{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9874476987447699,
  "eval_steps": 500,
  "global_step": 59,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016736401673640166,
      "grad_norm": 54.77672900099115,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.5891661643981934,
      "logits/rejected": -2.6211607456207275,
      "logps/chosen": -235.66262817382812,
      "logps/pi_response": -137.99111938476562,
      "logps/ref_response": -137.99111938476562,
      "logps/rejected": -449.43157958984375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16736401673640167,
      "grad_norm": 28.748939461156713,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.614354133605957,
      "logits/rejected": -2.564743757247925,
      "logps/chosen": -256.384521484375,
      "logps/pi_response": -133.1466064453125,
      "logps/ref_response": -117.319091796875,
      "logps/rejected": -412.2139587402344,
      "loss": 0.6516,
      "rewards/accuracies": 0.5972222089767456,
      "rewards/chosen": -0.23861868679523468,
      "rewards/margins": 0.14193283021450043,
      "rewards/rejected": -0.3805515468120575,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 43.6088024700336,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.4717307090759277,
      "logits/rejected": -2.40873122215271,
      "logps/chosen": -349.28350830078125,
      "logps/pi_response": -157.78677368164062,
      "logps/ref_response": -126.720947265625,
      "logps/rejected": -584.6751098632812,
      "loss": 0.5726,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": -1.033881425857544,
      "rewards/margins": 0.9884095191955566,
      "rewards/rejected": -2.0222911834716797,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 26.405027337696136,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.3895068168640137,
      "logits/rejected": -2.3256490230560303,
      "logps/chosen": -333.0101623535156,
      "logps/pi_response": -143.91600036621094,
      "logps/ref_response": -122.86802673339844,
      "logps/rejected": -536.3897705078125,
      "loss": 0.4988,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.7796052694320679,
      "rewards/margins": 0.8185471296310425,
      "rewards/rejected": -1.5981523990631104,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 37.15297987623143,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.389011859893799,
      "logits/rejected": -2.298595905303955,
      "logps/chosen": -318.46966552734375,
      "logps/pi_response": -141.33489990234375,
      "logps/ref_response": -125.3458023071289,
      "logps/rejected": -560.50244140625,
      "loss": 0.4834,
      "rewards/accuracies": 0.765625,
      "rewards/chosen": -0.6531937718391418,
      "rewards/margins": 0.8870981931686401,
      "rewards/rejected": -1.5402920246124268,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 30.111485179213247,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.350783586502075,
      "logits/rejected": -2.2837891578674316,
      "logps/chosen": -334.4332275390625,
      "logps/pi_response": -146.03616333007812,
      "logps/ref_response": -125.48101806640625,
      "logps/rejected": -562.9764404296875,
      "loss": 0.485,
      "rewards/accuracies": 0.7593749761581421,
      "rewards/chosen": -0.7923403978347778,
      "rewards/margins": 0.860241711139679,
      "rewards/rejected": -1.6525821685791016,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5293374061584473,
      "train_runtime": 2648.5577,
      "train_samples_per_second": 5.77,
      "train_steps_per_second": 0.022
    }
  ],
  "logging_steps": 10,
  "max_steps": 59,
  "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
}