed_time": "16:55:18", "remaining_time": "2 days, 9:49:57"}
+{"current_steps": 499, "total_steps": 2200, "loss": 2.2477, "accuracy": 0.765625, "lr": 8.905077740913701e-06, "epoch": 0.9064486830154405, "percentage": 22.68, "elapsed_time": "16:57:17", "remaining_time": "2 days, 9:47:44"}
+{"current_steps": 500, "total_steps": 2200, "loss": 2.6346, "accuracy": 0.65625, "lr": 8.900873084594164e-06, "epoch": 0.9082652134423251, "percentage": 22.73, "elapsed_time": "16:59:33", "remaining_time": "2 days, 9:46:30"}
+{"current_steps": 501, "total_steps": 2200, "loss": 2.2969, "accuracy": 0.765625, "lr": 8.896661493302258e-06, "epoch": 0.9100817438692098, "percentage": 22.77, "elapsed_time": "17:01:41", "remaining_time": "2 days, 9:44:47"}
+{"current_steps": 502, "total_steps": 2200, "loss": 2.5968, "accuracy": 0.703125, "lr": 8.892442975626152e-06, "epoch": 0.9118982742960945, "percentage": 22.82, "elapsed_time": "17:03:33", "remaining_time": "2 days, 9:42:08"}
+{"current_steps": 503, "total_steps": 2200, "loss": 2.169, "accuracy": 0.71875, "lr": 8.888217540168139e-06, "epoch": 0.9137148047229791, "percentage": 22.86, "elapsed_time": "17:05:23", "remaining_time": "2 days, 9:39:25"}
+{"current_steps": 504, "total_steps": 2200, "loss": 2.5361, "accuracy": 0.625, "lr": 8.883985195544617e-06, "epoch": 0.9155313351498637, "percentage": 22.91, "elapsed_time": "17:07:18", "remaining_time": "2 days, 9:36:58"}
+{"current_steps": 505, "total_steps": 2200, "loss": 2.5452, "accuracy": 0.65625, "lr": 8.879745950386075e-06, "epoch": 0.9173478655767484, "percentage": 22.95, "elapsed_time": "17:09:08", "remaining_time": "2 days, 9:34:16"}
+{"current_steps": 506, "total_steps": 2200, "loss": 2.74, "accuracy": 0.59375, "lr": 8.87549981333707e-06, "epoch": 0.9191643960036331, "percentage": 23.0, "elapsed_time": "17:11:00", "remaining_time": "2 days, 9:31:37"}
+{"current_steps": 507, "total_steps": 2200, "loss": 2.7052, "accuracy": 0.609375, "lr": 8.871246793056215e-06, "epoch": 0.9209809264305178, "percentage": 23.05, "elapsed_time": "17:12:39", "remaining_time": "2 days, 9:28:18"}
+{"current_steps": 508, "total_steps": 2200, "loss": 2.2445, "accuracy": 0.703125, "lr": 8.866986898216157e-06, "epoch": 0.9227974568574023, "percentage": 23.09, "elapsed_time": "17:14:21", "remaining_time": "2 days, 9:25:08"}
+{"current_steps": 509, "total_steps": 2200, "loss": 2.1459, "accuracy": 0.71875, "lr": 8.862720137503568e-06, "epoch": 0.924613987284287, "percentage": 23.14, "elapsed_time": "17:16:30", "remaining_time": "2 days, 9:23:30"}
+{"current_steps": 510, "total_steps": 2200, "loss": 2.1311, "accuracy": 0.734375, "lr": 8.858446519619113e-06, "epoch": 0.9264305177111717, "percentage": 23.18, "elapsed_time": "17:18:42", "remaining_time": "2 days, 9:21:59"}
+{"current_steps": 511, "total_steps": 2200, "loss": 2.1092, "accuracy": 0.78125, "lr": 8.854166053277443e-06, "epoch": 0.9282470481380564, "percentage": 23.23, "elapsed_time": "17:21:05", "remaining_time": "2 days, 9:21:04"}
+{"current_steps": 512, "total_steps": 2200, "loss": 2.6399, "accuracy": 0.625, "lr": 8.849878747207175e-06, "epoch": 0.9300635785649409, "percentage": 23.27, "elapsed_time": "17:22:53", "remaining_time": "2 days, 9:18:16"}
+{"current_steps": 513, "total_steps": 2200, "loss": 1.9078, "accuracy": 0.796875, "lr": 8.845584610150871e-06, "epoch": 0.9318801089918256, "percentage": 23.32, "elapsed_time": "17:25:11", "remaining_time": "2 days, 9:17:06"}
+{"current_steps": 514, "total_steps": 2200, "loss": 2.1243, "accuracy": 0.75, "lr": 8.841283650865027e-06, "epoch": 0.9336966394187103, "percentage": 23.36, "elapsed_time": "17:27:32", "remaining_time": "2 days, 9:16:06"}
+{"current_steps": 515, "total_steps": 2200, "loss": 2.4256, "accuracy": 0.65625, "lr": 8.836975878120046e-06, "epoch": 0.935513169845595, "percentage": 23.41, "elapsed_time": "17:29:54", "remaining_time": "2 days, 9:15:09"}
+{"current_steps": 516, "total_steps": 2200, "loss": 2.3568, "accuracy": 0.671875, "lr": 8.832661300700228e-06, "epoch": 0.9373297002724795, "percentage": 23.45, "elapsed_time": "17:31:50", "remaining_time": "2 days, 9:12:46"}
+{"current_steps": 517, "total_steps": 2200, "loss": 2.3041, "accuracy": 0.65625, "lr": 8.828339927403745e-06, "epoch": 0.9391462306993642, "percentage": 23.5, "elapsed_time": "17:33:56", "remaining_time": "2 days, 9:10:55"}
+{"current_steps": 518, "total_steps": 2200, "loss": 2.3468, "accuracy": 0.703125, "lr": 8.824011767042631e-06, "epoch": 0.9409627611262489, "percentage": 23.55, "elapsed_time": "17:35:49", "remaining_time": "2 days, 9:08:24"}
+{"current_steps": 519, "total_steps": 2200, "loss": 2.8427, "accuracy": 0.609375, "lr": 8.819676828442758e-06, "epoch": 0.9427792915531336, "percentage": 23.59, "elapsed_time": "17:38:08", "remaining_time": "2 days, 9:07:13"}
+{"current_steps": 520, "total_steps": 2200, "loss": 2.6559, "accuracy": 0.59375, "lr": 8.815335120443822e-06, "epoch": 0.9445958219800181, "percentage": 23.64, "elapsed_time": "17:40:15", "remaining_time": "2 days, 9:05:25"}
+{"current_steps": 521, "total_steps": 2200, "loss": 2.1866, "accuracy": 0.828125, "lr": 8.810986651899322e-06, "epoch": 0.9464123524069028, "percentage": 23.68, "elapsed_time": "17:41:46", "remaining_time": "2 days, 9:01:44"}
+{"current_steps": 522, "total_steps": 2200, "loss": 2.1217, "accuracy": 0.75, "lr": 8.80663143167654e-06, "epoch": 0.9482288828337875, "percentage": 23.73, "elapsed_time": "17:44:03", "remaining_time": "2 days, 9:00:28"}
+{"current_steps": 523, "total_steps": 2200, "loss": 2.5629, "accuracy": 0.65625, "lr": 8.80226946865653e-06, "epoch": 0.9500454132606722, "percentage": 23.77, "elapsed_time": "17:46:02", "remaining_time": "2 days, 8:58:17"}
+{"current_steps": 524, "total_steps": 2200, "loss": 2.5327, "accuracy": 0.59375, "lr": 8.797900771734094e-06, "epoch": 0.9518619436875567, "percentage": 23.82, "elapsed_time": "17:47:58", "remaining_time": "2 days, 8:55:52"}
+{"current_steps": 525, "total_steps": 2200, "loss": 2.5188, "accuracy": 0.6875, "lr": 8.793525349817765e-06, "epoch": 0.9536784741144414, "percentage": 23.86, "elapsed_time": "17:50:01", "remaining_time": "2 days, 8:53:52"}
+{"current_steps": 526, "total_steps": 2200, "loss": 2.6082, "accuracy": 0.625, "lr": 8.78914321182979e-06, "epoch": 0.9554950045413261, "percentage": 23.91, "elapsed_time": "17:52:12", "remaining_time": "2 days, 8:52:19"}
+{"current_steps": 527, "total_steps": 2200, "loss": 2.2244, "accuracy": 0.703125, "lr": 8.784754366706115e-06, "epoch": 0.9573115349682108, "percentage": 23.95, "elapsed_time": "17:54:24", "remaining_time": "2 days, 8:50:45"}
+{"current_steps": 528, "total_steps": 2200, "loss": 2.0645, "accuracy": 0.71875, "lr": 8.780358823396352e-06, "epoch": 0.9591280653950953, "percentage": 24.0, "elapsed_time": "17:56:46", "remaining_time": "2 days, 8:49:46"}
+{"current_steps": 529, "total_steps": 2200, "loss": 1.8904, "accuracy": 0.859375, "lr": 8.775956590863785e-06, "epoch": 0.96094459582198, "percentage": 24.05, "elapsed_time": "17:58:22", "remaining_time": "2 days, 8:46:22"}
+{"current_steps": 530, "total_steps": 2200, "loss": 2.0136, "accuracy": 0.703125, "lr": 8.771547678085332e-06, "epoch": 0.9627611262488647, "percentage": 24.09, "elapsed_time": "18:00:05", "remaining_time": "2 days, 8:43:18"}
+{"current_steps": 531, "total_steps": 2200, "loss": 2.2551, "accuracy": 0.78125, "lr": 8.767132094051534e-06, "epoch": 0.9645776566757494, "percentage": 24.14, "elapsed_time": "18:01:56", "remaining_time": "2 days, 8:40:40"}
+{"current_steps": 532, "total_steps": 2200, "loss": 1.8607, "accuracy": 0.78125, "lr": 8.762709847766532e-06, "epoch": 0.9663941871026339, "percentage": 24.18, "elapsed_time": "18:03:49", "remaining_time": "2 days, 8:38:09"}
+{"current_steps": 533, "total_steps": 2200, "loss": 2.6128, "accuracy": 0.65625, "lr": 8.758280948248059e-06, "epoch": 0.9682107175295186, "percentage": 24.23, "elapsed_time": "18:05:48", "remaining_time": "2 days, 8:35:57"}
+{"current_steps": 534, "total_steps": 2200, "loss": 2.0608, "accuracy": 0.765625, "lr": 8.753845404527413e-06, "epoch": 0.9700272479564033, "percentage": 24.27, "elapsed_time": "18:07:39", "remaining_time": "2 days, 8:33:20"}
+{"current_steps": 535, "total_steps": 2200, "loss": 2.4587, "accuracy": 0.671875, "lr": 8.74940322564944e-06, "epoch": 0.971843778383288, "percentage": 24.32, "elapsed_time": "18:09:29", "remaining_time": "2 days, 8:30:39"}
+{"current_steps": 536, "total_steps": 2200, "loss": 2.3589, "accuracy": 0.734375, "lr": 8.744954420672514e-06, "epoch": 0.9736603088101726, "percentage": 24.36, "elapsed_time": "18:11:51", "remaining_time": "2 days, 8:29:39"}
+{"current_steps": 537, "total_steps": 2200, "loss": 2.3469, "accuracy": 0.75, "lr": 8.740498998668523e-06, "epoch": 0.9754768392370572, "percentage": 24.41, "elapsed_time": "18:13:40", "remaining_time": "2 days, 8:26:55"}
+{"current_steps": 538, "total_steps": 2200, "loss": 2.1889, "accuracy": 0.703125, "lr": 8.736036968722851e-06, "epoch": 0.9772933696639419, "percentage": 24.45, "elapsed_time": "18:15:56", "remaining_time": "2 days, 8:25:35"}
+{"current_steps": 539, "total_steps": 2200, "loss": 2.048, "accuracy": 0.8125, "lr": 8.73156833993435e-06, "epoch": 0.9791099000908265, "percentage": 24.5, "elapsed_time": "18:18:10", "remaining_time": "2 days, 8:24:09"}
+{"current_steps": 540, "total_steps": 2200, "loss": 2.2089, "accuracy": 0.703125, "lr": 8.727093121415338e-06, "epoch": 0.9809264305177112, "percentage": 24.55, "elapsed_time": "18:20:11", "remaining_time": "2 days, 8:22:03"}
+{"current_steps": 541, "total_steps": 2200, "loss": 2.3303, "accuracy": 0.59375, "lr": 8.722611322291558e-06, "epoch": 0.9827429609445958, "percentage": 24.59, "elapsed_time": "18:22:18", "remaining_time": "2 days, 8:20:16"}
+{"current_steps": 542, "total_steps": 2200, "loss": 2.1393, "accuracy": 0.734375, "lr": 8.718122951702183e-06, "epoch": 0.9845594913714805, "percentage": 24.64, "elapsed_time": "18:23:56", "remaining_time": "2 days, 8:17:00"}
+{"current_steps": 543, "total_steps": 2200, "loss": 2.4923, "accuracy": 0.734375, "lr": 8.713628018799782e-06, "epoch": 0.9863760217983651, "percentage": 24.68, "elapsed_time": "18:26:00", "remaining_time": "2 days, 8:15:03"}
+{"current_steps": 544, "total_steps": 2200, "loss": 2.0945, "accuracy": 0.734375, "lr": 8.709126532750304e-06, "epoch": 0.9881925522252498, "percentage": 24.73, "elapsed_time": "18:28:04", "remaining_time": "2 days, 8:13:07"}
+{"current_steps": 545, "total_steps": 2200, "loss": 2.1642, "accuracy": 0.75, "lr": 8.704618502733064e-06, "epoch": 0.9900090826521344, "percentage": 24.77, "elapsed_time": "18:30:05", "remaining_time": "2 days, 8:11:01"}
+{"current_steps": 546, "total_steps": 2200, "loss": 2.0951, "accuracy": 0.6875, "lr": 8.70010393794072e-06, "epoch": 0.9918256130790191, "percentage": 24.82, "elapsed_time": "18:32:05", "remaining_time": "2 days, 8:08:51"}
+{"current_steps": 547, "total_steps": 2200, "loss": 2.1101, "accuracy": 0.734375, "lr": 8.695582847579254e-06, "epoch": 0.9936421435059037, "percentage": 24.86, "elapsed_time": "18:34:09", "remaining_time": "2 days, 8:06:56"}
+{"current_steps": 548, "total_steps": 2200, "loss": 2.2745, "accuracy": 0.703125, "lr": 8.69105524086796e-06, "epoch": 0.9954586739327884, "percentage": 24.91, "elapsed_time": "18:36:05", "remaining_time": "2 days, 8:04:33"}
+{"current_steps": 549, "total_steps": 2200, "loss": 2.5954, "accuracy": 0.65625, "lr": 8.686521127039414e-06, "epoch": 0.997275204359673, "percentage": 24.95, "elapsed_time": "18:39:12", "remaining_time": "2 days, 8:05:46"}
+{"current_steps": 550, "total_steps": 2200, "loss": 2.7544, "accuracy": 0.546875, "lr": 8.681980515339464e-06, "epoch": 0.9990917347865577, "percentage": 25.0, "elapsed_time": "18:41:07", "remaining_time": "2 days, 8:03:23"}
+{"current_steps": 551, "total_steps": 2200, "loss": 1.1517, "accuracy": 0.75, "lr": 8.677433415027209e-06, "epoch": 1.0, "percentage": 25.05, "elapsed_time": "18:42:18", "remaining_time": "2 days, 7:58:45"}
+{"current_steps": 552, "total_steps": 2200, "loss": 2.1275, "accuracy": 0.765625, "lr": 8.672879835374976e-06, "epoch": 1.0018165304268847, "percentage": 25.09, "elapsed_time": "18:44:19", "remaining_time": "2 days, 7:56:42"}
+{"current_steps": 553, "total_steps": 2200, "loss": 2.3209, "accuracy": 0.75, "lr": 8.668319785668307e-06, "epoch": 1.0036330608537694, "percentage": 25.14, "elapsed_time": "18:46:23", "remaining_time": "2 days, 7:54:43"}
+{"current_steps": 554, "total_steps": 2200, "loss": 1.7106, "accuracy": 0.828125, "lr": 8.66375327520594e-06, "epoch": 1.005449591280654, "percentage": 25.18, "elapsed_time": "18:48:42", "remaining_time": "2 days, 7:53:31"}
+{"current_steps": 555, "total_steps": 2200, "loss": 1.8965, "accuracy": 0.84375, "lr": 8.659180313299783e-06, "epoch": 1.0072661217075387, "percentage": 25.23, "elapsed_time": "18:50:59", "remaining_time": "2 days, 7:52:13"}
+{"current_steps": 556, "total_steps": 2200, "loss": 2.027, "accuracy": 0.78125, "lr": 8.6546009092749e-06, "epoch": 1.0090826521344232, "percentage": 25.27, "elapsed_time": "18:53:19", "remaining_time": "2 days, 7:51:03"}
+{"current_steps": 557, "total_steps": 2200, "loss": 1.966, "accuracy": 0.78125, "lr": 8.650015072469496e-06, "epoch": 1.0108991825613078, "percentage": 25.32, "elapsed_time": "18:55:40", "remaining_time": "2 days, 7:49:57"}
+{"current_steps": 558, "total_steps": 2200, "loss": 2.1115, "accuracy": 0.703125, "lr": 8.645422812234888e-06, "epoch": 1.0127157129881925, "percentage": 25.36, "elapsed_time": "18:57:06", "remaining_time": "2 days, 7:46:07"}
+{"current_steps": 559, "total_steps": 2200, "loss": 2.2691, "accuracy": 0.78125, "lr": 8.640824137935498e-06, "epoch": 1.0145322434150772, "percentage": 25.41, "elapsed_time": "18:59:28", "remaining_time": "2 days, 7:45:03"}
+{"current_steps": 560, "total_steps": 2200, "loss": 1.974, "accuracy": 0.8125, "lr": 8.636219058948823e-06, "epoch": 1.0163487738419619, "percentage": 25.45, "elapsed_time": "19:01:18", "remaining_time": "2 days, 7:42:23"}
+{"current_steps": 561, "total_steps": 2200, "loss": 1.9963, "accuracy": 0.8125, "lr": 8.631607584665415e-06, "epoch": 1.0181653042688465, "percentage": 25.5, "elapsed_time": "19:03:26", "remaining_time": "2 days, 7:40:39"}
+{"current_steps": 562, "total_steps": 2200, "loss": 2.0399, "accuracy": 0.765625, "lr": 8.62698972448888e-06, "epoch": 1.0199818346957312, "percentage": 25.55, "elapsed_time": "19:05:35", "remaining_time": "2 days, 7:38:55"}
+{"current_steps": 563, "total_steps": 2200, "loss": 2.3184, "accuracy": 0.65625, "lr": 8.622365487835836e-06, "epoch": 1.021798365122616, "percentage": 25.59, "elapsed_time": "19:07:41", "remaining_time": "2 days, 7:37:03"}
+{"current_steps": 564, "total_steps": 2200, "loss": 2.2, "accuracy": 0.734375, "lr": 8.617734884135906e-06, "epoch": 1.0236148955495004, "percentage": 25.64, "elapsed_time": "19:09:31", "remaining_time": "2 days, 7:34:26"}
+{"current_steps": 565, "total_steps": 2200, "loss": 1.8705, "accuracy": 0.796875, "lr": 8.613097922831699e-06, "epoch": 1.025431425976385, "percentage": 25.68, "elapsed_time": "19:11:44", "remaining_time": "2 days, 7:32:55"}
+{"current_steps": 566, "total_steps": 2200, "loss": 1.8687, "accuracy": 0.828125, "lr": 8.608454613378783e-06, "epoch": 1.0272479564032697, "percentage": 25.73, "elapsed_time": "19:13:34", "remaining_time": "2 days, 7:30:15"}
+{"current_steps": 567, "total_steps": 2200, "loss": 2.0884, "accuracy": 0.78125, "lr": 8.603804965245678e-06, "epoch": 1.0290644868301544, "percentage": 25.77, "elapsed_time": "19:15:26", "remaining_time": "2 days, 7:27:46"}
+{"current_steps": 568, "total_steps": 2200, "loss": 2.3203, "accuracy": 0.625, "lr": 8.599148987913829e-06, "epoch": 1.030881017257039, "percentage": 25.82, "elapsed_time": "19:17:38", "remaining_time": "2 days, 7:26:10"}
+{"current_steps": 569, "total_steps": 2200, "loss": 2.1789, "accuracy": 0.734375, "lr": 8.594486690877577e-06, "epoch": 1.0326975476839237, "percentage": 25.86, "elapsed_time": "19:19:51", "remaining_time": "2 days, 7:24:39"}
+{"current_steps": 570, "total_steps": 2200, "loss": 2.2068, "accuracy": 0.703125, "lr": 8.589818083644161e-06, "epoch": 1.0345140781108084, "percentage": 25.91, "elapsed_time": "19:21:59", "remaining_time": "2 days, 7:22:54"}
+{"current_steps": 571, "total_steps": 2200, "loss": 1.9628, "accuracy": 0.796875, "lr": 8.585143175733686e-06, "epoch": 1.036330608537693, "percentage": 25.95, "elapsed_time": "19:23:44", "remaining_time": "2 days, 7:20:00"}
+{"current_steps": 572, "total_steps": 2200, "loss": 2.2717, "accuracy": 0.734375, "lr": 8.5804619766791e-06, "epoch": 1.0381471389645776, "percentage": 26.0, "elapsed_time": "19:26:03", "remaining_time": "2 days, 7:18:46"}
+{"current_steps": 573, "total_steps": 2200, "loss": 2.1781, "accuracy": 0.765625, "lr": 8.575774496026184e-06, "epoch": 1.0399636693914622, "percentage": 26.05, "elapsed_time": "19:28:06", "remaining_time": "2 days, 7:16:46"}
+{"current_steps": 574, "total_steps": 2200, "loss": 2.1358, "accuracy": 0.8125, "lr": 8.571080743333528e-06, "epoch": 1.041780199818347, "percentage": 26.09, "elapsed_time": "19:30:21", "remaining_time": "2 days, 7:15:18"}
+{"current_steps": 575, "total_steps": 2200, "loss": 1.9609, "accuracy": 0.8125, "lr": 8.566380728172512e-06, "epoch": 1.0435967302452316, "percentage": 26.14, "elapsed_time": "19:32:36", "remaining_time": "2 days, 7:13:53"}
+{"current_steps": 576, "total_steps": 2200, "loss": 2.311, "accuracy": 0.765625, "lr": 8.561674460127281e-06, "epoch": 1.0454132606721163, "percentage": 26.18, "elapsed_time": "19:34:43", "remaining_time": "2 days, 7:12:05"}
+{"current_steps": 577, "total_steps": 2200, "loss": 1.9574, "accuracy": 0.71875, "lr": 8.556961948794738e-06, "epoch": 1.047229791099001, "percentage": 26.23, "elapsed_time": "19:36:33", "remaining_time": "2 days, 7:09:28"}
+{"current_steps": 578, "total_steps": 2200, "loss": 2.0244, "accuracy": 0.78125, "lr": 8.552243203784514e-06, "epoch": 1.0490463215258856, "percentage": 26.27, "elapsed_time": "19:38:43", "remaining_time": "2 days, 7:07:45"}
+{"current_steps": 579, "total_steps": 2200, "loss": 1.7738, "accuracy": 0.8125, "lr": 8.547518234718954e-06, "epoch": 1.0508628519527703, "percentage": 26.32, "elapsed_time": "19:40:49", "remaining_time": "2 days, 7:05:54"}
+{"current_steps": 580, "total_steps": 2200, "loss": 1.8789, "accuracy": 0.8125, "lr": 8.542787051233088e-06, "epoch": 1.052679382379655, "percentage": 26.36, "elapsed_time": "19:43:17", "remaining_time": "2 days, 7:05:03"}
+{"current_steps": 581, "total_steps": 2200, "loss": 2.2372, "accuracy": 0.6875, "lr": 8.538049662974625e-06, "epoch": 1.0544959128065394, "percentage": 26.41, "elapsed_time": "19:45:14", "remaining_time": "2 days, 7:02:46"}
+{"current_steps": 582, "total_steps": 2200, "loss": 1.831, "accuracy": 0.84375, "lr": 8.533306079603928e-06, "epoch": 1.056312443233424, "percentage": 26.45, "elapsed_time": "19:47:37", "remaining_time": "2 days, 7:01:41"}
+{"current_steps": 583, "total_steps": 2200, "loss": 2.1295, "accuracy": 0.765625, "lr": 8.52855631079398e-06, "epoch": 1.0581289736603088, "percentage": 26.5, "elapsed_time": "19:49:28", "remaining_time": "2 days, 6:59:05"}
+{"current_steps": 584, "total_steps": 2200, "loss": 2.783, "accuracy": 0.640625, "lr": 8.523800366230397e-06, "epoch": 1.0599455040871935, "percentage": 26.55, "elapsed_time": "19:51:42", "remaining_time": "2 days, 6:57:36"}
+{"current_steps": 585, "total_steps": 2200, "loss": 2.1046, "accuracy": 0.75, "lr": 8.519038255611372e-06, "epoch": 1.0617620345140781, "percentage": 26.59, "elapsed_time": "19:53:55", "remaining_time": "2 days, 6:56:02"}
+{"current_steps": 586, "total_steps": 2200, "loss": 2.1365, "accuracy": 0.75, "lr": 8.51426998864768e-06, "epoch": 1.0635785649409628, "percentage": 26.64, "elapsed_time": "19:55:55", "remaining_time": "2 days, 6:53:52"}
+{"current_steps": 587, "total_steps": 2200, "loss": 2.1912, "accuracy": 0.734375, "lr": 8.509495575062647e-06, "epoch": 1.0653950953678475, "percentage": 26.68, "elapsed_time": "19:57:42", "remaining_time": "2 days, 6:51:07"}
+{"current_steps": 588, "total_steps": 2200, "loss": 2.1753, "accuracy": 0.75, "lr": 8.504715024592132e-06, "epoch": 1.067211625794732, "percentage": 26.73, "elapsed_time": "20:00:00", "remaining_time": "2 days, 6:49:49"}
+{"current_steps": 589, "total_steps": 2200, "loss": 2.3005, "accuracy": 0.671875, "lr": 8.499928346984512e-06, "epoch": 1.0690281562216166, "percentage": 26.77, "elapsed_time": "20:01:52", "remaining_time": "2 days, 6:47:17"}
+{"current_steps": 590, "total_steps": 2200, "loss": 2.1574, "accuracy": 0.734375, "lr": 8.495135552000657e-06, "epoch": 1.0708446866485013, "percentage": 26.82, "elapsed_time": "20:04:22", "remaining_time": "2 days, 6:46:30"}
+{"current_steps": 591, "total_steps": 2200, "loss": 2.0387, "accuracy": 0.78125, "lr": 8.490336649413909e-06, "epoch": 1.072661217075386, "percentage": 26.86, "elapsed_time": "20:06:17", "remaining_time": "2 days, 6:44:08"}
+{"current_steps": 592, "total_steps": 2200, "loss": 2.3886, "accuracy": 0.640625, "lr": 8.485531649010063e-06, "epoch": 1.0744777475022707, "percentage": 26.91, "elapsed_time": "20:08:16", "remaining_time": "2 days, 6:41:55"}
+{"current_steps": 593, "total_steps": 2200, "loss": 2.2219, "accuracy": 0.6875, "lr": 8.480720560587356e-06, "epoch": 1.0762942779291553, "percentage": 26.95, "elapsed_time": "20:09:52", "remaining_time": "2 days, 6:38:41"}
+{"current_steps": 594, "total_steps": 2200, "loss": 2.1268, "accuracy": 0.796875, "lr": 8.475903393956434e-06, "epoch": 1.07811080835604, "percentage": 27.0, "elapsed_time": "20:12:08", "remaining_time": "2 days, 6:37:15"}
+{"current_steps": 595, "total_steps": 2200, "loss": 1.9352, "accuracy": 0.78125, "lr": 8.471080158940336e-06, "epoch": 1.0799273387829247, "percentage": 27.05, "elapsed_time": "20:14:27", "remaining_time": "2 days, 6:35:58"}
+{"current_steps": 596, "total_steps": 2200, "loss": 2.1384, "accuracy": 0.6875, "lr": 8.46625086537448e-06, "epoch": 1.0817438692098094, "percentage": 27.09, "elapsed_time": "20:16:10", "remaining_time": "2 days, 6:33:04"}
+{"current_steps": 597, "total_steps": 2200, "loss": 1.9695, "accuracy": 0.828125, "lr": 8.46141552310664e-06, "epoch": 1.0835603996366938, "percentage": 27.14, "elapsed_time": "20:17:51", "remaining_time": "2 days, 6:30:03"}
+{"current_steps": 598, "total_steps": 2200, "loss": 1.9499, "accuracy": 0.765625, "lr": 8.456574141996913e-06, "epoch": 1.0853769300635785, "percentage": 27.18, "elapsed_time": "20:20:07", "remaining_time": "2 days, 6:28:37"}
+{"current_steps": 599, "total_steps": 2200, "loss": 1.993, "accuracy": 0.734375, "lr": 8.451726731917722e-06, "epoch": 1.0871934604904632, "percentage": 27.23, "elapsed_time": "20:22:17", "remaining_time": "2 days, 6:26:54"}
+{"current_steps": 600, "total_steps": 2200, "loss": 2.5042, "accuracy": 0.703125, "lr": 8.446873302753783e-06, "epoch": 1.0890099909173478, "percentage": 27.27, "elapsed_time": "20:24:32", "remaining_time": "2 days, 6:25:25"}
+{"current_steps": 601, "total_steps": 2200, "loss": 2.0308, "accuracy": 0.78125, "lr": 8.44201386440208e-06, "epoch": 1.0908265213442325, "percentage": 27.32, "elapsed_time": "20:26:30", "remaining_time": "2 days, 6:23:11"}
+{"current_steps": 602, "total_steps": 2200, "loss": 1.8605, "accuracy": 0.78125, "lr": 8.437148426771852e-06, "epoch": 1.0926430517711172, "percentage": 27.36, "elapsed_time": "20:28:11", "remaining_time": "2 days, 6:20:11"}
+{"current_steps": 603, "total_steps": 2200, "loss": 1.9416, "accuracy": 0.828125, "lr": 8.432276999784578e-06, "epoch": 1.0944595821980019, "percentage": 27.41, "elapsed_time": "20:29:45", "remaining_time": "2 days, 6:16:54"}
+{"current_steps": 604, "total_steps": 2200, "loss": 2.0358, "accuracy": 0.8125, "lr": 8.427399593373941e-06, "epoch": 1.0962761126248866, "percentage": 27.45, "elapsed_time": "20:31:59", "remaining_time": "2 days, 6:15:23"}
+{"current_steps": 605, "total_steps": 2200, "loss": 2.6379, "accuracy": 0.734375, "lr": 8.422516217485826e-06, "epoch": 1.0980926430517712, "percentage": 27.5, "elapsed_time": "20:34:13", "remaining_time": "2 days, 6:13:52"}
+{"current_steps": 606, "total_steps": 2200, "loss": 1.8705, "accuracy": 0.765625, "lr": 8.417626882078287e-06, "epoch": 1.0999091734786557, "percentage": 27.55, "elapsed_time": "20:36:23", "remaining_time": "2 days, 6:12:10"}
+{"current_steps": 607, "total_steps": 2200, "loss": 2.3611, "accuracy": 0.71875, "lr": 8.412731597121527e-06, "epoch": 1.1017257039055404, "percentage": 27.59, "elapsed_time": "20:38:23", "remaining_time": "2 days, 6:10:01"}
+{"current_steps": 608, "total_steps": 2200, "loss": 2.0535, "accuracy": 0.734375, "lr": 8.407830372597884e-06, "epoch": 1.103542234332425, "percentage": 27.64, "elapsed_time": "20:40:17", "remaining_time": "2 days, 6:07:36"}
+{"current_steps": 609, "total_steps": 2200, "loss": 2.1751, "accuracy": 0.78125, "lr": 8.402923218501813e-06, "epoch": 1.1053587647593097, "percentage": 27.68, "elapsed_time": "20:42:22", "remaining_time": "2 days, 6:05:40"}
+{"current_steps": 610, "total_steps": 2200, "loss": 2.2522, "accuracy": 0.78125, "lr": 8.39801014483985e-06, "epoch": 1.1071752951861944, "percentage": 27.73, "elapsed_time": "20:44:38", "remaining_time": "2 days, 6:04:13"}
+{"current_steps": 611, "total_steps": 2200, "loss": 1.9467, "accuracy": 0.8125, "lr": 8.393091161630612e-06, "epoch": 1.108991825613079, "percentage": 27.77, "elapsed_time": "20:46:34", "remaining_time": "2 days, 6:01:53"}
+{"current_steps": 612, "total_steps": 2200, "loss": 2.041, "accuracy": 0.78125, "lr": 8.388166278904759e-06, "epoch": 1.1108083560399638, "percentage": 27.82, "elapsed_time": "20:48:03", "remaining_time": "2 days, 5:58:24"}
+{"current_steps": 613, "total_steps": 2200, "loss": 2.0593, "accuracy": 0.6875, "lr": 8.383235506704986e-06, "epoch": 1.1126248864668482, "percentage": 27.86, "elapsed_time": "20:49:53", "remaining_time": "2 days, 5:55:51"}
+{"current_steps": 614, "total_steps": 2200, "loss": 2.1543, "accuracy": 0.828125, "lr": 8.378298855085996e-06, "epoch": 1.1144414168937329, "percentage": 27.91, "elapsed_time": "20:51:56", "remaining_time": "2 days, 5:53:49"}
+{"current_steps": 615, "total_steps": 2200, "loss": 2.2401, "accuracy": 0.796875, "lr": 8.373356334114484e-06, "epoch": 1.1162579473206176, "percentage": 27.95, "elapsed_time": "20:54:01", "remaining_time": "2 days, 5:51:54"}
+{"current_steps": 616, "total_steps": 2200, "loss": 1.8293, "accuracy": 0.828125, "lr": 8.368407953869105e-06, "epoch": 1.1180744777475022, "percentage": 28.0, "elapsed_time": "20:56:10", "remaining_time": "2 days, 5:50:11"}
+{"current_steps": 617, "total_steps": 2200, "loss": 1.9608, "accuracy": 0.796875, "lr": 8.363453724440471e-06, "epoch": 1.119891008174387, "percentage": 28.05, "elapsed_time": "20:58:08", "remaining_time": "2 days, 5:47:57"}
+{"current_steps": 618, "total_steps": 2200, "loss": 1.8448, "accuracy": 0.796875, "lr": 8.358493655931119e-06, "epoch": 1.1217075386012716, "percentage": 28.09, "elapsed_time": "21:00:18", "remaining_time": "2 days, 5:46:13"}
+{"current_steps": 619, "total_steps": 2200, "loss": 2.2057, "accuracy": 0.734375, "lr": 8.353527758455491e-06, "epoch": 1.1235240690281563, "percentage": 28.14, "elapsed_time": "21:02:33", "remaining_time": "2 days, 5:44:42"}
+{"current_steps": 620, "total_steps": 2200, "loss": 2.4366, "accuracy": 0.65625, "lr": 8.348556042139918e-06, "epoch": 1.125340599455041, "percentage": 28.18, "elapsed_time": "21:04:27", "remaining_time": "2 days, 5:42:20"}
+{"current_steps": 621, "total_steps": 2200, "loss": 2.126, "accuracy": 0.71875, "lr": 8.343578517122594e-06, "epoch": 1.1271571298819256, "percentage": 28.23, "elapsed_time": "21:06:27", "remaining_time": "2 days, 5:40:11"}
+{"current_steps": 622, "total_steps": 2200, "loss": 2.084, "accuracy": 0.703125, "lr": 8.338595193553559e-06, "epoch": 1.12897366030881, "percentage": 28.27, "elapsed_time": "21:08:25", "remaining_time": "2 days, 5:37:59"}
+{"current_steps": 623, "total_steps": 2200, "loss": 1.7885, "accuracy": 0.828125, "lr": 8.333606081594678e-06, "epoch": 1.1307901907356948, "percentage": 28.32, "elapsed_time": "21:10:13", "remaining_time": "2 days, 5:35:20"}
+{"current_steps": 624, "total_steps": 2200, "loss": 1.7624, "accuracy": 0.75, "lr": 8.32861119141962e-06, "epoch": 1.1326067211625794, "percentage": 28.36, "elapsed_time": "21:12:10", "remaining_time": "2 days, 5:33:02"}
+{"current_steps": 625, "total_steps": 2200, "loss": 2.1956, "accuracy": 0.65625, "lr": 8.323610533213835e-06, "epoch": 1.134423251589464, "percentage": 28.41, "elapsed_time": "21:13:57", "remaining_time": "2 days, 5:30:22"}
+{"current_steps": 626, "total_steps": 2200, "loss": 1.675, "accuracy": 0.828125, "lr": 8.318604117174533e-06, "epoch": 1.1362397820163488, "percentage": 28.45, "elapsed_time": "21:15:59", "remaining_time": "2 days, 5:28:19"}
+{"current_steps": 627, "total_steps": 2200, "loss": 2.3784, "accuracy": 0.671875, "lr": 8.313591953510674e-06, "epoch": 1.1380563124432335, "percentage": 28.5, "elapsed_time": "21:17:53", "remaining_time": "2 days, 5:25:56"}
+{"current_steps": 628, "total_steps": 2200, "loss": 2.1623, "accuracy": 0.71875, "lr": 8.308574052442928e-06, "epoch": 1.1398728428701181, "percentage": 28.55, "elapsed_time": "21:19:58", "remaining_time": "2 days, 5:24:01"}
+{"current_steps": 629, "total_steps": 2200, "loss": 1.8051, "accuracy": 0.828125, "lr": 8.30355042420367e-06, "epoch": 1.1416893732970028, "percentage": 28.59, "elapsed_time": "21:21:57", "remaining_time": "2 days, 5:21:49"}
+{"current_steps": 630, "total_steps": 2200, "loss": 2.3401, "accuracy": 0.71875, "lr": 8.29852107903695e-06, "epoch": 1.1435059037238875, "percentage": 28.64, "elapsed_time": "21:23:43", "remaining_time": "2 days, 5:19:08"}
+{"current_steps": 631, "total_steps": 2200, "loss": 2.2151, "accuracy": 0.75, "lr": 8.293486027198483e-06, "epoch": 1.145322434150772, "percentage": 28.68, "elapsed_time": "21:25:45", "remaining_time": "2 days, 5:17:04"}
+{"current_steps": 632, "total_steps": 2200, "loss": 2.6777, "accuracy": 0.6875, "lr": 8.288445278955615e-06, "epoch": 1.1471389645776566, "percentage": 28.73, "elapsed_time": "21:27:53", "remaining_time": "2 days, 5:15:15"}
+{"current_steps": 633, "total_steps": 2200, "loss": 1.9289, "accuracy": 0.765625, "lr": 8.283398844587308e-06, "epoch": 1.1489554950045413, "percentage": 28.77, "elapsed_time": "21:30:13", "remaining_time": "2 days, 5:13:59"}
+{"current_steps": 634, "total_steps": 2200, "loss": 1.8554, "accuracy": 0.796875, "lr": 8.278346734384122e-06, "epoch": 1.150772025431426, "percentage": 28.82, "elapsed_time": "21:32:15", "remaining_time": "2 days, 5:11:56"}
+{"current_steps": 635, "total_steps": 2200, "loss": 2.3917, "accuracy": 0.6875, "lr": 8.27328895864819e-06, "epoch": 1.1525885558583107, "percentage": 28.86, "elapsed_time": "21:34:18", "remaining_time": "2 days, 5:09:53"}
+{"current_steps": 636, "total_steps": 2200, "loss": 1.8909, "accuracy": 0.875, "lr": 8.268225527693193e-06, "epoch": 1.1544050862851953, "percentage": 28.91, "elapsed_time": "21:36:20", "remaining_time": "2 days, 5:07:51"}
+{"current_steps": 637, "total_steps": 2200, "loss": 1.9508, "accuracy": 0.828125, "lr": 8.263156451844353e-06, "epoch": 1.15622161671208, "percentage": 28.95, "elapsed_time": "21:38:11", "remaining_time": "2 days, 5:05:20"}
+{"current_steps": 638, "total_steps": 2200, "loss": 2.1933, "accuracy": 0.734375, "lr": 8.258081741438396e-06, "epoch": 1.1580381471389645, "percentage": 29.0, "elapsed_time": "21:40:14", "remaining_time": "2 days, 5:03:21"}
+{"current_steps": 639, "total_steps": 2200, "loss": 2.0025, "accuracy": 0.8125, "lr": 8.25300140682354e-06, "epoch": 1.1598546775658491, "percentage": 29.05, "elapsed_time": "21:42:09", "remaining_time": "2 days, 5:01:01"}
+{"current_steps": 640, "total_steps": 2200, "loss": 2.3348, "accuracy": 0.671875, "lr": 8.247915458359473e-06, "epoch": 1.1616712079927338, "percentage": 29.09, "elapsed_time": "21:44:11", "remaining_time": "2 days, 4:58:56"}
+{"current_steps": 641, "total_steps": 2200, "loss": 2.2004, "accuracy": 0.75, "lr": 8.242823906417329e-06, "epoch": 1.1634877384196185, "percentage": 29.14, "elapsed_time": "21:46:19", "remaining_time": "2 days, 4:57:08"}
+{"current_steps": 642, "total_steps": 2200, "loss": 2.0052, "accuracy": 0.75, "lr": 8.237726761379668e-06, "epoch": 1.1653042688465032, "percentage": 29.18, "elapsed_time": "21:48:39", "remaining_time": "2 days, 4:55:50"}
+{"current_steps": 643, "total_steps": 2200, "loss": 2.5072, "accuracy": 0.6875, "lr": 8.232624033640458e-06, "epoch": 1.1671207992733879, "percentage": 29.23, "elapsed_time": "21:50:56", "remaining_time": "2 days, 4:54:24"}
+{"current_steps": 644, "total_steps": 2200, "loss": 2.0759, "accuracy": 0.75, "lr": 8.227515733605049e-06, "epoch": 1.1689373297002725, "percentage": 29.27, "elapsed_time": "21:52:55", "remaining_time": "2 days, 4:52:13"}
+{"current_steps": 645, "total_steps": 2200, "loss": 1.9992, "accuracy": 0.78125, "lr": 8.222401871690153e-06, "epoch": 1.1707538601271572, "percentage": 29.32, "elapsed_time": "21:54:37", "remaining_time": "2 days, 4:49:22"}
+{"current_steps": 646, "total_steps": 2200, "loss": 2.0353, "accuracy": 0.75, "lr": 8.217282458323825e-06, "epoch": 1.1725703905540419, "percentage": 29.36, "elapsed_time": "21:56:49", "remaining_time": "2 days, 4:47:44"}
+{"current_steps": 647, "total_steps": 2200, "loss": 2.0684, "accuracy": 0.734375, "lr": 8.21215750394544e-06, "epoch": 1.1743869209809263, "percentage": 29.41, "elapsed_time": "21:58:39", "remaining_time": "2 days, 4:45:10"}
+{"current_steps": 648, "total_steps": 2200, "loss": 2.1319, "accuracy": 0.765625, "lr": 8.207027019005675e-06, "epoch": 1.176203451407811, "percentage": 29.45, "elapsed_time": "22:00:54", "remaining_time": "2 days, 4:43:38"}
+{"current_steps": 649, "total_steps": 2200, "loss": 1.7919, "accuracy": 0.78125, "lr": 8.20189101396648e-06, "epoch": 1.1780199818346957, "percentage": 29.5, "elapsed_time": "22:02:49", "remaining_time": "2 days, 4:41:20"}
+{"current_steps": 650, "total_steps": 2200, "loss": 1.8932, "accuracy": 0.71875, "lr": 8.196749499301062e-06, "epoch": 1.1798365122615804, "percentage": 29.55, "elapsed_time": "22:04:48", "remaining_time": "2 days, 4:39:08"}
+{"current_steps": 651, "total_steps": 2200, "loss": 2.0615, "accuracy": 0.75, "lr": 8.191602485493868e-06, "epoch": 1.181653042688465, "percentage": 29.59, "elapsed_time": "22:07:08", "remaining_time": "2 days, 4:37:50"}
+{"current_steps": 652, "total_steps": 2200, "loss": 2.0064, "accuracy": 0.71875, "lr": 8.186449983040552e-06, "epoch": 1.1834695731153497, "percentage": 29.64, "elapsed_time": "22:09:08", "remaining_time": "2 days, 4:35:42"}
+{"current_steps": 653, "total_steps": 2200, "loss": 2.0473, "accuracy": 0.828125, "lr": 8.181292002447966e-06, "epoch": 1.1852861035422344, "percentage": 29.68, "elapsed_time": "22:11:10", "remaining_time": "2 days, 4:33:39"}
+{"current_steps": 654, "total_steps": 2200, "loss": 1.9884, "accuracy": 0.765625, "lr": 8.17612855423413e-06, "epoch": 1.1871026339691189, "percentage": 29.73, "elapsed_time": "22:13:24", "remaining_time": "2 days, 4:32:04"}
+{"current_steps": 655, "total_steps": 2200, "loss": 2.0203, "accuracy": 0.765625, "lr": 8.170959648928214e-06, "epoch": 1.1889191643960035, "percentage": 29.77, "elapsed_time": "22:15:04", "remaining_time": "2 days, 4:29:09"}
+{"current_steps": 656, "total_steps": 2200, "loss": 1.9023, "accuracy": 0.75, "lr": 8.165785297070516e-06, "epoch": 1.1907356948228882, "percentage": 29.82, "elapsed_time": "22:16:59", "remaining_time": "2 days, 4:26:50"}
+{"current_steps": 657, "total_steps": 2200, "loss": 2.0105, "accuracy": 0.78125, "lr": 8.16060550921244e-06, "epoch": 1.192552225249773, "percentage": 29.86, "elapsed_time": "22:19:13", "remaining_time": "2 days, 4:25:14"}
+{"current_steps": 658, "total_steps": 2200, "loss": 1.9895, "accuracy": 0.828125, "lr": 8.155420295916474e-06, "epoch": 1.1943687556766576, "percentage": 29.91, "elapsed_time": "22:21:23", "remaining_time": "2 days, 4:23:30"}
+{"current_steps": 659, "total_steps": 2200, "loss": 2.2687, "accuracy": 0.671875, "lr": 8.150229667756172e-06, "epoch": 1.1961852861035422, "percentage": 29.95, "elapsed_time": "22:23:05", "remaining_time": "2 days, 4:20:41"}
+{"current_steps": 660, "total_steps": 2200, "loss": 2.1371, "accuracy": 0.734375, "lr": 8.14503363531613e-06, "epoch": 1.198001816530427, "percentage": 30.0, "elapsed_time": "22:25:13", "remaining_time": "2 days, 4:18:50"}
+{"current_steps": 661, "total_steps": 2200, "loss": 2.4151, "accuracy": 0.671875, "lr": 8.139832209191961e-06, "epoch": 1.1998183469573116, "percentage": 30.05, "elapsed_time": "22:27:05", "remaining_time": "2 days, 4:16:25"}
+{"current_steps": 662, "total_steps": 2200, "loss": 2.1216, "accuracy": 0.75, "lr": 8.13462539999028e-06, "epoch": 1.2016348773841963, "percentage": 30.09, "elapsed_time": "22:28:45", "remaining_time": "2 days, 4:13:31"}
+{"current_steps": 663, "total_steps": 2200, "loss": 2.1232, "accuracy": 0.734375, "lr": 8.129413218328674e-06, "epoch": 1.2034514078110807, "percentage": 30.14, "elapsed_time": "22:30:34", "remaining_time": "2 days, 4:10:58"}
+{"current_steps": 664, "total_steps": 2200, "loss": 1.7544, "accuracy": 0.828125, "lr": 8.124195674835695e-06, "epoch": 1.2052679382379654, "percentage": 30.18, "elapsed_time": "22:32:35", "remaining_time": "2 days, 4:08:53"}
+{"current_steps": 665, "total_steps": 2200, "loss": 2.0524, "accuracy": 0.703125, "lr": 8.118972780150817e-06, "epoch": 1.20708446866485, "percentage": 30.23, "elapsed_time": "22:34:19", "remaining_time": "2 days, 4:06:09"}
+{"current_steps": 666, "total_steps": 2200, "loss": 1.8973, "accuracy": 0.78125, "lr": 8.113744544924434e-06, "epoch": 1.2089009990917348, "percentage": 30.27, "elapsed_time": "22:36:25", "remaining_time": "2 days, 4:04:14"}
+{"current_steps": 667, "total_steps": 2200, "loss": 2.0627, "accuracy": 0.78125, "lr": 8.108510979817828e-06, "epoch": 1.2107175295186194, "percentage": 30.32, "elapsed_time": "22:38:43", "remaining_time": "2 days, 4:02:50"}
+{"current_steps": 668, "total_steps": 2200, "loss": 1.9091, "accuracy": 0.75, "lr": 8.10327209550315e-06, "epoch": 1.2125340599455041, "percentage": 30.36, "elapsed_time": "22:41:04", "remaining_time": "2 days, 4:01:30"}
+{"current_steps": 669, "total_steps": 2200, "loss": 1.8977, "accuracy": 0.8125, "lr": 8.098027902663396e-06, "epoch": 1.2143505903723888, "percentage": 30.41, "elapsed_time": "22:43:08", "remaining_time": "2 days, 3:59:31"}
+{"current_steps": 670, "total_steps": 2200, "loss": 1.9182, "accuracy": 0.765625, "lr": 8.092778411992388e-06, "epoch": 1.2161671207992735, "percentage": 30.45, "elapsed_time": "22:45:03", "remaining_time": "2 days, 3:57:13"}
+{"current_steps": 671, "total_steps": 2200, "loss": 2.032, "accuracy": 0.703125, "lr": 8.087523634194755e-06, "epoch": 1.2179836512261581, "percentage": 30.5, "elapsed_time": "22:47:17", "remaining_time": "2 days, 3:55:37"}
+{"current_steps": 672, "total_steps": 2200, "loss": 1.7671, "accuracy": 0.828125, "lr": 8.082263579985898e-06, "epoch": 1.2198001816530426, "percentage": 30.55, "elapsed_time": "22:49:29", "remaining_time": "2 days, 3:53:57"}
+{"current_steps": 673, "total_steps": 2200, "loss": 2.2589, "accuracy": 0.703125, "lr": 8.076998260091989e-06, "epoch": 1.2216167120799273, "percentage": 30.59, "elapsed_time": "22:51:25", "remaining_time": "2 days, 3:51:42"}
+{"current_steps": 674, "total_steps": 2200, "loss": 2.2534, "accuracy": 0.71875, "lr": 8.071727685249929e-06, "epoch": 1.223433242506812, "percentage": 30.64, "elapsed_time": "22:53:29", "remaining_time": "2 days, 3:49:43"}
+{"current_steps": 675, "total_steps": 2200, "loss": 1.8979, "accuracy": 0.828125, "lr": 8.06645186620734e-06, "epoch": 1.2252497729336966, "percentage": 30.68, "elapsed_time": "22:55:51", "remaining_time": "2 days, 3:48:26"}
+{"current_steps": 676, "total_steps": 2200, "loss": 2.4093, "accuracy": 0.625, "lr": 8.061170813722533e-06, "epoch": 1.2270663033605813, "percentage": 30.73, "elapsed_time": "22:58:04", "remaining_time": "2 days, 3:46:46"}
+{"current_steps": 677, "total_steps": 2200, "loss": 1.7849, "accuracy": 0.796875, "lr": 8.055884538564493e-06, "epoch": 1.228882833787466, "percentage": 30.77, "elapsed_time": "23:00:08", "remaining_time": "2 days, 3:44:48"}
+{"current_steps": 678, "total_steps": 2200, "loss": 2.162, "accuracy": 0.8125, "lr": 8.050593051512859e-06, "epoch": 1.2306993642143507, "percentage": 30.82, "elapsed_time": "23:02:23", "remaining_time": "2 days, 3:43:13"}
+{"current_steps": 679, "total_steps": 2200, "loss": 2.0311, "accuracy": 0.8125, "lr": 8.045296363357891e-06, "epoch": 1.2325158946412351, "percentage": 30.86, "elapsed_time": "23:04:39", "remaining_time": "2 days, 3:41:42"}
+{"current_steps": 680, "total_steps": 2200, "loss": 1.8343, "accuracy": 0.796875, "lr": 8.039994484900463e-06, "epoch": 1.2343324250681198, "percentage": 30.91, "elapsed_time": "23:07:05", "remaining_time": "2 days, 3:40:32"}
+{"current_steps": 681, "total_steps": 2200, "loss": 2.0966, "accuracy": 0.671875, "lr": 8.034687426952023e-06, "epoch": 1.2361489554950045, "percentage": 30.95, "elapsed_time": "23:08:57", "remaining_time": "2 days, 3:38:08"}
+{"current_steps": 682, "total_steps": 2200, "loss": 1.7236, "accuracy": 0.875, "lr": 8.029375200334588e-06, "epoch": 1.2379654859218892, "percentage": 31.0, "elapsed_time": "23:10:43", "remaining_time": "2 days, 3:35:28"}
+{"current_steps": 683, "total_steps": 2200, "loss": 2.0931, "accuracy": 0.75, "lr": 8.024057815880716e-06, "epoch": 1.2397820163487738, "percentage": 31.05, "elapsed_time": "23:12:51", "remaining_time": "2 days, 3:33:38"}
+{"current_steps": 684, "total_steps": 2200, "loss": 1.9353, "accuracy": 0.796875, "lr": 8.018735284433475e-06, "epoch": 1.2415985467756585, "percentage": 31.09, "elapsed_time": "23:14:50", "remaining_time": "2 days, 3:31:29"}
+{"current_steps": 685, "total_steps": 2200, "loss": 1.9716, "accuracy": 0.796875, "lr": 8.013407616846436e-06, "epoch": 1.2434150772025432, "percentage": 31.14, "elapsed_time": "23:16:53", "remaining_time": "2 days, 3:29:27"}
+{"current_steps": 686, "total_steps": 2200, "loss": 1.7989, "accuracy": 0.765625, "lr": 8.008074823983642e-06, "epoch": 1.2452316076294279, "percentage": 31.18, "elapsed_time": "23:19:57", "remaining_time": "2 days, 3:29:42"}
+{"current_steps": 687, "total_steps": 2200, "loss": 2.3677, "accuracy": 0.765625, "lr": 8.002736916719582e-06, "epoch": 1.2470481380563125, "percentage": 31.23, "elapsed_time": "23:22:04", "remaining_time": "2 days, 3:27:50"}
+{"current_steps": 688, "total_steps": 2200, "loss": 2.2395, "accuracy": 0.78125, "lr": 7.997393905939183e-06, "epoch": 1.248864668483197, "percentage": 31.27, "elapsed_time": "23:24:02", "remaining_time": "2 days, 3:25:38"}
+{"current_steps": 689, "total_steps": 2200, "loss": 2.0968, "accuracy": 0.75, "lr": 7.992045802537772e-06, "epoch": 1.2506811989100817, "percentage": 31.32, "elapsed_time": "23:25:57", "remaining_time": "2 days, 3:23:18"}
+{"current_steps": 690, "total_steps": 2200, "loss": 2.3605, "accuracy": 0.71875, "lr": 7.986692617421063e-06, "epoch": 1.2524977293369663, "percentage": 31.36, "elapsed_time": "23:28:12", "remaining_time": "2 days, 3:21:43"}
+{"current_steps": 691, "total_steps": 2200, "loss": 1.9464, "accuracy": 0.765625, "lr": 7.981334361505131e-06, "epoch": 1.254314259763851, "percentage": 31.41, "elapsed_time": "23:30:03", "remaining_time": "2 days, 3:19:17"}
+{"current_steps": 692, "total_steps": 2200, "loss": 1.8599, "accuracy": 0.765625, "lr": 7.975971045716398e-06, "epoch": 1.2561307901907357, "percentage": 31.45, "elapsed_time": "23:32:11", "remaining_time": "2 days, 3:17:26"}
+{"current_steps": 693, "total_steps": 2200, "loss": 2.0777, "accuracy": 0.765625, "lr": 7.970602680991594e-06, "epoch": 1.2579473206176204, "percentage": 31.5, "elapsed_time": "23:34:32", "remaining_time": "2 days, 3:16:04"}
+{"current_steps": 694, "total_steps": 2200, "loss": 2.1322, "accuracy": 0.71875, "lr": 7.965229278277749e-06, "epoch": 1.259763851044505, "percentage": 31.55, "elapsed_time": "23:36:33", "remaining_time": "2 days, 3:13:57"}
+{"current_steps": 695, "total_steps": 2200, "loss": 1.8157, "accuracy": 0.828125, "lr": 7.959850848532172e-06, "epoch": 1.2615803814713895, "percentage": 31.59, "elapsed_time": "23:38:21", "remaining_time": "2 days, 3:11:24"}
+{"current_steps": 696, "total_steps": 2200, "loss": 2.3068, "accuracy": 0.640625, "lr": 7.95446740272241e-06, "epoch": 1.2633969118982744, "percentage": 31.64, "elapsed_time": "23:40:37", "remaining_time": "2 days, 3:09:50"}
+{"current_steps": 697, "total_steps": 2200, "loss": 1.9729, "accuracy": 0.84375, "lr": 7.949078951826254e-06, "epoch": 1.2652134423251589, "percentage": 31.68, "elapsed_time": "23:42:52", "remaining_time": "2 days, 3:08:16"}
+{"current_steps": 698, "total_steps": 2200, "loss": 2.1536, "accuracy": 0.8125, "lr": 7.943685506831689e-06, "epoch": 1.2670299727520435, "percentage": 31.73, "elapsed_time": "23:44:45", "remaining_time": "2 days, 3:05:52"}
+{"current_steps": 699, "total_steps": 2200, "loss": 2.0396, "accuracy": 0.71875, "lr": 7.938287078736889e-06, "epoch": 1.2688465031789282, "percentage": 31.77, "elapsed_time": "23:46:48", "remaining_time": "2 days, 3:03:51"}
+{"current_steps": 700, "total_steps": 2200, "loss": 1.9223, "accuracy": 0.828125, "lr": 7.932883678550191e-06, "epoch": 1.270663033605813, "percentage": 31.82, "elapsed_time": "23:48:52", "remaining_time": "2 days, 3:01:52"}
+{"current_steps": 701, "total_steps": 2200, "loss": 1.9669, "accuracy": 0.84375, "lr": 7.927475317290065e-06, "epoch": 1.2724795640326976, "percentage": 31.86, "elapsed_time": "23:51:01", "remaining_time": "2 days, 3:00:02"}
+{"current_steps": 702, "total_steps": 2200, "loss": 2.2422, "accuracy": 0.6875, "lr": 7.922062005985105e-06, "epoch": 1.2742960944595823, "percentage": 31.91, "elapsed_time": "23:52:50", "remaining_time": "2 days, 2:57:32"}
+{"current_steps": 703, "total_steps": 2200, "loss": 2.286, "accuracy": 0.75, "lr": 7.916643755673994e-06, "epoch": 1.276112624886467, "percentage": 31.95, "elapsed_time": "23:54:46", "remaining_time": "2 days, 2:55:17"}
+{"current_steps": 704, "total_steps": 2200, "loss": 2.2664, "accuracy": 0.71875, "lr": 7.911220577405485e-06, "epoch": 1.2779291553133514, "percentage": 32.0, "elapsed_time": "23:57:02", "remaining_time": "2 days, 2:53:43"}
+{"current_steps": 705, "total_steps": 2200, "loss": 2.2835, "accuracy": 0.703125, "lr": 7.905792482238387e-06, "epoch": 1.2797456857402363, "percentage": 32.05, "elapsed_time": "23:59:14", "remaining_time": "2 days, 2:52:01"}
+{"current_steps": 706, "total_steps": 2200, "loss": 1.9752, "accuracy": 0.796875, "lr": 7.900359481241531e-06, "epoch": 1.2815622161671207, "percentage": 32.09, "elapsed_time": "1 day, 0:00:47", "remaining_time": "2 days, 2:48:55"}
+{"current_steps": 707, "total_steps": 2200, "loss": 2.0422, "accuracy": 0.796875, "lr": 7.89492158549375e-06, "epoch": 1.2833787465940054, "percentage": 32.14, "elapsed_time": "1 day, 0:03:00", "remaining_time": "2 days, 2:47:16"}
+{"current_steps": 708, "total_steps": 2200, "loss": 2.1817, "accuracy": 0.8125, "lr": 7.889478806083859e-06, "epoch": 1.28519527702089, "percentage": 32.18, "elapsed_time": "1 day, 0:05:24", "remaining_time": "2 days, 2:45:58"}
+{"current_steps": 709, "total_steps": 2200, "loss": 2.1021, "accuracy": 0.828125, "lr": 7.884031154110636e-06, "epoch": 1.2870118074477748, "percentage": 32.23, "elapsed_time": "1 day, 0:07:20", "remaining_time": "2 days, 2:43:42"}
+{"current_steps": 710, "total_steps": 2200, "loss": 1.8779, "accuracy": 0.734375, "lr": 7.87857864068279e-06, "epoch": 1.2888283378746594, "percentage": 32.27, "elapsed_time": "1 day, 0:09:24", "remaining_time": "2 days, 2:41:42"}
+{"current_steps": 711, "total_steps": 2200, "loss": 2.2478, "accuracy": 0.703125, "lr": 7.873121276918947e-06, "epoch": 1.2906448683015441, "percentage": 32.32, "elapsed_time": "1 day, 0:11:29", "remaining_time": "2 days, 2:39:45"}
+{"current_steps": 712, "total_steps": 2200, "loss": 2.2313, "accuracy": 0.671875, "lr": 7.867659073947618e-06, "epoch": 1.2924613987284288, "percentage": 32.36, "elapsed_time": "1 day, 0:13:32", "remaining_time": "2 days, 2:37:45"}
+{"current_steps": 713, "total_steps": 2200, "loss": 1.8824, "accuracy": 0.75, "lr": 7.862192042907191e-06, "epoch": 1.2942779291553133, "percentage": 32.41, "elapsed_time": "1 day, 0:15:34", "remaining_time": "2 days, 2:35:40"}
+{"current_steps": 714, "total_steps": 2200, "loss": 2.0765, "accuracy": 0.765625, "lr": 7.856720194945892e-06, "epoch": 1.296094459582198, "percentage": 32.45, "elapsed_time": "1 day, 0:17:18", "remaining_time": "2 days, 2:32:59"}
+{"current_steps": 715, "total_steps": 2200, "loss": 2.0101, "accuracy": 0.8125, "lr": 7.85124354122177e-06, "epoch": 1.2979109900090826, "percentage": 32.5, "elapsed_time": "1 day, 0:19:17", "remaining_time": "2 days, 2:30:49"}
+{"current_steps": 716, "total_steps": 2200, "loss": 1.9544, "accuracy": 0.78125, "lr": 7.845762092902678e-06, "epoch": 1.2997275204359673, "percentage": 32.55, "elapsed_time": "1 day, 0:21:12", "remaining_time": "2 days, 2:28:31"}
+{"current_steps": 717, "total_steps": 2200, "loss": 2.3578, "accuracy": 0.71875, "lr": 7.840275861166243e-06, "epoch": 1.301544050862852, "percentage": 32.59, "elapsed_time": "1 day, 0:23:29", "remaining_time": "2 days, 2:26:59"}
+{"current_steps": 718, "total_steps": 2200, "loss": 2.0909, "accuracy": 0.75, "lr": 7.834784857199846e-06, "epoch": 1.3033605812897366, "percentage": 32.64, "elapsed_time": "1 day, 0:25:11", "remaining_time": "2 days, 2:24:15"}
+{"current_steps": 719, "total_steps": 2200, "loss": 2.1544, "accuracy": 0.78125, "lr": 7.8292890922006e-06, "epoch": 1.3051771117166213, "percentage": 32.68, "elapsed_time": "1 day, 0:26:54", "remaining_time": "2 days, 2:21:32"}
+{"current_steps": 720, "total_steps": 2200, "loss": 2.0783, "accuracy": 0.75, "lr": 7.82378857737533e-06, "epoch": 1.3069936421435058, "percentage": 32.73, "elapsed_time": "1 day, 0:29:15", "remaining_time": "2 days, 2:20:09"}
+{"current_steps": 721, "total_steps": 2200, "loss": 2.0252, "accuracy": 0.71875, "lr": 7.818283323940539e-06, "epoch": 1.3088101725703907, "percentage": 32.77, "elapsed_time": "1 day, 0:31:19", "remaining_time": "2 days, 2:18:09"}
+{"current_steps": 722, "total_steps": 2200, "loss": 1.9435, "accuracy": 0.796875, "lr": 7.8127733431224e-06, "epoch": 1.3106267029972751, "percentage": 32.82, "elapsed_time": "1 day, 0:33:31", "remaining_time": "2 days, 2:16:26"}
+{"current_steps": 723, "total_steps": 2200, "loss": 2.0554, "accuracy": 0.828125, "lr": 7.80725864615672e-06, "epoch": 1.3124432334241598, "percentage": 32.86, "elapsed_time": "1 day, 0:35:42", "remaining_time": "2 days, 2:14:40"}
+{"current_steps": 724, "total_steps": 2200, "loss": 1.8288, "accuracy": 0.796875, "lr": 7.801739244288929e-06, "epoch": 1.3142597638510445, "percentage": 32.91, "elapsed_time": "1 day, 0:37:37", "remaining_time": "2 days, 2:12:23"}
+{"current_steps": 725, "total_steps": 2200, "loss": 1.8875, "accuracy": 0.796875, "lr": 7.79621514877405e-06, "epoch": 1.3160762942779292, "percentage": 32.95, "elapsed_time": "1 day, 0:40:01", "remaining_time": "2 days, 2:11:04"}
+{"current_steps": 726, "total_steps": 2200, "loss": 1.7664, "accuracy": 0.78125, "lr": 7.79068637087667e-06, "epoch": 1.3178928247048138, "percentage": 33.0, "elapsed_time": "1 day, 0:42:02", "remaining_time": "2 days, 2:08:59"}
+{"current_steps": 727, "total_steps": 2200, "loss": 2.3856, "accuracy": 0.765625, "lr": 7.785152921870939e-06, "epoch": 1.3197093551316985, "percentage": 33.05, "elapsed_time": "1 day, 0:44:00", "remaining_time": "2 days, 2:06:48"}
+{"current_steps": 728, "total_steps": 2200, "loss": 1.9723, "accuracy": 0.75, "lr": 7.779614813040513e-06, "epoch": 1.3215258855585832, "percentage": 33.09, "elapsed_time": "1 day, 0:45:52", "remaining_time": "2 days, 2:04:24"}
+{"current_steps": 729, "total_steps": 2200, "loss": 2.1344, "accuracy": 0.75, "lr": 7.774072055678565e-06, "epoch": 1.3233424159854676, "percentage": 33.14, "elapsed_time": "1 day, 0:47:51", "remaining_time": "2 days, 2:02:15"}
+{"current_steps": 730, "total_steps": 2200, "loss": 2.1545, "accuracy": 0.8125, "lr": 7.768524661087744e-06, "epoch": 1.3251589464123525, "percentage": 33.18, "elapsed_time": "1 day, 0:50:01", "remaining_time": "2 days, 2:00:27"}
+{"current_steps": 731, "total_steps": 2200, "loss": 2.0652, "accuracy": 0.734375, "lr": 7.76297264058015e-06, "epoch": 1.326975476839237, "percentage": 33.23, "elapsed_time": "1 day, 0:52:26", "remaining_time": "2 days, 1:59:10"}
+{"current_steps": 732, "total_steps": 2200, "loss": 2.0932, "accuracy": 0.6875, "lr": 7.757416005477323e-06, "epoch": 1.3287920072661217, "percentage": 33.27, "elapsed_time": "1 day, 0:54:16", "remaining_time": "2 days, 1:56:42"}
+{"current_steps": 733, "total_steps": 2200, "loss": 1.9757, "accuracy": 0.71875, "lr": 7.751854767110205e-06, "epoch": 1.3306085376930064, "percentage": 33.32, "elapsed_time": "1 day, 0:55:58", "remaining_time": "2 days, 1:54:00"}
+{"current_steps": 734, "total_steps": 2200, "loss": 1.999, "accuracy": 0.734375, "lr": 7.746288936819134e-06, "epoch": 1.332425068119891, "percentage": 33.36, "elapsed_time": "1 day, 0:57:48", "remaining_time": "2 days, 1:51:32"}
+{"current_steps": 735, "total_steps": 2200, "loss": 1.8674, "accuracy": 0.78125, "lr": 7.740718525953804e-06, "epoch": 1.3342415985467757, "percentage": 33.41, "elapsed_time": "1 day, 1:00:11", "remaining_time": "2 days, 1:50:11"}
+{"current_steps": 736, "total_steps": 2200, "loss": 1.8622, "accuracy": 0.78125, "lr": 7.735143545873252e-06, "epoch": 1.3360581289736604, "percentage": 33.45, "elapsed_time": "1 day, 1:02:03", "remaining_time": "2 days, 1:47:48"}
+{"current_steps": 737, "total_steps": 2200, "loss": 2.2686, "accuracy": 0.65625, "lr": 7.729564007945834e-06, "epoch": 1.337874659400545, "percentage": 33.5, "elapsed_time": "1 day, 1:04:13", "remaining_time": "2 days, 1:46:00"}
+{"current_steps": 738, "total_steps": 2200, "loss": 2.4316, "accuracy": 0.6875, "lr": 7.723979923549201e-06, "epoch": 1.3396911898274295, "percentage": 33.55, "elapsed_time": "1 day, 1:05:54", "remaining_time": "2 days, 1:43:15"}
+{"current_steps": 739, "total_steps": 2200, "loss": 2.1929, "accuracy": 0.78125, "lr": 7.718391304070274e-06, "epoch": 1.3415077202543142, "percentage": 33.59, "elapsed_time": "1 day, 1:08:21", "remaining_time": "2 days, 1:42:01"}
+{"current_steps": 740, "total_steps": 2200, "loss": 2.2098, "accuracy": 0.75, "lr": 7.712798160905218e-06, "epoch": 1.3433242506811989, "percentage": 33.64, "elapsed_time": "1 day, 1:10:26", "remaining_time": "2 days, 1:40:03"}
+{"current_steps": 741, "total_steps": 2200, "loss": 2.2352, "accuracy": 0.6875, "lr": 7.707200505459427e-06, "epoch": 1.3451407811080835, "percentage": 33.68, "elapsed_time": "1 day, 1:12:34", "remaining_time": "2 days, 1:38:12"}
+{"current_steps": 742, "total_steps": 2200, "loss": 2.0034, "accuracy": 0.8125, "lr": 7.701598349147496e-06, "epoch": 1.3469573115349682, "percentage": 33.73, "elapsed_time": "1 day, 1:14:48", "remaining_time": "2 days, 1:36:31"}
+{"current_steps": 743, "total_steps": 2200, "loss": 2.3257, "accuracy": 0.703125, "lr": 7.695991703393196e-06, "epoch": 1.348773841961853, "percentage": 33.77, "elapsed_time": "1 day, 1:17:07", "remaining_time": "2 days, 1:35:02"}
+{"current_steps": 744, "total_steps": 2200, "loss": 1.9158, "accuracy": 0.859375, "lr": 7.690380579629454e-06, "epoch": 1.3505903723887376, "percentage": 33.82, "elapsed_time": "1 day, 1:18:45", "remaining_time": "2 days, 1:32:11"}
+{"current_steps": 745, "total_steps": 2200, "loss": 2.405, "accuracy": 0.6875, "lr": 7.684764989298329e-06, "epoch": 1.352406902815622, "percentage": 33.86, "elapsed_time": "1 day, 1:20:59", "remaining_time": "2 days, 1:30:31"}
+{"current_steps": 746, "total_steps": 2200, "loss": 2.4038, "accuracy": 0.78125, "lr": 7.679144943850986e-06, "epoch": 1.354223433242507, "percentage": 33.91, "elapsed_time": "1 day, 1:23:01", "remaining_time": "2 days, 1:28:28"}
+{"current_steps": 747, "total_steps": 2200, "loss": 1.6934, "accuracy": 0.8125, "lr": 7.673520454747681e-06, "epoch": 1.3560399636693914, "percentage": 33.95, "elapsed_time": "1 day, 1:25:19", "remaining_time": "2 days, 1:26:55"}
+{"current_steps": 748, "total_steps": 2200, "loss": 2.0308, "accuracy": 0.71875, "lr": 7.667891533457718e-06, "epoch": 1.357856494096276, "percentage": 34.0, "elapsed_time": "1 day, 1:27:07", "remaining_time": "2 days, 1:24:24"}
+{"current_steps": 749, "total_steps": 2200, "loss": 2.2639, "accuracy": 0.703125, "lr": 7.662258191459456e-06, "epoch": 1.3596730245231607, "percentage": 34.05, "elapsed_time": "1 day, 1:29:17", "remaining_time": "2 days, 1:22:37"}
+{"current_steps": 750, "total_steps": 2200, "loss": 2.0835, "accuracy": 0.75, "lr": 7.656620440240257e-06, "epoch": 1.3614895549500454, "percentage": 34.09, "elapsed_time": "1 day, 1:31:16", "remaining_time": "2 days, 1:20:28"}
+{"current_steps": 751, "total_steps": 2200, "loss": 2.1026, "accuracy": 0.71875, "lr": 7.650978291296475e-06, "epoch": 1.36330608537693, "percentage": 34.14, "elapsed_time": "1 day, 1:33:09", "remaining_time": "2 days, 1:18:06"}
+{"current_steps": 752, "total_steps": 2200, "loss": 1.8117, "accuracy": 0.8125, "lr": 7.645331756133438e-06, "epoch": 1.3651226158038148, "percentage": 34.18, "elapsed_time": "1 day, 1:35:00", "remaining_time": "2 days, 1:15:42"}
+{"current_steps": 753, "total_steps": 2200, "loss": 1.9148, "accuracy": 0.734375, "lr": 7.639680846265411e-06, "epoch": 1.3669391462306995, "percentage": 34.23, "elapsed_time": "1 day, 1:36:40", "remaining_time": "2 days, 1:12:56"}
+{"current_steps": 754, "total_steps": 2200, "loss": 1.9067, "accuracy": 0.78125, "lr": 7.634025573215585e-06, "epoch": 1.368755676657584, "percentage": 34.27, "elapsed_time": "1 day, 1:39:02", "remaining_time": "2 days, 1:11:31"}
+{"current_steps": 755, "total_steps": 2200, "loss": 2.195, "accuracy": 0.75, "lr": 7.628365948516047e-06, "epoch": 1.3705722070844686, "percentage": 34.32, "elapsed_time": "1 day, 1:41:19", "remaining_time": "2 days, 1:09:56"}
+{"current_steps": 756, "total_steps": 2200, "loss": 2.1411, "accuracy": 0.75, "lr": 7.622701983707756e-06, "epoch": 1.3723887375113533, "percentage": 34.36, "elapsed_time": "1 day, 1:42:58", "remaining_time": "2 days, 1:07:09"}
+{"current_steps": 757, "total_steps": 2200, "loss": 2.203, "accuracy": 0.8125, "lr": 7.617033690340523e-06, "epoch": 1.374205267938238, "percentage": 34.41, "elapsed_time": "1 day, 1:44:42", "remaining_time": "2 days, 1:04:32"}
+{"current_steps": 758, "total_steps": 2200, "loss": 1.5709, "accuracy": 0.859375, "lr": 7.611361079972984e-06, "epoch": 1.3760217983651226, "percentage": 34.45, "elapsed_time": "1 day, 1:46:23", "remaining_time": "2 days, 1:01:49"}
+{"current_steps": 759, "total_steps": 2200, "loss": 2.2132, "accuracy": 0.640625, "lr": 7.605684164172581e-06, "epoch": 1.3778383287920073, "percentage": 34.5, "elapsed_time": "1 day, 1:48:07", "remaining_time": "2 days, 0:59:11"}
+{"current_steps": 760, "total_steps": 2200, "loss": 1.9836, "accuracy": 0.734375, "lr": 7.600002954515532e-06, "epoch": 1.379654859218892, "percentage": 34.55, "elapsed_time": "1 day, 1:50:00", "remaining_time": "2 days, 0:56:52"}
+{"current_steps": 761, "total_steps": 2200, "loss": 2.4998, "accuracy": 0.640625, "lr": 7.5943174625868146e-06, "epoch": 1.3814713896457766, "percentage": 34.59, "elapsed_time": "1 day, 1:51:56", "remaining_time": "2 days, 0:54:36"}
+{"current_steps": 762, "total_steps": 2200, "loss": 1.8657, "accuracy": 0.765625, "lr": 7.588627699980134e-06, "epoch": 1.3832879200726613, "percentage": 34.64, "elapsed_time": "1 day, 1:53:47", "remaining_time": "2 days, 0:52:12"}
+{"current_steps": 763, "total_steps": 2200, "loss": 2.3292, "accuracy": 0.671875, "lr": 7.582933678297911e-06, "epoch": 1.3851044504995458, "percentage": 34.68, "elapsed_time": "1 day, 1:55:52", "remaining_time": "2 days, 0:50:16"}
+{"current_steps": 764, "total_steps": 2200, "loss": 2.5069, "accuracy": 0.71875, "lr": 7.577235409151242e-06, "epoch": 1.3869209809264305, "percentage": 34.73, "elapsed_time": "1 day, 1:57:48", "remaining_time": "2 days, 0:48:01"}
+{"current_steps": 765, "total_steps": 2200, "loss": 2.1158, "accuracy": 0.703125, "lr": 7.5715329041598955e-06, "epoch": 1.3887375113533151, "percentage": 34.77, "elapsed_time": "1 day, 1:59:50", "remaining_time": "2 days, 0:45:58"}
+{"current_steps": 766, "total_steps": 2200, "loss": 2.0158, "accuracy": 0.78125, "lr": 7.565826174952267e-06, "epoch": 1.3905540417801998, "percentage": 34.82, "elapsed_time": "1 day, 2:02:01", "remaining_time": "2 days, 0:44:12"}
+{"current_steps": 767, "total_steps": 2200, "loss": 1.9386, "accuracy": 0.796875, "lr": 7.560115233165375e-06, "epoch": 1.3923705722070845, "percentage": 34.86, "elapsed_time": "1 day, 2:03:53", "remaining_time": "2 days, 0:41:51"}
+{"current_steps": 768, "total_steps": 2200, "loss": 1.9933, "accuracy": 0.8125, "lr": 7.554400090444822e-06, "epoch": 1.3941871026339692, "percentage": 34.91, "elapsed_time": "1 day, 2:05:36", "remaining_time": "2 days, 0:39:11"}
+{"current_steps": 769, "total_steps": 2200, "loss": 1.784, "accuracy": 0.84375, "lr": 7.54868075844478e-06, "epoch": 1.3960036330608538, "percentage": 34.95, "elapsed_time": "1 day, 2:07:38", "remaining_time": "2 days, 0:37:10"}
+{"current_steps": 770, "total_steps": 2200, "loss": 2.2637, "accuracy": 0.703125, "lr": 7.5429572488279615e-06, "epoch": 1.3978201634877383, "percentage": 35.0, "elapsed_time": "1 day, 2:09:23", "remaining_time": "2 days, 0:34:34"}
+{"current_steps": 771, "total_steps": 2200, "loss": 2.412, "accuracy": 0.671875, "lr": 7.5372295732656e-06, "epoch": 1.3996366939146232, "percentage": 35.05, "elapsed_time": "1 day, 2:11:32", "remaining_time": "2 days, 0:32:45"}
+{"current_steps": 772, "total_steps": 2200, "loss": 2.1841, "accuracy": 0.734375, "lr": 7.531497743437424e-06, "epoch": 1.4014532243415077, "percentage": 35.09, "elapsed_time": "1 day, 2:13:29", "remaining_time": "2 days, 0:30:32"}
+{"current_steps": 773, "total_steps": 2200, "loss": 2.064, "accuracy": 0.78125, "lr": 7.525761771031632e-06, "epoch": 1.4032697547683923, "percentage": 35.14, "elapsed_time": "1 day, 2:15:11", "remaining_time": "2 days, 0:27:52"}
+{"current_steps": 774, "total_steps": 2200, "loss": 2.077, "accuracy": 0.703125, "lr": 7.520021667744869e-06, "epoch": 1.405086285195277, "percentage": 35.18, "elapsed_time": "1 day, 2:17:20", "remaining_time": "2 days, 0:26:03"}
+{"current_steps": 775, "total_steps": 2200, "loss": 1.8298, "accuracy": 0.8125, "lr": 7.514277445282206e-06, "epoch": 1.4069028156221617, "percentage": 35.23, "elapsed_time": "1 day, 2:19:26", "remaining_time": "2 days, 0:24:08"}
+{"current_steps": 776, "total_steps": 2200, "loss": 2.1025, "accuracy": 0.78125, "lr": 7.508529115357115e-06, "epoch": 1.4087193460490464, "percentage": 35.27, "elapsed_time": "1 day, 2:21:14", "remaining_time": "2 days, 0:21:39"}
+{"current_steps": 777, "total_steps": 2200, "loss": 2.032, "accuracy": 0.8125, "lr": 7.502776689691439e-06, "epoch": 1.410535876475931, "percentage": 35.32, "elapsed_time": "1 day, 2:23:39", "remaining_time": "2 days, 0:20:18"}
+{"current_steps": 778, "total_steps": 2200, "loss": 2.0954, "accuracy": 0.71875, "lr": 7.4970201800153765e-06, "epoch": 1.4123524069028157, "percentage": 35.36, "elapsed_time": "1 day, 2:25:24", "remaining_time": "2 days, 0:17:45"}
+{"current_steps": 779, "total_steps": 2200, "loss": 1.8012, "accuracy": 0.8125, "lr": 7.491259598067452e-06, "epoch": 1.4141689373297002, "percentage": 35.41, "elapsed_time": "1 day, 2:27:43", "remaining_time": "2 days, 0:16:13"}
+{"current_steps": 780, "total_steps": 2200, "loss": 2.215, "accuracy": 0.71875, "lr": 7.485494955594495e-06, "epoch": 1.4159854677565848, "percentage": 35.45, "elapsed_time": "1 day, 2:29:48", "remaining_time": "2 days, 0:14:15"}
+{"current_steps": 781, "total_steps": 2200, "loss": 1.9022, "accuracy": 0.8125, "lr": 7.479726264351619e-06, "epoch": 1.4178019981834695, "percentage": 35.5, "elapsed_time": "1 day, 2:31:43", "remaining_time": "2 days, 0:12:00"}
+{"current_steps": 782, "total_steps": 2200, "loss": 1.9222, "accuracy": 0.765625, "lr": 7.473953536102185e-06, "epoch": 1.4196185286103542, "percentage": 35.55, "elapsed_time": "1 day, 2:33:33", "remaining_time": "2 days, 0:09:36"}
+{"current_steps": 783, "total_steps": 2200, "loss": 1.7606, "accuracy": 0.84375, "lr": 7.468176782617793e-06, "epoch": 1.4214350590372389, "percentage": 35.59, "elapsed_time": "1 day, 2:35:36", "remaining_time": "2 days, 0:07:34"}
+{"current_steps": 784, "total_steps": 2200, "loss": 1.892, "accuracy": 0.828125, "lr": 7.462396015678249e-06, "epoch": 1.4232515894641236, "percentage": 35.64, "elapsed_time": "1 day, 2:37:08", "remaining_time": "2 days, 0:04:38"}
+{"current_steps": 785, "total_steps": 2200, "loss": 2.1411, "accuracy": 0.8125, "lr": 7.456611247071544e-06, "epoch": 1.4250681198910082, "percentage": 35.68, "elapsed_time": "1 day, 2:39:02", "remaining_time": "2 days, 0:02:21"}
+{"current_steps": 786, "total_steps": 2200, "loss": 1.8014, "accuracy": 0.8125, "lr": 7.450822488593829e-06, "epoch": 1.4268846503178927, "percentage": 35.73, "elapsed_time": "1 day, 2:41:10", "remaining_time": "2 days, 0:00:29"}
+{"current_steps": 787, "total_steps": 2200, "loss": 1.7508, "accuracy": 0.84375, "lr": 7.445029752049387e-06, "epoch": 1.4287011807447776, "percentage": 35.77, "elapsed_time": "1 day, 2:43:10", "remaining_time": "1 day, 23:58:23"}
+{"current_steps": 788, "total_steps": 2200, "loss": 1.8207, "accuracy": 0.734375, "lr": 7.439233049250622e-06, "epoch": 1.430517711171662, "percentage": 35.82, "elapsed_time": "1 day, 2:44:45", "remaining_time": "1 day, 23:55:31"}
+{"current_steps": 789, "total_steps": 2200, "loss": 1.781, "accuracy": 0.84375, "lr": 7.433432392018013e-06, "epoch": 1.4323342415985467, "percentage": 35.86, "elapsed_time": "1 day, 2:46:57", "remaining_time": "1 day, 23:53:46"}
+{"current_steps": 790, "total_steps": 2200, "loss": 2.2499, "accuracy": 0.703125, "lr": 7.427627792180116e-06, "epoch": 1.4341507720254314, "percentage": 35.91, "elapsed_time": "1 day, 2:48:35", "remaining_time": "1 day, 23:51:01"}
+{"current_steps": 791, "total_steps": 2200, "loss": 1.8155, "accuracy": 0.828125, "lr": 7.4218192615735185e-06, "epoch": 1.435967302452316, "percentage": 35.95, "elapsed_time": "1 day, 2:51:04", "remaining_time": "1 day, 23:49:46"}
+{"current_steps": 792, "total_steps": 2200, "loss": 1.9892, "accuracy": 0.8125, "lr": 7.416006812042827e-06, "epoch": 1.4377838328792008, "percentage": 36.0, "elapsed_time": "1 day, 2:53:13", "remaining_time": "1 day, 23:47:56"}
+{"current_steps": 793, "total_steps": 2200, "loss": 1.8609, "accuracy": 0.796875, "lr": 7.410190455440638e-06, "epoch": 1.4396003633060854, "percentage": 36.05, "elapsed_time": "1 day, 2:55:07", "remaining_time": "1 day, 23:45:40"}
+{"current_steps": 794, "total_steps": 2200, "loss": 2.4321, "accuracy": 0.59375, "lr": 7.4043702036275136e-06, "epoch": 1.44141689373297, "percentage": 36.09, "elapsed_time": "1 day, 2:57:01", "remaining_time": "1 day, 23:43:23"}
+{"current_steps": 795, "total_steps": 2200, "loss": 2.1999, "accuracy": 0.734375, "lr": 7.398546068471961e-06, "epoch": 1.4432334241598546, "percentage": 36.14, "elapsed_time": "1 day, 2:58:53", "remaining_time": "1 day, 23:41:03"}
+{"current_steps": 796, "total_steps": 2200, "loss": 1.9802, "accuracy": 0.75, "lr": 7.392718061850407e-06, "epoch": 1.4450499545867395, "percentage": 36.18, "elapsed_time": "1 day, 3:01:10", "remaining_time": "1 day, 23:39:28"}
+{"current_steps": 797, "total_steps": 2200, "loss": 1.7361, "accuracy": 0.859375, "lr": 7.3868861956471725e-06, "epoch": 1.446866485013624, "percentage": 36.23, "elapsed_time": "1 day, 3:02:49", "remaining_time": "1 day, 23:36:43"}
+{"current_steps": 798, "total_steps": 2200, "loss": 2.3738, "accuracy": 0.6875, "lr": 7.381050481754447e-06, "epoch": 1.4486830154405086, "percentage": 36.27, "elapsed_time": "1 day, 3:04:53", "remaining_time": "1 day, 23:34:44"}
+{"current_steps": 799, "total_steps": 2200, "loss": 2.4222, "accuracy": 0.703125, "lr": 7.375210932072265e-06, "epoch": 1.4504995458673933, "percentage": 36.32, "elapsed_time": "1 day, 3:07:05", "remaining_time": "1 day, 23:33:00"}
+{"current_steps": 800, "total_steps": 2200, "loss": 1.8844, "accuracy": 0.796875, "lr": 7.36936755850849e-06, "epoch": 1.452316076294278, "percentage": 36.36, "elapsed_time": "1 day, 3:09:01", "remaining_time": "1 day, 23:30:47"}
+{"current_steps": 801, "total_steps": 2200, "loss": 2.5897, "accuracy": 0.625, "lr": 7.363520372978774e-06, "epoch": 1.4541326067211626, "percentage": 36.41, "elapsed_time": "1 day, 3:10:58", "remaining_time": "1 day, 23:28:36"}
+{"current_steps": 802, "total_steps": 2200, "loss": 1.8134, "accuracy": 0.796875, "lr": 7.357669387406548e-06, "epoch": 1.4559491371480473, "percentage": 36.45, "elapsed_time": "1 day, 3:12:45", "remaining_time": "1 day, 23:26:07"}
+{"current_steps": 803, "total_steps": 2200, "loss": 2.0345, "accuracy": 0.78125, "lr": 7.351814613722991e-06, "epoch": 1.457765667574932, "percentage": 36.5, "elapsed_time": "1 day, 3:14:55", "remaining_time": "1 day, 23:24:19"}
+{"current_steps": 804, "total_steps": 2200, "loss": 1.7825, "accuracy": 0.859375, "lr": 7.3459560638670035e-06, "epoch": 1.4595821980018164, "percentage": 36.55, "elapsed_time": "1 day, 3:17:09", "remaining_time": "1 day, 23:22:38"}
+{"current_steps": 805, "total_steps": 2200, "loss": 1.9808, "accuracy": 0.8125, "lr": 7.3400937497851906e-06, "epoch": 1.461398728428701, "percentage": 36.59, "elapsed_time": "1 day, 3:18:52", "remaining_time": "1 day, 23:20:03"}
+{"current_steps": 806, "total_steps": 2200, "loss": 1.8753, "accuracy": 0.796875, "lr": 7.334227683431832e-06, "epoch": 1.4632152588555858, "percentage": 36.64, "elapsed_time": "1 day, 3:20:56", "remaining_time": "1 day, 23:18:03"}
+{"current_steps": 807, "total_steps": 2200, "loss": 2.3324, "accuracy": 0.703125, "lr": 7.328357876768858e-06, "epoch": 1.4650317892824705, "percentage": 36.68, "elapsed_time": "1 day, 3:22:44", "remaining_time": "1 day, 23:15:37"}
+{"current_steps": 808, "total_steps": 2200, "loss": 2.0898, "accuracy": 0.765625, "lr": 7.322484341765824e-06, "epoch": 1.4668483197093551, "percentage": 36.73, "elapsed_time": "1 day, 3:24:49", "remaining_time": "1 day, 23:13:38"}
+{"current_steps": 809, "total_steps": 2200, "loss": 2.0743, "accuracy": 0.765625, "lr": 7.316607090399894e-06, "epoch": 1.4686648501362398, "percentage": 36.77, "elapsed_time": "1 day, 3:26:26", "remaining_time": "1 day, 23:10:53"}
+{"current_steps": 810, "total_steps": 2200, "loss": 1.7474, "accuracy": 0.859375, "lr": 7.310726134655807e-06, "epoch": 1.4704813805631245, "percentage": 36.82, "elapsed_time": "1 day, 3:28:10", "remaining_time": "1 day, 23:08:21"}
+{"current_steps": 811, "total_steps": 2200, "loss": 2.0516, "accuracy": 0.78125, "lr": 7.30484148652585e-06, "epoch": 1.472297910990009, "percentage": 36.86, "elapsed_time": "1 day, 3:30:20", "remaining_time": "1 day, 23:06:31"}
+{"current_steps": 812, "total_steps": 2200, "loss": 1.9513, "accuracy": 0.765625, "lr": 7.298953158009852e-06, "epoch": 1.4741144414168939, "percentage": 36.91, "elapsed_time": "1 day, 3:32:34", "remaining_time": "1 day, 23:04:51"}
+{"current_steps": 813, "total_steps": 2200, "loss": 1.9357, "accuracy": 0.796875, "lr": 7.293061161115138e-06, "epoch": 1.4759309718437783, "percentage": 36.95, "elapsed_time": "1 day, 3:34:26", "remaining_time": "1 day, 23:02:30"}
+{"current_steps": 814, "total_steps": 2200, "loss": 1.9108, "accuracy": 0.828125, "lr": 7.287165507856513e-06, "epoch": 1.477747502270663, "percentage": 37.0, "elapsed_time": "1 day, 3:36:36", "remaining_time": "1 day, 23:00:43"}
+{"current_steps": 815, "total_steps": 2200, "loss": 1.9764, "accuracy": 0.828125, "lr": 7.281266210256244e-06, "epoch": 1.4795640326975477, "percentage": 37.05, "elapsed_time": "1 day, 3:38:51", "remaining_time": "1 day, 22:59:02"}
+{"current_steps": 816, "total_steps": 2200, "loss": 2.1791, "accuracy": 0.71875, "lr": 7.275363280344023e-06, "epoch": 1.4813805631244323, "percentage": 37.09, "elapsed_time": "1 day, 3:41:01", "remaining_time": "1 day, 22:57:14"}
+{"current_steps": 817, "total_steps": 2200, "loss": 2.0077, "accuracy": 0.71875, "lr": 7.269456730156954e-06, "epoch": 1.483197093551317, "percentage": 37.14, "elapsed_time": "1 day, 3:42:35", "remaining_time": "1 day, 22:54:23"}
+{"current_steps": 818, "total_steps": 2200, "loss": 2.1468, "accuracy": 0.734375, "lr": 7.26354657173952e-06, "epoch": 1.4850136239782017, "percentage": 37.18, "elapsed_time": "1 day, 3:44:13", "remaining_time": "1 day, 22:51:41"}
+{"current_steps": 819, "total_steps": 2200, "loss": 2.0372, "accuracy": 0.703125, "lr": 7.257632817143566e-06, "epoch": 1.4868301544050864, "percentage": 37.23, "elapsed_time": "1 day, 3:46:09", "remaining_time": "1 day, 22:49:29"}
+{"current_steps": 820, "total_steps": 2200, "loss": 1.786, "accuracy": 0.734375, "lr": 7.251715478428263e-06, "epoch": 1.4886466848319708, "percentage": 37.27, "elapsed_time": "1 day, 3:48:25", "remaining_time": "1 day, 22:47:50"}
+{"current_steps": 821, "total_steps": 2200, "loss": 2.1653, "accuracy": 0.71875, "lr": 7.2457945676600975e-06, "epoch": 1.4904632152588557, "percentage": 37.32, "elapsed_time": "1 day, 3:49:59", "remaining_time": "1 day, 22:45:01"}
+{"current_steps": 822, "total_steps": 2200, "loss": 2.0422, "accuracy": 0.671875, "lr": 7.239870096912835e-06, "epoch": 1.4922797456857402, "percentage": 37.36, "elapsed_time": "1 day, 3:51:52", "remaining_time": "1 day, 22:42:43"}
+{"current_steps": 823, "total_steps": 2200, "loss": 1.9344, "accuracy": 0.796875, "lr": 7.233942078267507e-06, "epoch": 1.4940962761126249, "percentage": 37.41, "elapsed_time": "1 day, 3:54:45", "remaining_time": "1 day, 22:42:06"}
+{"current_steps": 824, "total_steps": 2200, "loss": 2.2769, "accuracy": 0.78125, "lr": 7.228010523812372e-06, "epoch": 1.4959128065395095, "percentage": 37.45, "elapsed_time": "1 day, 3:56:54", "remaining_time": "1 day, 22:40:16"}
+{"current_steps": 825, "total_steps": 2200, "loss": 2.0686, "accuracy": 0.765625, "lr": 7.222075445642904e-06, "epoch": 1.4977293369663942, "percentage": 37.5, "elapsed_time": "1 day, 3:59:10", "remaining_time": "1 day, 22:38:38"}
+{"current_steps": 826, "total_steps": 2200, "loss": 1.4687, "accuracy": 0.90625, "lr": 7.216136855861763e-06, "epoch": 1.4995458673932789, "percentage": 37.55, "elapsed_time": "1 day, 4:01:00", "remaining_time": "1 day, 22:36:15"}
+{"current_steps": 827, "total_steps": 2200, "loss": 1.9276, "accuracy": 0.78125, "lr": 7.2101947665787605e-06, "epoch": 1.5013623978201633, "percentage": 37.59, "elapsed_time": "1 day, 4:03:15", "remaining_time": "1 day, 22:34:35"}
+{"current_steps": 828, "total_steps": 2200, "loss": 1.8234, "accuracy": 0.78125, "lr": 7.20424918991086e-06, "epoch": 1.5031789282470482, "percentage": 37.64, "elapsed_time": "1 day, 4:05:11", "remaining_time": "1 day, 22:32:22"}
+{"current_steps": 829, "total_steps": 2200, "loss": 1.8965, "accuracy": 0.828125, "lr": 7.198300137982121e-06, "epoch": 1.5049954586739327, "percentage": 37.68, "elapsed_time": "1 day, 4:07:12", "remaining_time": "1 day, 22:30:18"}
+{"current_steps": 830, "total_steps": 2200, "loss": 1.8165, "accuracy": 0.78125, "lr": 7.1923476229237e-06, "epoch": 1.5068119891008176, "percentage": 37.73, "elapsed_time": "1 day, 4:09:23", "remaining_time": "1 day, 22:28:31"}
+{"current_steps": 831, "total_steps": 2200, "loss": 1.8323, "accuracy": 0.859375, "lr": 7.186391656873809e-06, "epoch": 1.508628519527702, "percentage": 37.77, "elapsed_time": "1 day, 4:11:41", "remaining_time": "1 day, 22:26:54"}
+{"current_steps": 832, "total_steps": 2200, "loss": 2.0109, "accuracy": 0.734375, "lr": 7.180432251977702e-06, "epoch": 1.5104450499545867, "percentage": 37.82, "elapsed_time": "1 day, 4:13:27", "remaining_time": "1 day, 22:24:26"}
+{"current_steps": 833, "total_steps": 2200, "loss": 1.8424, "accuracy": 0.8125, "lr": 7.1744694203876405e-06, "epoch": 1.5122615803814714, "percentage": 37.86, "elapsed_time": "1 day, 4:15:46", "remaining_time": "1 day, 22:22:51"}
+{"current_steps": 834, "total_steps": 2200, "loss": 1.8164, "accuracy": 0.78125, "lr": 7.168503174262881e-06, "epoch": 1.514078110808356, "percentage": 37.91, "elapsed_time": "1 day, 4:17:24", "remaining_time": "1 day, 22:20:10"}
+{"current_steps": 835, "total_steps": 2200, "loss": 2.0479, "accuracy": 0.734375, "lr": 7.162533525769633e-06, "epoch": 1.5158946412352408, "percentage": 37.95, "elapsed_time": "1 day, 4:19:32", "remaining_time": "1 day, 22:18:17"}
+{"current_steps": 836, "total_steps": 2200, "loss": 1.83, "accuracy": 0.75, "lr": 7.156560487081052e-06, "epoch": 1.5177111716621252, "percentage": 38.0, "elapsed_time": "1 day, 4:21:33", "remaining_time": "1 day, 22:16:13"}
+{"current_steps": 837, "total_steps": 2200, "loss": 1.9901, "accuracy": 0.8125, "lr": 7.150584070377202e-06, "epoch": 1.5195277020890101, "percentage": 38.05, "elapsed_time": "1 day, 4:23:26", "remaining_time": "1 day, 22:13:56"}
+{"current_steps": 838, "total_steps": 2200, "loss": 1.7565, "accuracy": 0.84375, "lr": 7.144604287845041e-06, "epoch": 1.5213442325158946, "percentage": 38.09, "elapsed_time": "1 day, 4:25:22", "remaining_time": "1 day, 22:11:45"}
+{"current_steps": 839, "total_steps": 2200, "loss": 1.8708, "accuracy": 0.8125, "lr": 7.1386211516783835e-06, "epoch": 1.5231607629427792, "percentage": 38.14, "elapsed_time": "1 day, 4:27:39", "remaining_time": "1 day, 22:10:06"}
+{"current_steps": 840, "total_steps": 2200, "loss": 1.8225, "accuracy": 0.78125, "lr": 7.132634674077884e-06, "epoch": 1.524977293369664, "percentage": 38.18, "elapsed_time": "1 day, 4:29:21", "remaining_time": "1 day, 22:07:32"}
+{"current_steps": 841, "total_steps": 2200, "loss": 2.2036, "accuracy": 0.6875, "lr": 7.12664486725102e-06, "epoch": 1.5267938237965486, "percentage": 38.23, "elapsed_time": "1 day, 4:31:17", "remaining_time": "1 day, 22:05:20"}
+{"current_steps": 842, "total_steps": 2200, "loss": 2.0745, "accuracy": 0.796875, "lr": 7.120651743412043e-06, "epoch": 1.5286103542234333, "percentage": 38.27, "elapsed_time": "1 day, 4:33:27", "remaining_time": "1 day, 22:03:31"}
+{"current_steps": 843, "total_steps": 2200, "loss": 2.9054, "accuracy": 0.6875, "lr": 7.114655314781979e-06, "epoch": 1.5304268846503177, "percentage": 38.32, "elapsed_time": "1 day, 4:35:23", "remaining_time": "1 day, 22:01:18"}
+{"current_steps": 844, "total_steps": 2200, "loss": 2.0337, "accuracy": 0.765625, "lr": 7.10865559358859e-06, "epoch": 1.5322434150772026, "percentage": 38.36, "elapsed_time": "1 day, 4:37:23", "remaining_time": "1 day, 21:59:12"}
+{"current_steps": 845, "total_steps": 2200, "loss": 2.0159, "accuracy": 0.78125, "lr": 7.102652592066354e-06, "epoch": 1.534059945504087, "percentage": 38.41, "elapsed_time": "1 day, 4:38:57", "remaining_time": "1 day, 21:56:26"}
+{"current_steps": 846, "total_steps": 2200, "loss": 2.2342, "accuracy": 0.75, "lr": 7.096646322456433e-06, "epoch": 1.535876475930972, "percentage": 38.45, "elapsed_time": "1 day, 4:40:55", "remaining_time": "1 day, 21:54:17"}
+{"current_steps": 847, "total_steps": 2200, "loss": 1.9024, "accuracy": 0.796875, "lr": 7.090636797006658e-06, "epoch": 1.5376930063578564, "percentage": 38.5, "elapsed_time": "1 day, 4:42:51", "remaining_time": "1 day, 21:52:05"}
+{"current_steps": 848, "total_steps": 2200, "loss": 1.8777, "accuracy": 0.75, "lr": 7.0846240279714984e-06, "epoch": 1.5395095367847411, "percentage": 38.55, "elapsed_time": "1 day, 4:44:47", "remaining_time": "1 day, 21:49:54"}
+{"current_steps": 849, "total_steps": 2200, "loss": 2.0238, "accuracy": 0.765625, "lr": 7.078608027612035e-06, "epoch": 1.5413260672116258, "percentage": 38.59, "elapsed_time": "1 day, 4:47:07", "remaining_time": "1 day, 21:48:20"}
+{"current_steps": 850, "total_steps": 2200, "loss": 2.2203, "accuracy": 0.8125, "lr": 7.072588808195944e-06, "epoch": 1.5431425976385105, "percentage": 38.64, "elapsed_time": "1 day, 4:48:47", "remaining_time": "1 day, 21:45:43"}
+{"current_steps": 851, "total_steps": 2200, "loss": 1.7834, "accuracy": 0.78125, "lr": 7.066566381997457e-06, "epoch": 1.5449591280653951, "percentage": 38.68, "elapsed_time": "1 day, 4:51:09", "remaining_time": "1 day, 21:44:13"}
+{"current_steps": 852, "total_steps": 2200, "loss": 1.7877, "accuracy": 0.765625, "lr": 7.060540761297353e-06, "epoch": 1.5467756584922796, "percentage": 38.73, "elapsed_time": "1 day, 4:53:14", "remaining_time": "1 day, 21:42:15"}
+{"current_steps": 853, "total_steps": 2200, "loss": 2.1282, "accuracy": 0.75, "lr": 7.05451195838292e-06, "epoch": 1.5485921889191645, "percentage": 38.77, "elapsed_time": "1 day, 4:55:13", "remaining_time": "1 day, 21:40:08"}
+{"current_steps": 854, "total_steps": 2200, "loss": 1.8773, "accuracy": 0.796875, "lr": 7.048479985547941e-06, "epoch": 1.550408719346049, "percentage": 38.82, "elapsed_time": "1 day, 4:57:10", "remaining_time": "1 day, 21:37:58"}
+{"current_steps": 855, "total_steps": 2200, "loss": 2.144, "accuracy": 0.65625, "lr": 7.042444855092653e-06, "epoch": 1.5522252497729339, "percentage": 38.86, "elapsed_time": "1 day, 4:59:04", "remaining_time": "1 day, 21:35:44"}
+{"current_steps": 856, "total_steps": 2200, "loss": 1.8283, "accuracy": 0.8125, "lr": 7.0364065793237406e-06, "epoch": 1.5540417801998183, "percentage": 38.91, "elapsed_time": "1 day, 5:01:10", "remaining_time": "1 day, 21:33:48"}
+{"current_steps": 857, "total_steps": 2200, "loss": 2.1762, "accuracy": 0.703125, "lr": 7.0303651705542985e-06, "epoch": 1.555858310626703, "percentage": 38.95, "elapsed_time": "1 day, 5:03:25", "remaining_time": "1 day, 21:32:07"}
+{"current_steps": 858, "total_steps": 2200, "loss": 2.4176, "accuracy": 0.65625, "lr": 7.024320641103812e-06, "epoch": 1.5576748410535877, "percentage": 39.0, "elapsed_time": "1 day, 5:05:34", "remaining_time": "1 day, 21:30:15"}
+{"current_steps": 859, "total_steps": 2200, "loss": 2.0407, "accuracy": 0.703125, "lr": 7.018273003298129e-06, "epoch": 1.5594913714804723, "percentage": 39.05, "elapsed_time": "1 day, 5:07:55", "remaining_time": "1 day, 21:28:42"}
+{"current_steps": 860, "total_steps": 2200, "loss": 1.7125, "accuracy": 0.828125, "lr": 7.0122222694694356e-06, "epoch": 1.561307901907357, "percentage": 39.09, "elapsed_time": "1 day, 5:09:53", "remaining_time": "1 day, 21:26:34"}
+{"current_steps": 861, "total_steps": 2200, "loss": 2.0229, "accuracy": 0.75, "lr": 7.006168451956233e-06, "epoch": 1.5631244323342415, "percentage": 39.14, "elapsed_time": "1 day, 5:12:00", "remaining_time": "1 day, 21:24:39"}
+{"current_steps": 862, "total_steps": 2200, "loss": 1.9925, "accuracy": 0.78125, "lr": 7.0001115631033065e-06, "epoch": 1.5649409627611264, "percentage": 39.18, "elapsed_time": "1 day, 5:13:57", "remaining_time": "1 day, 21:22:29"}
+{"current_steps": 863, "total_steps": 2200, "loss": 1.8486, "accuracy": 0.828125, "lr": 6.994051615261711e-06, "epoch": 1.5667574931880108, "percentage": 39.23, "elapsed_time": "1 day, 5:16:06", "remaining_time": "1 day, 21:20:39"}
+{"current_steps": 864, "total_steps": 2200, "loss": 1.9414, "accuracy": 0.75, "lr": 6.987988620788731e-06, "epoch": 1.5685740236148955, "percentage": 39.27, "elapsed_time": "1 day, 5:18:10", "remaining_time": "1 day, 21:18:39"}
+{"current_steps": 865, "total_steps": 2200, "loss": 1.7377, "accuracy": 0.8125, "lr": 6.98192259204787e-06, "epoch": 1.5703905540417802, "percentage": 39.32, "elapsed_time": "1 day, 5:20:13", "remaining_time": "1 day, 21:16:39"}
+{"current_steps": 866, "total_steps": 2200, "loss": 2.3553, "accuracy": 0.765625, "lr": 6.9758535414088166e-06, "epoch": 1.5722070844686649, "percentage": 39.36, "elapsed_time": "1 day, 5:22:21", "remaining_time": "1 day, 21:14:45"}
+{"current_steps": 867, "total_steps": 2200, "loss": 2.0972, "accuracy": 0.78125, "lr": 6.969781481247423e-06, "epoch": 1.5740236148955495, "percentage": 39.41, "elapsed_time": "1 day, 5:24:07", "remaining_time": "1 day, 21:12:18"}
+{"current_steps": 868, "total_steps": 2200, "loss": 2.0153, "accuracy": 0.84375, "lr": 6.963706423945674e-06, "epoch": 1.575840145322434, "percentage": 39.45, "elapsed_time": "1 day, 5:26:21", "remaining_time": "1 day, 21:10:34"}
+{"current_steps": 869, "total_steps": 2200, "loss": 2.1137, "accuracy": 0.765625, "lr": 6.957628381891674e-06, "epoch": 1.577656675749319, "percentage": 39.5, "elapsed_time": "1 day, 5:28:11", "remaining_time": "1 day, 21:08:14"}
+{"current_steps": 870, "total_steps": 2200, "loss": 2.0419, "accuracy": 0.765625, "lr": 6.951547367479604e-06, "epoch": 1.5794732061762033, "percentage": 39.55, "elapsed_time": "1 day, 5:30:23", "remaining_time": "1 day, 21:06:28"}
+{"current_steps": 871, "total_steps": 2200, "loss": 2.0281, "accuracy": 0.734375, "lr": 6.945463393109712e-06, "epoch": 1.5812897366030882, "percentage": 39.59, "elapsed_time": "1 day, 5:32:37", "remaining_time": "1 day, 21:04:43"}
+{"current_steps": 872, "total_steps": 2200, "loss": 2.2254, "accuracy": 0.71875, "lr": 6.939376471188285e-06, "epoch": 1.5831062670299727, "percentage": 39.64, "elapsed_time": "1 day, 5:34:14", "remaining_time": "1 day, 21:02:02"}
+{"current_steps": 873, "total_steps": 2200, "loss": 2.0876, "accuracy": 0.75, "lr": 6.933286614127613e-06, "epoch": 1.5849227974568574, "percentage": 39.68, "elapsed_time": "1 day, 5:35:40", "remaining_time": "1 day, 20:59:06"}
+{"current_steps": 874, "total_steps": 2200, "loss": 2.0501, "accuracy": 0.8125, "lr": 6.927193834345976e-06, "epoch": 1.586739327883742, "percentage": 39.73, "elapsed_time": "1 day, 5:38:04", "remaining_time": "1 day, 20:57:37"}
+{"current_steps": 875, "total_steps": 2200, "loss": 1.9226, "accuracy": 0.765625, "lr": 6.921098144267613e-06, "epoch": 1.5885558583106267, "percentage": 39.77, "elapsed_time": "1 day, 5:40:15", "remaining_time": "1 day, 20:55:49"}
+{"current_steps": 876, "total_steps": 2200, "loss": 2.3308, "accuracy": 0.6875, "lr": 6.914999556322697e-06, "epoch": 1.5903723887375114, "percentage": 39.82, "elapsed_time": "1 day, 5:42:20", "remaining_time": "1 day, 20:53:51"}
+{"current_steps": 877, "total_steps": 2200, "loss": 2.297, "accuracy": 0.6875, "lr": 6.90889808294731e-06, "epoch": 1.5921889191643959, "percentage": 39.86, "elapsed_time": "1 day, 5:44:07", "remaining_time": "1 day, 20:51:27"}
+{"current_steps": 878, "total_steps": 2200, "loss": 1.9587, "accuracy": 0.78125, "lr": 6.9027937365834206e-06, "epoch": 1.5940054495912808, "percentage": 39.91, "elapsed_time": "1 day, 5:46:17", "remaining_time": "1 day, 20:49:37"}
+{"current_steps": 879, "total_steps": 2200, "loss": 2.0544, "accuracy": 0.78125, "lr": 6.896686529678853e-06, "epoch": 1.5958219800181652, "percentage": 39.95, "elapsed_time": "1 day, 5:48:09", "remaining_time": "1 day, 20:47:19"}
+{"current_steps": 880, "total_steps": 2200, "loss": 2.1884, "accuracy": 0.71875, "lr": 6.890576474687264e-06, "epoch": 1.5976385104450501, "percentage": 40.0, "elapsed_time": "1 day, 5:50:09", "remaining_time": "1 day, 20:45:14"}
+{"current_steps": 881, "total_steps": 2200, "loss": 1.8273, "accuracy": 0.78125, "lr": 6.884463584068121e-06, "epoch": 1.5994550408719346, "percentage": 40.05, "elapsed_time": "1 day, 5:52:08", "remaining_time": "1 day, 20:43:06"}
+{"current_steps": 882, "total_steps": 2200, "loss": 1.9555, "accuracy": 0.875, "lr": 6.878347870286673e-06, "epoch": 1.6012715712988193, "percentage": 40.09, "elapsed_time": "1 day, 5:53:56", "remaining_time": "1 day, 20:40:44"}
+{"current_steps": 883, "total_steps": 2200, "loss": 1.7596, "accuracy": 0.859375, "lr": 6.872229345813924e-06, "epoch": 1.603088101725704, "percentage": 40.14, "elapsed_time": "1 day, 5:55:31", "remaining_time": "1 day, 20:38:02"}
+{"current_steps": 884, "total_steps": 2200, "loss": 2.176, "accuracy": 0.8125, "lr": 6.8661080231266104e-06, "epoch": 1.6049046321525886, "percentage": 40.18, "elapsed_time": "1 day, 5:57:25", "remaining_time": "1 day, 20:35:47"}
+{"current_steps": 885, "total_steps": 2200, "loss": 2.0558, "accuracy": 0.6875, "lr": 6.859983914707177e-06, "epoch": 1.6067211625794733, "percentage": 40.23, "elapsed_time": "1 day, 5:59:32", "remaining_time": "1 day, 20:33:53"}
+{"current_steps": 886, "total_steps": 2200, "loss": 2.1291, "accuracy": 0.75, "lr": 6.853857033043745e-06, "epoch": 1.6085376930063577, "percentage": 40.27, "elapsed_time": "1 day, 6:01:54", "remaining_time": "1 day, 20:32:20"}
+{"current_steps": 887, "total_steps": 2200, "loss": 1.9655, "accuracy": 0.765625, "lr": 6.847727390630093e-06, "epoch": 1.6103542234332426, "percentage": 40.32, "elapsed_time": "1 day, 6:03:54", "remaining_time": "1 day, 20:30:15"}
+{"current_steps": 888, "total_steps": 2200, "loss": 1.9864, "accuracy": 0.8125, "lr": 6.84159499996563e-06, "epoch": 1.612170753860127, "percentage": 40.36, "elapsed_time": "1 day, 6:05:48", "remaining_time": "1 day, 20:28:02"}
+{"current_steps": 889, "total_steps": 2200, "loss": 1.7634, "accuracy": 0.796875, "lr": 6.83545987355537e-06, "epoch": 1.6139872842870118, "percentage": 40.41, "elapsed_time": "1 day, 6:07:38", "remaining_time": "1 day, 20:25:42"}
+{"current_steps": 890, "total_steps": 2200, "loss": 2.3534, "accuracy": 0.703125, "lr": 6.829322023909901e-06, "epoch": 1.6158038147138964, "percentage": 40.45, "elapsed_time": "1 day, 6:10:04", "remaining_time": "1 day, 20:24:16"}
+{"current_steps": 891, "total_steps": 2200, "loss": 1.9453, "accuracy": 0.765625, "lr": 6.823181463545368e-06, "epoch": 1.6176203451407811, "percentage": 40.5, "elapsed_time": "1 day, 6:12:13", "remaining_time": "1 day, 20:22:24"}
+{"current_steps": 892, "total_steps": 2200, "loss": 1.9703, "accuracy": 0.765625, "lr": 6.817038204983444e-06, "epoch": 1.6194368755676658, "percentage": 40.55, "elapsed_time": "1 day, 6:14:08", "remaining_time": "1 day, 20:20:12"}
+{"current_steps": 893, "total_steps": 2200, "loss": 1.9296, "accuracy": 0.703125, "lr": 6.810892260751302e-06, "epoch": 1.6212534059945503, "percentage": 40.59, "elapsed_time": "1 day, 6:16:01", "remaining_time": "1 day, 20:17:57"}
+{"current_steps": 894, "total_steps": 2200, "loss": 1.857, "accuracy": 0.765625, "lr": 6.804743643381592e-06, "epoch": 1.6230699364214352, "percentage": 40.64, "elapsed_time": "1 day, 6:18:07", "remaining_time": "1 day, 20:16:00"}
+{"current_steps": 895, "total_steps": 2200, "loss": 1.8795, "accuracy": 0.734375, "lr": 6.798592365412415e-06, "epoch": 1.6248864668483196, "percentage": 40.68, "elapsed_time": "1 day, 6:20:16", "remaining_time": "1 day, 20:14:09"}
+{"current_steps": 896, "total_steps": 2200, "loss": 2.4462, "accuracy": 0.71875, "lr": 6.792438439387299e-06, "epoch": 1.6267029972752045, "percentage": 40.73, "elapsed_time": "1 day, 6:22:25", "remaining_time": "1 day, 20:12:16"}
+{"current_steps": 897, "total_steps": 2200, "loss": 2.0741, "accuracy": 0.75, "lr": 6.7862818778551695e-06, "epoch": 1.628519527702089, "percentage": 40.77, "elapsed_time": "1 day, 6:24:21", "remaining_time": "1 day, 20:10:06"}
+{"current_steps": 898, "total_steps": 2200, "loss": 2.1919, "accuracy": 0.78125, "lr": 6.780122693370329e-06, "epoch": 1.6303360581289736, "percentage": 40.82, "elapsed_time": "1 day, 6:26:11", "remaining_time": "1 day, 20:07:46"}
+{"current_steps": 899, "total_steps": 2200, "loss": 1.8887, "accuracy": 0.75, "lr": 6.773960898492428e-06, "epoch": 1.6321525885558583, "percentage": 40.86, "elapsed_time": "1 day, 6:28:09", "remaining_time": "1 day, 20:05:38"}
+{"current_steps": 900, "total_steps": 2200, "loss": 2.2367, "accuracy": 0.734375, "lr": 6.767796505786435e-06, "epoch": 1.633969118982743, "percentage": 40.91, "elapsed_time": "1 day, 6:30:33", "remaining_time": "1 day, 20:04:08"}
+{"current_steps": 901, "total_steps": 2200, "loss": 2.0995, "accuracy": 0.734375, "lr": 6.761629527822622e-06, "epoch": 1.6357856494096277, "percentage": 40.95, "elapsed_time": "1 day, 6:32:27", "remaining_time": "1 day, 20:01:54"}
+{"current_steps": 902, "total_steps": 2200, "loss": 1.8919, "accuracy": 0.828125, "lr": 6.755459977176533e-06, "epoch": 1.6376021798365121, "percentage": 41.0, "elapsed_time": "1 day, 6:34:52", "remaining_time": "1 day, 20:00:25"}
+{"current_steps": 903, "total_steps": 2200, "loss": 1.7682, "accuracy": 0.828125, "lr": 6.749287866428953e-06, "epoch": 1.639418710263397, "percentage": 41.05, "elapsed_time": "1 day, 6:36:43", "remaining_time": "1 day, 19:58:08"}
+{"current_steps": 904, "total_steps": 2200, "loss": 2.5709, "accuracy": 0.65625, "lr": 6.743113208165893e-06, "epoch": 1.6412352406902815, "percentage": 41.09, "elapsed_time": "1 day, 6:38:51", "remaining_time": "1 day, 19:56:14"}
+{"current_steps": 905, "total_steps": 2200, "loss": 1.9292, "accuracy": 0.796875, "lr": 6.736936014978552e-06, "epoch": 1.6430517711171662, "percentage": 41.14, "elapsed_time": "1 day, 6:41:05", "remaining_time": "1 day, 19:54:29"}
+{"current_steps": 906, "total_steps": 2200, "loss": 2.2204, "accuracy": 0.703125, "lr": 6.730756299463308e-06, "epoch": 1.6448683015440508, "percentage": 41.18, "elapsed_time": "1 day, 6:42:39", "remaining_time": "1 day, 19:51:46"}
+{"current_steps": 907, "total_steps": 2200, "loss": 1.9456, "accuracy": 0.78125, "lr": 6.724574074221673e-06, "epoch": 1.6466848319709355, "percentage": 41.23, "elapsed_time": "1 day, 6:44:42", "remaining_time": "1 day, 19:49:47"}
+{"current_steps": 908, "total_steps": 2200, "loss": 1.7985, "accuracy": 0.84375, "lr": 6.718389351860285e-06, "epoch": 1.6485013623978202, "percentage": 41.27, "elapsed_time": "1 day, 6:47:04", "remaining_time": "1 day, 19:48:12"}
+{"current_steps": 909, "total_steps": 2200, "loss": 2.1353, "accuracy": 0.765625, "lr": 6.712202144990863e-06, "epoch": 1.6503178928247049, "percentage": 41.32, "elapsed_time": "1 day, 6:49:00", "remaining_time": "1 day, 19:46:02"}
+{"current_steps": 910, "total_steps": 2200, "loss": 2.0659, "accuracy": 0.8125, "lr": 6.7060124662302066e-06, "epoch": 1.6521344232515895, "percentage": 41.36, "elapsed_time": "1 day, 6:50:42", "remaining_time": "1 day, 19:43:32"}
+{"current_steps": 911, "total_steps": 2200, "loss": 2.0403, "accuracy": 0.765625, "lr": 6.699820328200143e-06, "epoch": 1.653950953678474, "percentage": 41.41, "elapsed_time": "1 day, 6:52:43", "remaining_time": "1 day, 19:41:27"}
+{"current_steps": 912, "total_steps": 2200, "loss": 2.091, "accuracy": 0.734375, "lr": 6.69362574352752e-06, "epoch": 1.655767484105359, "percentage": 41.45, "elapsed_time": "1 day, 6:54:30", "remaining_time": "1 day, 19:39:05"}
+{"current_steps": 913, "total_steps": 2200, "loss": 2.0593, "accuracy": 0.796875, "lr": 6.6874287248441794e-06, "epoch": 1.6575840145322434, "percentage": 41.5, "elapsed_time": "1 day, 6:56:45", "remaining_time": "1 day, 19:37:21"}
+{"current_steps": 914, "total_steps": 2200, "loss": 2.2115, "accuracy": 0.75, "lr": 6.681229284786915e-06, "epoch": 1.659400544959128, "percentage": 41.55, "elapsed_time": "1 day, 6:58:57", "remaining_time": "1 day, 19:35:33"}
+{"current_steps": 915, "total_steps": 2200, "loss": 1.988, "accuracy": 0.734375, "lr": 6.675027435997469e-06, "epoch": 1.6612170753860127, "percentage": 41.59, "elapsed_time": "1 day, 7:00:35", "remaining_time": "1 day, 19:32:57"}
+{"current_steps": 916, "total_steps": 2200, "loss": 1.9141, "accuracy": 0.8125, "lr": 6.668823191122487e-06, "epoch": 1.6630336058128974, "percentage": 41.64, "elapsed_time": "1 day, 7:02:43", "remaining_time": "1 day, 19:31:04"}
+{"current_steps": 917, "total_steps": 2200, "loss": 1.7811, "accuracy": 0.8125, "lr": 6.662616562813512e-06, "epoch": 1.664850136239782, "percentage": 41.68, "elapsed_time": "1 day, 7:04:48", "remaining_time": "1 day, 19:29:05"}
+{"current_steps": 918, "total_steps": 2200, "loss": 1.9802, "accuracy": 0.75, "lr": 6.6564075637269295e-06, "epoch": 1.6666666666666665, "percentage": 41.73, "elapsed_time": "1 day, 7:06:38", "remaining_time": "1 day, 19:26:47"}
+{"current_steps": 919, "total_steps": 2200, "loss": 2.0028, "accuracy": 0.78125, "lr": 6.650196206523977e-06, "epoch": 1.6684831970935514, "percentage": 41.77, "elapsed_time": "1 day, 7:08:41", "remaining_time": "1 day, 19:24:47"}
+{"current_steps": 920, "total_steps": 2200, "loss": 2.0059, "accuracy": 0.796875, "lr": 6.643982503870693e-06, "epoch": 1.6702997275204359, "percentage": 41.82, "elapsed_time": "1 day, 7:10:50", "remaining_time": "1 day, 19:22:54"}
+{"current_steps": 921, "total_steps": 2200, "loss": 1.9826, "accuracy": 0.75, "lr": 6.637766468437894e-06, "epoch": 1.6721162579473208, "percentage": 41.86, "elapsed_time": "1 day, 7:12:34", "remaining_time": "1 day, 19:20:27"}
+{"current_steps": 922, "total_steps": 2200, "loss": 1.9747, "accuracy": 0.765625, "lr": 6.631548112901164e-06, "epoch": 1.6739327883742052, "percentage": 41.91, "elapsed_time": "1 day, 7:14:24", "remaining_time": "1 day, 19:18:08"}
+{"current_steps": 923, "total_steps": 2200, "loss": 1.7486, "accuracy": 0.828125, "lr": 6.6253274499408095e-06, "epoch": 1.67574931880109, "percentage": 41.95, "elapsed_time": "1 day, 7:16:41", "remaining_time": "1 day, 19:16:27"}
+{"current_steps": 924, "total_steps": 2200, "loss": 1.7618, "accuracy": 0.734375, "lr": 6.619104492241848e-06, "epoch": 1.6775658492279746, "percentage": 42.0, "elapsed_time": "1 day, 7:18:40", "remaining_time": "1 day, 19:14:21"}
+{"current_steps": 925, "total_steps": 2200, "loss": 2.1665, "accuracy": 0.78125, "lr": 6.61287925249397e-06, "epoch": 1.6793823796548593, "percentage": 42.05, "elapsed_time": "1 day, 7:20:44", "remaining_time": "1 day, 19:12:22"}
+{"current_steps": 926, "total_steps": 2200, "loss": 2.0174, "accuracy": 0.78125, "lr": 6.606651743391528e-06, "epoch": 1.681198910081744, "percentage": 42.09, "elapsed_time": "1 day, 7:23:01", "remaining_time": "1 day, 19:10:41"}
+{"current_steps": 927, "total_steps": 2200, "loss": 2.0636, "accuracy": 0.78125, "lr": 6.6004219776334925e-06, "epoch": 1.6830154405086284, "percentage": 42.14, "elapsed_time": "1 day, 7:25:21", "remaining_time": "1 day, 19:09:03"}
+{"current_steps": 928, "total_steps": 2200, "loss": 1.8203, "accuracy": 0.84375, "lr": 6.594189967923445e-06, "epoch": 1.6848319709355133, "percentage": 42.18, "elapsed_time": "1 day, 7:27:14", "remaining_time": "1 day, 19:06:49"}
+{"current_steps": 929, "total_steps": 2200, "loss": 2.5874, "accuracy": 0.671875, "lr": 6.587955726969533e-06, "epoch": 1.6866485013623977, "percentage": 42.23, "elapsed_time": "1 day, 7:29:33", "remaining_time": "1 day, 19:05:11"}
+{"current_steps": 930, "total_steps": 2200, "loss": 1.7334, "accuracy": 0.828125, "lr": 6.581719267484466e-06, "epoch": 1.6884650317892824, "percentage": 42.27, "elapsed_time": "1 day, 7:31:24", "remaining_time": "1 day, 19:02:53"}
+{"current_steps": 931, "total_steps": 2200, "loss": 2.2827, "accuracy": 0.71875, "lr": 6.575480602185468e-06, "epoch": 1.690281562216167, "percentage": 42.32, "elapsed_time": "1 day, 7:33:39", "remaining_time": "1 day, 19:01:09"}
+{"current_steps": 932, "total_steps": 2200, "loss": 2.1175, "accuracy": 0.75, "lr": 6.56923974379426e-06, "epoch": 1.6920980926430518, "percentage": 42.36, "elapsed_time": "1 day, 7:35:50", "remaining_time": "1 day, 18:59:19"}
+{"current_steps": 933, "total_steps": 2200, "loss": 2.384, "accuracy": 0.734375, "lr": 6.562996705037046e-06, "epoch": 1.6939146230699365, "percentage": 42.41, "elapsed_time": "1 day, 7:37:50", "remaining_time": "1 day, 18:57:13"}
+{"current_steps": 934, "total_steps": 2200, "loss": 2.0037, "accuracy": 0.75, "lr": 6.556751498644461e-06, "epoch": 1.695731153496821, "percentage": 42.45, "elapsed_time": "1 day, 7:39:33", "remaining_time": "1 day, 18:54:46"}
+{"current_steps": 935, "total_steps": 2200, "loss": 2.3693, "accuracy": 0.671875, "lr": 6.550504137351576e-06, "epoch": 1.6975476839237058, "percentage": 42.5, "elapsed_time": "1 day, 7:41:55", "remaining_time": "1 day, 18:53:10"}
+{"current_steps": 936, "total_steps": 2200, "loss": 2.02, "accuracy": 0.71875, "lr": 6.544254633897843e-06, "epoch": 1.6993642143505903, "percentage": 42.55, "elapsed_time": "1 day, 7:44:13", "remaining_time": "1 day, 18:51:30"}
+{"current_steps": 937, "total_steps": 2200, "loss": 1.8375, "accuracy": 0.78125, "lr": 6.538003001027089e-06, "epoch": 1.7011807447774752, "percentage": 42.59, "elapsed_time": "1 day, 7:45:59", "remaining_time": "1 day, 18:49:07"}
+{"current_steps": 938, "total_steps": 2200, "loss": 2.6753, "accuracy": 0.609375, "lr": 6.531749251487482e-06, "epoch": 1.7029972752043596, "percentage": 42.64, "elapsed_time": "1 day, 7:48:10", "remaining_time": "1 day, 18:47:16"}
+{"current_steps": 939, "total_steps": 2200, "loss": 2.2164, "accuracy": 0.6875, "lr": 6.525493398031508e-06, "epoch": 1.7048138056312443, "percentage": 42.68, "elapsed_time": "1 day, 7:50:07", "remaining_time": "1 day, 18:45:08"}
+{"current_steps": 940, "total_steps": 2200, "loss": 1.7358, "accuracy": 0.828125, "lr": 6.5192354534159406e-06, "epoch": 1.706630336058129, "percentage": 42.73, "elapsed_time": "1 day, 7:52:18", "remaining_time": "1 day, 18:43:18"}
+{"current_steps": 941, "total_steps": 2200, "loss": 2.3819, "accuracy": 0.703125, "lr": 6.5129754304018165e-06, "epoch": 1.7084468664850136, "percentage": 42.77, "elapsed_time": "1 day, 7:54:36", "remaining_time": "1 day, 18:41:37"}
+{"current_steps": 942, "total_steps": 2200, "loss": 1.7958, "accuracy": 0.765625, "lr": 6.506713341754415e-06, "epoch": 1.7102633969118983, "percentage": 42.82, "elapsed_time": "1 day, 7:57:01", "remaining_time": "1 day, 18:40:06"}
+{"current_steps": 943, "total_steps": 2200, "loss": 2.1943, "accuracy": 0.734375, "lr": 6.500449200243224e-06, "epoch": 1.7120799273387828, "percentage": 42.86, "elapsed_time": "1 day, 7:58:56", "remaining_time": "1 day, 18:37:54"}
+{"current_steps": 944, "total_steps": 2200, "loss": 1.9642, "accuracy": 0.78125, "lr": 6.494183018641919e-06, "epoch": 1.7138964577656677, "percentage": 42.91, "elapsed_time": "1 day, 8:00:25", "remaining_time": "1 day, 18:35:08"}
+{"current_steps": 945, "total_steps": 2200, "loss": 2.102, "accuracy": 0.703125, "lr": 6.487914809728335e-06, "epoch": 1.7157129881925521, "percentage": 42.95, "elapsed_time": "1 day, 8:02:54", "remaining_time": "1 day, 18:33:42"}
+{"current_steps": 946, "total_steps": 2200, "loss": 1.8887, "accuracy": 0.78125, "lr": 6.481644586284443e-06, "epoch": 1.717529518619437, "percentage": 43.0, "elapsed_time": "1 day, 8:04:48", "remaining_time": "1 day, 18:31:29"}
+{"current_steps": 947, "total_steps": 2200, "loss": 2.1631, "accuracy": 0.703125, "lr": 6.475372361096316e-06, "epoch": 1.7193460490463215, "percentage": 43.05, "elapsed_time": "1 day, 8:06:41", "remaining_time": "1 day, 18:29:15"}
+{"current_steps": 948, "total_steps": 2200, "loss": 1.6936, "accuracy": 0.859375, "lr": 6.469098146954119e-06, "epoch": 1.7211625794732062, "percentage": 43.09, "elapsed_time": "1 day, 8:08:58", "remaining_time": "1 day, 18:27:33"}
+{"current_steps": 949, "total_steps": 2200, "loss": 2.1523, "accuracy": 0.796875, "lr": 6.462821956652064e-06, "epoch": 1.7229791099000908, "percentage": 43.14, "elapsed_time": "1 day, 8:11:08", "remaining_time": "1 day, 18:25:41"}
+{"current_steps": 950, "total_steps": 2200, "loss": 1.9428, "accuracy": 0.765625, "lr": 6.456543802988395e-06, "epoch": 1.7247956403269755, "percentage": 43.18, "elapsed_time": "1 day, 8:12:50", "remaining_time": "1 day, 18:23:12"}
+{"current_steps": 951, "total_steps": 2200, "loss": 1.7506, "accuracy": 0.828125, "lr": 6.450263698765363e-06, "epoch": 1.7266121707538602, "percentage": 43.23, "elapsed_time": "1 day, 8:14:56", "remaining_time": "1 day, 18:21:15"}
+{"current_steps": 952, "total_steps": 2200, "loss": 1.6958, "accuracy": 0.859375, "lr": 6.4439816567891925e-06, "epoch": 1.7284287011807447, "percentage": 43.27, "elapsed_time": "1 day, 8:17:05", "remaining_time": "1 day, 18:19:22"}
+{"current_steps": 953, "total_steps": 2200, "loss": 1.9554, "accuracy": 0.796875, "lr": 6.43769768987006e-06, "epoch": 1.7302452316076296, "percentage": 43.32, "elapsed_time": "1 day, 8:18:57", "remaining_time": "1 day, 18:17:07"}
+{"current_steps": 954, "total_steps": 2200, "loss": 2.1501, "accuracy": 0.75, "lr": 6.431411810822069e-06, "epoch": 1.732061762034514, "percentage": 43.36, "elapsed_time": "1 day, 8:20:54", "remaining_time": "1 day, 18:14:58"}
+{"current_steps": 955, "total_steps": 2200, "loss": 2.1222, "accuracy": 0.734375, "lr": 6.425124032463223e-06, "epoch": 1.7338782924613987, "percentage": 43.41, "elapsed_time": "1 day, 8:22:37", "remaining_time": "1 day, 18:12:31"}
+{"current_steps": 956, "total_steps": 2200, "loss": 2.1891, "accuracy": 0.75, "lr": 6.418834367615395e-06, "epoch": 1.7356948228882834, "percentage": 43.45, "elapsed_time": "1 day, 8:24:30", "remaining_time": "1 day, 18:10:18"}
+{"current_steps": 957, "total_steps": 2200, "loss": 2.0766, "accuracy": 0.8125, "lr": 6.4125428291043064e-06, "epoch": 1.737511353315168, "percentage": 43.5, "elapsed_time": "1 day, 8:26:43", "remaining_time": "1 day, 18:08:30"}
+{"current_steps": 958, "total_steps": 2200, "loss": 2.0315, "accuracy": 0.796875, "lr": 6.4062494297595025e-06, "epoch": 1.7393278837420527, "percentage": 43.55, "elapsed_time": "1 day, 8:29:03", "remaining_time": "1 day, 18:06:51"}
+{"current_steps": 959, "total_steps": 2200, "loss": 1.9766, "accuracy": 0.8125, "lr": 6.399954182414317e-06, "epoch": 1.7411444141689372, "percentage": 43.59, "elapsed_time": "1 day, 8:31:02", "remaining_time": "1 day, 18:04:45"}
+{"current_steps": 960, "total_steps": 2200, "loss": 2.6866, "accuracy": 0.6875, "lr": 6.393657099905854e-06, "epoch": 1.742960944595822, "percentage": 43.64, "elapsed_time": "1 day, 8:33:51", "remaining_time": "1 day, 18:03:43"}
+{"current_steps": 961, "total_steps": 2200, "loss": 2.0836, "accuracy": 0.75, "lr": 6.387358195074968e-06, "epoch": 1.7447774750227065, "percentage": 43.68, "elapsed_time": "1 day, 8:35:49", "remaining_time": "1 day, 18:01:36"}
+{"current_steps": 962, "total_steps": 2200, "loss": 1.9472, "accuracy": 0.75, "lr": 6.381057480766217e-06, "epoch": 1.7465940054495914, "percentage": 43.73, "elapsed_time": "1 day, 8:37:58", "remaining_time": "1 day, 17:59:43"}
+{"current_steps": 963, "total_steps": 2200, "loss": 2.0027, "accuracy": 0.703125, "lr": 6.374754969827855e-06, "epoch": 1.7484105358764759, "percentage": 43.77, "elapsed_time": "1 day, 8:39:57", "remaining_time": "1 day, 17:57:37"}
+{"current_steps": 964, "total_steps": 2200, "loss": 2.0086, "accuracy": 0.78125, "lr": 6.3684506751118016e-06, "epoch": 1.7502270663033606, "percentage": 43.82, "elapsed_time": "1 day, 8:42:16", "remaining_time": "1 day, 17:55:56"}
+{"current_steps": 965, "total_steps": 2200, "loss": 2.3424, "accuracy": 0.65625, "lr": 6.362144609473609e-06, "epoch": 1.7520435967302452, "percentage": 43.86, "elapsed_time": "1 day, 8:44:18", "remaining_time": "1 day, 17:53:54"}
+{"current_steps": 966, "total_steps": 2200, "loss": 1.925, "accuracy": 0.84375, "lr": 6.355836785772444e-06, "epoch": 1.75386012715713, "percentage": 43.91, "elapsed_time": "1 day, 8:46:20", "remaining_time": "1 day, 17:51:52"}
+{"current_steps": 967, "total_steps": 2200, "loss": 1.8814, "accuracy": 0.8125, "lr": 6.349527216871059e-06, "epoch": 1.7556766575840146, "percentage": 43.95, "elapsed_time": "1 day, 8:48:12", "remaining_time": "1 day, 17:49:36"}
+{"current_steps": 968, "total_steps": 2200, "loss": 1.7882, "accuracy": 0.734375, "lr": 6.343215915635762e-06, "epoch": 1.757493188010899, "percentage": 44.0, "elapsed_time": "1 day, 8:49:49", "remaining_time": "1 day, 17:47:03"}
+{"current_steps": 969, "total_steps": 2200, "loss": 1.9777, "accuracy": 0.828125, "lr": 6.336902894936394e-06, "epoch": 1.759309718437784, "percentage": 44.05, "elapsed_time": "1 day, 8:52:00", "remaining_time": "1 day, 17:45:11"}
+{"current_steps": 970, "total_steps": 2200, "loss": 2.353, "accuracy": 0.71875, "lr": 6.330588167646305e-06, "epoch": 1.7611262488646684, "percentage": 44.09, "elapsed_time": "1 day, 8:53:52", "remaining_time": "1 day, 17:42:57"}
+{"current_steps": 971, "total_steps": 2200, "loss": 1.7381, "accuracy": 0.8125, "lr": 6.324271746642323e-06, "epoch": 1.7629427792915533, "percentage": 44.14, "elapsed_time": "1 day, 8:56:03", "remaining_time": "1 day, 17:41:06"}
+{"current_steps": 972, "total_steps": 2200, "loss": 2.2215, "accuracy": 0.71875, "lr": 6.317953644804731e-06, "epoch": 1.7647593097184378, "percentage": 44.18, "elapsed_time": "1 day, 8:58:19", "remaining_time": "1 day, 17:39:21"}
+{"current_steps": 973, "total_steps": 2200, "loss": 2.1095, "accuracy": 0.75, "lr": 6.311633875017237e-06, "epoch": 1.7665758401453224, "percentage": 44.23, "elapsed_time": "1 day, 9:00:14", "remaining_time": "1 day, 17:37:11"}
+{"current_steps": 974, "total_steps": 2200, "loss": 1.9114, "accuracy": 0.765625, "lr": 6.305312450166952e-06, "epoch": 1.768392370572207, "percentage": 44.27, "elapsed_time": "1 day, 9:02:02", "remaining_time": "1 day, 17:34:50"}
+{"current_steps": 975, "total_steps": 2200, "loss": 2.3069, "accuracy": 0.765625, "lr": 6.298989383144366e-06, "epoch": 1.7702089009990918, "percentage": 44.32, "elapsed_time": "1 day, 9:04:03", "remaining_time": "1 day, 17:32:46"}
+{"current_steps": 976, "total_steps": 2200, "loss": 2.0616, "accuracy": 0.75, "lr": 6.292664686843308e-06, "epoch": 1.7720254314259765, "percentage": 44.36, "elapsed_time": "1 day, 9:06:18", "remaining_time": "1 day, 17:31:01"}
+{"current_steps": 977, "total_steps": 2200, "loss": 2.0672, "accuracy": 0.6875, "lr": 6.286338374160939e-06, "epoch": 1.773841961852861, "percentage": 44.41, "elapsed_time": "1 day, 9:08:19", "remaining_time": "1 day, 17:28:57"}
+{"current_steps": 978, "total_steps": 2200, "loss": 1.7177, "accuracy": 0.84375, "lr": 6.280010457997711e-06, "epoch": 1.7756584922797458, "percentage": 44.45, "elapsed_time": "1 day, 9:10:34", "remaining_time": "1 day, 17:27:11"}
+{"current_steps": 979, "total_steps": 2200, "loss": 1.9793, "accuracy": 0.71875, "lr": 6.273680951257343e-06, "epoch": 1.7774750227066303, "percentage": 44.5, "elapsed_time": "1 day, 9:12:41", "remaining_time": "1 day, 17:25:15"}
+{"current_steps": 980, "total_steps": 2200, "loss": 1.9557, "accuracy": 0.84375, "lr": 6.267349866846808e-06, "epoch": 1.779291553133515, "percentage": 44.55, "elapsed_time": "1 day, 9:14:25", "remaining_time": "1 day, 17:22:50"}
+{"current_steps": 981, "total_steps": 2200, "loss": 1.6879, "accuracy": 0.859375, "lr": 6.2610172176762855e-06, "epoch": 1.7811080835603996, "percentage": 44.59, "elapsed_time": "1 day, 9:16:19", "remaining_time": "1 day, 17:20:39"}
+{"current_steps": 982, "total_steps": 2200, "loss": 1.9226, "accuracy": 0.78125, "lr": 6.25468301665915e-06, "epoch": 1.7829246139872843, "percentage": 44.64, "elapsed_time": "1 day, 9:18:21", "remaining_time": "1 day, 17:18:36"}
+{"current_steps": 983, "total_steps": 2200, "loss": 2.334, "accuracy": 0.71875, "lr": 6.248347276711941e-06, "epoch": 1.784741144414169, "percentage": 44.68, "elapsed_time": "1 day, 9:20:40", "remaining_time": "1 day, 17:16:56"}
+{"current_steps": 984, "total_steps": 2200, "loss": 1.8675, "accuracy": 0.75, "lr": 6.242010010754336e-06, "epoch": 1.7865576748410534, "percentage": 44.73, "elapsed_time": "1 day, 9:22:42", "remaining_time": "1 day, 17:14:53"}
+{"current_steps": 985, "total_steps": 2200, "loss": 1.8697, "accuracy": 0.796875, "lr": 6.2356712317091205e-06, "epoch": 1.7883742052679383, "percentage": 44.77, "elapsed_time": "1 day, 9:24:31", "remaining_time": "1 day, 17:12:34"}
+{"current_steps": 986, "total_steps": 2200, "loss": 1.7414, "accuracy": 0.859375, "lr": 6.229330952502174e-06, "epoch": 1.7901907356948228, "percentage": 44.82, "elapsed_time": "1 day, 9:26:21", "remaining_time": "1 day, 17:10:17"}
+{"current_steps": 987, "total_steps": 2200, "loss": 2.1574, "accuracy": 0.8125, "lr": 6.222989186062426e-06, "epoch": 1.7920072661217077, "percentage": 44.86, "elapsed_time": "1 day, 9:28:34", "remaining_time": "1 day, 17:08:29"}
+{"current_steps": 988, "total_steps": 2200, "loss": 2.388, "accuracy": 0.6875, "lr": 6.216645945321841e-06, "epoch": 1.7938237965485921, "percentage": 44.91, "elapsed_time": "1 day, 9:30:32", "remaining_time": "1 day, 17:06:22"}
+{"current_steps": 989, "total_steps": 2200, "loss": 2.0893, "accuracy": 0.78125, "lr": 6.210301243215396e-06, "epoch": 1.7956403269754768, "percentage": 44.95, "elapsed_time": "1 day, 9:32:38", "remaining_time": "1 day, 17:04:24"}
+{"current_steps": 990, "total_steps": 2200, "loss": 2.0696, "accuracy": 0.75, "lr": 6.20395509268104e-06, "epoch": 1.7974568574023615, "percentage": 45.0, "elapsed_time": "1 day, 9:34:59", "remaining_time": "1 day, 17:02:46"}
+{"current_steps": 991, "total_steps": 2200, "loss": 2.2475, "accuracy": 0.75, "lr": 6.197607506659678e-06, "epoch": 1.7992733878292462, "percentage": 45.05, "elapsed_time": "1 day, 9:37:10", "remaining_time": "1 day, 17:00:54"}
+{"current_steps": 992, "total_steps": 2200, "loss": 2.046, "accuracy": 0.65625, "lr": 6.191258498095148e-06, "epoch": 1.8010899182561309, "percentage": 45.09, "elapsed_time": "1 day, 9:39:23", "remaining_time": "1 day, 16:59:06"}
+{"current_steps": 993, "total_steps": 2200, "loss": 1.7434, "accuracy": 0.859375, "lr": 6.184908079934181e-06, "epoch": 1.8029064486830153, "percentage": 45.14, "elapsed_time": "1 day, 9:40:46", "remaining_time": "1 day, 16:56:15"}
+{"current_steps": 994, "total_steps": 2200, "loss": 2.1035, "accuracy": 0.71875, "lr": 6.178556265126383e-06, "epoch": 1.8047229791099002, "percentage": 45.18, "elapsed_time": "1 day, 9:42:39", "remaining_time": "1 day, 16:54:02"}
+{"current_steps": 995, "total_steps": 2200, "loss": 1.8209, "accuracy": 0.8125, "lr": 6.172203066624213e-06, "epoch": 1.8065395095367847, "percentage": 45.23, "elapsed_time": "1 day, 9:44:54", "remaining_time": "1 day, 16:52:16"}
+{"current_steps": 996, "total_steps": 2200, "loss": 1.8418, "accuracy": 0.71875, "lr": 6.16584849738295e-06, "epoch": 1.8083560399636693, "percentage": 45.27, "elapsed_time": "1 day, 9:47:06", "remaining_time": "1 day, 16:50:26"}
+{"current_steps": 997, "total_steps": 2200, "loss": 2.3627, "accuracy": 0.734375, "lr": 6.159492570360664e-06, "epoch": 1.810172570390554, "percentage": 45.32, "elapsed_time": "1 day, 9:49:22", "remaining_time": "1 day, 16:48:41"}
+{"current_steps": 998, "total_steps": 2200, "loss": 1.9746, "accuracy": 0.75, "lr": 6.1531352985182e-06, "epoch": 1.8119891008174387, "percentage": 45.36, "elapsed_time": "1 day, 9:51:39", "remaining_time": "1 day, 16:46:56"}
+{"current_steps": 999, "total_steps": 2200, "loss": 1.881, "accuracy": 0.796875, "lr": 6.146776694819139e-06, "epoch": 1.8138056312443234, "percentage": 45.41, "elapsed_time": "1 day, 9:53:59", "remaining_time": "1 day, 16:45:15"}
+{"current_steps": 1000, "total_steps": 2200, "loss": 1.8357, "accuracy": 0.78125, "lr": 6.140416772229785e-06, "epoch": 1.815622161671208, "percentage": 45.45, "elapsed_time": "1 day, 9:56:02", "remaining_time": "1 day, 16:43:14"}
+{"current_steps": 1001, "total_steps": 2200, "loss": 1.9502, "accuracy": 0.78125, "lr": 6.134055543719122e-06, "epoch": 1.8174386920980927, "percentage": 45.5, "elapsed_time": "1 day, 9:57:59", "remaining_time": "1 day, 16:41:07"}
+{"current_steps": 1002, "total_steps": 2200, "loss": 2.1536, "accuracy": 0.75, "lr": 6.1276930222588085e-06, "epoch": 1.8192552225249772, "percentage": 45.55, "elapsed_time": "1 day, 10:00:14", "remaining_time": "1 day, 16:39:20"}
+{"current_steps": 1003, "total_steps": 2200, "loss": 1.7472, "accuracy": 0.796875, "lr": 6.12132922082313e-06, "epoch": 1.821071752951862, "percentage": 45.59, "elapsed_time": "1 day, 10:02:03", "remaining_time": "1 day, 16:37:02"}
+{"current_steps": 1004, "total_steps": 2200, "loss": 1.9234, "accuracy": 0.75, "lr": 6.114964152388986e-06, "epoch": 1.8228882833787465, "percentage": 45.64, "elapsed_time": "1 day, 10:04:11", "remaining_time": "1 day, 16:35:07"}
+{"current_steps": 1005, "total_steps": 2200, "loss": 1.9861, "accuracy": 0.75, "lr": 6.10859782993586e-06, "epoch": 1.8247048138056312, "percentage": 45.68, "elapsed_time": "1 day, 10:06:09", "remaining_time": "1 day, 16:32:59"}
+{"current_steps": 1006, "total_steps": 2200, "loss": 1.828, "accuracy": 0.796875, "lr": 6.1022302664457935e-06, "epoch": 1.8265213442325159, "percentage": 45.73, "elapsed_time": "1 day, 10:07:58", "remaining_time": "1 day, 16:30:42"}
+{"current_steps": 1007, "total_steps": 2200, "loss": 2.4612, "accuracy": 0.703125, "lr": 6.095861474903354e-06, "epoch": 1.8283378746594006, "percentage": 45.77, "elapsed_time": "1 day, 10:10:22", "remaining_time": "1 day, 16:29:05"}
+{"current_steps": 1008, "total_steps": 2200, "loss": 2.119, "accuracy": 0.796875, "lr": 6.089491468295619e-06, "epoch": 1.8301544050862852, "percentage": 45.82, "elapsed_time": "1 day, 10:12:29", "remaining_time": "1 day, 16:27:08"}
+{"current_steps": 1009, "total_steps": 2200, "loss": 1.8928, "accuracy": 0.75, "lr": 6.083120259612139e-06, "epoch": 1.8319709355131697, "percentage": 45.86, "elapsed_time": "1 day, 10:14:27", "remaining_time": "1 day, 16:25:01"}
+{"current_steps": 1010, "total_steps": 2200, "loss": 1.7299, "accuracy": 0.796875, "lr": 6.076747861844919e-06, "epoch": 1.8337874659400546, "percentage": 45.91, "elapsed_time": "1 day, 10:16:48", "remaining_time": "1 day, 16:23:21"}
+{"current_steps": 1011, "total_steps": 2200, "loss": 1.937, "accuracy": 0.765625, "lr": 6.070374287988387e-06, "epoch": 1.835603996366939, "percentage": 45.95, "elapsed_time": "1 day, 10:19:05", "remaining_time": "1 day, 16:21:37"}
+{"current_steps": 1012, "total_steps": 2200, "loss": 2.2507, "accuracy": 0.6875, "lr": 6.06399955103937e-06, "epoch": 1.837420526793824, "percentage": 46.0, "elapsed_time": "1 day, 10:21:24", "remaining_time": "1 day, 16:19:55"}
+{"current_steps": 1013, "total_steps": 2200, "loss": 1.8517, "accuracy": 0.8125, "lr": 6.057623663997067e-06, "epoch": 1.8392370572207084, "percentage": 46.05, "elapsed_time": "1 day, 10:23:15", "remaining_time": "1 day, 16:17:39"}
+{"current_steps": 1014, "total_steps": 2200, "loss": 2.0057, "accuracy": 0.765625, "lr": 6.051246639863021e-06, "epoch": 1.841053587647593, "percentage": 46.09, "elapsed_time": "1 day, 10:25:22", "remaining_time": "1 day, 16:15:42"}
+{"current_steps": 1015, "total_steps": 2200, "loss": 1.9055, "accuracy": 0.78125, "lr": 6.044868491641097e-06, "epoch": 1.8428701180744778, "percentage": 46.14, "elapsed_time": "1 day, 10:27:42", "remaining_time": "1 day, 16:14:01"}
+{"current_steps": 1016, "total_steps": 2200, "loss": 1.9628, "accuracy": 0.6875, "lr": 6.038489232337447e-06, "epoch": 1.8446866485013624, "percentage": 46.18, "elapsed_time": "1 day, 10:29:55", "remaining_time": "1 day, 16:12:11"}
+{"current_steps": 1017, "total_steps": 2200, "loss": 1.856, "accuracy": 0.828125, "lr": 6.032108874960491e-06, "epoch": 1.8465031789282471, "percentage": 46.23, "elapsed_time": "1 day, 10:32:05", "remaining_time": "1 day, 16:10:18"}
+{"current_steps": 1018, "total_steps": 2200, "loss": 1.8242, "accuracy": 0.84375, "lr": 6.025727432520892e-06, "epoch": 1.8483197093551316, "percentage": 46.27, "elapsed_time": "1 day, 10:34:04", "remaining_time": "1 day, 16:08:12"}
+{"current_steps": 1019, "total_steps": 2200, "loss": 1.9857, "accuracy": 0.703125, "lr": 6.01934491803152e-06, "epoch": 1.8501362397820165, "percentage": 46.32, "elapsed_time": "1 day, 10:36:24", "remaining_time": "1 day, 16:06:31"}
+{"current_steps": 1020, "total_steps": 2200, "loss": 2.1156, "accuracy": 0.75, "lr": 6.012961344507433e-06, "epoch": 1.851952770208901, "percentage": 46.36, "elapsed_time": "1 day, 10:38:17", "remaining_time": "1 day, 16:04:18"}
+{"current_steps": 1021, "total_steps": 2200, "loss": 1.759, "accuracy": 0.8125, "lr": 6.0065767249658506e-06, "epoch": 1.8537693006357856, "percentage": 46.41, "elapsed_time": "1 day, 10:40:43", "remaining_time": "1 day, 16:02:42"}
+{"current_steps": 1022, "total_steps": 2200, "loss": 2.3511, "accuracy": 0.75, "lr": 6.0001910724261215e-06, "epoch": 1.8555858310626703, "percentage": 46.45, "elapsed_time": "1 day, 10:42:38", "remaining_time": "1 day, 16:00:32"}
+{"current_steps": 1023, "total_steps": 2200, "loss": 2.058, "accuracy": 0.703125, "lr": 5.993804399909704e-06, "epoch": 1.857402361489555, "percentage": 46.5, "elapsed_time": "1 day, 10:45:03", "remaining_time": "1 day, 15:58:55"}
+{"current_steps": 1024, "total_steps": 2200, "loss": 2.0877, "accuracy": 0.734375, "lr": 5.987416720440137e-06, "epoch": 1.8592188919164396, "percentage": 46.55, "elapsed_time": "1 day, 10:47:21", "remaining_time": "1 day, 15:57:12"}
+{"current_steps": 1025, "total_steps": 2200, "loss": 1.9073, "accuracy": 0.765625, "lr": 5.9810280470430096e-06, "epoch": 1.861035422343324, "percentage": 46.59, "elapsed_time": "1 day, 10:49:51", "remaining_time": "1 day, 15:55:41"}
+{"current_steps": 1026, "total_steps": 2200, "loss": 2.5596, "accuracy": 0.65625, "lr": 5.974638392745937e-06, "epoch": 1.862851952770209, "percentage": 46.64, "elapsed_time": "1 day, 10:51:56", "remaining_time": "1 day, 15:53:42"}
+{"current_steps": 1027, "total_steps": 2200, "loss": 1.9632, "accuracy": 0.765625, "lr": 5.9682477705785415e-06, "epoch": 1.8646684831970934, "percentage": 46.68, "elapsed_time": "1 day, 10:53:42", "remaining_time": "1 day, 15:51:21"}
+{"current_steps": 1028, "total_steps": 2200, "loss": 2.0445, "accuracy": 0.703125, "lr": 5.961856193572411e-06, "epoch": 1.8664850136239783, "percentage": 46.73, "elapsed_time": "1 day, 10:55:28", "remaining_time": "1 day, 15:49:00"}
+{"current_steps": 1029, "total_steps": 2200, "loss": 1.8366, "accuracy": 0.78125, "lr": 5.955463674761085e-06, "epoch": 1.8683015440508628, "percentage": 46.77, "elapsed_time": "1 day, 10:57:33", "remaining_time": "1 day, 15:47:01"}
+{"current_steps": 1030, "total_steps": 2200, "loss": 1.7379, "accuracy": 0.78125, "lr": 5.9490702271800225e-06, "epoch": 1.8701180744777475, "percentage": 46.82, "elapsed_time": "1 day, 10:59:18", "remaining_time": "1 day, 15:44:38"}
+{"current_steps": 1031, "total_steps": 2200, "loss": 2.1195, "accuracy": 0.734375, "lr": 5.9426758638665775e-06, "epoch": 1.8719346049046321, "percentage": 46.86, "elapsed_time": "1 day, 11:01:38", "remaining_time": "1 day, 15:42:56"}
+{"current_steps": 1032, "total_steps": 2200, "loss": 1.7604, "accuracy": 0.84375, "lr": 5.936280597859968e-06, "epoch": 1.8737511353315168, "percentage": 46.91, "elapsed_time": "1 day, 11:03:39", "remaining_time": "1 day, 15:40:53"}
+{"current_steps": 1033, "total_steps": 2200, "loss": 1.7271, "accuracy": 0.828125, "lr": 5.929884442201255e-06, "epoch": 1.8755676657584015, "percentage": 46.95, "elapsed_time": "1 day, 11:05:30", "remaining_time": "1 day, 15:38:38"}
+{"current_steps": 1034, "total_steps": 2200, "loss": 1.9179, "accuracy": 0.828125, "lr": 5.9234874099333165e-06, "epoch": 1.877384196185286, "percentage": 47.0, "elapsed_time": "1 day, 11:07:55", "remaining_time": "1 day, 15:37:01"}
+{"current_steps": 1035, "total_steps": 2200, "loss": 2.0308, "accuracy": 0.75, "lr": 5.91708951410081e-06, "epoch": 1.8792007266121709, "percentage": 47.05, "elapsed_time": "1 day, 11:09:48", "remaining_time": "1 day, 15:34:49"}
+{"current_steps": 1036, "total_steps": 2200, "loss": 2.1059, "accuracy": 0.703125, "lr": 5.910690767750164e-06, "epoch": 1.8810172570390553, "percentage": 47.09, "elapsed_time": "1 day, 11:11:48", "remaining_time": "1 day, 15:32:43"}
+{"current_steps": 1037, "total_steps": 2200, "loss": 1.7585, "accuracy": 0.84375, "lr": 5.904291183929533e-06, "epoch": 1.8828337874659402, "percentage": 47.14, "elapsed_time": "1 day, 11:13:46", "remaining_time": "1 day, 15:30:36"}
+{"current_steps": 1038, "total_steps": 2200, "loss": 2.1306, "accuracy": 0.703125, "lr": 5.897890775688787e-06, "epoch": 1.8846503178928247, "percentage": 47.18, "elapsed_time": "1 day, 11:15:32", "remaining_time": "1 day, 15:28:16"}
+{"current_steps": 1039, "total_steps": 2200, "loss": 1.6835, "accuracy": 0.859375, "lr": 5.891489556079467e-06, "epoch": 1.8864668483197093, "percentage": 47.23, "elapsed_time": "1 day, 11:17:37", "remaining_time": "1 day, 15:26:16"}
+{"current_steps": 1040, "total_steps": 2200, "loss": 1.7787, "accuracy": 0.84375, "lr": 5.88508753815478e-06, "epoch": 1.888283378746594, "percentage": 47.27, "elapsed_time": "1 day, 11:19:42", "remaining_time": "1 day, 15:24:17"}
+{"current_steps": 1041, "total_steps": 2200, "loss": 2.0775, "accuracy": 0.734375, "lr": 5.878684734969551e-06, "epoch": 1.8900999091734787, "percentage": 47.32, "elapsed_time": "1 day, 11:21:40", "remaining_time": "1 day, 15:22:10"}
+{"current_steps": 1042, "total_steps": 2200, "loss": 2.4721, "accuracy": 0.6875, "lr": 5.872281159580212e-06, "epoch": 1.8919164396003634, "percentage": 47.36, "elapsed_time": "1 day, 11:23:37", "remaining_time": "1 day, 15:20:01"}
+{"current_steps": 1043, "total_steps": 2200, "loss": 1.7064, "accuracy": 0.828125, "lr": 5.865876825044768e-06, "epoch": 1.8937329700272478, "percentage": 47.41, "elapsed_time": "1 day, 11:25:34", "remaining_time": "1 day, 15:17:54"}
+{"current_steps": 1044, "total_steps": 2200, "loss": 2.1084, "accuracy": 0.734375, "lr": 5.859471744422775e-06, "epoch": 1.8955495004541327, "percentage": 47.45, "elapsed_time": "1 day, 11:27:42", "remaining_time": "1 day, 15:15:58"}
+{"current_steps": 1045, "total_steps": 2200, "loss": 1.9188, "accuracy": 0.75, "lr": 5.8530659307753034e-06, "epoch": 1.8973660308810172, "percentage": 47.5, "elapsed_time": "1 day, 11:29:51", "remaining_time": "1 day, 15:14:02"}
+{"current_steps": 1046, "total_steps": 2200, "loss": 1.6973, "accuracy": 0.828125, "lr": 5.8466593971649235e-06, "epoch": 1.8991825613079019, "percentage": 47.55, "elapsed_time": "1 day, 11:32:16", "remaining_time": "1 day, 15:12:26"}
+{"current_steps": 1047, "total_steps": 2200, "loss": 1.7217, "accuracy": 0.859375, "lr": 5.840252156655676e-06, "epoch": 1.9009990917347865, "percentage": 47.59, "elapsed_time": "1 day, 11:34:12", "remaining_time": "1 day, 15:10:16"}
+{"current_steps": 1048, "total_steps": 2200, "loss": 1.9223, "accuracy": 0.765625, "lr": 5.833844222313035e-06, "epoch": 1.9028156221616712, "percentage": 47.64, "elapsed_time": "1 day, 11:35:47", "remaining_time": "1 day, 15:07:44"}
+{"current_steps": 1049, "total_steps": 2200, "loss": 1.8614, "accuracy": 0.78125, "lr": 5.827435607203898e-06, "epoch": 1.904632152588556, "percentage": 47.68, "elapsed_time": "1 day, 11:37:31", "remaining_time": "1 day, 15:05:22"}
+{"current_steps": 1050, "total_steps": 2200, "loss": 2.0506, "accuracy": 0.734375, "lr": 5.821026324396546e-06, "epoch": 1.9064486830154403, "percentage": 47.73, "elapsed_time": "1 day, 11:39:08", "remaining_time": "1 day, 15:02:52"}
+{"current_steps": 1051, "total_steps": 2200, "loss": 2.2419, "accuracy": 0.703125, "lr": 5.814616386960623e-06, "epoch": 1.9082652134423252, "percentage": 47.77, "elapsed_time": "1 day, 11:41:06", "remaining_time": "1 day, 15:00:45"}
+{"current_steps": 1052, "total_steps": 2200, "loss": 2.0351, "accuracy": 0.765625, "lr": 5.808205807967106e-06, "epoch": 1.9100817438692097, "percentage": 47.82, "elapsed_time": "1 day, 11:43:14", "remaining_time": "1 day, 14:58:49"}
+{"current_steps": 1053, "total_steps": 2200, "loss": 1.8518, "accuracy": 0.78125, "lr": 5.801794600488284e-06, "epoch": 1.9118982742960946, "percentage": 47.86, "elapsed_time": "1 day, 11:45:22", "remaining_time": "1 day, 14:56:53"}
+{"current_steps": 1054, "total_steps": 2200, "loss": 1.8466, "accuracy": 0.828125, "lr": 5.795382777597724e-06, "epoch": 1.913714804722979, "percentage": 47.91, "elapsed_time": "1 day, 11:47:18", "remaining_time": "1 day, 14:54:44"}
+{"current_steps": 1055, "total_steps": 2200, "loss": 1.583, "accuracy": 0.90625, "lr": 5.788970352370248e-06, "epoch": 1.9155313351498637, "percentage": 47.95, "elapsed_time": "1 day, 11:49:25", "remaining_time": "1 day, 14:52:47"}
+{"current_steps": 1056, "total_steps": 2200, "loss": 2.325, "accuracy": 0.671875, "lr": 5.782557337881911e-06, "epoch": 1.9173478655767484, "percentage": 48.0, "elapsed_time": "1 day, 11:51:47", "remaining_time": "1 day, 14:51:06"}
+{"current_steps": 1057, "total_steps": 2200, "loss": 1.8354, "accuracy": 0.8125, "lr": 5.776143747209963e-06, "epoch": 1.919164396003633, "percentage": 48.05, "elapsed_time": "1 day, 11:53:21", "remaining_time": "1 day, 14:48:33"}
+{"current_steps": 1058, "total_steps": 2200, "loss": 1.908, "accuracy": 0.734375, "lr": 5.769729593432835e-06, "epoch": 1.9209809264305178, "percentage": 48.09, "elapsed_time": "1 day, 11:55:23", "remaining_time": "1 day, 14:46:31"}
+{"current_steps": 1059, "total_steps": 2200, "loss": 2.3566, "accuracy": 0.734375, "lr": 5.763314889630098e-06, "epoch": 1.9227974568574022, "percentage": 48.14, "elapsed_time": "1 day, 11:57:15", "remaining_time": "1 day, 14:44:18"}
+{"current_steps": 1060, "total_steps": 2200, "loss": 1.9618, "accuracy": 0.734375, "lr": 5.756899648882457e-06, "epoch": 1.9246139872842871, "percentage": 48.18, "elapsed_time": "1 day, 11:59:18", "remaining_time": "1 day, 14:42:16"}
+{"current_steps": 1061, "total_steps": 2200, "loss": 2.2817, "accuracy": 0.71875, "lr": 5.750483884271699e-06, "epoch": 1.9264305177111716, "percentage": 48.23, "elapsed_time": "1 day, 12:01:13", "remaining_time": "1 day, 14:40:06"}
+{"current_steps": 1062, "total_steps": 2200, "loss": 2.0882, "accuracy": 0.734375, "lr": 5.744067608880689e-06, "epoch": 1.9282470481380565, "percentage": 48.27, "elapsed_time": "1 day, 12:03:14", "remaining_time": "1 day, 14:38:02"}
+{"current_steps": 1063, "total_steps": 2200, "loss": 1.8218, "accuracy": 0.78125, "lr": 5.737650835793329e-06, "epoch": 1.930063578564941, "percentage": 48.32, "elapsed_time": "1 day, 12:04:39", "remaining_time": "1 day, 14:35:20"}
+{"current_steps": 1064, "total_steps": 2200, "loss": 1.6036, "accuracy": 0.875, "lr": 5.731233578094534e-06, "epoch": 1.9318801089918256, "percentage": 48.36, "elapsed_time": "1 day, 12:06:39", "remaining_time": "1 day, 14:33:16"}
+{"current_steps": 1065, "total_steps": 2200, "loss": 1.7351, "accuracy": 0.796875, "lr": 5.724815848870213e-06, "epoch": 1.9336966394187103, "percentage": 48.41, "elapsed_time": "1 day, 12:08:24", "remaining_time": "1 day, 14:30:55"}
+{"current_steps": 1066, "total_steps": 2200, "loss": 2.0475, "accuracy": 0.8125, "lr": 5.718397661207233e-06, "epoch": 1.935513169845595, "percentage": 48.45, "elapsed_time": "1 day, 12:10:37", "remaining_time": "1 day, 14:29:05"}
+{"current_steps": 1067, "total_steps": 2200, "loss": 1.9754, "accuracy": 0.765625, "lr": 5.7119790281933914e-06, "epoch": 1.9373297002724796, "percentage": 48.5, "elapsed_time": "1 day, 12:12:44", "remaining_time": "1 day, 14:27:07"}
+{"current_steps": 1068, "total_steps": 2200, "loss": 1.9098, "accuracy": 0.828125, "lr": 5.705559962917406e-06, "epoch": 1.939146230699364, "percentage": 48.55, "elapsed_time": "1 day, 12:14:50", "remaining_time": "1 day, 14:25:10"}
+{"current_steps": 1069, "total_steps": 2200, "loss": 1.7809, "accuracy": 0.796875, "lr": 5.699140478468864e-06, "epoch": 1.940962761126249, "percentage": 48.59, "elapsed_time": "1 day, 12:17:07", "remaining_time": "1 day, 14:23:23"}
+{"current_steps": 1070, "total_steps": 2200, "loss": 2.0455, "accuracy": 0.796875, "lr": 5.692720587938212e-06, "epoch": 1.9427792915531334, "percentage": 48.64, "elapsed_time": "1 day, 12:19:24", "remaining_time": "1 day, 14:21:37"}
+{"current_steps": 1071, "total_steps": 2200, "loss": 1.4992, "accuracy": 0.875, "lr": 5.686300304416725e-06, "epoch": 1.9445958219800181, "percentage": 48.68, "elapsed_time": "1 day, 12:21:17", "remaining_time": "1 day, 14:19:25"}
+{"current_steps": 1072, "total_steps": 2200, "loss": 2.0695, "accuracy": 0.765625, "lr": 5.67987964099648e-06, "epoch": 1.9464123524069028, "percentage": 48.73, "elapsed_time": "1 day, 12:23:35", "remaining_time": "1 day, 14:17:39"}
+{"current_steps": 1073, "total_steps": 2200, "loss": 1.7233, "accuracy": 0.78125, "lr": 5.673458610770327e-06, "epoch": 1.9482288828337875, "percentage": 48.77, "elapsed_time": "1 day, 12:25:19", "remaining_time": "1 day, 14:15:18"}
+{"current_steps": 1074, "total_steps": 2200, "loss": 2.0319, "accuracy": 0.78125, "lr": 5.667037226831865e-06, "epoch": 1.9500454132606722, "percentage": 48.82, "elapsed_time": "1 day, 12:27:25", "remaining_time": "1 day, 14:13:20"}
+{"current_steps": 1075, "total_steps": 2200, "loss": 1.7892, "accuracy": 0.796875, "lr": 5.660615502275413e-06, "epoch": 1.9518619436875566, "percentage": 48.86, "elapsed_time": "1 day, 12:29:41", "remaining_time": "1 day, 14:11:32"}
+{"current_steps": 1076, "total_steps": 2200, "loss": 2.3361, "accuracy": 0.734375, "lr": 5.654193450195986e-06, "epoch": 1.9536784741144415, "percentage": 48.91, "elapsed_time": "1 day, 12:31:52", "remaining_time": "1 day, 14:09:39"}
+{"current_steps": 1077, "total_steps": 2200, "loss": 2.1204, "accuracy": 0.75, "lr": 5.647771083689266e-06, "epoch": 1.955495004541326, "percentage": 48.95, "elapsed_time": "1 day, 12:33:46", "remaining_time": "1 day, 14:07:28"}
+{"current_steps": 1078, "total_steps": 2200, "loss": 2.3019, "accuracy": 0.71875, "lr": 5.641348415851578e-06, "epoch": 1.9573115349682109, "percentage": 49.0, "elapsed_time": "1 day, 12:35:46", "remaining_time": "1 day, 14:05:24"}
+{"current_steps": 1079, "total_steps": 2200, "loss": 1.8076, "accuracy": 0.8125, "lr": 5.634925459779858e-06, "epoch": 1.9591280653950953, "percentage": 49.05, "elapsed_time": "1 day, 12:38:11", "remaining_time": "1 day, 14:03:45"}
+{"current_steps": 1080, "total_steps": 2200, "loss": 1.6456, "accuracy": 0.828125, "lr": 5.6285022285716325e-06, "epoch": 1.96094459582198, "percentage": 49.09, "elapsed_time": "1 day, 12:40:15", "remaining_time": "1 day, 14:01:44"}
+{"current_steps": 1081, "total_steps": 2200, "loss": 2.4464, "accuracy": 0.71875, "lr": 5.622078735324991e-06, "epoch": 1.9627611262488647, "percentage": 49.14, "elapsed_time": "1 day, 12:42:24", "remaining_time": "1 day, 13:59:50"}
+{"current_steps": 1082, "total_steps": 2200, "loss": 2.2637, "accuracy": 0.671875, "lr": 5.615654993138551e-06, "epoch": 1.9645776566757494, "percentage": 49.18, "elapsed_time": "1 day, 12:44:16", "remaining_time": "1 day, 13:57:36"}
+{"current_steps": 1083, "total_steps": 2200, "loss": 1.9737, "accuracy": 0.796875, "lr": 5.6092310151114416e-06, "epoch": 1.966394187102634, "percentage": 49.23, "elapsed_time": "1 day, 12:46:08", "remaining_time": "1 day, 13:55:24"}
+{"current_steps": 1084, "total_steps": 2200, "loss": 1.7235, "accuracy": 0.828125, "lr": 5.602806814343275e-06, "epoch": 1.9682107175295185, "percentage": 49.27, "elapsed_time": "1 day, 12:47:51", "remaining_time": "1 day, 13:53:01"}
+{"current_steps": 1085, "total_steps": 2200, "loss": 1.7914, "accuracy": 0.84375, "lr": 5.596382403934112e-06, "epoch": 1.9700272479564034, "percentage": 49.32, "elapsed_time": "1 day, 12:49:53", "remaining_time": "1 day, 13:50:59"}
+{"current_steps": 1086, "total_steps": 2200, "loss": 1.7396, "accuracy": 0.84375, "lr": 5.589957796984445e-06, "epoch": 1.9718437783832878, "percentage": 49.36, "elapsed_time": "1 day, 12:52:09", "remaining_time": "1 day, 13:49:11"}
+{"current_steps": 1087, "total_steps": 2200, "loss": 1.8953, "accuracy": 0.75, "lr": 5.583533006595169e-06, "epoch": 1.9736603088101727, "percentage": 49.41, "elapsed_time": "1 day, 12:54:13", "remaining_time": "1 day, 13:47:10"}
+{"current_steps": 1088, "total_steps": 2200, "loss": 1.7933, "accuracy": 0.84375, "lr": 5.577108045867548e-06, "epoch": 1.9754768392370572, "percentage": 49.45, "elapsed_time": "1 day, 12:56:28", "remaining_time": "1 day, 13:45:22"}
+{"current_steps": 1089, "total_steps": 2200, "loss": 2.1234, "accuracy": 0.78125, "lr": 5.570682927903194e-06, "epoch": 1.9772933696639419, "percentage": 49.5, "elapsed_time": "1 day, 12:58:15", "remaining_time": "1 day, 13:43:04"}
+{"current_steps": 1090, "total_steps": 2200, "loss": 1.6846, "accuracy": 0.796875, "lr": 5.564257665804044e-06, "epoch": 1.9791099000908265, "percentage": 49.55, "elapsed_time": "1 day, 13:00:24", "remaining_time": "1 day, 13:41:09"}
+{"current_steps": 1091, "total_steps": 2200, "loss": 1.8208, "accuracy": 0.796875, "lr": 5.557832272672325e-06, "epoch": 1.9809264305177112, "percentage": 49.59, "elapsed_time": "1 day, 13:02:36", "remaining_time": "1 day, 13:39:17"}
+{"current_steps": 1092, "total_steps": 2200, "loss": 1.8884, "accuracy": 0.71875, "lr": 5.551406761610531e-06, "epoch": 1.982742960944596, "percentage": 49.64, "elapsed_time": "1 day, 13:04:50", "remaining_time": "1 day, 13:37:26"}
+{"current_steps": 1093, "total_steps": 2200, "loss": 1.7552, "accuracy": 0.796875, "lr": 5.544981145721401e-06, "epoch": 1.9845594913714804, "percentage": 49.68, "elapsed_time": "1 day, 13:06:55", "remaining_time": "1 day, 13:35:27"}
+{"current_steps": 1094, "total_steps": 2200, "loss": 1.7055, "accuracy": 0.8125, "lr": 5.538555438107883e-06, "epoch": 1.9863760217983653, "percentage": 49.73, "elapsed_time": "1 day, 13:09:00", "remaining_time": "1 day, 13:33:27"}
+{"current_steps": 1095, "total_steps": 2200, "loss": 1.8296, "accuracy": 0.78125, "lr": 5.532129651873112e-06, "epoch": 1.9881925522252497, "percentage": 49.77, "elapsed_time": "1 day, 13:10:36", "remaining_time": "1 day, 13:30:58"}
+{"current_steps": 1096, "total_steps": 2200, "loss": 1.9843, "accuracy": 0.8125, "lr": 5.5257038001203885e-06, "epoch": 1.9900090826521344, "percentage": 49.82, "elapsed_time": "1 day, 13:12:26", "remaining_time": "1 day, 13:28:44"}
+{"current_steps": 1097, "total_steps": 2200, "loss": 1.712, "accuracy": 0.859375, "lr": 5.51927789595314e-06, "epoch": 1.991825613079019, "percentage": 49.86, "elapsed_time": "1 day, 13:15:04", "remaining_time": "1 day, 13:27:17"}
+{"current_steps": 1098, "total_steps": 2200, "loss": 1.7041, "accuracy": 0.796875, "lr": 5.512851952474907e-06, "epoch": 1.9936421435059037, "percentage": 49.91, "elapsed_time": "1 day, 13:17:15", "remaining_time": "1 day, 13:25:24"}
+{"current_steps": 1099, "total_steps": 2200, "loss": 2.0985, "accuracy": 0.734375, "lr": 5.506425982789302e-06, "epoch": 1.9954586739327884, "percentage": 49.95, "elapsed_time": "1 day, 13:19:11", "remaining_time": "1 day, 13:23:16"}
+{"current_steps": 1100, "total_steps": 2200, "loss": 2.0178, "accuracy": 0.703125, "lr": 5.500000000000001e-06, "epoch": 1.9972752043596729, "percentage": 50.0, "elapsed_time": "1 day, 13:21:25", "remaining_time": "1 day, 13:21:25"}
+{"current_steps": 1101, "total_steps": 2200, "loss": 2.2207, "accuracy": 0.75, "lr": 5.4935740172107e-06, "epoch": 1.9990917347865578, "percentage": 50.05, "elapsed_time": "1 day, 13:23:29", "remaining_time": "1 day, 13:19:25"}
+{"current_steps": 1102, "total_steps": 2200, "loss": 0.7729, "accuracy": 0.84375, "lr": 5.487148047525097e-06, "epoch": 2.0, "percentage": 50.09, "elapsed_time": "1 day, 13:24:27", "remaining_time": "1 day, 13:16:18"}
+{"current_steps": 1103, "total_steps": 2200, "loss": 1.6624, "accuracy": 0.828125, "lr": 5.480722104046862e-06, "epoch": 2.0018165304268845, "percentage": 50.14, "elapsed_time": "1 day, 13:26:10", "remaining_time": "1 day, 13:13:57"}
+{"current_steps": 1104, "total_steps": 2200, "loss": 1.6736, "accuracy": 0.78125, "lr": 5.474296199879613e-06, "epoch": 2.0036330608537694, "percentage": 50.18, "elapsed_time": "1 day, 13:28:07", "remaining_time": "1 day, 13:11:49"}
+{"current_steps": 1105, "total_steps": 2200, "loss": 1.6968, "accuracy": 0.859375, "lr": 5.467870348126888e-06, "epoch": 2.005449591280654, "percentage": 50.23, "elapsed_time": "1 day, 13:30:06", "remaining_time": "1 day, 13:09:44"}
+{"current_steps": 1106, "total_steps": 2200, "loss": 1.5586, "accuracy": 0.875, "lr": 5.46144456189212e-06, "epoch": 2.0072661217075387, "percentage": 50.27, "elapsed_time": "1 day, 13:31:41", "remaining_time": "1 day, 13:07:15"}
+{"current_steps": 1107, "total_steps": 2200, "loss": 1.7186, "accuracy": 0.8125, "lr": 5.455018854278601e-06, "epoch": 2.009082652134423, "percentage": 50.32, "elapsed_time": "1 day, 13:33:38", "remaining_time": "1 day, 13:05:08"}
+{"current_steps": 1108, "total_steps": 2200, "loss": 1.5237, "accuracy": 0.859375, "lr": 5.44859323838947e-06, "epoch": 2.010899182561308, "percentage": 50.36, "elapsed_time": "1 day, 13:35:20", "remaining_time": "1 day, 13:02:46"}
+{"current_steps": 1109, "total_steps": 2200, "loss": 1.5114, "accuracy": 0.84375, "lr": 5.442167727327676e-06, "epoch": 2.0127157129881925, "percentage": 50.41, "elapsed_time": "1 day, 13:37:07", "remaining_time": "1 day, 13:00:29"}
+{"current_steps": 1110, "total_steps": 2200, "loss": 1.7694, "accuracy": 0.765625, "lr": 5.435742334195959e-06, "epoch": 2.0145322434150774, "percentage": 50.45, "elapsed_time": "1 day, 13:39:02", "remaining_time": "1 day, 12:58:20"}
+{"current_steps": 1111, "total_steps": 2200, "loss": 1.5144, "accuracy": 0.890625, "lr": 5.429317072096808e-06, "epoch": 2.016348773841962, "percentage": 50.5, "elapsed_time": "1 day, 13:41:15", "remaining_time": "1 day, 12:56:28"}
+{"current_steps": 1112, "total_steps": 2200, "loss": 1.669, "accuracy": 0.8125, "lr": 5.422891954132454e-06, "epoch": 2.0181653042688463, "percentage": 50.55, "elapsed_time": "1 day, 13:43:33", "remaining_time": "1 day, 12:54:42"}
+{"current_steps": 1113, "total_steps": 2200, "loss": 1.6566, "accuracy": 0.84375, "lr": 5.416466993404833e-06, "epoch": 2.0199818346957312, "percentage": 50.59, "elapsed_time": "1 day, 13:45:06", "remaining_time": "1 day, 12:52:11"}
+{"current_steps": 1114, "total_steps": 2200, "loss": 1.5899, "accuracy": 0.796875, "lr": 5.410042203015556e-06, "epoch": 2.0217983651226157, "percentage": 50.64, "elapsed_time": "1 day, 13:47:13", "remaining_time": "1 day, 12:50:14"}
+{"current_steps": 1115, "total_steps": 2200, "loss": 1.5857, "accuracy": 0.84375, "lr": 5.40361759606589e-06, "epoch": 2.0236148955495006, "percentage": 50.68, "elapsed_time": "1 day, 13:49:25", "remaining_time": "1 day, 12:48:21"}
+{"current_steps": 1116, "total_steps": 2200, "loss": 1.7371, "accuracy": 0.828125, "lr": 5.3971931856567274e-06, "epoch": 2.025431425976385, "percentage": 50.73, "elapsed_time": "1 day, 13:51:46", "remaining_time": "1 day, 12:46:38"}
+{"current_steps": 1117, "total_steps": 2200, "loss": 1.3727, "accuracy": 0.96875, "lr": 5.39076898488856e-06, "epoch": 2.02724795640327, "percentage": 50.77, "elapsed_time": "1 day, 13:53:54", "remaining_time": "1 day, 12:44:41"}
+{"current_steps": 1118, "total_steps": 2200, "loss": 2.0189, "accuracy": 0.796875, "lr": 5.384345006861451e-06, "epoch": 2.0290644868301544, "percentage": 50.82, "elapsed_time": "1 day, 13:55:45", "remaining_time": "1 day, 12:42:28"}
+{"current_steps": 1119, "total_steps": 2200, "loss": 1.5383, "accuracy": 0.890625, "lr": 5.37792126467501e-06, "epoch": 2.030881017257039, "percentage": 50.86, "elapsed_time": "1 day, 13:57:45", "remaining_time": "1 day, 12:40:24"}
+{"current_steps": 1120, "total_steps": 2200, "loss": 1.517, "accuracy": 0.859375, "lr": 5.371497771428368e-06, "epoch": 2.0326975476839237, "percentage": 50.91, "elapsed_time": "1 day, 13:59:18", "remaining_time": "1 day, 12:37:53"}
+{"current_steps": 1121, "total_steps": 2200, "loss": 1.6882, "accuracy": 0.875, "lr": 5.365074540220143e-06, "epoch": 2.034514078110808, "percentage": 50.95, "elapsed_time": "1 day, 14:01:36", "remaining_time": "1 day, 12:36:07"}
+{"current_steps": 1122, "total_steps": 2200, "loss": 1.4543, "accuracy": 0.875, "lr": 5.358651584148424e-06, "epoch": 2.036330608537693, "percentage": 51.0, "elapsed_time": "1 day, 14:04:07", "remaining_time": "1 day, 12:34:32"}
+{"current_steps": 1123, "total_steps": 2200, "loss": 1.3534, "accuracy": 0.90625, "lr": 5.3522289163107345e-06, "epoch": 2.0381471389645776, "percentage": 51.05, "elapsed_time": "1 day, 14:05:54", "remaining_time": "1 day, 12:32:16"}
+{"current_steps": 1124, "total_steps": 2200, "loss": 1.4851, "accuracy": 0.8125, "lr": 5.345806549804016e-06, "epoch": 2.0399636693914625, "percentage": 51.09, "elapsed_time": "1 day, 14:08:04", "remaining_time": "1 day, 12:30:21"}
+{"current_steps": 1125, "total_steps": 2200, "loss": 1.4804, "accuracy": 0.90625, "lr": 5.339384497724589e-06, "epoch": 2.041780199818347, "percentage": 51.14, "elapsed_time": "1 day, 14:09:35", "remaining_time": "1 day, 12:27:50"}
+{"current_steps": 1126, "total_steps": 2200, "loss": 1.4414, "accuracy": 0.890625, "lr": 5.332962773168137e-06, "epoch": 2.043596730245232, "percentage": 51.18, "elapsed_time": "1 day, 14:11:26", "remaining_time": "1 day, 12:25:36"}
+{"current_steps": 1127, "total_steps": 2200, "loss": 1.3415, "accuracy": 0.859375, "lr": 5.326541389229674e-06, "epoch": 2.0454132606721163, "percentage": 51.23, "elapsed_time": "1 day, 14:13:28", "remaining_time": "1 day, 12:23:34"}
+{"current_steps": 1128, "total_steps": 2200, "loss": 1.5028, "accuracy": 0.890625, "lr": 5.320120359003521e-06, "epoch": 2.0472297910990007, "percentage": 51.27, "elapsed_time": "1 day, 14:15:37", "remaining_time": "1 day, 12:21:39"}
+{"current_steps": 1129, "total_steps": 2200, "loss": 1.46, "accuracy": 0.84375, "lr": 5.313699695583276e-06, "epoch": 2.0490463215258856, "percentage": 51.32, "elapsed_time": "1 day, 14:17:40", "remaining_time": "1 day, 12:19:37"}
+{"current_steps": 1130, "total_steps": 2200, "loss": 1.6726, "accuracy": 0.828125, "lr": 5.30727941206179e-06, "epoch": 2.05086285195277, "percentage": 51.36, "elapsed_time": "1 day, 14:19:48", "remaining_time": "1 day, 12:17:41"}
+{"current_steps": 1131, "total_steps": 2200, "loss": 1.6628, "accuracy": 0.828125, "lr": 5.300859521531138e-06, "epoch": 2.052679382379655, "percentage": 51.41, "elapsed_time": "1 day, 14:21:56", "remaining_time": "1 day, 12:15:44"}
+{"current_steps": 1132, "total_steps": 2200, "loss": 1.4598, "accuracy": 0.875, "lr": 5.294440037082596e-06, "epoch": 2.0544959128065394, "percentage": 51.45, "elapsed_time": "1 day, 14:24:05", "remaining_time": "1 day, 12:13:49"}
+{"current_steps": 1133, "total_steps": 2200, "loss": 1.429, "accuracy": 0.875, "lr": 5.288020971806609e-06, "epoch": 2.0563124432334243, "percentage": 51.5, "elapsed_time": "1 day, 14:25:58", "remaining_time": "1 day, 12:11:38"}
+{"current_steps": 1134, "total_steps": 2200, "loss": 1.652, "accuracy": 0.765625, "lr": 5.281602338792769e-06, "epoch": 2.058128973660309, "percentage": 51.55, "elapsed_time": "1 day, 14:28:08", "remaining_time": "1 day, 12:09:44"}
+{"current_steps": 1135, "total_steps": 2200, "loss": 1.51, "accuracy": 0.84375, "lr": 5.275184151129789e-06, "epoch": 2.0599455040871932, "percentage": 51.59, "elapsed_time": "1 day, 14:29:52", "remaining_time": "1 day, 12:07:24"}
+{"current_steps": 1136, "total_steps": 2200, "loss": 1.3913, "accuracy": 0.875, "lr": 5.268766421905467e-06, "epoch": 2.061762034514078, "percentage": 51.64, "elapsed_time": "1 day, 14:32:08", "remaining_time": "1 day, 12:05:36"}
+{"current_steps": 1137, "total_steps": 2200, "loss": 1.6787, "accuracy": 0.84375, "lr": 5.262349164206671e-06, "epoch": 2.0635785649409626, "percentage": 51.68, "elapsed_time": "1 day, 14:34:15", "remaining_time": "1 day, 12:03:38"}
+{"current_steps": 1138, "total_steps": 2200, "loss": 1.9326, "accuracy": 0.734375, "lr": 5.2559323911193124e-06, "epoch": 2.0653950953678475, "percentage": 51.73, "elapsed_time": "1 day, 14:36:31", "remaining_time": "1 day, 12:01:49"}
+{"current_steps": 1139, "total_steps": 2200, "loss": 1.6712, "accuracy": 0.8125, "lr": 5.249516115728302e-06, "epoch": 2.067211625794732, "percentage": 51.77, "elapsed_time": "1 day, 14:38:14", "remaining_time": "1 day, 11:59:29"}
+{"current_steps": 1140, "total_steps": 2200, "loss": 1.7355, "accuracy": 0.75, "lr": 5.2431003511175435e-06, "epoch": 2.069028156221617, "percentage": 51.82, "elapsed_time": "1 day, 14:39:56", "remaining_time": "1 day, 11:57:08"}
+{"current_steps": 1141, "total_steps": 2200, "loss": 1.6904, "accuracy": 0.828125, "lr": 5.2366851103699024e-06, "epoch": 2.0708446866485013, "percentage": 51.86, "elapsed_time": "1 day, 14:41:51", "remaining_time": "1 day, 11:55:00"}
+{"current_steps": 1142, "total_steps": 2200, "loss": 1.6106, "accuracy": 0.828125, "lr": 5.230270406567169e-06, "epoch": 2.072661217075386, "percentage": 51.91, "elapsed_time": "1 day, 14:44:04", "remaining_time": "1 day, 11:53:07"}
+{"current_steps": 1143, "total_steps": 2200, "loss": 1.6687, "accuracy": 0.859375, "lr": 5.223856252790038e-06, "epoch": 2.0744777475022707, "percentage": 51.95, "elapsed_time": "1 day, 14:46:07", "remaining_time": "1 day, 11:51:05"}
+{"current_steps": 1144, "total_steps": 2200, "loss": 1.2402, "accuracy": 0.90625, "lr": 5.217442662118091e-06, "epoch": 2.076294277929155, "percentage": 52.0, "elapsed_time": "1 day, 14:47:55", "remaining_time": "1 day, 11:48:51"}
+{"current_steps": 1145, "total_steps": 2200, "loss": 1.6288, "accuracy": 0.828125, "lr": 5.2110296476297515e-06, "epoch": 2.07811080835604, "percentage": 52.05, "elapsed_time": "1 day, 14:49:47", "remaining_time": "1 day, 11:46:39"}
+{"current_steps": 1146, "total_steps": 2200, "loss": 1.7521, "accuracy": 0.8125, "lr": 5.204617222402278e-06, "epoch": 2.0799273387829245, "percentage": 52.09, "elapsed_time": "1 day, 14:51:12", "remaining_time": "1 day, 11:44:03"}
+{"current_steps": 1147, "total_steps": 2200, "loss": 1.6775, "accuracy": 0.8125, "lr": 5.1982053995117174e-06, "epoch": 2.0817438692098094, "percentage": 52.14, "elapsed_time": "1 day, 14:53:29", "remaining_time": "1 day, 11:42:14"}
+{"current_steps": 1148, "total_steps": 2200, "loss": 2.2784, "accuracy": 0.78125, "lr": 5.1917941920328955e-06, "epoch": 2.083560399636694, "percentage": 52.18, "elapsed_time": "1 day, 14:55:02", "remaining_time": "1 day, 11:39:46"}
+{"current_steps": 1149, "total_steps": 2200, "loss": 1.5316, "accuracy": 0.8125, "lr": 5.185383613039379e-06, "epoch": 2.0853769300635787, "percentage": 52.23, "elapsed_time": "1 day, 14:56:56", "remaining_time": "1 day, 11:37:36"}
+{"current_steps": 1150, "total_steps": 2200, "loss": 1.8403, "accuracy": 0.796875, "lr": 5.178973675603456e-06, "epoch": 2.087193460490463, "percentage": 52.27, "elapsed_time": "1 day, 14:59:13", "remaining_time": "1 day, 11:35:48"}
+{"current_steps": 1151, "total_steps": 2200, "loss": 1.5046, "accuracy": 0.8125, "lr": 5.172564392796103e-06, "epoch": 2.089009990917348, "percentage": 52.32, "elapsed_time": "1 day, 15:01:12", "remaining_time": "1 day, 11:33:44"}
+{"current_steps": 1152, "total_steps": 2200, "loss": 1.665, "accuracy": 0.796875, "lr": 5.166155777686966e-06, "epoch": 2.0908265213442325, "percentage": 52.36, "elapsed_time": "1 day, 15:03:16", "remaining_time": "1 day, 11:31:43"}
+{"current_steps": 1153, "total_steps": 2200, "loss": 1.4786, "accuracy": 0.828125, "lr": 5.159747843344327e-06, "epoch": 2.092643051771117, "percentage": 52.41, "elapsed_time": "1 day, 15:04:47", "remaining_time": "1 day, 11:29:13"}
+{"current_steps": 1154, "total_steps": 2200, "loss": 1.6674, "accuracy": 0.78125, "lr": 5.153340602835077e-06, "epoch": 2.094459582198002, "percentage": 52.45, "elapsed_time": "1 day, 15:07:14", "remaining_time": "1 day, 11:27:33"}
+{"current_steps": 1155, "total_steps": 2200, "loss": 1.3256, "accuracy": 0.921875, "lr": 5.1469340692247e-06, "epoch": 2.0962761126248863, "percentage": 52.5, "elapsed_time": "1 day, 15:09:20", "remaining_time": "1 day, 11:25:35"}
+{"current_steps": 1156, "total_steps": 2200, "loss": 1.44, "accuracy": 0.84375, "lr": 5.1405282555772265e-06, "epoch": 2.0980926430517712, "percentage": 52.55, "elapsed_time": "1 day, 15:11:24", "remaining_time": "1 day, 11:23:35"}
+{"current_steps": 1157, "total_steps": 2200, "loss": 1.3248, "accuracy": 0.921875, "lr": 5.134123174955233e-06, "epoch": 2.0999091734786557, "percentage": 52.59, "elapsed_time": "1 day, 15:13:38", "remaining_time": "1 day, 11:21:44"}
+{"current_steps": 1158, "total_steps": 2200, "loss": 1.6194, "accuracy": 0.859375, "lr": 5.127718840419789e-06, "epoch": 2.1017257039055406, "percentage": 52.64, "elapsed_time": "1 day, 15:15:22", "remaining_time": "1 day, 11:19:25"}
+{"current_steps": 1159, "total_steps": 2200, "loss": 1.6539, "accuracy": 0.84375, "lr": 5.12131526503045e-06, "epoch": 2.103542234332425, "percentage": 52.68, "elapsed_time": "1 day, 15:17:08", "remaining_time": "1 day, 11:17:09"}
+{"current_steps": 1160, "total_steps": 2200, "loss": 1.5297, "accuracy": 0.84375, "lr": 5.114912461845223e-06, "epoch": 2.10535876475931, "percentage": 52.73, "elapsed_time": "1 day, 15:19:04", "remaining_time": "1 day, 11:15:02"}
+{"current_steps": 1161, "total_steps": 2200, "loss": 1.8857, "accuracy": 0.796875, "lr": 5.108510443920534e-06, "epoch": 2.1071752951861944, "percentage": 52.77, "elapsed_time": "1 day, 15:21:17", "remaining_time": "1 day, 11:13:10"}
+{"current_steps": 1162, "total_steps": 2200, "loss": 1.8805, "accuracy": 0.796875, "lr": 5.1021092243112146e-06, "epoch": 2.108991825613079, "percentage": 52.82, "elapsed_time": "1 day, 15:23:35", "remaining_time": "1 day, 11:11:22"}
+{"current_steps": 1163, "total_steps": 2200, "loss": 1.4499, "accuracy": 0.921875, "lr": 5.095708816070466e-06, "epoch": 2.1108083560399638, "percentage": 52.86, "elapsed_time": "1 day, 15:25:54", "remaining_time": "1 day, 11:09:35"}
+{"current_steps": 1164, "total_steps": 2200, "loss": 1.4273, "accuracy": 0.859375, "lr": 5.0893092322498385e-06, "epoch": 2.112624886466848, "percentage": 52.91, "elapsed_time": "1 day, 15:28:08", "remaining_time": "1 day, 11:07:43"}
+{"current_steps": 1165, "total_steps": 2200, "loss": 1.5834, "accuracy": 0.8125, "lr": 5.082910485899191e-06, "epoch": 2.114441416893733, "percentage": 52.95, "elapsed_time": "1 day, 15:30:14", "remaining_time": "1 day, 11:05:44"}
+{"current_steps": 1166, "total_steps": 2200, "loss": 1.495, "accuracy": 0.859375, "lr": 5.076512590066686e-06, "epoch": 2.1162579473206176, "percentage": 53.0, "elapsed_time": "1 day, 15:32:36", "remaining_time": "1 day, 11:04:00"}
+{"current_steps": 1167, "total_steps": 2200, "loss": 1.8571, "accuracy": 0.796875, "lr": 5.070115557798748e-06, "epoch": 2.1180744777475025, "percentage": 53.05, "elapsed_time": "1 day, 15:34:58", "remaining_time": "1 day, 11:02:16"}
+{"current_steps": 1168, "total_steps": 2200, "loss": 1.4526, "accuracy": 0.875, "lr": 5.063719402140033e-06, "epoch": 2.119891008174387, "percentage": 53.09, "elapsed_time": "1 day, 15:37:16", "remaining_time": "1 day, 11:00:27"}
+{"current_steps": 1169, "total_steps": 2200, "loss": 1.9912, "accuracy": 0.8125, "lr": 5.057324136133425e-06, "epoch": 2.1217075386012714, "percentage": 53.14, "elapsed_time": "1 day, 15:39:09", "remaining_time": "1 day, 10:58:17"}
+{"current_steps": 1170, "total_steps": 2200, "loss": 1.5323, "accuracy": 0.859375, "lr": 5.050929772819978e-06, "epoch": 2.1235240690281563, "percentage": 53.18, "elapsed_time": "1 day, 15:41:18", "remaining_time": "1 day, 10:56:22"}
+{"current_steps": 1171, "total_steps": 2200, "loss": 1.4151, "accuracy": 0.875, "lr": 5.044536325238917e-06, "epoch": 2.1253405994550407, "percentage": 53.23, "elapsed_time": "1 day, 15:43:31", "remaining_time": "1 day, 10:54:29"}
+{"current_steps": 1172, "total_steps": 2200, "loss": 1.4471, "accuracy": 0.875, "lr": 5.03814380642759e-06, "epoch": 2.1271571298819256, "percentage": 53.27, "elapsed_time": "1 day, 15:45:35", "remaining_time": "1 day, 10:52:28"}
+{"current_steps": 1173, "total_steps": 2200, "loss": 1.4202, "accuracy": 0.90625, "lr": 5.031752229421461e-06, "epoch": 2.12897366030881, "percentage": 53.32, "elapsed_time": "1 day, 15:47:31", "remaining_time": "1 day, 10:50:21"}
+{"current_steps": 1174, "total_steps": 2200, "loss": 1.4605, "accuracy": 0.828125, "lr": 5.025361607254064e-06, "epoch": 2.130790190735695, "percentage": 53.36, "elapsed_time": "1 day, 15:49:24", "remaining_time": "1 day, 10:48:10"}
+{"current_steps": 1175, "total_steps": 2200, "loss": 1.5052, "accuracy": 0.828125, "lr": 5.0189719529569935e-06, "epoch": 2.1326067211625794, "percentage": 53.41, "elapsed_time": "1 day, 15:51:24", "remaining_time": "1 day, 10:46:07"}
+{"current_steps": 1176, "total_steps": 2200, "loss": 1.4768, "accuracy": 0.859375, "lr": 5.012583279559865e-06, "epoch": 2.134423251589464, "percentage": 53.45, "elapsed_time": "1 day, 15:53:35", "remaining_time": "1 day, 10:44:12"}
+{"current_steps": 1177, "total_steps": 2200, "loss": 1.7802, "accuracy": 0.765625, "lr": 5.006195600090297e-06, "epoch": 2.136239782016349, "percentage": 53.5, "elapsed_time": "1 day, 15:55:46", "remaining_time": "1 day, 10:42:18"}
+{"current_steps": 1178, "total_steps": 2200, "loss": 1.6805, "accuracy": 0.765625, "lr": 4.999808927573881e-06, "epoch": 2.1380563124432332, "percentage": 53.55, "elapsed_time": "1 day, 15:57:59", "remaining_time": "1 day, 10:40:25"}
+{"current_steps": 1179, "total_steps": 2200, "loss": 1.4665, "accuracy": 0.90625, "lr": 4.993423275034152e-06, "epoch": 2.139872842870118, "percentage": 53.59, "elapsed_time": "1 day, 16:00:14", "remaining_time": "1 day, 10:38:35"}
+{"current_steps": 1180, "total_steps": 2200, "loss": 1.5666, "accuracy": 0.828125, "lr": 4.9870386554925674e-06, "epoch": 2.1416893732970026, "percentage": 53.64, "elapsed_time": "1 day, 16:02:26", "remaining_time": "1 day, 10:36:41"}
+{"current_steps": 1181, "total_steps": 2200, "loss": 1.4335, "accuracy": 0.84375, "lr": 4.980655081968481e-06, "epoch": 2.1435059037238875, "percentage": 53.68, "elapsed_time": "1 day, 16:04:17", "remaining_time": "1 day, 10:34:29"}
+{"current_steps": 1182, "total_steps": 2200, "loss": 1.544, "accuracy": 0.828125, "lr": 4.974272567479109e-06, "epoch": 2.145322434150772, "percentage": 53.73, "elapsed_time": "1 day, 16:06:37", "remaining_time": "1 day, 10:32:42"}
+{"current_steps": 1183, "total_steps": 2200, "loss": 1.4489, "accuracy": 0.90625, "lr": 4.96789112503951e-06, "epoch": 2.147138964577657, "percentage": 53.77, "elapsed_time": "1 day, 16:08:35", "remaining_time": "1 day, 10:30:37"}
+{"current_steps": 1184, "total_steps": 2200, "loss": 1.6567, "accuracy": 0.84375, "lr": 4.961510767662555e-06, "epoch": 2.1489554950045413, "percentage": 53.82, "elapsed_time": "1 day, 16:10:50", "remaining_time": "1 day, 10:28:46"}
+{"current_steps": 1185, "total_steps": 2200, "loss": 1.3804, "accuracy": 0.90625, "lr": 4.955131508358904e-06, "epoch": 2.1507720254314258, "percentage": 53.86, "elapsed_time": "1 day, 16:12:08", "remaining_time": "1 day, 10:26:06"}
+{"current_steps": 1186, "total_steps": 2200, "loss": 1.8752, "accuracy": 0.8125, "lr": 4.948753360136981e-06, "epoch": 2.1525885558583107, "percentage": 53.91, "elapsed_time": "1 day, 16:14:04", "remaining_time": "1 day, 10:23:58"}
+{"current_steps": 1187, "total_steps": 2200, "loss": 1.5727, "accuracy": 0.875, "lr": 4.942376336002934e-06, "epoch": 2.154405086285195, "percentage": 53.95, "elapsed_time": "1 day, 16:16:09", "remaining_time": "1 day, 10:21:58"}
+{"current_steps": 1188, "total_steps": 2200, "loss": 1.3016, "accuracy": 0.84375, "lr": 4.936000448960631e-06, "epoch": 2.15622161671208, "percentage": 54.0, "elapsed_time": "1 day, 16:18:01", "remaining_time": "1 day, 10:19:47"}
+{"current_steps": 1189, "total_steps": 2200, "loss": 1.5729, "accuracy": 0.890625, "lr": 4.9296257120116165e-06, "epoch": 2.1580381471389645, "percentage": 54.05, "elapsed_time": "1 day, 16:19:52", "remaining_time": "1 day, 10:17:36"}
+{"current_steps": 1190, "total_steps": 2200, "loss": 1.8277, "accuracy": 0.75, "lr": 4.923252138155084e-06, "epoch": 2.1598546775658494, "percentage": 54.09, "elapsed_time": "1 day, 16:21:34", "remaining_time": "1 day, 10:15:17"}
+{"current_steps": 1191, "total_steps": 2200, "loss": 1.4208, "accuracy": 0.875, "lr": 4.916879740387862e-06, "epoch": 2.161671207992734, "percentage": 54.14, "elapsed_time": "1 day, 16:23:27", "remaining_time": "1 day, 10:13:07"}
+{"current_steps": 1192, "total_steps": 2200, "loss": 1.6284, "accuracy": 0.84375, "lr": 4.910508531704382e-06, "epoch": 2.1634877384196187, "percentage": 54.18, "elapsed_time": "1 day, 16:25:33", "remaining_time": "1 day, 10:11:09"}
+{"current_steps": 1193, "total_steps": 2200, "loss": 1.474, "accuracy": 0.875, "lr": 4.904138525096647e-06, "epoch": 2.165304268846503, "percentage": 54.23, "elapsed_time": "1 day, 16:27:44", "remaining_time": "1 day, 10:09:14"}
+{"current_steps": 1194, "total_steps": 2200, "loss": 1.733, "accuracy": 0.8125, "lr": 4.897769733554209e-06, "epoch": 2.1671207992733876, "percentage": 54.27, "elapsed_time": "1 day, 16:29:35", "remaining_time": "1 day, 10:07:02"}
+{"current_steps": 1195, "total_steps": 2200, "loss": 1.3834, "accuracy": 0.875, "lr": 4.89140217006414e-06, "epoch": 2.1689373297002725, "percentage": 54.32, "elapsed_time": "1 day, 16:31:52", "remaining_time": "1 day, 10:05:12"}
+{"current_steps": 1196, "total_steps": 2200, "loss": 1.3927, "accuracy": 0.90625, "lr": 4.885035847611015e-06, "epoch": 2.170753860127157, "percentage": 54.36, "elapsed_time": "1 day, 16:33:45", "remaining_time": "1 day, 10:03:03"}
+{"current_steps": 1197, "total_steps": 2200, "loss": 1.513, "accuracy": 0.796875, "lr": 4.878670779176872e-06, "epoch": 2.172570390554042, "percentage": 54.41, "elapsed_time": "1 day, 16:35:48", "remaining_time": "1 day, 10:01:01"}
+{"current_steps": 1198, "total_steps": 2200, "loss": 1.5972, "accuracy": 0.828125, "lr": 4.872306977741192e-06, "epoch": 2.1743869209809263, "percentage": 54.45, "elapsed_time": "1 day, 16:38:05", "remaining_time": "1 day, 9:59:12"}
+{"current_steps": 1199, "total_steps": 2200, "loss": 1.6461, "accuracy": 0.859375, "lr": 4.865944456280879e-06, "epoch": 2.1762034514078112, "percentage": 54.5, "elapsed_time": "1 day, 16:40:08", "remaining_time": "1 day, 9:57:11"}
diff --git a/trainer_state.json b/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dd3dfdd7e6e889bdf1e10def94c5835fdfa31c99
--- /dev/null
+++ b/trainer_state.json
@@ -0,0 +1,162 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 84,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 16.365549600718158,
+      "learning_rate": 9e-06,
+      "logits/chosen": 0.43800339102745056,
+      "logits/rejected": 0.4281293749809265,
+      "logps/chosen": -324.7308349609375,
+      "logps/rejected": -337.3280029296875,
+      "loss": 0.6876,
+      "rewards/accuracies": 0.4374999701976776,
+      "rewards/chosen": 0.009319745004177094,
+      "rewards/margins": 0.012683698907494545,
+      "rewards/rejected": -0.0033639525063335896,
+      "step": 10
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 6.549199651973184,
+      "learning_rate": 8.64047538607195e-06,
+      "logits/chosen": 0.4200649559497833,
+      "logits/rejected": 0.41489219665527344,
+      "logps/chosen": -321.9690246582031,
+      "logps/rejected": -355.4017639160156,
+      "loss": 0.6962,
+      "rewards/accuracies": 0.48125001788139343,
+      "rewards/chosen": 0.034620046615600586,
+      "rewards/margins": -0.00358342076651752,
+      "rewards/rejected": 0.038203466683626175,
+      "step": 20
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 7.086531400587909,
+      "learning_rate": 7.62573262972983e-06,
+      "logits/chosen": 0.4370676875114441,
+      "logits/rejected": 0.4306566119194031,
+      "logps/chosen": -315.1405944824219,
+      "logps/rejected": -343.5477600097656,
+      "loss": 0.6646,
+      "rewards/accuracies": 0.6187500357627869,
+      "rewards/chosen": 0.1420479416847229,
+      "rewards/margins": 0.06437937170267105,
+      "rewards/rejected": 0.07766857743263245,
+      "step": 30
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 5.431551347783104,
+      "learning_rate": 6.135932223670029e-06,
+      "logits/chosen": 0.4090261161327362,
+      "logits/rejected": 0.40613797307014465,
+      "logps/chosen": -317.56817626953125,
+      "logps/rejected": -354.5175476074219,
+      "loss": 0.5202,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.44594669342041016,
+      "rewards/margins": 0.39714840054512024,
+      "rewards/rejected": 0.04879830405116081,
+      "step": 40
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 5.871421809973983,
+      "learning_rate": 4.4355778300251406e-06,
+      "logits/chosen": 0.4284099042415619,
+      "logits/rejected": 0.4222380816936493,
+      "logps/chosen": -319.2432861328125,
+      "logps/rejected": -343.3741760253906,
+      "loss": 0.4481,
+      "rewards/accuracies": 0.9124999642372131,
+      "rewards/chosen": 0.6467863321304321,
+      "rewards/margins": 0.6407225131988525,
+      "rewards/rejected": 0.0060638319700956345,
+      "step": 50
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 5.194288258129358,
+      "learning_rate": 2.8265555015932123e-06,
+      "logits/chosen": 0.4114301800727844,
+      "logits/rejected": 0.4095650017261505,
+      "logps/chosen": -312.0003662109375,
+      "logps/rejected": -348.30206298828125,
+      "loss": 0.3622,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.7869285345077515,
+      "rewards/margins": 0.9796440601348877,
+      "rewards/rejected": -0.19271555542945862,
+      "step": 60
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 3.569068921757812,
+      "learning_rate": 1.5945359205624398e-06,
+      "logits/chosen": 0.4047119617462158,
+      "logits/rejected": 0.4038873612880707,
+      "logps/chosen": -307.7536926269531,
+      "logps/rejected": -354.0033264160156,
+      "loss": 0.2611,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0587913990020752,
+      "rewards/margins": 1.4121346473693848,
+      "rewards/rejected": -0.35334333777427673,
+      "step": 70
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 4.489026775368698,
+      "learning_rate": 9.58255563091781e-07,
+      "logits/chosen": 0.3864372968673706,
+      "logits/rejected": 0.3842785954475403,
+      "logps/chosen": -312.03338623046875,
+      "logps/rejected": -350.1846923828125,
+      "loss": 0.2398,
+      "rewards/accuracies": 0.9937499761581421,
+      "rewards/chosen": 1.117773413658142,
+      "rewards/margins": 1.5565866231918335,
+      "rewards/rejected": -0.43881309032440186,
+      "step": 80
+    },
+    {
+      "epoch": 3.0,
+      "step": 84,
+      "total_flos": 216736934658048.0,
+      "train_loss": 0.47654047750291373,
+      "train_runtime": 2544.5527,
+      "train_samples_per_second": 0.519,
+      "train_steps_per_second": 0.033
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 84,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 115,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 216736934658048.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/training_args.bin b/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..689f92ab91fc248999cca8078aa5f2d327b997ef
--- /dev/null
+++ b/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:540287b777c24cb36855ecf7a73f9ff8f66586431561f6b8a7857beb5a25ff53
+size 7544
diff --git a/training_loss.png b/training_loss.png
new file mode 100644
index 0000000000000000000000000000000000000000..b5cdda27633a40a60ec07e812f7291fee2e3ffd3
--- /dev/null
+++ b/training_loss.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9aac53ef48f77828aa6099c5bb7f1f33629a859cfed712837c662f1a9b8860a
+size 33551
diff --git a/training_rewards_accuracies.png b/training_rewards_accuracies.png
new file mode 100644
index 0000000000000000000000000000000000000000..ed82535841b4d812e28421965f853e342e1b3652
--- /dev/null
+++ b/training_rewards_accuracies.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de2f407213980e5d408f7ede6ec7d61c123e83881b7184f90d057f54aeea67ef
+size 36529