{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 385,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 262.0,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 1.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 282.0,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.07872009277344,
      "logits/rejected": 80.78690338134766,
      "logps/chosen": -34.225643157958984,
      "logps/rejected": -32.97146987915039,
      "loss": 1.4705,
      "rewards/accuracies": 0.4027777910232544,
      "rewards/chosen": 0.00803940836340189,
      "rewards/margins": 0.01457094494253397,
      "rewards/rejected": -0.00653153657913208,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 316.0,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.64378356933594,
      "logits/rejected": 80.53114318847656,
      "logps/chosen": -33.554771423339844,
      "logps/rejected": -30.747882843017578,
      "loss": 1.6639,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.06756959110498428,
      "rewards/margins": 0.05276557803153992,
      "rewards/rejected": 0.014804007485508919,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 209.0,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.3618392944336,
      "logits/rejected": 82.39045715332031,
      "logps/chosen": -33.964210510253906,
      "logps/rejected": -31.346271514892578,
      "loss": 1.9077,
      "rewards/accuracies": 0.375,
      "rewards/chosen": 0.02430456317961216,
      "rewards/margins": -0.01447761058807373,
      "rewards/rejected": 0.03878217190504074,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 708.0,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 80.66716003417969,
      "logits/rejected": 80.66474151611328,
      "logps/chosen": -33.04372787475586,
      "logps/rejected": -33.376953125,
      "loss": 2.3531,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.06516679376363754,
      "rewards/margins": 0.06938251852989197,
      "rewards/rejected": -0.004215720109641552,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 684.0,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.12054443359375,
      "logits/rejected": 78.13239288330078,
      "logps/chosen": -31.36956787109375,
      "logps/rejected": -31.187658309936523,
      "loss": 2.5686,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.08215981721878052,
      "rewards/margins": -0.024972762912511826,
      "rewards/rejected": -0.057187050580978394,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 326.0,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 82.62169647216797,
      "logits/rejected": 82.67021179199219,
      "logps/chosen": -31.33014488220215,
      "logps/rejected": -29.82718849182129,
      "loss": 2.2332,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": -0.0774935632944107,
      "rewards/margins": 0.04084175080060959,
      "rewards/rejected": -0.11833532154560089,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 796.0,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 83.52657318115234,
      "logits/rejected": 83.54931640625,
      "logps/chosen": -30.717477798461914,
      "logps/rejected": -32.99928665161133,
      "loss": 2.7176,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.04422692582011223,
      "rewards/margins": -0.06343810260295868,
      "rewards/rejected": 0.019211167469620705,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 342.0,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.1966781616211,
      "logits/rejected": 81.19418334960938,
      "logps/chosen": -31.369304656982422,
      "logps/rejected": -30.800960540771484,
      "loss": 2.4943,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.06781856715679169,
      "rewards/margins": 0.08947588503360748,
      "rewards/rejected": -0.021657321602106094,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 502.0,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.3118896484375,
      "logits/rejected": 78.27383422851562,
      "logps/chosen": -32.46575164794922,
      "logps/rejected": -31.248449325561523,
      "loss": 2.2351,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.08621449768543243,
      "rewards/margins": 0.2347041368484497,
      "rewards/rejected": -0.14848963916301727,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 296.0,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.64895629882812,
      "logits/rejected": 83.68805694580078,
      "logps/chosen": -34.173667907714844,
      "logps/rejected": -31.847728729248047,
      "loss": 2.7048,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.06566362082958221,
      "rewards/margins": 0.11425628513097763,
      "rewards/rejected": -0.04859266057610512,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.86346435546875,
      "eval_logits/rejected": 98.85167694091797,
      "eval_logps/chosen": -32.50230407714844,
      "eval_logps/rejected": -36.137577056884766,
      "eval_loss": 2.2451484203338623,
      "eval_rewards/accuracies": 0.5336378812789917,
      "eval_rewards/chosen": -0.02956419251859188,
      "eval_rewards/margins": 0.05596020445227623,
      "eval_rewards/rejected": -0.08552439510822296,
      "eval_runtime": 104.0782,
      "eval_samples_per_second": 3.296,
      "eval_steps_per_second": 0.413,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 492.0,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.73783874511719,
      "logits/rejected": 83.64616394042969,
      "logps/chosen": -32.79869842529297,
      "logps/rejected": -32.73226547241211,
      "loss": 2.701,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.11205414682626724,
      "rewards/margins": 0.15024258196353912,
      "rewards/rejected": -0.03818843513727188,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 324.0,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 84.05355072021484,
      "logits/rejected": 84.1574478149414,
      "logps/chosen": -28.724782943725586,
      "logps/rejected": -35.51475143432617,
      "loss": 2.094,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.10240070521831512,
      "rewards/margins": 0.13020212948322296,
      "rewards/rejected": -0.027801427990198135,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 185.0,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 81.37618255615234,
      "logits/rejected": 81.40374755859375,
      "logps/chosen": -30.55348777770996,
      "logps/rejected": -32.026222229003906,
      "loss": 2.2639,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.17693254351615906,
      "rewards/margins": 0.22423109412193298,
      "rewards/rejected": -0.04729856550693512,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 356.0,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 82.8284912109375,
      "logits/rejected": 82.82528686523438,
      "logps/chosen": -27.208200454711914,
      "logps/rejected": -32.791831970214844,
      "loss": 2.1202,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.11956217139959335,
      "rewards/margins": 0.3037911057472229,
      "rewards/rejected": -0.18422891199588776,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 376.0,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 81.57420349121094,
      "logits/rejected": 81.54073333740234,
      "logps/chosen": -29.004501342773438,
      "logps/rejected": -32.630916595458984,
      "loss": 2.179,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.19230429828166962,
      "rewards/margins": 0.22451286017894745,
      "rewards/rejected": -0.032208558171987534,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 464.0,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 83.5596923828125,
      "logits/rejected": 83.56761169433594,
      "logps/chosen": -33.89311981201172,
      "logps/rejected": -29.929208755493164,
      "loss": 3.3447,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.12181256711483002,
      "rewards/margins": 0.11014753580093384,
      "rewards/rejected": 0.011665038764476776,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 296.0,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 84.21192169189453,
      "logits/rejected": 84.15568542480469,
      "logps/chosen": -30.989816665649414,
      "logps/rejected": -32.166038513183594,
      "loss": 2.6535,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.13473577797412872,
      "rewards/margins": 0.20772309601306915,
      "rewards/rejected": -0.07298731058835983,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 520.0,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 81.9034423828125,
      "logits/rejected": 81.88355255126953,
      "logps/chosen": -30.792713165283203,
      "logps/rejected": -31.40665626525879,
      "loss": 2.561,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.20221999287605286,
      "rewards/margins": 0.3269447684288025,
      "rewards/rejected": -0.12472478300333023,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 181.0,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 83.61729431152344,
      "logits/rejected": 83.60554504394531,
      "logps/chosen": -30.603130340576172,
      "logps/rejected": -30.67363929748535,
      "loss": 2.4852,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.08164240419864655,
      "rewards/margins": 0.1844494342803955,
      "rewards/rejected": -0.10280702263116837,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 384.0,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 79.16850280761719,
      "logits/rejected": 79.1117935180664,
      "logps/chosen": -34.15711212158203,
      "logps/rejected": -32.46302032470703,
      "loss": 3.9434,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.2978156805038452,
      "rewards/margins": 0.28262609243392944,
      "rewards/rejected": 0.015189537778496742,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 98.95487213134766,
      "eval_logits/rejected": 98.95005798339844,
      "eval_logps/chosen": -32.4478645324707,
      "eval_logps/rejected": -36.15636444091797,
      "eval_loss": 1.9716209173202515,
      "eval_rewards/accuracies": 0.5930232405662537,
      "eval_rewards/chosen": -0.002344276988878846,
      "eval_rewards/margins": 0.09257492423057556,
      "eval_rewards/rejected": -0.09491920471191406,
      "eval_runtime": 104.0484,
      "eval_samples_per_second": 3.297,
      "eval_steps_per_second": 0.413,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 756.0,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 81.85539245605469,
      "logits/rejected": 81.76030731201172,
      "logps/chosen": -33.80899429321289,
      "logps/rejected": -35.018226623535156,
      "loss": 2.8449,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.06836456805467606,
      "rewards/margins": 0.08442962169647217,
      "rewards/rejected": -0.01606505550444126,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 290.0,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 84.03166198730469,
      "logits/rejected": 84.112060546875,
      "logps/chosen": -31.503955841064453,
      "logps/rejected": -30.661306381225586,
      "loss": 3.4983,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.2224985659122467,
      "rewards/margins": 0.1886742264032364,
      "rewards/rejected": 0.03382434695959091,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 354.0,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 81.38267517089844,
      "logits/rejected": 81.43292999267578,
      "logps/chosen": -32.516029357910156,
      "logps/rejected": -33.79094314575195,
      "loss": 2.6207,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.1448572874069214,
      "rewards/margins": 0.034253668040037155,
      "rewards/rejected": 0.11060361564159393,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 364.0,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 83.76815795898438,
      "logits/rejected": 84.01641082763672,
      "logps/chosen": -31.310298919677734,
      "logps/rejected": -31.66178321838379,
      "loss": 2.0842,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.18888285756111145,
      "rewards/margins": 0.25375330448150635,
      "rewards/rejected": -0.06487040966749191,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 298.0,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 82.50977325439453,
      "logits/rejected": 82.58284759521484,
      "logps/chosen": -27.186697006225586,
      "logps/rejected": -29.944438934326172,
      "loss": 2.3496,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.20459958910942078,
      "rewards/margins": 0.22644969820976257,
      "rewards/rejected": -0.02185012958943844,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 976.0,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 79.92189025878906,
      "logits/rejected": 80.03739166259766,
      "logps/chosen": -30.968799591064453,
      "logps/rejected": -36.08448028564453,
      "loss": 2.944,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.29609745740890503,
      "rewards/margins": 0.28752270340919495,
      "rewards/rejected": 0.008574736304581165,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 197.0,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 79.33562469482422,
      "logits/rejected": 79.36061096191406,
      "logps/chosen": -30.985218048095703,
      "logps/rejected": -31.533437728881836,
      "loss": 2.6828,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.39685314893722534,
      "rewards/margins": 0.41086119413375854,
      "rewards/rejected": -0.014008039608597755,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 388.0,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 81.81623840332031,
      "logits/rejected": 81.6318588256836,
      "logps/chosen": -31.43568992614746,
      "logps/rejected": -29.53765296936035,
      "loss": 2.8578,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.1871013194322586,
      "rewards/margins": 0.12242940813302994,
      "rewards/rejected": 0.06467190384864807,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 270.0,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 82.0903549194336,
      "logits/rejected": 82.0211181640625,
      "logps/chosen": -33.648677825927734,
      "logps/rejected": -31.893665313720703,
      "loss": 2.8329,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.22778387367725372,
      "rewards/margins": 0.1661832630634308,
      "rewards/rejected": 0.061600614339113235,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 230.0,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 77.76834869384766,
      "logits/rejected": 77.82737731933594,
      "logps/chosen": -32.740943908691406,
      "logps/rejected": -28.960865020751953,
      "loss": 2.0991,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.33238038420677185,
      "rewards/margins": 0.26180073618888855,
      "rewards/rejected": 0.07057970762252808,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 99.00546264648438,
      "eval_logits/rejected": 98.99657440185547,
      "eval_logps/chosen": -32.35263442993164,
      "eval_logps/rejected": -35.902122497558594,
      "eval_loss": 2.260390281677246,
      "eval_rewards/accuracies": 0.48795682191848755,
      "eval_rewards/chosen": 0.045268867164850235,
      "eval_rewards/margins": 0.013066344894468784,
      "eval_rewards/rejected": 0.03220251947641373,
      "eval_runtime": 104.0306,
      "eval_samples_per_second": 3.297,
      "eval_steps_per_second": 0.413,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 640.0,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 84.65354919433594,
      "logits/rejected": 84.689208984375,
      "logps/chosen": -30.26632308959961,
      "logps/rejected": -32.187259674072266,
      "loss": 2.4257,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.2986040413379669,
      "rewards/margins": 0.2923537790775299,
      "rewards/rejected": 0.0062502650544047356,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 438.0,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 82.1827621459961,
      "logits/rejected": 82.17810821533203,
      "logps/chosen": -30.906885147094727,
      "logps/rejected": -28.876495361328125,
      "loss": 4.1182,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.3458386957645416,
      "rewards/margins": 0.3243403136730194,
      "rewards/rejected": 0.021498391404747963,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 336.0,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 79.60859680175781,
      "logits/rejected": 79.64774322509766,
      "logps/chosen": -29.61593246459961,
      "logps/rejected": -32.607086181640625,
      "loss": 2.7354,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.3683364987373352,
      "rewards/margins": 0.3544464707374573,
      "rewards/rejected": 0.013890000991523266,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 552.0,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 83.55511474609375,
      "logits/rejected": 83.56120300292969,
      "logps/chosen": -32.71057891845703,
      "logps/rejected": -33.46552276611328,
      "loss": 2.9727,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.2507854104042053,
      "rewards/margins": 0.2991998791694641,
      "rewards/rejected": -0.048414476215839386,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 416.0,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 82.80843353271484,
      "logits/rejected": 82.7990951538086,
      "logps/chosen": -32.887977600097656,
      "logps/rejected": -33.06817626953125,
      "loss": 2.7683,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.37787994742393494,
      "rewards/margins": 0.3660680055618286,
      "rewards/rejected": 0.011811929754912853,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 141.0,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 84.11600494384766,
      "logits/rejected": 84.12767791748047,
      "logps/chosen": -28.876087188720703,
      "logps/rejected": -31.597909927368164,
      "loss": 1.7283,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.3380628228187561,
      "rewards/margins": 0.3120931088924408,
      "rewards/rejected": 0.025969672948122025,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 402.0,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 83.5467529296875,
      "logits/rejected": 83.55477142333984,
      "logps/chosen": -32.44352340698242,
      "logps/rejected": -35.13766860961914,
      "loss": 3.1738,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.2163194715976715,
      "rewards/margins": 0.23046875,
      "rewards/rejected": -0.014149266295135021,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 322.0,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 77.60735321044922,
      "logits/rejected": 77.49246215820312,
      "logps/chosen": -30.158939361572266,
      "logps/rejected": -28.004974365234375,
      "loss": 2.7343,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.20859412848949432,
      "rewards/margins": 0.1307946741580963,
      "rewards/rejected": 0.07779943943023682,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 2.572123973400562,
      "train_runtime": 2556.3668,
      "train_samples_per_second": 1.204,
      "train_steps_per_second": 0.151
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "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
}