{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.992,
  "eval_steps": 100,
  "global_step": 62,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 7.120798164408008,
      "learning_rate": 7.142857142857142e-08,
      "logits/chosen": -1.457259178161621,
      "logits/rejected": -1.0621511936187744,
      "logps/chosen": -272.0050354003906,
      "logps/rejected": -816.85107421875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16,
      "grad_norm": 6.902950407858967,
      "learning_rate": 4.963384589619232e-07,
      "logits/chosen": -1.310307502746582,
      "logits/rejected": -1.0846761465072632,
      "logps/chosen": -535.0698852539062,
      "logps/rejected": -898.6776733398438,
      "loss": 0.6917,
      "rewards/accuracies": 0.4444444477558136,
      "rewards/chosen": -0.001652209903113544,
      "rewards/margins": 0.001886376878246665,
      "rewards/rejected": -0.003538586664944887,
      "step": 10
    },
    {
      "epoch": 0.32,
      "grad_norm": 6.3100220055883565,
      "learning_rate": 4.341852844691012e-07,
      "logits/chosen": -1.3299431800842285,
      "logits/rejected": -1.0885355472564697,
      "logps/chosen": -514.69140625,
      "logps/rejected": -881.6802978515625,
      "loss": 0.6634,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.011395025067031384,
      "rewards/margins": 0.07305588573217392,
      "rewards/rejected": -0.08445090055465698,
      "step": 20
    },
    {
      "epoch": 0.48,
      "grad_norm": 7.330007892766956,
      "learning_rate": 3.135545835483718e-07,
      "logits/chosen": -1.2920414209365845,
      "logits/rejected": -1.104931116104126,
      "logps/chosen": -543.4114990234375,
      "logps/rejected": -906.7586059570312,
      "loss": 0.593,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -0.043302834033966064,
      "rewards/margins": 0.253420889377594,
      "rewards/rejected": -0.29672369360923767,
      "step": 30
    },
    {
      "epoch": 0.64,
      "grad_norm": 7.8988027878682825,
      "learning_rate": 1.7274575140626315e-07,
      "logits/chosen": -1.2756078243255615,
      "logits/rejected": -1.09669029712677,
      "logps/chosen": -593.3638305664062,
      "logps/rejected": -903.4348754882812,
      "loss": 0.5274,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": -0.08866459131240845,
      "rewards/margins": 0.4585743844509125,
      "rewards/rejected": -0.5472390651702881,
      "step": 40
    },
    {
      "epoch": 0.8,
      "grad_norm": 7.092952620264841,
      "learning_rate": 5.6464597340229375e-08,
      "logits/chosen": -1.1793540716171265,
      "logits/rejected": -1.0933005809783936,
      "logps/chosen": -541.88671875,
      "logps/rejected": -1034.5299072265625,
      "loss": 0.4638,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -0.13559284806251526,
      "rewards/margins": 0.9057224988937378,
      "rewards/rejected": -1.0413153171539307,
      "step": 50
    },
    {
      "epoch": 0.96,
      "grad_norm": 7.683839890605678,
      "learning_rate": 1.6295661628624447e-09,
      "logits/chosen": -1.1111315488815308,
      "logits/rejected": -1.098163366317749,
      "logps/chosen": -511.08526611328125,
      "logps/rejected": -985.9429931640625,
      "loss": 0.4288,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.20404699444770813,
      "rewards/margins": 0.9808656573295593,
      "rewards/rejected": -1.1849125623703003,
      "step": 60
    },
    {
      "epoch": 0.99,
      "step": 62,
      "total_flos": 0.0,
      "train_loss": 0.5556549103029312,
      "train_runtime": 1091.242,
      "train_samples_per_second": 3.642,
      "train_steps_per_second": 0.057
    }
  ],
  "logging_steps": 10,
  "max_steps": 62,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}