checkpoint-500/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.4103405826836274,
  "eval_steps": 50,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.008206811653672548,
      "grad_norm": 0.07778492569923401,
      "learning_rate": 4.999451708687114e-06,
      "logits/chosen": -2.053281307220459,
      "logits/rejected": -2.495474338531494,
      "logps/chosen": -0.3126755356788635,
      "logps/rejected": -0.3312620520591736,
      "loss": 7.6211,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": -0.4690132737159729,
      "rewards/margins": 0.027879873290657997,
      "rewards/rejected": -0.49689316749572754,
      "step": 10
    },
    {
      "epoch": 0.016413623307345096,
      "grad_norm": 0.07773654907941818,
      "learning_rate": 4.997807075247147e-06,
      "logits/chosen": -2.0624098777770996,
      "logits/rejected": -2.4424185752868652,
      "logps/chosen": -0.26926660537719727,
      "logps/rejected": -0.2978014051914215,
      "loss": 7.5195,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": -0.4038998484611511,
      "rewards/margins": 0.04280223697423935,
      "rewards/rejected": -0.44670209288597107,
      "step": 20
    },
    {
      "epoch": 0.024620434961017644,
      "grad_norm": 0.07357177883386612,
      "learning_rate": 4.9950668210706795e-06,
      "logits/chosen": -2.068427562713623,
      "logits/rejected": -2.486642360687256,
      "logps/chosen": -0.29993391036987305,
      "logps/rejected": -0.34360918402671814,
      "loss": 7.4913,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.4499008059501648,
      "rewards/margins": 0.06551288068294525,
      "rewards/rejected": -0.5154137015342712,
      "step": 30
    },
    {
      "epoch": 0.03282724661469019,
      "grad_norm": 0.14212799072265625,
      "learning_rate": 4.9912321481237616e-06,
      "logits/chosen": -2.015650987625122,
      "logits/rejected": -2.3838727474212646,
      "logps/chosen": -0.2911723852157593,
      "logps/rejected": -0.30521970987319946,
      "loss": 7.5217,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": -0.4367586076259613,
      "rewards/margins": 0.021070968359708786,
      "rewards/rejected": -0.4578295648097992,
      "step": 40
    },
    {
      "epoch": 0.04103405826836274,
      "grad_norm": 0.08107248693704605,
      "learning_rate": 4.986304738420684e-06,
      "logits/chosen": -2.1150989532470703,
      "logits/rejected": -2.4338631629943848,
      "logps/chosen": -0.26249754428863525,
      "logps/rejected": -0.3132360577583313,
      "loss": 7.519,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.3937462866306305,
      "rewards/margins": 0.07610772550106049,
      "rewards/rejected": -0.4698540270328522,
      "step": 50
    },
    {
      "epoch": 0.04103405826836274,
      "eval_logits/chosen": -2.0232737064361572,
      "eval_logits/rejected": -2.4952735900878906,
      "eval_logps/chosen": -0.27974528074264526,
      "eval_logps/rejected": -0.3420677185058594,
      "eval_loss": 0.9291417598724365,
      "eval_rewards/accuracies": 0.49494948983192444,
      "eval_rewards/chosen": -0.41961798071861267,
      "eval_rewards/margins": 0.09348361939191818,
      "eval_rewards/rejected": -0.5131015777587891,
      "eval_runtime": 26.0563,
      "eval_samples_per_second": 30.242,
      "eval_steps_per_second": 3.799,
      "step": 50
    },
    {
      "epoch": 0.04924086992203529,
      "grad_norm": 0.06815352290868759,
      "learning_rate": 4.980286753286196e-06,
      "logits/chosen": -1.9890680313110352,
      "logits/rejected": -2.3848204612731934,
      "logps/chosen": -0.26213228702545166,
      "logps/rejected": -0.31342557072639465,
      "loss": 7.432,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.3931984603404999,
      "rewards/margins": 0.0769399031996727,
      "rewards/rejected": -0.4701383709907532,
      "step": 60
    },
    {
      "epoch": 0.057447681575707836,
      "grad_norm": 0.06748568266630173,
      "learning_rate": 4.973180832407471e-06,
      "logits/chosen": -2.070542812347412,
      "logits/rejected": -2.3977038860321045,
      "logps/chosen": -0.24570491909980774,
      "logps/rejected": -0.3655605912208557,
      "loss": 7.35,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.3685573935508728,
      "rewards/margins": 0.17978355288505554,
      "rewards/rejected": -0.548340916633606,
      "step": 70
    },
    {
      "epoch": 0.06565449322938038,
      "grad_norm": 0.10909309983253479,
      "learning_rate": 4.964990092676263e-06,
      "logits/chosen": -2.2012317180633545,
      "logits/rejected": -2.346029758453369,
      "logps/chosen": -0.2279246598482132,
      "logps/rejected": -0.35396742820739746,
      "loss": 7.5082,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -0.341886967420578,
      "rewards/margins": 0.18906418979167938,
      "rewards/rejected": -0.5309511423110962,
      "step": 80
    },
    {
      "epoch": 0.07386130488305294,
      "grad_norm": 0.05977805703878403,
      "learning_rate": 4.9557181268217225e-06,
      "logits/chosen": -2.0719449520111084,
      "logits/rejected": -2.4491190910339355,
      "logps/chosen": -0.2503294348716736,
      "logps/rejected": -0.29939892888069153,
      "loss": 7.5129,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.37549418210983276,
      "rewards/margins": 0.07360419631004333,
      "rewards/rejected": -0.4490983486175537,
      "step": 90
    },
    {
      "epoch": 0.08206811653672548,
      "grad_norm": 0.051751479506492615,
      "learning_rate": 4.9453690018345144e-06,
      "logits/chosen": -2.0634045600891113,
      "logits/rejected": -2.458428382873535,
      "logps/chosen": -0.24033495783805847,
      "logps/rejected": -0.29080909490585327,
      "loss": 7.4432,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.3605024516582489,
      "rewards/margins": 0.07571124285459518,
      "rewards/rejected": -0.4362136721611023,
      "step": 100
    },
    {
      "epoch": 0.08206811653672548,
      "eval_logits/chosen": -2.0207154750823975,
      "eval_logits/rejected": -2.486215353012085,
      "eval_logps/chosen": -0.2376101016998291,
      "eval_logps/rejected": -0.32593628764152527,
      "eval_loss": 0.9085211753845215,
      "eval_rewards/accuracies": 0.5353535413742065,
      "eval_rewards/chosen": -0.35641518235206604,
      "eval_rewards/margins": 0.13248924911022186,
      "eval_rewards/rejected": -0.4889043867588043,
      "eval_runtime": 26.0119,
      "eval_samples_per_second": 30.294,
      "eval_steps_per_second": 3.806,
      "step": 100
    },
    {
      "epoch": 0.09027492819039803,
      "grad_norm": 0.06007291004061699,
      "learning_rate": 4.933947257182901e-06,
      "logits/chosen": -2.1248741149902344,
      "logits/rejected": -2.409808874130249,
      "logps/chosen": -0.2354653775691986,
      "logps/rejected": -0.30269068479537964,
      "loss": 7.317,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.3531980812549591,
      "rewards/margins": 0.10083796828985214,
      "rewards/rejected": -0.45403605699539185,
      "step": 110
    },
    {
      "epoch": 0.09848173984407058,
      "grad_norm": 0.055738095194101334,
      "learning_rate": 4.921457902821578e-06,
      "logits/chosen": -2.0635311603546143,
      "logits/rejected": -2.4297730922698975,
      "logps/chosen": -0.2315257489681244,
      "logps/rejected": -0.33639490604400635,
      "loss": 7.2775,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.3472886383533478,
      "rewards/margins": 0.15730372071266174,
      "rewards/rejected": -0.5045923590660095,
      "step": 120
    },
    {
      "epoch": 0.10668855149774313,
      "grad_norm": 0.07971248030662537,
      "learning_rate": 4.907906416994146e-06,
      "logits/chosen": -2.07852840423584,
      "logits/rejected": -2.4043469429016113,
      "logps/chosen": -0.20596058666706085,
      "logps/rejected": -0.33416762948036194,
      "loss": 7.336,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -0.3089408874511719,
      "rewards/margins": 0.19231058657169342,
      "rewards/rejected": -0.5012514591217041,
      "step": 130
    },
    {
      "epoch": 0.11489536315141567,
      "grad_norm": 0.08581534773111343,
      "learning_rate": 4.893298743830168e-06,
      "logits/chosen": -2.115981340408325,
      "logits/rejected": -2.5363636016845703,
      "logps/chosen": -0.22111928462982178,
      "logps/rejected": -0.3136863708496094,
      "loss": 7.2892,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.33167898654937744,
      "rewards/margins": 0.1388506144285202,
      "rewards/rejected": -0.47052955627441406,
      "step": 140
    },
    {
      "epoch": 0.12310217480508823,
      "grad_norm": 0.06293604522943497,
      "learning_rate": 4.8776412907378845e-06,
      "logits/chosen": -2.070842742919922,
      "logits/rejected": -2.4669342041015625,
      "logps/chosen": -0.20812074840068817,
      "logps/rejected": -0.29536327719688416,
      "loss": 7.26,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.31218111515045166,
      "rewards/margins": 0.13086381554603577,
      "rewards/rejected": -0.4430449604988098,
      "step": 150
    },
    {
      "epoch": 0.12310217480508823,
      "eval_logits/chosen": -2.062544822692871,
      "eval_logits/rejected": -2.5318312644958496,
      "eval_logps/chosen": -0.2108660489320755,
      "eval_logps/rejected": -0.3196176588535309,
      "eval_loss": 0.8929102420806885,
      "eval_rewards/accuracies": 0.5555555820465088,
      "eval_rewards/chosen": -0.31629908084869385,
      "eval_rewards/margins": 0.1631273776292801,
      "eval_rewards/rejected": -0.47942644357681274,
      "eval_runtime": 26.0407,
      "eval_samples_per_second": 30.26,
      "eval_steps_per_second": 3.802,
      "step": 150
    },
    {
      "epoch": 0.13130898645876077,
      "grad_norm": 0.06755395233631134,
      "learning_rate": 4.860940925593703e-06,
      "logits/chosen": -2.187638998031616,
      "logits/rejected": -2.4928510189056396,
      "logps/chosen": -0.2070399969816208,
      "logps/rejected": -0.30727890133857727,
      "loss": 7.1947,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.3105599880218506,
      "rewards/margins": 0.1503583937883377,
      "rewards/rejected": -0.4609183669090271,
      "step": 160
    },
    {
      "epoch": 0.1395157981124333,
      "grad_norm": 0.08956371247768402,
      "learning_rate": 4.84320497372973e-06,
      "logits/chosen": -2.0751285552978516,
      "logits/rejected": -2.478673219680786,
      "logps/chosen": -0.18197472393512726,
      "logps/rejected": -0.2756109833717346,
      "loss": 7.1774,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.2729620933532715,
      "rewards/margins": 0.14045441150665283,
      "rewards/rejected": -0.41341647505760193,
      "step": 170
    },
    {
      "epoch": 0.14772260976610588,
      "grad_norm": 0.07708129286766052,
      "learning_rate": 4.824441214720629e-06,
      "logits/chosen": -2.113537311553955,
      "logits/rejected": -2.530677556991577,
      "logps/chosen": -0.20599500834941864,
      "logps/rejected": -0.2911488711833954,
      "loss": 7.1722,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.308992475271225,
      "rewards/margins": 0.1277308166027069,
      "rewards/rejected": -0.4367233216762543,
      "step": 180
    },
    {
      "epoch": 0.15592942141977842,
      "grad_norm": 0.0884585976600647,
      "learning_rate": 4.804657878971252e-06,
      "logits/chosen": -2.151444673538208,
      "logits/rejected": -2.559861898422241,
      "logps/chosen": -0.2093551605939865,
      "logps/rejected": -0.2878231108188629,
      "loss": 7.1186,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.31403273344039917,
      "rewards/margins": 0.11770190298557281,
      "rewards/rejected": -0.4317346513271332,
      "step": 190
    },
    {
      "epoch": 0.16413623307345096,
      "grad_norm": 0.09445559978485107,
      "learning_rate": 4.783863644106502e-06,
      "logits/chosen": -2.278620481491089,
      "logits/rejected": -2.5897645950317383,
      "logps/chosen": -0.18631704151630402,
      "logps/rejected": -0.3201253116130829,
      "loss": 7.082,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -0.2794755697250366,
      "rewards/margins": 0.2007124423980713,
      "rewards/rejected": -0.4801879823207855,
      "step": 200
    },
    {
      "epoch": 0.16413623307345096,
      "eval_logits/chosen": -2.1718955039978027,
      "eval_logits/rejected": -2.6710257530212402,
      "eval_logps/chosen": -0.20382821559906006,
      "eval_logps/rejected": -0.3390556573867798,
      "eval_loss": 0.8775798678398132,
      "eval_rewards/accuracies": 0.5858585834503174,
      "eval_rewards/chosen": -0.3057423532009125,
      "eval_rewards/margins": 0.20284107327461243,
      "eval_rewards/rejected": -0.5085834264755249,
      "eval_runtime": 26.0531,
      "eval_samples_per_second": 30.246,
      "eval_steps_per_second": 3.8,
      "step": 200
    },
    {
      "epoch": 0.1723430447271235,
      "grad_norm": 0.11077430099248886,
      "learning_rate": 4.762067631165049e-06,
      "logits/chosen": -2.2566323280334473,
      "logits/rejected": -2.621065378189087,
      "logps/chosen": -0.18663282692432404,
      "logps/rejected": -0.290865957736969,
      "loss": 7.1321,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.27994924783706665,
      "rewards/margins": 0.15634974837303162,
      "rewards/rejected": -0.43629899621009827,
      "step": 210
    },
    {
      "epoch": 0.18054985638079607,
      "grad_norm": 0.15500974655151367,
      "learning_rate": 4.7392794005985324e-06,
      "logits/chosen": -2.27781343460083,
      "logits/rejected": -2.700369358062744,
      "logps/chosen": -0.21367880702018738,
      "logps/rejected": -0.31559067964553833,
      "loss": 6.9886,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.3205181956291199,
      "rewards/margins": 0.15286779403686523,
      "rewards/rejected": -0.4733859896659851,
      "step": 220
    },
    {
      "epoch": 0.1887566680344686,
      "grad_norm": 0.12770676612854004,
      "learning_rate": 4.715508948078037e-06,
      "logits/chosen": -2.216815710067749,
      "logits/rejected": -2.759458541870117,
      "logps/chosen": -0.21546092629432678,
      "logps/rejected": -0.34664005041122437,
      "loss": 6.966,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -0.32319143414497375,
      "rewards/margins": 0.196768656373024,
      "rewards/rejected": -0.5199600458145142,
      "step": 230
    },
    {
      "epoch": 0.19696347968814115,
      "grad_norm": 0.15062908828258514,
      "learning_rate": 4.690766700109659e-06,
      "logits/chosen": -2.2262110710144043,
      "logits/rejected": -2.7840607166290283,
      "logps/chosen": -0.2078159749507904,
      "logps/rejected": -0.4006090760231018,
      "loss": 7.012,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.3117239773273468,
      "rewards/margins": 0.2891896665096283,
      "rewards/rejected": -0.6009136438369751,
      "step": 240
    },
    {
      "epoch": 0.2051702913418137,
      "grad_norm": 0.24995267391204834,
      "learning_rate": 4.665063509461098e-06,
      "logits/chosen": -2.4198288917541504,
      "logits/rejected": -2.8148205280303955,
      "logps/chosen": -0.23191122710704803,
      "logps/rejected": -0.38251757621765137,
      "loss": 6.7812,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.34786686301231384,
      "rewards/margins": 0.2259095013141632,
      "rewards/rejected": -0.573776364326477,
      "step": 250
    },
    {
      "epoch": 0.2051702913418137,
      "eval_logits/chosen": -2.3532145023345947,
      "eval_logits/rejected": -2.9015841484069824,
      "eval_logps/chosen": -0.22620753943920135,
      "eval_logps/rejected": -0.4290919005870819,
      "eval_loss": 0.8434350490570068,
      "eval_rewards/accuracies": 0.5959596037864685,
      "eval_rewards/chosen": -0.33931130170822144,
      "eval_rewards/margins": 0.304326593875885,
      "eval_rewards/rejected": -0.6436378955841064,
      "eval_runtime": 26.012,
      "eval_samples_per_second": 30.294,
      "eval_steps_per_second": 3.806,
      "step": 250
    },
    {
      "epoch": 0.21337710299548626,
      "grad_norm": 0.22134838998317719,
      "learning_rate": 4.638410650401267e-06,
      "logits/chosen": -2.4054629802703857,
      "logits/rejected": -2.8117618560791016,
      "logps/chosen": -0.20882606506347656,
      "logps/rejected": -0.41059261560440063,
      "loss": 6.6024,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.31323909759521484,
      "rewards/margins": 0.3026497960090637,
      "rewards/rejected": -0.6158889532089233,
      "step": 260
    },
    {
      "epoch": 0.2215839146491588,
      "grad_norm": 0.23838171362876892,
      "learning_rate": 4.610819813755038e-06,
      "logits/chosen": -2.534034490585327,
      "logits/rejected": -2.846797227859497,
      "logps/chosen": -0.2372482568025589,
      "logps/rejected": -0.493452787399292,
      "loss": 6.7377,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.35587236285209656,
      "rewards/margins": 0.38430681824684143,
      "rewards/rejected": -0.740179181098938,
      "step": 270
    },
    {
      "epoch": 0.22979072630283134,
      "grad_norm": 0.2903882563114166,
      "learning_rate": 4.582303101775249e-06,
      "logits/chosen": -2.4838695526123047,
      "logits/rejected": -2.8494999408721924,
      "logps/chosen": -0.23487380146980286,
      "logps/rejected": -0.5447143316268921,
      "loss": 6.6463,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.3523106873035431,
      "rewards/margins": 0.4647606909275055,
      "rewards/rejected": -0.8170714378356934,
      "step": 280
    },
    {
      "epoch": 0.23799753795650389,
      "grad_norm": 0.2977660596370697,
      "learning_rate": 4.55287302283426e-06,
      "logits/chosen": -2.539785861968994,
      "logits/rejected": -2.922631025314331,
      "logps/chosen": -0.2749824821949005,
      "logps/rejected": -0.5817859768867493,
      "loss": 6.2424,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.41247373819351196,
      "rewards/margins": 0.4602052569389343,
      "rewards/rejected": -0.8726789355278015,
      "step": 290
    },
    {
      "epoch": 0.24620434961017645,
      "grad_norm": 0.388954222202301,
      "learning_rate": 4.522542485937369e-06,
      "logits/chosen": -2.6039352416992188,
      "logits/rejected": -2.8011627197265625,
      "logps/chosen": -0.37122753262519836,
      "logps/rejected": -0.6376734972000122,
      "loss": 6.3458,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.5568413138389587,
      "rewards/margins": 0.39966899156570435,
      "rewards/rejected": -0.9565103650093079,
      "step": 300
    },
    {
      "epoch": 0.24620434961017645,
      "eval_logits/chosen": -2.448641061782837,
      "eval_logits/rejected": -2.8639307022094727,
      "eval_logps/chosen": -0.35992980003356934,
      "eval_logps/rejected": -0.7961164712905884,
      "eval_loss": 0.7593368887901306,
      "eval_rewards/accuracies": 0.6060606241226196,
      "eval_rewards/chosen": -0.5398945808410645,
      "eval_rewards/margins": 0.6542800664901733,
      "eval_rewards/rejected": -1.1941747665405273,
      "eval_runtime": 26.0378,
      "eval_samples_per_second": 30.264,
      "eval_steps_per_second": 3.802,
      "step": 300
    },
    {
      "epoch": 0.254411161263849,
      "grad_norm": 0.41667941212654114,
      "learning_rate": 4.491324795060491e-06,
      "logits/chosen": -2.596318483352661,
      "logits/rejected": -2.8140273094177246,
      "logps/chosen": -0.3550270199775696,
      "logps/rejected": -0.673725426197052,
      "loss": 6.2201,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.532540500164032,
      "rewards/margins": 0.47804751992225647,
      "rewards/rejected": -1.0105879306793213,
      "step": 310
    },
    {
      "epoch": 0.26261797291752154,
      "grad_norm": 0.5356656312942505,
      "learning_rate": 4.4592336433146e-06,
      "logits/chosen": -2.5131359100341797,
      "logits/rejected": -2.7125701904296875,
      "logps/chosen": -0.4790285527706146,
      "logps/rejected": -0.9030188322067261,
      "loss": 5.9656,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.7185428738594055,
      "rewards/margins": 0.6359853744506836,
      "rewards/rejected": -1.3545281887054443,
      "step": 320
    },
    {
      "epoch": 0.2708247845711941,
      "grad_norm": 0.6395894885063171,
      "learning_rate": 4.426283106939474e-06,
      "logits/chosen": -2.408463954925537,
      "logits/rejected": -2.784302234649658,
      "logps/chosen": -0.43364158272743225,
      "logps/rejected": -0.8759375810623169,
      "loss": 5.6972,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.6504624485969543,
      "rewards/margins": 0.6634438037872314,
      "rewards/rejected": -1.313906192779541,
      "step": 330
    },
    {
      "epoch": 0.2790315962248666,
      "grad_norm": 0.5482347011566162,
      "learning_rate": 4.3924876391293915e-06,
      "logits/chosen": -2.4591078758239746,
      "logits/rejected": -2.7352993488311768,
      "logps/chosen": -0.4786604344844818,
      "logps/rejected": -1.337053894996643,
      "loss": 5.5123,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.7179905772209167,
      "rewards/margins": 1.2875900268554688,
      "rewards/rejected": -2.0055809020996094,
      "step": 340
    },
    {
      "epoch": 0.2872384078785392,
      "grad_norm": 0.44611233472824097,
      "learning_rate": 4.357862063693486e-06,
      "logits/chosen": -2.426030397415161,
      "logits/rejected": -2.7272305488586426,
      "logps/chosen": -0.5190738439559937,
      "logps/rejected": -1.3207228183746338,
      "loss": 5.242,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.7786108255386353,
      "rewards/margins": 1.2024734020233154,
      "rewards/rejected": -1.9810841083526611,
      "step": 350
    },
    {
      "epoch": 0.2872384078785392,
      "eval_logits/chosen": -2.5548582077026367,
      "eval_logits/rejected": -2.8742663860321045,
      "eval_logps/chosen": -0.5429244637489319,
      "eval_logps/rejected": -1.556882619857788,
      "eval_loss": 0.6406257748603821,
      "eval_rewards/accuracies": 0.6464646458625793,
      "eval_rewards/chosen": -0.8143868446350098,
      "eval_rewards/margins": 1.520937204360962,
      "eval_rewards/rejected": -2.335324287414551,
      "eval_runtime": 26.0367,
      "eval_samples_per_second": 30.265,
      "eval_steps_per_second": 3.802,
      "step": 350
    },
    {
      "epoch": 0.29544521953221176,
      "grad_norm": 0.6435768008232117,
      "learning_rate": 4.322421568553529e-06,
      "logits/chosen": -2.5521063804626465,
      "logits/rejected": -2.8884310722351074,
      "logps/chosen": -0.5124669075012207,
      "logps/rejected": -1.4227123260498047,
      "loss": 5.2755,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -0.768700361251831,
      "rewards/margins": 1.365368127822876,
      "rewards/rejected": -2.134068489074707,
      "step": 360
    },
    {
      "epoch": 0.3036520311858843,
      "grad_norm": 0.7270930409431458,
      "learning_rate": 4.286181699082008e-06,
      "logits/chosen": -2.613312244415283,
      "logits/rejected": -2.8482322692871094,
      "logps/chosen": -0.7397282719612122,
      "logps/rejected": -1.7584373950958252,
      "loss": 4.6592,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -1.1095924377441406,
      "rewards/margins": 1.528063416481018,
      "rewards/rejected": -2.6376559734344482,
      "step": 370
    },
    {
      "epoch": 0.31185884283955684,
      "grad_norm": 0.417350172996521,
      "learning_rate": 4.249158351283414e-06,
      "logits/chosen": -2.6574740409851074,
      "logits/rejected": -2.8720998764038086,
      "logps/chosen": -0.7095499634742737,
      "logps/rejected": -2.2104392051696777,
      "loss": 4.4512,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -1.0643248558044434,
      "rewards/margins": 2.2513341903686523,
      "rewards/rejected": -3.3156590461730957,
      "step": 380
    },
    {
      "epoch": 0.3200656544932294,
      "grad_norm": 0.5288704633712769,
      "learning_rate": 4.211367764821722e-06,
      "logits/chosen": -2.7438926696777344,
      "logits/rejected": -2.9578990936279297,
      "logps/chosen": -0.7221090793609619,
      "logps/rejected": -2.169327735900879,
      "loss": 4.8545,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -1.0831634998321533,
      "rewards/margins": 2.170828104019165,
      "rewards/rejected": -3.2539916038513184,
      "step": 390
    },
    {
      "epoch": 0.3282724661469019,
      "grad_norm": 0.5342369675636292,
      "learning_rate": 4.172826515897146e-06,
      "logits/chosen": -2.451097011566162,
      "logits/rejected": -2.8298330307006836,
      "logps/chosen": -0.6682878732681274,
      "logps/rejected": -2.4527621269226074,
      "loss": 4.5987,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -1.0024317502975464,
      "rewards/margins": 2.676711320877075,
      "rewards/rejected": -3.679143190383911,
      "step": 400
    },
    {
      "epoch": 0.3282724661469019,
      "eval_logits/chosen": -2.486518621444702,
      "eval_logits/rejected": -2.884521722793579,
      "eval_logps/chosen": -0.6234225630760193,
      "eval_logps/rejected": -2.3772060871124268,
      "eval_loss": 0.5354036688804626,
      "eval_rewards/accuracies": 0.6464646458625793,
      "eval_rewards/chosen": -0.9351338744163513,
      "eval_rewards/margins": 2.6306753158569336,
      "eval_rewards/rejected": -3.5658092498779297,
      "eval_runtime": 26.0392,
      "eval_samples_per_second": 30.262,
      "eval_steps_per_second": 3.802,
      "step": 400
    },
    {
      "epoch": 0.33647927780057446,
      "grad_norm": 0.3894485831260681,
      "learning_rate": 4.133551509975264e-06,
      "logits/chosen": -2.498328685760498,
      "logits/rejected": -2.832383394241333,
      "logps/chosen": -0.8072039484977722,
      "logps/rejected": -3.0143685340881348,
      "loss": 4.1825,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -1.210806131362915,
      "rewards/margins": 3.310746669769287,
      "rewards/rejected": -4.521553039550781,
      "step": 410
    },
    {
      "epoch": 0.344686089454247,
      "grad_norm": 1.1384832859039307,
      "learning_rate": 4.093559974371725e-06,
      "logits/chosen": -2.3939032554626465,
      "logits/rejected": -2.736551284790039,
      "logps/chosen": -0.8623531460762024,
      "logps/rejected": -3.228794813156128,
      "loss": 4.0383,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -1.293529748916626,
      "rewards/margins": 3.5496623516082764,
      "rewards/rejected": -4.843192100524902,
      "step": 420
    },
    {
      "epoch": 0.3528929011079196,
      "grad_norm": 0.5636156797409058,
      "learning_rate": 4.052869450695776e-06,
      "logits/chosen": -2.466431140899658,
      "logits/rejected": -2.7423882484436035,
      "logps/chosen": -0.867216944694519,
      "logps/rejected": -2.9713022708892822,
      "loss": 3.9764,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -1.3008254766464233,
      "rewards/margins": 3.156127691268921,
      "rewards/rejected": -4.456953525543213,
      "step": 430
    },
    {
      "epoch": 0.36109971276159214,
      "grad_norm": 1.3254886865615845,
      "learning_rate": 4.011497787155938e-06,
      "logits/chosen": -2.438335418701172,
      "logits/rejected": -2.716845989227295,
      "logps/chosen": -0.9805113673210144,
      "logps/rejected": -3.646573543548584,
      "loss": 4.0643,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -1.4707671403884888,
      "rewards/margins": 3.9990932941436768,
      "rewards/rejected": -5.469860553741455,
      "step": 440
    },
    {
      "epoch": 0.3693065244152647,
      "grad_norm": 1.0398495197296143,
      "learning_rate": 3.969463130731183e-06,
      "logits/chosen": -2.3812689781188965,
      "logits/rejected": -2.7526910305023193,
      "logps/chosen": -1.0975981950759888,
      "logps/rejected": -3.9439749717712402,
      "loss": 3.9749,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -1.6463972330093384,
      "rewards/margins": 4.269565105438232,
      "rewards/rejected": -5.915962219238281,
      "step": 450
    },
    {
      "epoch": 0.3693065244152647,
      "eval_logits/chosen": -2.4038827419281006,
      "eval_logits/rejected": -2.7936673164367676,
      "eval_logps/chosen": -1.0481945276260376,
      "eval_logps/rejected": -3.5816102027893066,
      "eval_loss": 0.47537824511528015,
      "eval_rewards/accuracies": 0.7373737096786499,
      "eval_rewards/chosen": -1.5722918510437012,
      "eval_rewards/margins": 3.800123691558838,
      "eval_rewards/rejected": -5.372415542602539,
      "eval_runtime": 26.0116,
      "eval_samples_per_second": 30.294,
      "eval_steps_per_second": 3.806,
      "step": 450
    },
    {
      "epoch": 0.3775133360689372,
      "grad_norm": 1.091323971748352,
      "learning_rate": 3.92678391921108e-06,
      "logits/chosen": -2.5687508583068848,
      "logits/rejected": -2.766624927520752,
      "logps/chosen": -1.1868960857391357,
      "logps/rejected": -3.4050726890563965,
      "loss": 4.114,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -1.780344009399414,
      "rewards/margins": 3.3272647857666016,
      "rewards/rejected": -5.107609272003174,
      "step": 460
    },
    {
      "epoch": 0.38572014772260976,
      "grad_norm": 0.8932163715362549,
      "learning_rate": 3.88347887310836e-06,
      "logits/chosen": -2.438856601715088,
      "logits/rejected": -2.8660335540771484,
      "logps/chosen": -1.420062780380249,
      "logps/rejected": -4.232975006103516,
      "loss": 3.9707,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -2.130094051361084,
      "rewards/margins": 4.219367980957031,
      "rewards/rejected": -6.349462032318115,
      "step": 470
    },
    {
      "epoch": 0.3939269593762823,
      "grad_norm": 1.110568642616272,
      "learning_rate": 3.839566987447492e-06,
      "logits/chosen": -2.5447945594787598,
      "logits/rejected": -2.720705032348633,
      "logps/chosen": -1.9743589162826538,
      "logps/rejected": -3.963395595550537,
      "loss": 3.7559,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -2.961538314819336,
      "rewards/margins": 2.983555316925049,
      "rewards/rejected": -5.945094108581543,
      "step": 480
    },
    {
      "epoch": 0.40213377102995485,
      "grad_norm": 1.1167631149291992,
      "learning_rate": 3.795067523432826e-06,
      "logits/chosen": -2.6660027503967285,
      "logits/rejected": -2.839933395385742,
      "logps/chosen": -2.3371267318725586,
      "logps/rejected": -4.315327167510986,
      "loss": 3.5333,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -3.505690097808838,
      "rewards/margins": 2.9673006534576416,
      "rewards/rejected": -6.4729905128479,
      "step": 490
    },
    {
      "epoch": 0.4103405826836274,
      "grad_norm": 1.7681843042373657,
      "learning_rate": 3.7500000000000005e-06,
      "logits/chosen": -2.4737250804901123,
      "logits/rejected": -2.7522482872009277,
      "logps/chosen": -2.802852153778076,
      "logps/rejected": -5.4331955909729,
      "loss": 3.1958,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -4.204278469085693,
      "rewards/margins": 3.9455153942108154,
      "rewards/rejected": -8.14979362487793,
      "step": 500
    },
    {
      "epoch": 0.4103405826836274,
      "eval_logits/chosen": -2.464381456375122,
      "eval_logits/rejected": -2.8693907260894775,
      "eval_logps/chosen": -2.83821177482605,
      "eval_logps/rejected": -5.77597188949585,
      "eval_loss": 0.3702296018600464,
      "eval_rewards/accuracies": 0.8383838534355164,
      "eval_rewards/chosen": -4.257318019866943,
      "eval_rewards/margins": 4.406640529632568,
      "eval_rewards/rejected": -8.663958549499512,
      "eval_runtime": 26.0401,
      "eval_samples_per_second": 30.261,
      "eval_steps_per_second": 3.802,
      "step": 500
    }
  ],
  "logging_steps": 10,
  "max_steps": 1500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.816762881216086e+18,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}