{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9960474308300395,
  "eval_steps": 500,
  "global_step": 189,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 26.908694644642612,
      "learning_rate": 2.6315789473684208e-08,
      "logits/chosen": -4.638427734375,
      "logits/rejected": -4.891327857971191,
      "logps/chosen": -198.52749633789062,
      "logps/rejected": -147.3392791748047,
      "loss": 0.6929,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.05,
      "grad_norm": 27.07664091604203,
      "learning_rate": 2.631578947368421e-07,
      "logits/chosen": -4.496448993682861,
      "logits/rejected": -4.815927028656006,
      "logps/chosen": -224.28125,
      "logps/rejected": -167.94735717773438,
      "loss": 0.6915,
      "rewards/accuracies": 0.5034722089767456,
      "rewards/chosen": 0.0034646072890609503,
      "rewards/margins": 0.002639756305143237,
      "rewards/rejected": 0.0008248506928794086,
      "step": 10
    },
    {
      "epoch": 0.11,
      "grad_norm": 26.30402064096193,
      "learning_rate": 4.999573126145131e-07,
      "logits/chosen": -4.625959873199463,
      "logits/rejected": -4.94482421875,
      "logps/chosen": -231.04525756835938,
      "logps/rejected": -196.3661651611328,
      "loss": 0.6465,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.04837086424231529,
      "rewards/margins": 0.11359457671642303,
      "rewards/rejected": -0.16196544468402863,
      "step": 20
    },
    {
      "epoch": 0.16,
      "grad_norm": 25.70165553073792,
      "learning_rate": 4.948524419003415e-07,
      "logits/chosen": -4.869608402252197,
      "logits/rejected": -5.148451805114746,
      "logps/chosen": -273.7060241699219,
      "logps/rejected": -259.2108154296875,
      "loss": 0.5717,
      "rewards/accuracies": 0.778124988079071,
      "rewards/chosen": -0.5196550488471985,
      "rewards/margins": 0.3445150554180145,
      "rewards/rejected": -0.8641700744628906,
      "step": 30
    },
    {
      "epoch": 0.21,
      "grad_norm": 32.38040367732233,
      "learning_rate": 4.81409414945389e-07,
      "logits/chosen": -4.95624303817749,
      "logits/rejected": -5.334275245666504,
      "logps/chosen": -321.26739501953125,
      "logps/rejected": -317.9222106933594,
      "loss": 0.5311,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.9283856153488159,
      "rewards/margins": 0.48562851548194885,
      "rewards/rejected": -1.4140141010284424,
      "step": 40
    },
    {
      "epoch": 0.26,
      "grad_norm": 27.107807886309228,
      "learning_rate": 4.6008601790947314e-07,
      "logits/chosen": -5.323241233825684,
      "logits/rejected": -5.817015171051025,
      "logps/chosen": -357.8787536621094,
      "logps/rejected": -385.47576904296875,
      "loss": 0.4831,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -1.411299467086792,
      "rewards/margins": 0.7530988454818726,
      "rewards/rejected": -2.164398193359375,
      "step": 50
    },
    {
      "epoch": 0.32,
      "grad_norm": 32.232061879934236,
      "learning_rate": 4.3160839350405605e-07,
      "logits/chosen": -5.831389904022217,
      "logits/rejected": -6.2499542236328125,
      "logps/chosen": -395.7707824707031,
      "logps/rejected": -446.3265686035156,
      "loss": 0.4294,
      "rewards/accuracies": 0.778124988079071,
      "rewards/chosen": -1.821434736251831,
      "rewards/margins": 0.9432821273803711,
      "rewards/rejected": -2.7647171020507812,
      "step": 60
    },
    {
      "epoch": 0.37,
      "grad_norm": 35.03072007251475,
      "learning_rate": 3.9694631307311825e-07,
      "logits/chosen": -6.090306758880615,
      "logits/rejected": -6.541258335113525,
      "logps/chosen": -430.2369689941406,
      "logps/rejected": -496.2119140625,
      "loss": 0.424,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -2.1488282680511475,
      "rewards/margins": 1.0692826509475708,
      "rewards/rejected": -3.218111038208008,
      "step": 70
    },
    {
      "epoch": 0.42,
      "grad_norm": 30.667469826354093,
      "learning_rate": 3.572801521931522e-07,
      "logits/chosen": -6.3887619972229,
      "logits/rejected": -6.877404689788818,
      "logps/chosen": -439.2911071777344,
      "logps/rejected": -526.5487060546875,
      "loss": 0.4001,
      "rewards/accuracies": 0.815625011920929,
      "rewards/chosen": -2.3115358352661133,
      "rewards/margins": 1.2141239643096924,
      "rewards/rejected": -3.5256600379943848,
      "step": 80
    },
    {
      "epoch": 0.47,
      "grad_norm": 37.36819911889553,
      "learning_rate": 3.139606943986089e-07,
      "logits/chosen": -6.5696258544921875,
      "logits/rejected": -7.1035637855529785,
      "logps/chosen": -458.3387756347656,
      "logps/rejected": -556.1650390625,
      "loss": 0.3875,
      "rewards/accuracies": 0.7718750238418579,
      "rewards/chosen": -2.5067451000213623,
      "rewards/margins": 1.3494058847427368,
      "rewards/rejected": -3.8561508655548096,
      "step": 90
    },
    {
      "epoch": 0.53,
      "grad_norm": 33.15053822353323,
      "learning_rate": 2.684631318687185e-07,
      "logits/chosen": -6.621747016906738,
      "logits/rejected": -7.236710548400879,
      "logps/chosen": -467.0467834472656,
      "logps/rejected": -582.046142578125,
      "loss": 0.3867,
      "rewards/accuracies": 0.796875,
      "rewards/chosen": -2.4837827682495117,
      "rewards/margins": 1.5418504476547241,
      "rewards/rejected": -4.025633811950684,
      "step": 100
    },
    {
      "epoch": 0.58,
      "grad_norm": 37.45830028947681,
      "learning_rate": 2.2233682952712483e-07,
      "logits/chosen": -6.568659782409668,
      "logits/rejected": -7.284300327301025,
      "logps/chosen": -460.4766540527344,
      "logps/rejected": -578.6600341796875,
      "loss": 0.3771,
      "rewards/accuracies": 0.846875011920929,
      "rewards/chosen": -2.3609726428985596,
      "rewards/margins": 1.647943139076233,
      "rewards/rejected": -4.008915901184082,
      "step": 110
    },
    {
      "epoch": 0.63,
      "grad_norm": 33.8427535333109,
      "learning_rate": 1.7715256327766884e-07,
      "logits/chosen": -6.796021461486816,
      "logits/rejected": -7.497170925140381,
      "logps/chosen": -504.50543212890625,
      "logps/rejected": -621.22314453125,
      "loss": 0.3508,
      "rewards/accuracies": 0.815625011920929,
      "rewards/chosen": -2.883434295654297,
      "rewards/margins": 1.6248239278793335,
      "rewards/rejected": -4.50825834274292,
      "step": 120
    },
    {
      "epoch": 0.69,
      "grad_norm": 35.353347844932394,
      "learning_rate": 1.3444902911492174e-07,
      "logits/chosen": -6.833544731140137,
      "logits/rejected": -7.472651481628418,
      "logps/chosen": -521.9656372070312,
      "logps/rejected": -659.3110961914062,
      "loss": 0.3705,
      "rewards/accuracies": 0.840624988079071,
      "rewards/chosen": -2.989759922027588,
      "rewards/margins": 1.8119176626205444,
      "rewards/rejected": -4.801677227020264,
      "step": 130
    },
    {
      "epoch": 0.74,
      "grad_norm": 32.6045025544378,
      "learning_rate": 9.56804446775518e-08,
      "logits/chosen": -6.738868713378906,
      "logits/rejected": -7.498864650726318,
      "logps/chosen": -470.77337646484375,
      "logps/rejected": -584.4710083007812,
      "loss": 0.3591,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -2.558176040649414,
      "rewards/margins": 1.6057535409927368,
      "rewards/rejected": -4.163929462432861,
      "step": 140
    },
    {
      "epoch": 0.79,
      "grad_norm": 32.49183208247093,
      "learning_rate": 6.216702761078166e-08,
      "logits/chosen": -7.049106597900391,
      "logits/rejected": -7.772597312927246,
      "logps/chosen": -487.25726318359375,
      "logps/rejected": -619.6534423828125,
      "loss": 0.3576,
      "rewards/accuracies": 0.8343750238418579,
      "rewards/chosen": -2.82848858833313,
      "rewards/margins": 1.7726719379425049,
      "rewards/rejected": -4.601161003112793,
      "step": 150
    },
    {
      "epoch": 0.84,
      "grad_norm": 37.26747220029015,
      "learning_rate": 3.5050037137906885e-08,
      "logits/chosen": -6.9701337814331055,
      "logits/rejected": -7.731366157531738,
      "logps/chosen": -494.5716247558594,
      "logps/rejected": -623.4630737304688,
      "loss": 0.3502,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -2.750206470489502,
      "rewards/margins": 1.7980148792266846,
      "rewards/rejected": -4.548220634460449,
      "step": 160
    },
    {
      "epoch": 0.9,
      "grad_norm": 31.918546112926368,
      "learning_rate": 1.5252909846235894e-08,
      "logits/chosen": -7.007571220397949,
      "logits/rejected": -7.6982011795043945,
      "logps/chosen": -509.54388427734375,
      "logps/rejected": -666.7489624023438,
      "loss": 0.3631,
      "rewards/accuracies": 0.890625,
      "rewards/chosen": -2.9002063274383545,
      "rewards/margins": 2.0059866905212402,
      "rewards/rejected": -4.906193733215332,
      "step": 170
    },
    {
      "epoch": 0.95,
      "grad_norm": 29.32551345390984,
      "learning_rate": 3.4498131616493565e-09,
      "logits/chosen": -6.939836025238037,
      "logits/rejected": -7.576680660247803,
      "logps/chosen": -514.7128295898438,
      "logps/rejected": -656.9924926757812,
      "loss": 0.3518,
      "rewards/accuracies": 0.8031250238418579,
      "rewards/chosen": -2.932391881942749,
      "rewards/margins": 1.7769733667373657,
      "rewards/rejected": -4.709364891052246,
      "step": 180
    },
    {
      "epoch": 1.0,
      "step": 189,
      "total_flos": 0.0,
      "train_loss": 0.42979541909757746,
      "train_runtime": 5368.3646,
      "train_samples_per_second": 9.04,
      "train_steps_per_second": 0.035
    }
  ],
  "logging_steps": 10,
  "max_steps": 189,
  "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
}