{
  "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": 50.84480396591844,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.5962367057800293,
      "logits/rejected": -2.5291247367858887,
      "logps/chosen": -313.06549072265625,
      "logps/pi_response": -192.275634765625,
      "logps/ref_response": -192.275634765625,
      "logps/rejected": -467.38446044921875,
      "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": 42.41677392718424,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.5569324493408203,
      "logits/rejected": -2.5175046920776367,
      "logps/chosen": -302.5186767578125,
      "logps/pi_response": -169.27621459960938,
      "logps/ref_response": -160.4980010986328,
      "logps/rejected": -473.1149597167969,
      "loss": 0.6544,
      "rewards/accuracies": 0.5694444179534912,
      "rewards/chosen": -0.23846334218978882,
      "rewards/margins": 0.1443914771080017,
      "rewards/rejected": -0.3828548491001129,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 39.98829098029278,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.4396791458129883,
      "logits/rejected": -2.3597466945648193,
      "logps/chosen": -386.8421325683594,
      "logps/pi_response": -202.8572998046875,
      "logps/ref_response": -167.9676055908203,
      "logps/rejected": -638.3150024414062,
      "loss": 0.5895,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -1.0396512746810913,
      "rewards/margins": 0.9388242959976196,
      "rewards/rejected": -1.978475570678711,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 33.31734090524923,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.3278262615203857,
      "logits/rejected": -2.257200241088867,
      "logps/chosen": -412.71881103515625,
      "logps/pi_response": -218.89254760742188,
      "logps/ref_response": -161.32485961914062,
      "logps/rejected": -613.4661865234375,
      "loss": 0.5116,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": -1.1704436540603638,
      "rewards/margins": 0.7398843169212341,
      "rewards/rejected": -1.9103281497955322,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 29.897018751189872,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.344947338104248,
      "logits/rejected": -2.2430291175842285,
      "logps/chosen": -405.5809020996094,
      "logps/pi_response": -244.9563751220703,
      "logps/ref_response": -167.4927520751953,
      "logps/rejected": -667.6639404296875,
      "loss": 0.4908,
      "rewards/accuracies": 0.784375011920929,
      "rewards/chosen": -1.212226390838623,
      "rewards/margins": 0.8576709032058716,
      "rewards/rejected": -2.069897413253784,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 33.345127493561584,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.3016886711120605,
      "logits/rejected": -2.236023187637329,
      "logps/chosen": -455.349365234375,
      "logps/pi_response": -263.25799560546875,
      "logps/ref_response": -165.34706115722656,
      "logps/rejected": -681.8700561523438,
      "loss": 0.4903,
      "rewards/accuracies": 0.7718750238418579,
      "rewards/chosen": -1.555412769317627,
      "rewards/margins": 0.8201801180839539,
      "rewards/rejected": -2.3755927085876465,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5340633715613413,
      "train_runtime": 2647.515,
      "train_samples_per_second": 5.773,
      "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
}