{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 50,
  "global_step": 321,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04672897196261682,
      "grad_norm": 61.970988574529464,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7293832302093506,
      "logits/rejected": -2.7098002433776855,
      "logps/chosen": -282.619384765625,
      "logps/rejected": -220.626708984375,
      "loss": 0.6898,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": 0.016255810856819153,
      "rewards/margins": 0.00715771596878767,
      "rewards/rejected": 0.009098095819354057,
      "step": 5
    },
    {
      "epoch": 0.09345794392523364,
      "grad_norm": 59.149513121461624,
      "learning_rate": 1e-06,
      "logits/chosen": -2.6811907291412354,
      "logits/rejected": -2.6537957191467285,
      "logps/chosen": -256.4613037109375,
      "logps/rejected": -214.9097442626953,
      "loss": 0.6465,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.46442437171936035,
      "rewards/margins": 0.15177568793296814,
      "rewards/rejected": 0.3126486837863922,
      "step": 10
    },
    {
      "epoch": 0.14018691588785046,
      "grad_norm": 43.868507115118234,
      "learning_rate": 9.993623730611148e-07,
      "logits/chosen": -2.4990592002868652,
      "logits/rejected": -2.4942288398742676,
      "logps/chosen": -252.8370361328125,
      "logps/rejected": -212.9906768798828,
      "loss": 0.6365,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 1.3768469095230103,
      "rewards/margins": 0.5569905042648315,
      "rewards/rejected": 0.8198563456535339,
      "step": 15
    },
    {
      "epoch": 0.18691588785046728,
      "grad_norm": 40.58644668089117,
      "learning_rate": 9.97451118516912e-07,
      "logits/chosen": -2.357009172439575,
      "logits/rejected": -2.3200223445892334,
      "logps/chosen": -245.38623046875,
      "logps/rejected": -190.89620971679688,
      "loss": 0.6273,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 1.5086402893066406,
      "rewards/margins": 0.8777653574943542,
      "rewards/rejected": 0.6308748722076416,
      "step": 20
    },
    {
      "epoch": 0.2336448598130841,
      "grad_norm": 46.4139582288858,
      "learning_rate": 9.94271111036929e-07,
      "logits/chosen": -2.2126636505126953,
      "logits/rejected": -2.2065536975860596,
      "logps/chosen": -242.53775024414062,
      "logps/rejected": -225.9445343017578,
      "loss": 0.6348,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 1.3302034139633179,
      "rewards/margins": 0.9245456457138062,
      "rewards/rejected": 0.4056577682495117,
      "step": 25
    },
    {
      "epoch": 0.2803738317757009,
      "grad_norm": 64.74245067517272,
      "learning_rate": 9.898304612549066e-07,
      "logits/chosen": -2.183245897293091,
      "logits/rejected": -2.1688549518585205,
      "logps/chosen": -233.8648223876953,
      "logps/rejected": -205.1978759765625,
      "loss": 0.5938,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 1.1376529932022095,
      "rewards/margins": 0.8598777651786804,
      "rewards/rejected": 0.27777519822120667,
      "step": 30
    },
    {
      "epoch": 0.32710280373831774,
      "grad_norm": 51.84730256605764,
      "learning_rate": 9.841404950825536e-07,
      "logits/chosen": -2.227497100830078,
      "logits/rejected": -2.1977126598358154,
      "logps/chosen": -237.3373565673828,
      "logps/rejected": -212.15512084960938,
      "loss": 0.5785,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": 1.2570827007293701,
      "rewards/margins": 0.9708169102668762,
      "rewards/rejected": 0.2862659692764282,
      "step": 35
    },
    {
      "epoch": 0.37383177570093457,
      "grad_norm": 40.60340275031408,
      "learning_rate": 9.77215724822721e-07,
      "logits/chosen": -2.243427276611328,
      "logits/rejected": -2.2088351249694824,
      "logps/chosen": -248.1513671875,
      "logps/rejected": -229.47531127929688,
      "loss": 0.5829,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.9161311984062195,
      "rewards/margins": 1.0557386875152588,
      "rewards/rejected": -0.13960735499858856,
      "step": 40
    },
    {
      "epoch": 0.4205607476635514,
      "grad_norm": 34.21102418831377,
      "learning_rate": 9.69073812155662e-07,
      "logits/chosen": -2.2237961292266846,
      "logits/rejected": -2.2018837928771973,
      "logps/chosen": -269.6021423339844,
      "logps/rejected": -224.58645629882812,
      "loss": 0.5896,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 1.0438940525054932,
      "rewards/margins": 0.9452205896377563,
      "rewards/rejected": 0.09867370873689651,
      "step": 45
    },
    {
      "epoch": 0.4672897196261682,
      "grad_norm": 42.29092168681447,
      "learning_rate": 9.597355230927788e-07,
      "logits/chosen": -2.0892410278320312,
      "logits/rejected": -2.0624351501464844,
      "logps/chosen": -240.1529083251953,
      "logps/rejected": -198.04843139648438,
      "loss": 0.5529,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 1.232414722442627,
      "rewards/margins": 1.3885271549224854,
      "rewards/rejected": -0.156112402677536,
      "step": 50
    },
    {
      "epoch": 0.4672897196261682,
      "eval_logits/chosen": -2.0593109130859375,
      "eval_logits/rejected": -2.02968692779541,
      "eval_logps/chosen": -239.67454528808594,
      "eval_logps/rejected": -216.6856689453125,
      "eval_loss": 0.5947180390357971,
      "eval_rewards/accuracies": 0.7317708134651184,
      "eval_rewards/chosen": 0.8738771080970764,
      "eval_rewards/margins": 1.1332703828811646,
      "eval_rewards/rejected": -0.2593933641910553,
      "eval_runtime": 202.7096,
      "eval_samples_per_second": 14.997,
      "eval_steps_per_second": 0.237,
      "step": 50
    },
    {
      "epoch": 0.514018691588785,
      "grad_norm": 35.93324426868974,
      "learning_rate": 9.4922467501275e-07,
      "logits/chosen": -2.0398037433624268,
      "logits/rejected": -2.0083518028259277,
      "logps/chosen": -261.76177978515625,
      "logps/rejected": -210.04684448242188,
      "loss": 0.5467,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.9814583659172058,
      "rewards/margins": 1.524784803390503,
      "rewards/rejected": -0.5433263778686523,
      "step": 55
    },
    {
      "epoch": 0.5607476635514018,
      "grad_norm": 50.252777089157334,
      "learning_rate": 9.375680759151206e-07,
      "logits/chosen": -2.092087984085083,
      "logits/rejected": -2.098419427871704,
      "logps/chosen": -252.8424530029297,
      "logps/rejected": -214.93258666992188,
      "loss": 0.5409,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 0.7585053443908691,
      "rewards/margins": 1.2850215435028076,
      "rewards/rejected": -0.5265161395072937,
      "step": 60
    },
    {
      "epoch": 0.6074766355140186,
      "grad_norm": 43.0211288247323,
      "learning_rate": 9.247954560462927e-07,
      "logits/chosen": -2.0893867015838623,
      "logits/rejected": -2.0975987911224365,
      "logps/chosen": -241.7879180908203,
      "logps/rejected": -239.7193145751953,
      "loss": 0.5337,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.8641234636306763,
      "rewards/margins": 1.493786096572876,
      "rewards/rejected": -0.6296626329421997,
      "step": 65
    },
    {
      "epoch": 0.6542056074766355,
      "grad_norm": 38.40101091003172,
      "learning_rate": 9.109393920723001e-07,
      "logits/chosen": -2.0060245990753174,
      "logits/rejected": -1.9721952676773071,
      "logps/chosen": -256.7455139160156,
      "logps/rejected": -214.79403686523438,
      "loss": 0.533,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.6738287210464478,
      "rewards/margins": 1.2377485036849976,
      "rewards/rejected": -0.563919723033905,
      "step": 70
    },
    {
      "epoch": 0.7009345794392523,
      "grad_norm": 40.40382308924294,
      "learning_rate": 8.960352239917699e-07,
      "logits/chosen": -1.9107002019882202,
      "logits/rejected": -1.8710416555404663,
      "logps/chosen": -246.01028442382812,
      "logps/rejected": -224.3345947265625,
      "loss": 0.5452,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.6804067492485046,
      "rewards/margins": 1.3526126146316528,
      "rewards/rejected": -0.6722058653831482,
      "step": 75
    },
    {
      "epoch": 0.7476635514018691,
      "grad_norm": 37.2632538141041,
      "learning_rate": 8.801209650009814e-07,
      "logits/chosen": -1.8342845439910889,
      "logits/rejected": -1.7722011804580688,
      "logps/chosen": -238.6627960205078,
      "logps/rejected": -204.77720642089844,
      "loss": 0.4954,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": 0.8959934115409851,
      "rewards/margins": 1.759472131729126,
      "rewards/rejected": -0.8634785413742065,
      "step": 80
    },
    {
      "epoch": 0.794392523364486,
      "grad_norm": 41.835599660746425,
      "learning_rate": 8.632372045409141e-07,
      "logits/chosen": -1.9798192977905273,
      "logits/rejected": -1.9717410802841187,
      "logps/chosen": -251.6102294921875,
      "logps/rejected": -240.1038360595703,
      "loss": 0.5641,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 1.3537020683288574,
      "rewards/margins": 1.3792067766189575,
      "rewards/rejected": -0.025504767894744873,
      "step": 85
    },
    {
      "epoch": 0.8411214953271028,
      "grad_norm": 33.80081107648868,
      "learning_rate": 8.454270047735642e-07,
      "logits/chosen": -2.012608051300049,
      "logits/rejected": -2.0120816230773926,
      "logps/chosen": -253.6774139404297,
      "logps/rejected": -194.4556884765625,
      "loss": 0.5553,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.8388773798942566,
      "rewards/margins": 1.2849448919296265,
      "rewards/rejected": -0.4460674822330475,
      "step": 90
    },
    {
      "epoch": 0.8878504672897196,
      "grad_norm": 29.38305905850477,
      "learning_rate": 8.267357907515661e-07,
      "logits/chosen": -1.9744971990585327,
      "logits/rejected": -1.9980430603027344,
      "logps/chosen": -264.6416320800781,
      "logps/rejected": -246.7568817138672,
      "loss": 0.538,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.25706252455711365,
      "rewards/margins": 2.062177896499634,
      "rewards/rejected": -1.8051154613494873,
      "step": 95
    },
    {
      "epoch": 0.9345794392523364,
      "grad_norm": 33.767105871104185,
      "learning_rate": 8.072112345612433e-07,
      "logits/chosen": -1.9495391845703125,
      "logits/rejected": -1.8999382257461548,
      "logps/chosen": -248.95272827148438,
      "logps/rejected": -255.79019165039062,
      "loss": 0.5159,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.18092432618141174,
      "rewards/margins": 1.6218032836914062,
      "rewards/rejected": -1.8027276992797852,
      "step": 100
    },
    {
      "epoch": 0.9345794392523364,
      "eval_logits/chosen": -1.9645830392837524,
      "eval_logits/rejected": -1.9168211221694946,
      "eval_logps/chosen": -250.5727081298828,
      "eval_logps/rejected": -234.28236389160156,
      "eval_loss": 0.5285552144050598,
      "eval_rewards/accuracies": 0.78125,
      "eval_rewards/chosen": -0.21594171226024628,
      "eval_rewards/margins": 1.8031220436096191,
      "eval_rewards/rejected": -2.019063711166382,
      "eval_runtime": 202.4154,
      "eval_samples_per_second": 15.019,
      "eval_steps_per_second": 0.237,
      "step": 100
    },
    {
      "epoch": 0.9813084112149533,
      "grad_norm": 31.738611864064374,
      "learning_rate": 7.869031337345827e-07,
      "logits/chosen": -1.9719873666763306,
      "logits/rejected": -1.923525094985962,
      "logps/chosen": -260.2001953125,
      "logps/rejected": -236.92495727539062,
      "loss": 0.4808,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.06123792380094528,
      "rewards/margins": 1.9958425760269165,
      "rewards/rejected": -2.0570805072784424,
      "step": 105
    },
    {
      "epoch": 1.02803738317757,
      "grad_norm": 20.57706834965762,
      "learning_rate": 7.658632842402432e-07,
      "logits/chosen": -2.0060174465179443,
      "logits/rejected": -1.9614614248275757,
      "logps/chosen": -249.8937530517578,
      "logps/rejected": -242.28305053710938,
      "loss": 0.3194,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": 0.40553492307662964,
      "rewards/margins": 2.3744890689849854,
      "rewards/rejected": -1.968954086303711,
      "step": 110
    },
    {
      "epoch": 1.074766355140187,
      "grad_norm": 35.46781940178886,
      "learning_rate": 7.441453483775353e-07,
      "logits/chosen": -2.0191903114318848,
      "logits/rejected": -1.9993083477020264,
      "logps/chosen": -243.10147094726562,
      "logps/rejected": -233.6897735595703,
      "loss": 0.2252,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.1334301233291626,
      "rewards/margins": 3.1529507637023926,
      "rewards/rejected": -2.0195209980010986,
      "step": 115
    },
    {
      "epoch": 1.1214953271028036,
      "grad_norm": 21.651387151346523,
      "learning_rate": 7.218047179103112e-07,
      "logits/chosen": -2.0854923725128174,
      "logits/rejected": -2.022040605545044,
      "logps/chosen": -228.9936981201172,
      "logps/rejected": -211.9679412841797,
      "loss": 0.2292,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.1401965618133545,
      "rewards/margins": 2.9597201347351074,
      "rewards/rejected": -1.8195232152938843,
      "step": 120
    },
    {
      "epoch": 1.1682242990654206,
      "grad_norm": 24.010613449905463,
      "learning_rate": 6.988983727898413e-07,
      "logits/chosen": -2.136890411376953,
      "logits/rejected": -2.12919282913208,
      "logps/chosen": -243.7528076171875,
      "logps/rejected": -251.5740203857422,
      "loss": 0.233,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.6651722192764282,
      "rewards/margins": 3.5824947357177734,
      "rewards/rejected": -1.9173227548599243,
      "step": 125
    },
    {
      "epoch": 1.2149532710280373,
      "grad_norm": 19.19894650607632,
      "learning_rate": 6.754847358270066e-07,
      "logits/chosen": -2.1517224311828613,
      "logits/rejected": -2.156635046005249,
      "logps/chosen": -250.22463989257812,
      "logps/rejected": -209.24722290039062,
      "loss": 0.2381,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.3966515064239502,
      "rewards/margins": 3.149792432785034,
      "rewards/rejected": -1.7531406879425049,
      "step": 130
    },
    {
      "epoch": 1.2616822429906542,
      "grad_norm": 23.637404006148603,
      "learning_rate": 6.516235236844661e-07,
      "logits/chosen": -2.180138349533081,
      "logits/rejected": -2.123582363128662,
      "logps/chosen": -249.1603546142578,
      "logps/rejected": -240.6315460205078,
      "loss": 0.2453,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.434802770614624,
      "rewards/margins": 3.7537569999694824,
      "rewards/rejected": -2.3189542293548584,
      "step": 135
    },
    {
      "epoch": 1.308411214953271,
      "grad_norm": 21.150266522840564,
      "learning_rate": 6.273755945688457e-07,
      "logits/chosen": -2.1802070140838623,
      "logits/rejected": -2.1404995918273926,
      "logps/chosen": -261.7569274902344,
      "logps/rejected": -238.82333374023438,
      "loss": 0.2056,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.6147171258926392,
      "rewards/margins": 3.631941318511963,
      "rewards/rejected": -2.017223596572876,
      "step": 140
    },
    {
      "epoch": 1.355140186915888,
      "grad_norm": 19.216251285680404,
      "learning_rate": 6.02802793011411e-07,
      "logits/chosen": -2.2070083618164062,
      "logits/rejected": -2.18471360206604,
      "logps/chosen": -250.37966918945312,
      "logps/rejected": -233.73361206054688,
      "loss": 0.2691,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 0.6839503049850464,
      "rewards/margins": 3.4379615783691406,
      "rewards/rejected": -2.7540111541748047,
      "step": 145
    },
    {
      "epoch": 1.4018691588785046,
      "grad_norm": 23.267149481370343,
      "learning_rate": 5.779677921331093e-07,
      "logits/chosen": -2.293395757675171,
      "logits/rejected": -2.2372868061065674,
      "logps/chosen": -240.39102172851562,
      "logps/rejected": -243.92245483398438,
      "loss": 0.2666,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.220241665840149,
      "rewards/margins": 3.838820219039917,
      "rewards/rejected": -2.6185781955718994,
      "step": 150
    },
    {
      "epoch": 1.4018691588785046,
      "eval_logits/chosen": -2.284703254699707,
      "eval_logits/rejected": -2.2442755699157715,
      "eval_logps/chosen": -240.5095977783203,
      "eval_logps/rejected": -230.9029083251953,
      "eval_loss": 0.5667340159416199,
      "eval_rewards/accuracies": 0.7890625,
      "eval_rewards/chosen": 0.7903707027435303,
      "eval_rewards/margins": 2.4714887142181396,
      "eval_rewards/rejected": -1.6811178922653198,
      "eval_runtime": 202.5833,
      "eval_samples_per_second": 15.006,
      "eval_steps_per_second": 0.237,
      "step": 150
    },
    {
      "epoch": 1.4485981308411215,
      "grad_norm": 22.64610474864586,
      "learning_rate": 5.529339337962897e-07,
      "logits/chosen": -2.320802927017212,
      "logits/rejected": -2.2888898849487305,
      "logps/chosen": -232.2943878173828,
      "logps/rejected": -249.5731658935547,
      "loss": 0.2286,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.71249258518219,
      "rewards/margins": 3.7487540245056152,
      "rewards/rejected": -2.036261796951294,
      "step": 155
    },
    {
      "epoch": 1.4953271028037383,
      "grad_norm": 17.267431239861295,
      "learning_rate": 5.277650670507915e-07,
      "logits/chosen": -2.3630456924438477,
      "logits/rejected": -2.334383487701416,
      "logps/chosen": -239.9236297607422,
      "logps/rejected": -220.5778350830078,
      "loss": 0.233,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.5703070163726807,
      "rewards/margins": 3.3855056762695312,
      "rewards/rejected": -1.815198540687561,
      "step": 160
    },
    {
      "epoch": 1.542056074766355,
      "grad_norm": 28.93807721843343,
      "learning_rate": 5.025253852864471e-07,
      "logits/chosen": -2.456749439239502,
      "logits/rejected": -2.4346892833709717,
      "logps/chosen": -237.8841552734375,
      "logps/rejected": -234.561767578125,
      "loss": 0.3141,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.509019136428833,
      "rewards/margins": 3.890598773956299,
      "rewards/rejected": -2.3815793991088867,
      "step": 165
    },
    {
      "epoch": 1.588785046728972,
      "grad_norm": 21.109754151195787,
      "learning_rate": 4.77279262507344e-07,
      "logits/chosen": -2.4918441772460938,
      "logits/rejected": -2.464447021484375,
      "logps/chosen": -242.0265350341797,
      "logps/rejected": -236.3219757080078,
      "loss": 0.275,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.0706355571746826,
      "rewards/margins": 3.679105043411255,
      "rewards/rejected": -2.6084697246551514,
      "step": 170
    },
    {
      "epoch": 1.6355140186915889,
      "grad_norm": 21.650032107276107,
      "learning_rate": 4.5209108914542714e-07,
      "logits/chosen": -2.494457721710205,
      "logits/rejected": -2.5054736137390137,
      "logps/chosen": -242.4899139404297,
      "logps/rejected": -258.30059814453125,
      "loss": 0.2774,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.321101188659668,
      "rewards/margins": 4.072310447692871,
      "rewards/rejected": -2.7512094974517822,
      "step": 175
    },
    {
      "epoch": 1.6822429906542056,
      "grad_norm": 21.356500983747313,
      "learning_rate": 4.2702510783220475e-07,
      "logits/chosen": -2.48954701423645,
      "logits/rejected": -2.4850687980651855,
      "logps/chosen": -227.01760864257812,
      "logps/rejected": -231.8510284423828,
      "loss": 0.2598,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 0.9325019121170044,
      "rewards/margins": 3.5170364379882812,
      "rewards/rejected": -2.5845344066619873,
      "step": 180
    },
    {
      "epoch": 1.7289719626168223,
      "grad_norm": 23.401658906609267,
      "learning_rate": 4.0214524954741586e-07,
      "logits/chosen": -2.4770901203155518,
      "logits/rejected": -2.449385404586792,
      "logps/chosen": -249.5960235595703,
      "logps/rejected": -240.88900756835938,
      "loss": 0.2955,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.3769264221191406,
      "rewards/margins": 3.714228868484497,
      "rewards/rejected": -2.3373022079467773,
      "step": 185
    },
    {
      "epoch": 1.7757009345794392,
      "grad_norm": 21.273588354036704,
      "learning_rate": 3.7751497056257305e-07,
      "logits/chosen": -2.4559922218322754,
      "logits/rejected": -2.4245378971099854,
      "logps/chosen": -249.3975372314453,
      "logps/rejected": -253.1663818359375,
      "loss": 0.2401,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.1007009744644165,
      "rewards/margins": 3.69682240486145,
      "rewards/rejected": -2.596121311187744,
      "step": 190
    },
    {
      "epoch": 1.8224299065420562,
      "grad_norm": 26.50900109669566,
      "learning_rate": 3.531970905952478e-07,
      "logits/chosen": -2.4134464263916016,
      "logits/rejected": -2.398714542388916,
      "logps/chosen": -243.4962615966797,
      "logps/rejected": -259.9742736816406,
      "loss": 0.3042,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.2408723831176758,
      "rewards/margins": 3.5389716625213623,
      "rewards/rejected": -2.2980995178222656,
      "step": 195
    },
    {
      "epoch": 1.8691588785046729,
      "grad_norm": 20.04208979237232,
      "learning_rate": 3.2925363258689553e-07,
      "logits/chosen": -2.4002745151519775,
      "logits/rejected": -2.381742000579834,
      "logps/chosen": -227.89956665039062,
      "logps/rejected": -238.51956176757812,
      "loss": 0.3127,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 1.159445881843567,
      "rewards/margins": 3.3063011169433594,
      "rewards/rejected": -2.146855115890503,
      "step": 200
    },
    {
      "epoch": 1.8691588785046729,
      "eval_logits/chosen": -2.4148459434509277,
      "eval_logits/rejected": -2.3878657817840576,
      "eval_logps/chosen": -241.9330291748047,
      "eval_logps/rejected": -232.25018310546875,
      "eval_loss": 0.5355702042579651,
      "eval_rewards/accuracies": 0.8046875,
      "eval_rewards/chosen": 0.6480298638343811,
      "eval_rewards/margins": 2.463876962661743,
      "eval_rewards/rejected": -1.8158468008041382,
      "eval_runtime": 202.4674,
      "eval_samples_per_second": 15.015,
      "eval_steps_per_second": 0.237,
      "step": 200
    },
    {
      "epoch": 1.9158878504672896,
      "grad_norm": 20.064900060265114,
      "learning_rate": 3.0574566451286086e-07,
      "logits/chosen": -2.4189305305480957,
      "logits/rejected": -2.40187406539917,
      "logps/chosen": -265.58294677734375,
      "logps/rejected": -251.65847778320312,
      "loss": 0.2513,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.1464284658432007,
      "rewards/margins": 3.5169920921325684,
      "rewards/rejected": -2.3705639839172363,
      "step": 205
    },
    {
      "epoch": 1.9626168224299065,
      "grad_norm": 22.038470491161178,
      "learning_rate": 2.8273314362803333e-07,
      "logits/chosen": -2.4120736122131348,
      "logits/rejected": -2.4032583236694336,
      "logps/chosen": -282.6405334472656,
      "logps/rejected": -240.655029296875,
      "loss": 0.2764,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.6487640142440796,
      "rewards/margins": 3.7647461891174316,
      "rewards/rejected": -2.1159825325012207,
      "step": 210
    },
    {
      "epoch": 2.0093457943925235,
      "grad_norm": 12.721605555302636,
      "learning_rate": 2.602747635454047e-07,
      "logits/chosen": -2.4080586433410645,
      "logits/rejected": -2.3771650791168213,
      "logps/chosen": -258.2095947265625,
      "logps/rejected": -212.6840362548828,
      "loss": 0.2701,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.2774152755737305,
      "rewards/margins": 3.8353512287139893,
      "rewards/rejected": -2.557935953140259,
      "step": 215
    },
    {
      "epoch": 2.05607476635514,
      "grad_norm": 14.30139581032378,
      "learning_rate": 2.384278045375523e-07,
      "logits/chosen": -2.3718018531799316,
      "logits/rejected": -2.343240261077881,
      "logps/chosen": -236.43167114257812,
      "logps/rejected": -254.13290405273438,
      "loss": 0.1493,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 1.3953884840011597,
      "rewards/margins": 4.300187110900879,
      "rewards/rejected": -2.9047982692718506,
      "step": 220
    },
    {
      "epoch": 2.102803738317757,
      "grad_norm": 17.63579128313249,
      "learning_rate": 2.1724798744286071e-07,
      "logits/chosen": -2.3075740337371826,
      "logits/rejected": -2.2964611053466797,
      "logps/chosen": -247.1775360107422,
      "logps/rejected": -232.0567169189453,
      "loss": 0.1552,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.4666545391082764,
      "rewards/margins": 4.158020496368408,
      "rewards/rejected": -2.691365957260132,
      "step": 225
    },
    {
      "epoch": 2.149532710280374,
      "grad_norm": 13.340130576113058,
      "learning_rate": 1.9678933154909095e-07,
      "logits/chosen": -2.3146255016326904,
      "logits/rejected": -2.270624876022339,
      "logps/chosen": -234.5230712890625,
      "logps/rejected": -253.0650634765625,
      "loss": 0.1433,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.5044587850570679,
      "rewards/margins": 4.230677604675293,
      "rewards/rejected": -2.7262187004089355,
      "step": 230
    },
    {
      "epoch": 2.196261682242991,
      "grad_norm": 14.326810577042355,
      "learning_rate": 1.77104016816768e-07,
      "logits/chosen": -2.29292631149292,
      "logits/rejected": -2.2760777473449707,
      "logps/chosen": -235.0215606689453,
      "logps/rejected": -248.5024871826172,
      "loss": 0.1523,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.5524753332138062,
      "rewards/margins": 4.352492332458496,
      "rewards/rejected": -2.8000173568725586,
      "step": 235
    },
    {
      "epoch": 2.2429906542056073,
      "grad_norm": 10.917474403037492,
      "learning_rate": 1.5824225079378684e-07,
      "logits/chosen": -2.287108898162842,
      "logits/rejected": -2.2589364051818848,
      "logps/chosen": -212.53848266601562,
      "logps/rejected": -262.1377258300781,
      "loss": 0.1379,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.49385404586792,
      "rewards/margins": 4.177613735198975,
      "rewards/rejected": -2.6837594509124756,
      "step": 240
    },
    {
      "epoch": 2.289719626168224,
      "grad_norm": 18.70156944510107,
      "learning_rate": 1.4025214056067237e-07,
      "logits/chosen": -2.3043479919433594,
      "logits/rejected": -2.2352633476257324,
      "logps/chosen": -237.90090942382812,
      "logps/rejected": -255.3131103515625,
      "loss": 0.1504,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.958906888961792,
      "rewards/margins": 4.553770065307617,
      "rewards/rejected": -2.594862461090088,
      "step": 245
    },
    {
      "epoch": 2.336448598130841,
      "grad_norm": 18.58957134665332,
      "learning_rate": 1.2317957003309725e-07,
      "logits/chosen": -2.28039813041687,
      "logits/rejected": -2.269214630126953,
      "logps/chosen": -242.89535522460938,
      "logps/rejected": -228.6979217529297,
      "loss": 0.152,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 1.4991402626037598,
      "rewards/margins": 4.210090637207031,
      "rewards/rejected": -2.7109508514404297,
      "step": 250
    },
    {
      "epoch": 2.336448598130841,
      "eval_logits/chosen": -2.28143310546875,
      "eval_logits/rejected": -2.2442617416381836,
      "eval_logps/chosen": -242.3255157470703,
      "eval_logps/rejected": -234.9364776611328,
      "eval_loss": 0.5442168116569519,
      "eval_rewards/accuracies": 0.7890625,
      "eval_rewards/chosen": 0.6087805032730103,
      "eval_rewards/margins": 2.6932570934295654,
      "eval_rewards/rejected": -2.0844767093658447,
      "eval_runtime": 202.4297,
      "eval_samples_per_second": 15.018,
      "eval_steps_per_second": 0.237,
      "step": 250
    },
    {
      "epoch": 2.383177570093458,
      "grad_norm": 12.504767667176505,
      "learning_rate": 1.0706808293459873e-07,
      "logits/chosen": -2.300673007965088,
      "logits/rejected": -2.2649149894714355,
      "logps/chosen": -240.6633758544922,
      "logps/rejected": -250.4842987060547,
      "loss": 0.1378,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.6441866159439087,
      "rewards/margins": 4.3823652267456055,
      "rewards/rejected": -2.7381789684295654,
      "step": 255
    },
    {
      "epoch": 2.4299065420560746,
      "grad_norm": 13.794920815357296,
      "learning_rate": 9.195877173797534e-08,
      "logits/chosen": -2.300203323364258,
      "logits/rejected": -2.258514881134033,
      "logps/chosen": -262.6177978515625,
      "logps/rejected": -264.5312194824219,
      "loss": 0.1359,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.8180078268051147,
      "rewards/margins": 4.726799964904785,
      "rewards/rejected": -2.908792018890381,
      "step": 260
    },
    {
      "epoch": 2.4766355140186915,
      "grad_norm": 16.46425878105061,
      "learning_rate": 7.789017285861438e-08,
      "logits/chosen": -2.2753031253814697,
      "logits/rejected": -2.2750420570373535,
      "logps/chosen": -253.7283935546875,
      "logps/rejected": -241.67990112304688,
      "loss": 0.1319,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.0793383121490479,
      "rewards/margins": 4.2754716873168945,
      "rewards/rejected": -3.1961333751678467,
      "step": 265
    },
    {
      "epoch": 2.5233644859813085,
      "grad_norm": 13.487096691897303,
      "learning_rate": 6.489816836706785e-08,
      "logits/chosen": -2.2921862602233887,
      "logits/rejected": -2.2367775440216064,
      "logps/chosen": -272.6159973144531,
      "logps/rejected": -244.61575317382812,
      "loss": 0.1606,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 1.8186426162719727,
      "rewards/margins": 4.736049175262451,
      "rewards/rejected": -2.9174060821533203,
      "step": 270
    },
    {
      "epoch": 2.5700934579439254,
      "grad_norm": 22.37915209288031,
      "learning_rate": 5.3015894471550914e-08,
      "logits/chosen": -2.2553036212921143,
      "logits/rejected": -2.249110460281372,
      "logps/chosen": -236.9984130859375,
      "logps/rejected": -265.2401123046875,
      "loss": 0.1559,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.1508326530456543,
      "rewards/margins": 4.554782390594482,
      "rewards/rejected": -3.403949737548828,
      "step": 275
    },
    {
      "epoch": 2.616822429906542,
      "grad_norm": 13.985165706457115,
      "learning_rate": 4.227365700378799e-08,
      "logits/chosen": -2.293137788772583,
      "logits/rejected": -2.2412238121032715,
      "logps/chosen": -241.9412841796875,
      "logps/rejected": -224.3672332763672,
      "loss": 0.127,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.2079182863235474,
      "rewards/margins": 4.391335964202881,
      "rewards/rejected": -3.183417558670044,
      "step": 280
    },
    {
      "epoch": 2.663551401869159,
      "grad_norm": 11.021506399671743,
      "learning_rate": 3.269885412375223e-08,
      "logits/chosen": -2.291661262512207,
      "logits/rejected": -2.2645070552825928,
      "logps/chosen": -234.7343292236328,
      "logps/rejected": -226.64303588867188,
      "loss": 0.124,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.284121036529541,
      "rewards/margins": 4.28745174407959,
      "rewards/rejected": -3.003331184387207,
      "step": 285
    },
    {
      "epoch": 2.710280373831776,
      "grad_norm": 15.975321991322621,
      "learning_rate": 2.4315906440446952e-08,
      "logits/chosen": -2.2947776317596436,
      "logits/rejected": -2.2545151710510254,
      "logps/chosen": -259.29241943359375,
      "logps/rejected": -244.57156372070312,
      "loss": 0.1606,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 1.2673962116241455,
      "rewards/margins": 4.276968002319336,
      "rewards/rejected": -3.0095717906951904,
      "step": 290
    },
    {
      "epoch": 2.7570093457943923,
      "grad_norm": 15.010245498160302,
      "learning_rate": 1.7146194726952778e-08,
      "logits/chosen": -2.317955732345581,
      "logits/rejected": -2.2629940509796143,
      "logps/chosen": -269.7888488769531,
      "logps/rejected": -267.582763671875,
      "loss": 0.1412,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.5963925123214722,
      "rewards/margins": 4.7224860191345215,
      "rewards/rejected": -3.126093626022339,
      "step": 295
    },
    {
      "epoch": 2.803738317757009,
      "grad_norm": 15.472628492526889,
      "learning_rate": 1.1208005388599951e-08,
      "logits/chosen": -2.2652642726898193,
      "logits/rejected": -2.266711950302124,
      "logps/chosen": -251.26358032226562,
      "logps/rejected": -226.0489044189453,
      "loss": 0.1431,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.2339963912963867,
      "rewards/margins": 4.164398193359375,
      "rewards/rejected": -2.9304018020629883,
      "step": 300
    },
    {
      "epoch": 2.803738317757009,
      "eval_logits/chosen": -2.282158374786377,
      "eval_logits/rejected": -2.2467496395111084,
      "eval_logps/chosen": -244.18870544433594,
      "eval_logps/rejected": -238.83473205566406,
      "eval_loss": 0.5449995398521423,
      "eval_rewards/accuracies": 0.7942708134651184,
      "eval_rewards/chosen": 0.42245936393737793,
      "eval_rewards/margins": 2.8967597484588623,
      "eval_rewards/rejected": -2.4743001461029053,
      "eval_runtime": 202.3815,
      "eval_samples_per_second": 15.021,
      "eval_steps_per_second": 0.237,
      "step": 300
    },
    {
      "epoch": 2.850467289719626,
      "grad_norm": 16.936006866528523,
      "learning_rate": 6.516483823349794e-09,
      "logits/chosen": -2.278529644012451,
      "logits/rejected": -2.2658190727233887,
      "logps/chosen": -254.26248168945312,
      "logps/rejected": -235.7794647216797,
      "loss": 0.1681,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.6976511478424072,
      "rewards/margins": 4.890439510345459,
      "rewards/rejected": -3.192788600921631,
      "step": 305
    },
    {
      "epoch": 2.897196261682243,
      "grad_norm": 14.960616808458298,
      "learning_rate": 3.0835957933397773e-09,
      "logits/chosen": -2.276339054107666,
      "logits/rejected": -2.2654058933258057,
      "logps/chosen": -237.9030303955078,
      "logps/rejected": -246.7018585205078,
      "loss": 0.1553,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.366158127784729,
      "rewards/margins": 4.665255069732666,
      "rewards/rejected": -3.2990965843200684,
      "step": 310
    },
    {
      "epoch": 2.94392523364486,
      "grad_norm": 17.16613844616987,
      "learning_rate": 9.180969061143851e-10,
      "logits/chosen": -2.2957701683044434,
      "logits/rejected": -2.253190279006958,
      "logps/chosen": -253.9457550048828,
      "logps/rejected": -231.51229858398438,
      "loss": 0.137,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.2239946126937866,
      "rewards/margins": 4.194064617156982,
      "rewards/rejected": -2.9700698852539062,
      "step": 315
    },
    {
      "epoch": 2.9906542056074765,
      "grad_norm": 14.055773300679348,
      "learning_rate": 2.5510283379992504e-11,
      "logits/chosen": -2.2696633338928223,
      "logits/rejected": -2.2665677070617676,
      "logps/chosen": -260.431640625,
      "logps/rejected": -258.1851806640625,
      "loss": 0.156,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.6931114196777344,
      "rewards/margins": 4.892333984375,
      "rewards/rejected": -3.1992225646972656,
      "step": 320
    },
    {
      "epoch": 3.0,
      "step": 321,
      "total_flos": 3785055088410624.0,
      "train_loss": 0.32465138984989156,
      "train_runtime": 12160.5022,
      "train_samples_per_second": 6.748,
      "train_steps_per_second": 0.026
    }
  ],
  "logging_steps": 5,
  "max_steps": 321,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3785055088410624.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}