{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.13675213675213677,
  "eval_steps": 40,
  "global_step": 40,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.017094017094017096,
      "grad_norm": 35.038580788061665,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7457876205444336,
      "logits/rejected": -2.7444841861724854,
      "logps/chosen": -164.26461791992188,
      "logps/rejected": -170.55870056152344,
      "loss": 0.6935,
      "rewards/accuracies": 0.26875001192092896,
      "rewards/chosen": 0.003455913159996271,
      "rewards/margins": -0.0019886991940438747,
      "rewards/rejected": 0.0054446132853627205,
      "step": 5
    },
    {
      "epoch": 0.03418803418803419,
      "grad_norm": 36.203903910498276,
      "learning_rate": 1e-06,
      "logits/chosen": -2.7106502056121826,
      "logits/rejected": -2.716397523880005,
      "logps/chosen": -171.80043029785156,
      "logps/rejected": -165.20602416992188,
      "loss": 0.6875,
      "rewards/accuracies": 0.5062500238418579,
      "rewards/chosen": 0.012000308372080326,
      "rewards/margins": 0.0025437879376113415,
      "rewards/rejected": 0.009456520900130272,
      "step": 10
    },
    {
      "epoch": 0.05128205128205128,
      "grad_norm": 33.9576577784673,
      "learning_rate": 9.999177507263144e-07,
      "logits/chosen": -2.651571750640869,
      "logits/rejected": -2.629457473754883,
      "logps/chosen": -174.04080200195312,
      "logps/rejected": -174.0542755126953,
      "loss": 0.6698,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.23909731209278107,
      "rewards/margins": 0.10868903249502182,
      "rewards/rejected": 0.13040827214717865,
      "step": 15
    },
    {
      "epoch": 0.06837606837606838,
      "grad_norm": 34.33646066636181,
      "learning_rate": 9.996710299650301e-07,
      "logits/chosen": -2.476440668106079,
      "logits/rejected": -2.450225353240967,
      "logps/chosen": -158.1311798095703,
      "logps/rejected": -158.0066680908203,
      "loss": 0.6613,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.4318675100803375,
      "rewards/margins": 0.14549395442008972,
      "rewards/rejected": 0.2863735556602478,
      "step": 20
    },
    {
      "epoch": 0.08547008547008547,
      "grad_norm": 33.16430522723429,
      "learning_rate": 9.992599188865604e-07,
      "logits/chosen": -2.3086318969726562,
      "logits/rejected": -2.3104796409606934,
      "logps/chosen": -150.59771728515625,
      "logps/rejected": -156.85037231445312,
      "loss": 0.6494,
      "rewards/accuracies": 0.581250011920929,
      "rewards/chosen": 0.5047669410705566,
      "rewards/margins": 0.16554531455039978,
      "rewards/rejected": 0.33922165632247925,
      "step": 25
    },
    {
      "epoch": 0.10256410256410256,
      "grad_norm": 34.52861424862365,
      "learning_rate": 9.98684552745256e-07,
      "logits/chosen": -2.217874050140381,
      "logits/rejected": -2.2254481315612793,
      "logps/chosen": -161.29412841796875,
      "logps/rejected": -161.40841674804688,
      "loss": 0.6295,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.4176379144191742,
      "rewards/margins": 0.26531916856765747,
      "rewards/rejected": 0.15231874585151672,
      "step": 30
    },
    {
      "epoch": 0.11965811965811966,
      "grad_norm": 31.455117829218544,
      "learning_rate": 9.979451208349055e-07,
      "logits/chosen": -2.2608728408813477,
      "logits/rejected": -2.246007204055786,
      "logps/chosen": -171.71456909179688,
      "logps/rejected": -174.46578979492188,
      "loss": 0.6305,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -0.01912705972790718,
      "rewards/margins": 0.31441593170166016,
      "rewards/rejected": -0.33354294300079346,
      "step": 35
    },
    {
      "epoch": 0.13675213675213677,
      "grad_norm": 31.67318837058587,
      "learning_rate": 9.970418664264595e-07,
      "logits/chosen": -2.345672130584717,
      "logits/rejected": -2.331491470336914,
      "logps/chosen": -171.24766540527344,
      "logps/rejected": -176.8189697265625,
      "loss": 0.5989,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.27867692708969116,
      "rewards/margins": 0.5290472507476807,
      "rewards/rejected": -0.8077241778373718,
      "step": 40
    },
    {
      "epoch": 0.13675213675213677,
      "eval_logits/chosen": -2.4102065563201904,
      "eval_logits/rejected": -2.401230573654175,
      "eval_logps/chosen": -162.36439514160156,
      "eval_logps/rejected": -167.4954071044922,
      "eval_loss": 0.6069236993789673,
      "eval_rewards/accuracies": 0.6365384459495544,
      "eval_rewards/chosen": -0.388705849647522,
      "eval_rewards/margins": 0.47280558943748474,
      "eval_rewards/rejected": -0.8615114688873291,
      "eval_runtime": 509.918,
      "eval_samples_per_second": 16.305,
      "eval_steps_per_second": 0.255,
      "step": 40
    }
  ],
  "logging_steps": 5,
  "max_steps": 876,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 40,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 471248375119872.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}