{
  "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": 5.777860228892019,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.832916021347046,
      "logits/rejected": -2.8954272270202637,
      "logps/chosen": -112.55461120605469,
      "logps/pi_response": -112.97522735595703,
      "logps/ref_response": -112.97522735595703,
      "logps/rejected": -126.53972625732422,
      "loss": 0.6918,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16736401673640167,
      "grad_norm": 5.866000234105043,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.791806697845459,
      "logits/rejected": -2.7901265621185303,
      "logps/chosen": -150.01193237304688,
      "logps/pi_response": -152.32586669921875,
      "logps/ref_response": -152.45172119140625,
      "logps/rejected": -154.6370849609375,
      "loss": 0.6917,
      "rewards/accuracies": 0.4305555522441864,
      "rewards/chosen": 0.0006289934390224516,
      "rewards/margins": -0.0005195397534407675,
      "rewards/rejected": 0.0011485331924632192,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.174800184771954,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.8215346336364746,
      "logits/rejected": -2.806915760040283,
      "logps/chosen": -148.59841918945312,
      "logps/pi_response": -149.57443237304688,
      "logps/ref_response": -147.01174926757812,
      "logps/rejected": -151.08834838867188,
      "loss": 0.692,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": -0.02874668501317501,
      "rewards/margins": -0.003781392704695463,
      "rewards/rejected": -0.02496529184281826,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 6.166347180577358,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.8225722312927246,
      "logits/rejected": -2.8092432022094727,
      "logps/chosen": -168.7109832763672,
      "logps/pi_response": -166.7081298828125,
      "logps/ref_response": -149.63009643554688,
      "logps/rejected": -168.52523803710938,
      "loss": 0.6913,
      "rewards/accuracies": 0.4937500059604645,
      "rewards/chosen": -0.17003677785396576,
      "rewards/margins": -0.0003072673571296036,
      "rewards/rejected": -0.16972950100898743,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 6.333082712906178,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.7256391048431396,
      "logits/rejected": -2.719358205795288,
      "logps/chosen": -158.1101837158203,
      "logps/pi_response": -156.38528442382812,
      "logps/ref_response": -143.953857421875,
      "logps/rejected": -155.69100952148438,
      "loss": 0.69,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.11886920034885406,
      "rewards/margins": 0.013960828073322773,
      "rewards/rejected": -0.13283005356788635,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 6.029091338880282,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7439560890197754,
      "logits/rejected": -2.743741750717163,
      "logps/chosen": -149.5755615234375,
      "logps/pi_response": -147.55282592773438,
      "logps/ref_response": -140.59339904785156,
      "logps/rejected": -145.41294860839844,
      "loss": 0.69,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.06889279931783676,
      "rewards/margins": 0.005167593248188496,
      "rewards/rejected": -0.07406039535999298,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6904175948288481,
      "train_runtime": 1301.3878,
      "train_samples_per_second": 11.744,
      "train_steps_per_second": 0.045
    }
  ],
  "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": 4,
  "trial_name": null,
  "trial_params": null
}