{
  "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": 56.26818620363976,
      "learning_rate": 5e-08,
      "logits/chosen": -2.545705795288086,
      "logits/rejected": -2.5019917488098145,
      "logps/chosen": -274.9046630859375,
      "logps/pi_response": -169.24205017089844,
      "logps/ref_response": -169.24205017089844,
      "logps/rejected": -458.52459716796875,
      "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": 54.427899684240565,
      "learning_rate": 2.9580343711206163e-07,
      "logits/chosen": -2.5388073921203613,
      "logits/rejected": -2.4851911067962646,
      "logps/chosen": -286.4378662109375,
      "logps/pi_response": -156.18502807617188,
      "logps/ref_response": -150.80105590820312,
      "logps/rejected": -439.7905578613281,
      "loss": 0.6708,
      "rewards/accuracies": 0.5729166865348816,
      "rewards/chosen": -0.10314265638589859,
      "rewards/margins": 0.07712497562170029,
      "rewards/rejected": -0.18026763200759888,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 54.980640270066274,
      "learning_rate": 2.512474502277316e-07,
      "logits/chosen": -2.4677674770355225,
      "logits/rejected": -2.4007956981658936,
      "logps/chosen": -322.31732177734375,
      "logps/pi_response": -178.76998901367188,
      "logps/ref_response": -160.28256225585938,
      "logps/rejected": -552.8850708007812,
      "loss": 0.5572,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.5668123364448547,
      "rewards/margins": 0.6728711128234863,
      "rewards/rejected": -1.2396835088729858,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 36.46322338041782,
      "learning_rate": 1.721469846003722e-07,
      "logits/chosen": -2.331505060195923,
      "logits/rejected": -2.259660243988037,
      "logps/chosen": -406.34173583984375,
      "logps/pi_response": -196.53720092773438,
      "logps/ref_response": -152.65841674804688,
      "logps/rejected": -605.3541259765625,
      "loss": 0.5347,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -1.1331523656845093,
      "rewards/margins": 0.8151086568832397,
      "rewards/rejected": -1.9482609033584595,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 36.52867903824387,
      "learning_rate": 8.549021965852197e-08,
      "logits/chosen": -2.3737633228302,
      "logits/rejected": -2.285168170928955,
      "logps/chosen": -365.64825439453125,
      "logps/pi_response": -197.98513793945312,
      "logps/ref_response": -156.18771362304688,
      "logps/rejected": -638.6150512695312,
      "loss": 0.5065,
      "rewards/accuracies": 0.7593749761581421,
      "rewards/chosen": -0.9340522885322571,
      "rewards/margins": 0.991562008857727,
      "rewards/rejected": -1.9256141185760498,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 34.881990414936325,
      "learning_rate": 2.0843458234896666e-08,
      "logits/chosen": -2.3647232055664062,
      "logits/rejected": -2.3200135231018066,
      "logps/chosen": -372.10552978515625,
      "logps/pi_response": -202.2058868408203,
      "logps/ref_response": -155.8397674560547,
      "logps/rejected": -605.3728637695312,
      "loss": 0.5094,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.969250500202179,
      "rewards/margins": 0.7957654595375061,
      "rewards/rejected": -1.765015959739685,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5495932223433155,
      "train_runtime": 2619.5522,
      "train_samples_per_second": 5.834,
      "train_steps_per_second": 0.023
    }
  ],
  "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
}