{
  "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": 11.057634185929018,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.832641363143921,
      "logits/rejected": -2.7308225631713867,
      "logps/chosen": -194.5078887939453,
      "logps/pi_response": -109.15676879882812,
      "logps/ref_response": -109.15676879882812,
      "logps/rejected": -199.842041015625,
      "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": 15.370165930636123,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.751119613647461,
      "logits/rejected": -2.7138187885284424,
      "logps/chosen": -213.61614990234375,
      "logps/pi_response": -115.39461517333984,
      "logps/ref_response": -114.79134368896484,
      "logps/rejected": -267.3037109375,
      "loss": 0.6847,
      "rewards/accuracies": 0.5763888955116272,
      "rewards/chosen": -0.03307180106639862,
      "rewards/margins": 0.02672014944255352,
      "rewards/rejected": -0.05979194864630699,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 30.006490022587922,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.7758188247680664,
      "logits/rejected": -2.7319676876068115,
      "logps/chosen": -258.4744567871094,
      "logps/pi_response": -113.13661193847656,
      "logps/ref_response": -115.21757507324219,
      "logps/rejected": -313.4935302734375,
      "loss": 0.6311,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -0.2657850384712219,
      "rewards/margins": 0.28187114000320435,
      "rewards/rejected": -0.5476561784744263,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 17.86539531293398,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7661945819854736,
      "logits/rejected": -2.7283473014831543,
      "logps/chosen": -281.109130859375,
      "logps/pi_response": -136.8625030517578,
      "logps/ref_response": -118.77275085449219,
      "logps/rejected": -353.06475830078125,
      "loss": 0.5838,
      "rewards/accuracies": 0.7093750238418579,
      "rewards/chosen": -0.4710591733455658,
      "rewards/margins": 0.4681057035923004,
      "rewards/rejected": -0.9391648173332214,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 16.943629934390696,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.6898460388183594,
      "logits/rejected": -2.651970624923706,
      "logps/chosen": -281.537841796875,
      "logps/pi_response": -143.15576171875,
      "logps/ref_response": -110.78315734863281,
      "logps/rejected": -359.834716796875,
      "loss": 0.5584,
      "rewards/accuracies": 0.6781250238418579,
      "rewards/chosen": -0.5756246447563171,
      "rewards/margins": 0.49899721145629883,
      "rewards/rejected": -1.0746219158172607,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 16.37436004148144,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.6713178157806396,
      "logits/rejected": -2.6225244998931885,
      "logps/chosen": -272.2515869140625,
      "logps/pi_response": -146.99404907226562,
      "logps/ref_response": -112.6738510131836,
      "logps/rejected": -340.9408874511719,
      "loss": 0.5611,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": -0.5049266815185547,
      "rewards/margins": 0.4622128903865814,
      "rewards/rejected": -0.9671396017074585,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5940128989138845,
      "train_runtime": 2688.5086,
      "train_samples_per_second": 5.685,
      "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
}