{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.985781990521327,
  "eval_steps": 500,
  "global_step": 315,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "grad_norm": 51.45086593171862,
      "learning_rate": 5e-08,
      "log_odds_chosen": 0.23327143490314484,
      "log_odds_ratio": -0.7173982262611389,
      "logits/chosen": -2.9019644260406494,
      "logits/rejected": -2.911267042160034,
      "logps/chosen": -1.2478530406951904,
      "logps/rejected": -1.4603310823440552,
      "loss": 1.1528,
      "nll_loss": 1.110867977142334,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.06239265203475952,
      "rewards/margins": 0.010623900219798088,
      "rewards/rejected": -0.07301654666662216,
      "step": 10
    },
    {
      "epoch": 0.19,
      "grad_norm": 43.38597658388667,
      "learning_rate": 1e-07,
      "log_odds_chosen": 0.025622714310884476,
      "log_odds_ratio": -0.803138256072998,
      "logits/chosen": -2.957723617553711,
      "logits/rejected": -2.9606709480285645,
      "logps/chosen": -1.2851994037628174,
      "logps/rejected": -1.2747653722763062,
      "loss": 1.109,
      "nll_loss": 1.0504928827285767,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.06425996124744415,
      "rewards/margins": -0.0005216972786001861,
      "rewards/rejected": -0.06373827159404755,
      "step": 20
    },
    {
      "epoch": 0.28,
      "grad_norm": 18.07276834258672,
      "learning_rate": 1.5e-07,
      "log_odds_chosen": 0.33495259284973145,
      "log_odds_ratio": -0.6577204465866089,
      "logits/chosen": -2.936415195465088,
      "logits/rejected": -2.939385175704956,
      "logps/chosen": -1.0753445625305176,
      "logps/rejected": -1.3320289850234985,
      "loss": 1.0256,
      "nll_loss": 0.9573599696159363,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.053767234086990356,
      "rewards/margins": 0.012834218330681324,
      "rewards/rejected": -0.06660144031047821,
      "step": 30
    },
    {
      "epoch": 0.38,
      "grad_norm": 14.138346458866787,
      "learning_rate": 2e-07,
      "log_odds_chosen": 0.06818895787000656,
      "log_odds_ratio": -0.7556289434432983,
      "logits/chosen": -2.979851484298706,
      "logits/rejected": -2.9792511463165283,
      "logps/chosen": -1.0226948261260986,
      "logps/rejected": -1.0867162942886353,
      "loss": 0.9841,
      "nll_loss": 0.9785293340682983,
      "rewards/accuracies": 0.518750011920929,
      "rewards/chosen": -0.05113474652171135,
      "rewards/margins": 0.003201070474460721,
      "rewards/rejected": -0.0543358139693737,
      "step": 40
    },
    {
      "epoch": 0.47,
      "grad_norm": 19.142475208524697,
      "learning_rate": 2.5e-07,
      "log_odds_chosen": 0.2577429711818695,
      "log_odds_ratio": -0.6650774478912354,
      "logits/chosen": -3.0103330612182617,
      "logits/rejected": -3.0095608234405518,
      "logps/chosen": -0.9924309849739075,
      "logps/rejected": -1.1665266752243042,
      "loss": 0.9833,
      "nll_loss": 0.965359091758728,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.04962155967950821,
      "rewards/margins": 0.008704781532287598,
      "rewards/rejected": -0.05832633376121521,
      "step": 50
    },
    {
      "epoch": 0.57,
      "grad_norm": 9.00103625601085,
      "learning_rate": 3e-07,
      "log_odds_chosen": 0.2601451277732849,
      "log_odds_ratio": -0.6529109477996826,
      "logits/chosen": -3.0405919551849365,
      "logits/rejected": -3.0266852378845215,
      "logps/chosen": -0.8860586881637573,
      "logps/rejected": -1.0516645908355713,
      "loss": 0.967,
      "nll_loss": 0.8791006207466125,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": -0.044302936643362045,
      "rewards/margins": 0.008280294016003609,
      "rewards/rejected": -0.0525832362473011,
      "step": 60
    },
    {
      "epoch": 0.66,
      "grad_norm": 8.208565662624949,
      "learning_rate": 3.5e-07,
      "log_odds_chosen": 0.24647900462150574,
      "log_odds_ratio": -0.6602240800857544,
      "logits/chosen": -2.9646213054656982,
      "logits/rejected": -2.9692559242248535,
      "logps/chosen": -0.8851203918457031,
      "logps/rejected": -1.04824697971344,
      "loss": 0.9521,
      "nll_loss": 0.8758776783943176,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.044256024062633514,
      "rewards/margins": 0.008156323805451393,
      "rewards/rejected": -0.05241234228014946,
      "step": 70
    },
    {
      "epoch": 0.76,
      "grad_norm": 9.949021333520767,
      "learning_rate": 4e-07,
      "log_odds_chosen": 0.3642485439777374,
      "log_odds_ratio": -0.6296299695968628,
      "logits/chosen": -2.9997572898864746,
      "logits/rejected": -2.9932103157043457,
      "logps/chosen": -0.8595079183578491,
      "logps/rejected": -1.0957872867584229,
      "loss": 0.8992,
      "nll_loss": 0.8687052726745605,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.042975399643182755,
      "rewards/margins": 0.011813966557383537,
      "rewards/rejected": -0.05478937178850174,
      "step": 80
    },
    {
      "epoch": 0.85,
      "grad_norm": 11.826771567950217,
      "learning_rate": 4.5e-07,
      "log_odds_chosen": 0.13775679469108582,
      "log_odds_ratio": -0.7174943685531616,
      "logits/chosen": -2.9466898441314697,
      "logits/rejected": -2.9638590812683105,
      "logps/chosen": -0.8956987261772156,
      "logps/rejected": -0.9907738566398621,
      "loss": 0.903,
      "nll_loss": 0.8976006507873535,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.04478494077920914,
      "rewards/margins": 0.0047537581995129585,
      "rewards/rejected": -0.04953869432210922,
      "step": 90
    },
    {
      "epoch": 0.95,
      "grad_norm": 7.898257944088662,
      "learning_rate": 5e-07,
      "log_odds_chosen": 0.19350430369377136,
      "log_odds_ratio": -0.6868109703063965,
      "logits/chosen": -2.932610034942627,
      "logits/rejected": -2.9326844215393066,
      "logps/chosen": -0.8639435768127441,
      "logps/rejected": -0.9883400797843933,
      "loss": 0.9159,
      "nll_loss": 0.8739891052246094,
      "rewards/accuracies": 0.606249988079071,
      "rewards/chosen": -0.04319717735052109,
      "rewards/margins": 0.0062198275700211525,
      "rewards/rejected": -0.049417007714509964,
      "step": 100
    },
    {
      "epoch": 1.0,
      "eval_log_odds_chosen": 0.3023655116558075,
      "eval_log_odds_ratio": -0.6428868770599365,
      "eval_logits/chosen": -2.887465238571167,
      "eval_logits/rejected": -2.8930952548980713,
      "eval_logps/chosen": -0.8412830829620361,
      "eval_logps/rejected": -0.9974649548530579,
      "eval_loss": 0.8794215321540833,
      "eval_nll_loss": 0.8561177253723145,
      "eval_rewards/accuracies": 0.6302083134651184,
      "eval_rewards/chosen": -0.042064156383275986,
      "eval_rewards/margins": 0.007809097412973642,
      "eval_rewards/rejected": -0.04987325146794319,
      "eval_runtime": 32.7887,
      "eval_samples_per_second": 22.874,
      "eval_steps_per_second": 0.732,
      "step": 105
    },
    {
      "epoch": 1.04,
      "grad_norm": 7.858898751829365,
      "learning_rate": 4.973358420187775e-07,
      "log_odds_chosen": 0.2784392237663269,
      "log_odds_ratio": -0.6759603023529053,
      "logits/chosen": -2.898139715194702,
      "logits/rejected": -2.888554096221924,
      "logps/chosen": -0.8769194483757019,
      "logps/rejected": -1.0477098226547241,
      "loss": 0.9267,
      "nll_loss": 0.9084149599075317,
      "rewards/accuracies": 0.5687500238418579,
      "rewards/chosen": -0.043845973908901215,
      "rewards/margins": 0.00853952020406723,
      "rewards/rejected": -0.05238549783825874,
      "step": 110
    },
    {
      "epoch": 1.14,
      "grad_norm": 8.696606134772816,
      "learning_rate": 4.894001499771015e-07,
      "log_odds_chosen": 0.408263623714447,
      "log_odds_ratio": -0.5970031023025513,
      "logits/chosen": -2.8863773345947266,
      "logits/rejected": -2.8754923343658447,
      "logps/chosen": -0.82969731092453,
      "logps/rejected": -1.0684994459152222,
      "loss": 0.8737,
      "nll_loss": 0.8546596765518188,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.04148487001657486,
      "rewards/margins": 0.011940106749534607,
      "rewards/rejected": -0.05342497676610947,
      "step": 120
    },
    {
      "epoch": 1.23,
      "grad_norm": 8.142229649165017,
      "learning_rate": 4.7636205937328664e-07,
      "log_odds_chosen": 0.24737975001335144,
      "log_odds_ratio": -0.6806861162185669,
      "logits/chosen": -2.8786978721618652,
      "logits/rejected": -2.869795083999634,
      "logps/chosen": -0.8045932054519653,
      "logps/rejected": -0.9472286105155945,
      "loss": 0.9147,
      "nll_loss": 0.8494957089424133,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.04022965580224991,
      "rewards/margins": 0.007131779100745916,
      "rewards/rejected": -0.04736143350601196,
      "step": 130
    },
    {
      "epoch": 1.33,
      "grad_norm": 8.40750837533949,
      "learning_rate": 4.584994544724695e-07,
      "log_odds_chosen": 0.2743343412876129,
      "log_odds_ratio": -0.6686180830001831,
      "logits/chosen": -2.866347312927246,
      "logits/rejected": -2.8765628337860107,
      "logps/chosen": -0.7923319935798645,
      "logps/rejected": -0.9410168528556824,
      "loss": 0.8631,
      "nll_loss": 0.8559592366218567,
      "rewards/accuracies": 0.606249988079071,
      "rewards/chosen": -0.039616603404283524,
      "rewards/margins": 0.007434243801981211,
      "rewards/rejected": -0.047050848603248596,
      "step": 140
    },
    {
      "epoch": 1.42,
      "grad_norm": 8.158461437554013,
      "learning_rate": 4.3619304568594546e-07,
      "log_odds_chosen": 0.3588339388370514,
      "log_odds_ratio": -0.6238493919372559,
      "logits/chosen": -2.869856119155884,
      "logits/rejected": -2.8966941833496094,
      "logps/chosen": -0.7852147817611694,
      "logps/rejected": -0.9708358645439148,
      "loss": 0.8549,
      "nll_loss": 0.8177824020385742,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.03926074132323265,
      "rewards/margins": 0.009281056001782417,
      "rewards/rejected": -0.04854179546236992,
      "step": 150
    },
    {
      "epoch": 1.52,
      "grad_norm": 8.09400448571549,
      "learning_rate": 4.099182553897228e-07,
      "log_odds_chosen": 0.47473040223121643,
      "log_odds_ratio": -0.5752421021461487,
      "logits/chosen": -2.8670287132263184,
      "logits/rejected": -2.849987506866455,
      "logps/chosen": -0.7626146078109741,
      "logps/rejected": -1.0768699645996094,
      "loss": 0.8506,
      "nll_loss": 0.7988383173942566,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.038130730390548706,
      "rewards/margins": 0.01571277156472206,
      "rewards/rejected": -0.05384349822998047,
      "step": 160
    },
    {
      "epoch": 1.61,
      "grad_norm": 6.8406261430616535,
      "learning_rate": 3.8023508512198257e-07,
      "log_odds_chosen": 0.35588333010673523,
      "log_odds_ratio": -0.6235536336898804,
      "logits/chosen": -2.8909614086151123,
      "logits/rejected": -2.8951175212860107,
      "logps/chosen": -0.7602669596672058,
      "logps/rejected": -0.9576327204704285,
      "loss": 0.8503,
      "nll_loss": 0.7980303764343262,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.03801335021853447,
      "rewards/margins": 0.009868290275335312,
      "rewards/rejected": -0.04788164049386978,
      "step": 170
    },
    {
      "epoch": 1.71,
      "grad_norm": 7.855569563727718,
      "learning_rate": 3.4777618012253895e-07,
      "log_odds_chosen": 0.243193581700325,
      "log_odds_ratio": -0.6670368909835815,
      "logits/chosen": -2.84794545173645,
      "logits/rejected": -2.8654625415802,
      "logps/chosen": -0.8232158422470093,
      "logps/rejected": -0.9566510915756226,
      "loss": 0.8633,
      "nll_loss": 0.8674771189689636,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": -0.041160788387060165,
      "rewards/margins": 0.006671770475804806,
      "rewards/rejected": -0.04783256724476814,
      "step": 180
    },
    {
      "epoch": 1.8,
      "grad_norm": 7.6860644181764375,
      "learning_rate": 3.1323334559792015e-07,
      "log_odds_chosen": 0.2766939699649811,
      "log_odds_ratio": -0.6430349349975586,
      "logits/chosen": -2.8580737113952637,
      "logits/rejected": -2.8492133617401123,
      "logps/chosen": -0.7749003171920776,
      "logps/rejected": -0.9475242495536804,
      "loss": 0.8497,
      "nll_loss": 0.7961811423301697,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.03874502331018448,
      "rewards/margins": 0.008631192147731781,
      "rewards/rejected": -0.047376204282045364,
      "step": 190
    },
    {
      "epoch": 1.9,
      "grad_norm": 7.778871538887919,
      "learning_rate": 2.7734280209446865e-07,
      "log_odds_chosen": 0.2929554879665375,
      "log_odds_ratio": -0.6454571485519409,
      "logits/chosen": -2.874631881713867,
      "logits/rejected": -2.872854232788086,
      "logps/chosen": -0.8560446500778198,
      "logps/rejected": -1.0225257873535156,
      "loss": 0.8742,
      "nll_loss": 0.8836696743965149,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.04280223324894905,
      "rewards/margins": 0.008324062451720238,
      "rewards/rejected": -0.05112629011273384,
      "step": 200
    },
    {
      "epoch": 1.99,
      "grad_norm": 8.608571431216085,
      "learning_rate": 2.4086949423558525e-07,
      "log_odds_chosen": 0.4544145464897156,
      "log_odds_ratio": -0.5953847169876099,
      "logits/chosen": -2.8644211292266846,
      "logits/rejected": -2.8725366592407227,
      "logps/chosen": -0.7150249481201172,
      "logps/rejected": -0.9680275917053223,
      "loss": 0.8397,
      "nll_loss": 0.7700489163398743,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.03575124591588974,
      "rewards/margins": 0.012650132179260254,
      "rewards/rejected": -0.048401378095149994,
      "step": 210
    },
    {
      "epoch": 2.0,
      "eval_log_odds_chosen": 0.3554897606372833,
      "eval_log_odds_ratio": -0.6257309317588806,
      "eval_logits/chosen": -2.8794047832489014,
      "eval_logits/rejected": -2.888169527053833,
      "eval_logps/chosen": -0.8071174621582031,
      "eval_logps/rejected": -0.9902352690696716,
      "eval_loss": 0.8611603379249573,
      "eval_nll_loss": 0.8365828990936279,
      "eval_rewards/accuracies": 0.6458333134651184,
      "eval_rewards/chosen": -0.040355876088142395,
      "eval_rewards/margins": 0.009155889973044395,
      "eval_rewards/rejected": -0.04951176047325134,
      "eval_runtime": 32.7226,
      "eval_samples_per_second": 22.92,
      "eval_steps_per_second": 0.733,
      "step": 211
    },
    {
      "epoch": 2.09,
      "grad_norm": 6.957321522636136,
      "learning_rate": 2.0459078725514089e-07,
      "log_odds_chosen": 0.5733749866485596,
      "log_odds_ratio": -0.5376938581466675,
      "logits/chosen": -2.876321315765381,
      "logits/rejected": -2.8998565673828125,
      "logps/chosen": -0.6645466089248657,
      "logps/rejected": -0.9755932092666626,
      "loss": 0.7779,
      "nll_loss": 0.7205012440681458,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.033227331936359406,
      "rewards/margins": 0.015552327036857605,
      "rewards/rejected": -0.04877965897321701,
      "step": 220
    },
    {
      "epoch": 2.18,
      "grad_norm": 7.383557049949752,
      "learning_rate": 1.692798988071385e-07,
      "log_odds_chosen": 0.6264925003051758,
      "log_odds_ratio": -0.5338762998580933,
      "logits/chosen": -2.9063897132873535,
      "logits/rejected": -2.9123330116271973,
      "logps/chosen": -0.6798352003097534,
      "logps/rejected": -1.0108721256256104,
      "loss": 0.7879,
      "nll_loss": 0.763096809387207,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.03399176150560379,
      "rewards/margins": 0.016551844775676727,
      "rewards/rejected": -0.05054359883069992,
      "step": 230
    },
    {
      "epoch": 2.27,
      "grad_norm": 7.205471545860071,
      "learning_rate": 1.3568941917384036e-07,
      "log_odds_chosen": 0.6044402122497559,
      "log_odds_ratio": -0.5256311297416687,
      "logits/chosen": -2.929206609725952,
      "logits/rejected": -2.9085187911987305,
      "logps/chosen": -0.6959676742553711,
      "logps/rejected": -1.0572305917739868,
      "loss": 0.7677,
      "nll_loss": 0.7402119636535645,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.034798379987478256,
      "rewards/margins": 0.018063146620988846,
      "rewards/rejected": -0.0528615303337574,
      "step": 240
    },
    {
      "epoch": 2.37,
      "grad_norm": 7.342052939736115,
      "learning_rate": 1.0453527111051183e-07,
      "log_odds_chosen": 0.6251620054244995,
      "log_odds_ratio": -0.5341116786003113,
      "logits/chosen": -2.8869872093200684,
      "logits/rejected": -2.912830352783203,
      "logps/chosen": -0.6446677446365356,
      "logps/rejected": -0.9978305101394653,
      "loss": 0.7682,
      "nll_loss": 0.6836553812026978,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.03223338723182678,
      "rewards/margins": 0.017658134922385216,
      "rewards/rejected": -0.04989152401685715,
      "step": 250
    },
    {
      "epoch": 2.46,
      "grad_norm": 7.710315387822491,
      "learning_rate": 7.648145119484151e-08,
      "log_odds_chosen": 0.582895815372467,
      "log_odds_ratio": -0.5480377078056335,
      "logits/chosen": -2.89570689201355,
      "logits/rejected": -2.8959426879882812,
      "logps/chosen": -0.7015441656112671,
      "logps/rejected": -0.9980028867721558,
      "loss": 0.7845,
      "nll_loss": 0.7611762881278992,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.035077206790447235,
      "rewards/margins": 0.014822937548160553,
      "rewards/rejected": -0.04990014433860779,
      "step": 260
    },
    {
      "epoch": 2.56,
      "grad_norm": 7.850416556443741,
      "learning_rate": 5.212587789268649e-08,
      "log_odds_chosen": 0.5626873970031738,
      "log_odds_ratio": -0.5825694799423218,
      "logits/chosen": -2.898651599884033,
      "logits/rejected": -2.9134531021118164,
      "logps/chosen": -0.713202714920044,
      "logps/rejected": -0.9749401211738586,
      "loss": 0.7837,
      "nll_loss": 0.7641984224319458,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.03566013649106026,
      "rewards/margins": 0.013086864724755287,
      "rewards/rejected": -0.04874700307846069,
      "step": 270
    },
    {
      "epoch": 2.65,
      "grad_norm": 7.144333751483519,
      "learning_rate": 3.198764796404807e-08,
      "log_odds_chosen": 0.6033560037612915,
      "log_odds_ratio": -0.5110877156257629,
      "logits/chosen": -2.9342381954193115,
      "logits/rejected": -2.9433302879333496,
      "logps/chosen": -0.6664480566978455,
      "logps/rejected": -0.9789284467697144,
      "loss": 0.7707,
      "nll_loss": 0.7404603362083435,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.03332240507006645,
      "rewards/margins": 0.015624021179974079,
      "rewards/rejected": -0.04894643276929855,
      "step": 280
    },
    {
      "epoch": 2.75,
      "grad_norm": 7.681448875266474,
      "learning_rate": 1.649597281686302e-08,
      "log_odds_chosen": 0.5207457542419434,
      "log_odds_ratio": -0.5457995533943176,
      "logits/chosen": -2.933040142059326,
      "logits/rejected": -2.942648410797119,
      "logps/chosen": -0.6956908106803894,
      "logps/rejected": -0.970605194568634,
      "loss": 0.7797,
      "nll_loss": 0.7457938194274902,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.03478454053401947,
      "rewards/margins": 0.013745710253715515,
      "rewards/rejected": -0.048530250787734985,
      "step": 290
    },
    {
      "epoch": 2.84,
      "grad_norm": 7.3441185490833805,
      "learning_rate": 5.981030611018234e-09,
      "log_odds_chosen": 0.643921434879303,
      "log_odds_ratio": -0.5143691897392273,
      "logits/chosen": -2.923649549484253,
      "logits/rejected": -2.937542676925659,
      "logps/chosen": -0.7117995619773865,
      "logps/rejected": -1.0296577215194702,
      "loss": 0.7861,
      "nll_loss": 0.7953752279281616,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.035589978098869324,
      "rewards/margins": 0.015892909839749336,
      "rewards/rejected": -0.05148288607597351,
      "step": 300
    },
    {
      "epoch": 2.94,
      "grad_norm": 7.411020953558967,
      "learning_rate": 6.66929084112089e-10,
      "log_odds_chosen": 0.531533420085907,
      "log_odds_ratio": -0.5645086765289307,
      "logits/chosen": -2.9212653636932373,
      "logits/rejected": -2.941697835922241,
      "logps/chosen": -0.6984909772872925,
      "logps/rejected": -0.9673376083374023,
      "loss": 0.7808,
      "nll_loss": 0.7526401877403259,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.034924548119306564,
      "rewards/margins": 0.013442324474453926,
      "rewards/rejected": -0.04836687445640564,
      "step": 310
    },
    {
      "epoch": 2.99,
      "eval_log_odds_chosen": 0.3802085816860199,
      "eval_log_odds_ratio": -0.6214679479598999,
      "eval_logits/chosen": -2.9039885997772217,
      "eval_logits/rejected": -2.914613723754883,
      "eval_logps/chosen": -0.8096261024475098,
      "eval_logps/rejected": -1.0036059617996216,
      "eval_loss": 0.8648061752319336,
      "eval_nll_loss": 0.8392227292060852,
      "eval_rewards/accuracies": 0.6458333134651184,
      "eval_rewards/chosen": -0.04048130661249161,
      "eval_rewards/margins": 0.009698997251689434,
      "eval_rewards/rejected": -0.05018030107021332,
      "eval_runtime": 32.6828,
      "eval_samples_per_second": 22.948,
      "eval_steps_per_second": 0.734,
      "step": 315
    },
    {
      "epoch": 2.99,
      "step": 315,
      "total_flos": 0.0,
      "train_loss": 0.877839538029262,
      "train_runtime": 3399.396,
      "train_samples_per_second": 5.957,
      "train_steps_per_second": 0.093
    }
  ],
  "logging_steps": 10,
  "max_steps": 315,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}