{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.888888888888889,
  "eval_steps": 50,
  "global_step": 39,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.37037037037037035,
      "grad_norm": 60.71038553758777,
      "learning_rate": 5e-07,
      "logits/chosen": -2.741605281829834,
      "logits/rejected": -2.7256243228912354,
      "logps/chosen": -245.3419952392578,
      "logps/rejected": -209.0109405517578,
      "loss": 0.6899,
      "rewards/accuracies": 0.3062500059604645,
      "rewards/chosen": 0.012878472916781902,
      "rewards/margins": 0.004363563377410173,
      "rewards/rejected": 0.008514910005033016,
      "step": 5
    },
    {
      "epoch": 0.7407407407407407,
      "grad_norm": 43.98222995029514,
      "learning_rate": 1e-06,
      "logits/chosen": -2.688591957092285,
      "logits/rejected": -2.6561312675476074,
      "logps/chosen": -280.5462951660156,
      "logps/rejected": -212.30868530273438,
      "loss": 0.6378,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.6305345296859741,
      "rewards/margins": 0.3155175745487213,
      "rewards/rejected": 0.3150169253349304,
      "step": 10
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 33.66088840001642,
      "learning_rate": 9.284285880837946e-07,
      "logits/chosen": -2.5620155334472656,
      "logits/rejected": -2.537381410598755,
      "logps/chosen": -232.41506958007812,
      "logps/rejected": -197.28775024414062,
      "loss": 0.585,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 1.4362003803253174,
      "rewards/margins": 0.5606983304023743,
      "rewards/rejected": 0.8755022287368774,
      "step": 15
    },
    {
      "epoch": 1.4814814814814814,
      "grad_norm": 25.573173438222078,
      "learning_rate": 7.342042203498951e-07,
      "logits/chosen": -2.4830667972564697,
      "logits/rejected": -2.4438910484313965,
      "logps/chosen": -245.80569458007812,
      "logps/rejected": -222.90194702148438,
      "loss": 0.3393,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 2.007744073867798,
      "rewards/margins": 2.3653132915496826,
      "rewards/rejected": -0.35756915807724,
      "step": 20
    },
    {
      "epoch": 1.8518518518518519,
      "grad_norm": 25.974140482052572,
      "learning_rate": 4.7293054570729126e-07,
      "logits/chosen": -2.445605754852295,
      "logits/rejected": -2.449216604232788,
      "logps/chosen": -227.45388793945312,
      "logps/rejected": -197.3377685546875,
      "loss": 0.2707,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 2.0573277473449707,
      "rewards/margins": 2.167480945587158,
      "rewards/rejected": -0.11015312373638153,
      "step": 25
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 13.846363309737754,
      "learning_rate": 2.1940646731880885e-07,
      "logits/chosen": -2.3977177143096924,
      "logits/rejected": -2.4030890464782715,
      "logps/chosen": -224.091796875,
      "logps/rejected": -220.98715209960938,
      "loss": 0.2184,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 2.3001487255096436,
      "rewards/margins": 3.433748722076416,
      "rewards/rejected": -1.1335997581481934,
      "step": 30
    },
    {
      "epoch": 2.5925925925925926,
      "grad_norm": 12.384491012990875,
      "learning_rate": 4.621229016452155e-08,
      "logits/chosen": -2.3934264183044434,
      "logits/rejected": -2.3859307765960693,
      "logps/chosen": -241.48324584960938,
      "logps/rejected": -235.92514038085938,
      "loss": 0.1547,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 2.813878297805786,
      "rewards/margins": 3.7714927196502686,
      "rewards/rejected": -0.9576143026351929,
      "step": 35
    },
    {
      "epoch": 2.888888888888889,
      "step": 39,
      "total_flos": 459455468666880.0,
      "train_loss": 0.3865767228297698,
      "train_runtime": 1711.4819,
      "train_samples_per_second": 5.993,
      "train_steps_per_second": 0.023
    }
  ],
  "logging_steps": 5,
  "max_steps": 39,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 459455468666880.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}