{
  "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": 11.914646684817926,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.8434743881225586,
      "logits/rejected": -2.769583225250244,
      "logps/chosen": -178.4530487060547,
      "logps/pi_response": -97.4625244140625,
      "logps/ref_response": -97.4625244140625,
      "logps/rejected": -233.52891540527344,
      "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": 11.557322184557252,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.7549960613250732,
      "logits/rejected": -2.7299859523773193,
      "logps/chosen": -225.90611267089844,
      "logps/pi_response": -120.62787628173828,
      "logps/ref_response": -119.74172973632812,
      "logps/rejected": -265.8953857421875,
      "loss": 0.6849,
      "rewards/accuracies": 0.5694444179534912,
      "rewards/chosen": -0.03211740776896477,
      "rewards/margins": 0.02102608233690262,
      "rewards/rejected": -0.05314348638057709,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 25.738246347544813,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.799567937850952,
      "logits/rejected": -2.7495579719543457,
      "logps/chosen": -259.5981750488281,
      "logps/pi_response": -116.07987213134766,
      "logps/ref_response": -118.667724609375,
      "logps/rejected": -312.01434326171875,
      "loss": 0.6336,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.24834828078746796,
      "rewards/margins": 0.2923477292060852,
      "rewards/rejected": -0.5406960248947144,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 15.788983217983654,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7804512977600098,
      "logits/rejected": -2.748396635055542,
      "logps/chosen": -290.38043212890625,
      "logps/pi_response": -150.2501678466797,
      "logps/ref_response": -124.9267807006836,
      "logps/rejected": -359.52850341796875,
      "loss": 0.5845,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.4211110472679138,
      "rewards/margins": 0.5139009356498718,
      "rewards/rejected": -0.9350120425224304,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 18.22123218196684,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.6774611473083496,
      "logits/rejected": -2.6376829147338867,
      "logps/chosen": -272.3063659667969,
      "logps/pi_response": -152.3288116455078,
      "logps/ref_response": -115.91612243652344,
      "logps/rejected": -363.81439208984375,
      "loss": 0.5576,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.5798826217651367,
      "rewards/margins": 0.5639127492904663,
      "rewards/rejected": -1.1437952518463135,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 20.96171170173109,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7017366886138916,
      "logits/rejected": -2.6490087509155273,
      "logps/chosen": -262.7826843261719,
      "logps/pi_response": -143.08602905273438,
      "logps/ref_response": -106.76103210449219,
      "logps/rejected": -336.96575927734375,
      "loss": 0.5617,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.5508919358253479,
      "rewards/margins": 0.521659791469574,
      "rewards/rejected": -1.0725517272949219,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.598152015168788,
      "train_runtime": 1828.3316,
      "train_samples_per_second": 8.359,
      "train_steps_per_second": 0.032
    }
  ],
  "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
}