{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 385,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 13.0625,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 1.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 10.875,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.074951171875,
      "logits/rejected": 80.7809829711914,
      "logps/chosen": -34.20733642578125,
      "logps/rejected": -32.97297668457031,
      "loss": 0.9951,
      "rewards/accuracies": 0.4583333432674408,
      "rewards/chosen": 0.003438829444348812,
      "rewards/margins": 0.004895869642496109,
      "rewards/rejected": -0.0014570390339940786,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 12.6875,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.64326477050781,
      "logits/rejected": 80.53416442871094,
      "logps/chosen": -33.720584869384766,
      "logps/rejected": -30.82167625427246,
      "loss": 0.9986,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": -0.003067571669816971,
      "rewards/margins": 0.0013512909645214677,
      "rewards/rejected": -0.004418861120939255,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 12.0,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.51445007324219,
      "logits/rejected": 82.54810333251953,
      "logps/chosen": -33.81728744506836,
      "logps/rejected": -31.204355239868164,
      "loss": 1.0024,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": 0.019553204998373985,
      "rewards/margins": -0.002394508570432663,
      "rewards/rejected": 0.0219477117061615,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 11.8125,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 81.101806640625,
      "logits/rejected": 81.09938049316406,
      "logps/chosen": -32.73223876953125,
      "logps/rejected": -33.143699645996094,
      "loss": 0.9783,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.04418279603123665,
      "rewards/margins": 0.02170029655098915,
      "rewards/rejected": 0.022482499480247498,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 11.9375,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.82136535644531,
      "logits/rejected": 78.82911682128906,
      "logps/chosen": -30.3783016204834,
      "logps/rejected": -30.641677856445312,
      "loss": 0.9605,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.082694411277771,
      "rewards/margins": 0.039533428847789764,
      "rewards/rejected": 0.04316098242998123,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 9.75,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 83.49494171142578,
      "logits/rejected": 83.55232238769531,
      "logps/chosen": -30.781469345092773,
      "logps/rejected": -29.190662384033203,
      "loss": 1.0006,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": 0.03936903923749924,
      "rewards/margins": -0.0006168211111798882,
      "rewards/rejected": 0.039985861629247665,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 12.0625,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 84.1231918334961,
      "logits/rejected": 84.15650939941406,
      "logps/chosen": -30.209863662719727,
      "logps/rejected": -32.619781494140625,
      "loss": 0.9999,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.041915904730558395,
      "rewards/margins": 0.000123101839562878,
      "rewards/rejected": 0.04179280251264572,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 12.375,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.8515853881836,
      "logits/rejected": 81.83155822753906,
      "logps/chosen": -30.982410430908203,
      "logps/rejected": -30.617040634155273,
      "loss": 0.9618,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.05225307494401932,
      "rewards/margins": 0.03819245845079422,
      "rewards/rejected": 0.014060619287192822,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 14.75,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.68502807617188,
      "logits/rejected": 78.65934753417969,
      "logps/chosen": -32.18014144897461,
      "logps/rejected": -30.878421783447266,
      "loss": 0.9627,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.04580371826887131,
      "rewards/margins": 0.03849860280752182,
      "rewards/rejected": 0.007305114530026913,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 12.375,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.6978530883789,
      "logits/rejected": 83.72080993652344,
      "logps/chosen": -33.769813537597656,
      "logps/rejected": -31.638240814208984,
      "loss": 0.9577,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.05351760983467102,
      "rewards/margins": 0.04228735715150833,
      "rewards/rejected": 0.01123025082051754,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.71279907226562,
      "eval_logits/rejected": 98.70589447021484,
      "eval_logps/chosen": -32.32750701904297,
      "eval_logps/rejected": -35.914642333984375,
      "eval_loss": 0.993482768535614,
      "eval_rewards/accuracies": 0.5066444873809814,
      "eval_rewards/chosen": 0.011567190289497375,
      "eval_rewards/margins": 0.006378817837685347,
      "eval_rewards/rejected": 0.005188372451812029,
      "eval_runtime": 104.0772,
      "eval_samples_per_second": 3.296,
      "eval_steps_per_second": 0.413,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 14.125,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.8444595336914,
      "logits/rejected": 83.72880554199219,
      "logps/chosen": -32.138458251953125,
      "logps/rejected": -32.649452209472656,
      "loss": 0.9122,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.08843465149402618,
      "rewards/margins": 0.0877910926938057,
      "rewards/rejected": 0.0006435603136196733,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 13.0,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 83.81868743896484,
      "logits/rejected": 83.92826080322266,
      "logps/chosen": -28.111730575561523,
      "logps/rejected": -35.34394454956055,
      "loss": 0.9297,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.08178504556417465,
      "rewards/margins": 0.07026515156030655,
      "rewards/rejected": 0.011519892141222954,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 9.375,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 80.93681335449219,
      "logits/rejected": 80.97130584716797,
      "logps/chosen": -30.195592880249023,
      "logps/rejected": -31.842870712280273,
      "loss": 0.939,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.07117608934640884,
      "rewards/margins": 0.06230046600103378,
      "rewards/rejected": 0.008875617757439613,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 11.0625,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 81.7886962890625,
      "logits/rejected": 81.80296325683594,
      "logps/chosen": -26.81143569946289,
      "logps/rejected": -32.920143127441406,
      "loss": 0.8867,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.0635889396071434,
      "rewards/margins": 0.11326569318771362,
      "rewards/rejected": -0.04967674985527992,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 12.1875,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 80.10134887695312,
      "logits/rejected": 80.06209564208984,
      "logps/chosen": -28.821029663085938,
      "logps/rejected": -33.226470947265625,
      "loss": 0.8772,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.05680803209543228,
      "rewards/margins": 0.1228049248456955,
      "rewards/rejected": -0.06599690765142441,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 13.4375,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 81.63700866699219,
      "logits/rejected": 81.66020202636719,
      "logps/chosen": -34.080284118652344,
      "logps/rejected": -30.817296981811523,
      "loss": 0.9079,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.005646559409797192,
      "rewards/margins": 0.09212217479944229,
      "rewards/rejected": -0.08647561073303223,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 15.125,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 82.27827453613281,
      "logits/rejected": 82.23299407958984,
      "logps/chosen": -31.016094207763672,
      "logps/rejected": -33.03407669067383,
      "loss": 0.8743,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.024319300428032875,
      "rewards/margins": 0.12572081387043,
      "rewards/rejected": -0.10140150785446167,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 12.5,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 79.44108581542969,
      "logits/rejected": 79.4173355102539,
      "logps/chosen": -30.948467254638672,
      "logps/rejected": -32.03376007080078,
      "loss": 0.8875,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.02486853487789631,
      "rewards/margins": 0.1125236377120018,
      "rewards/rejected": -0.08765510469675064,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 9.6875,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 80.90037536621094,
      "logits/rejected": 80.8733901977539,
      "logps/chosen": -30.740875244140625,
      "logps/rejected": -31.226177215576172,
      "loss": 0.9221,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.002553999889642,
      "rewards/margins": 0.07836906611919403,
      "rewards/rejected": -0.07581506669521332,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 14.4375,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 76.02571105957031,
      "logits/rejected": 75.9735107421875,
      "logps/chosen": -34.12778854370117,
      "logps/rejected": -33.30614471435547,
      "loss": 0.8562,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.0624953992664814,
      "rewards/margins": 0.14376921951770782,
      "rewards/rejected": -0.08127383887767792,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 97.97191619873047,
      "eval_logits/rejected": 97.95458984375,
      "eval_logps/chosen": -32.94685363769531,
      "eval_logps/rejected": -36.8216552734375,
      "eval_loss": 0.9646754860877991,
      "eval_rewards/accuracies": 0.5676910281181335,
      "eval_rewards/chosen": -0.05036771669983864,
      "eval_rewards/margins": 0.035145342350006104,
      "eval_rewards/rejected": -0.08551305532455444,
      "eval_runtime": 103.8205,
      "eval_samples_per_second": 3.304,
      "eval_steps_per_second": 0.414,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 18.625,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 78.56858825683594,
      "logits/rejected": 78.47276306152344,
      "logps/chosen": -33.586204528808594,
      "logps/rejected": -36.05295944213867,
      "loss": 0.8633,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0359516479074955,
      "rewards/margins": 0.1426382064819336,
      "rewards/rejected": -0.1066865548491478,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 14.6875,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 80.53726959228516,
      "logits/rejected": 80.62950134277344,
      "logps/chosen": -31.454355239868164,
      "logps/rejected": -32.02970504760742,
      "loss": 0.8286,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.049460187554359436,
      "rewards/margins": 0.17953529953956604,
      "rewards/rejected": -0.1300750970840454,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 14.0,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 77.57215881347656,
      "logits/rejected": 77.61759948730469,
      "logps/chosen": -32.74131393432617,
      "logps/rejected": -35.29151153564453,
      "loss": 0.8713,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.006443141493946314,
      "rewards/margins": 0.13437876105308533,
      "rewards/rejected": -0.12793561816215515,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 16.5,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 79.76509094238281,
      "logits/rejected": 80.07305908203125,
      "logps/chosen": -31.252161026000977,
      "logps/rejected": -32.75459671020508,
      "loss": 0.8345,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.04359050467610359,
      "rewards/margins": 0.16584597527980804,
      "rewards/rejected": -0.12225550413131714,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 15.0625,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 78.32225036621094,
      "logits/rejected": 78.37224578857422,
      "logps/chosen": -27.815731048583984,
      "logps/rejected": -31.077539443969727,
      "loss": 0.9043,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.02198370359838009,
      "rewards/margins": 0.09569612890481949,
      "rewards/rejected": -0.11767983436584473,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 13.75,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 75.31268310546875,
      "logits/rejected": 75.46823120117188,
      "logps/chosen": -31.16106605529785,
      "logps/rejected": -38.22980880737305,
      "loss": 0.7589,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.039992958307266235,
      "rewards/margins": 0.2528113126754761,
      "rewards/rejected": -0.21281830966472626,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 13.1875,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 74.42684173583984,
      "logits/rejected": 74.45647430419922,
      "logps/chosen": -31.796234130859375,
      "logps/rejected": -33.113197326660156,
      "loss": 0.8432,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.0017309554386883974,
      "rewards/margins": 0.15904627740383148,
      "rewards/rejected": -0.1607772409915924,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 19.75,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 77.391845703125,
      "logits/rejected": 77.1694564819336,
      "logps/chosen": -32.110328674316406,
      "logps/rejected": -30.959293365478516,
      "loss": 0.9109,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -0.030043601989746094,
      "rewards/margins": 0.09918614476919174,
      "rewards/rejected": -0.12922975420951843,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 13.1875,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 77.39460754394531,
      "logits/rejected": 77.31553649902344,
      "logps/chosen": -33.84746551513672,
      "logps/rejected": -34.28327178955078,
      "loss": 0.7586,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.02567802369594574,
      "rewards/margins": 0.2523185908794403,
      "rewards/rejected": -0.22664058208465576,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 14.8125,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 73.05286407470703,
      "logits/rejected": 73.1805648803711,
      "logps/chosen": -33.01818084716797,
      "logps/rejected": -30.520904541015625,
      "loss": 0.8271,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.03875284641981125,
      "rewards/margins": 0.18064062297344208,
      "rewards/rejected": -0.14188775420188904,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 97.71524047851562,
      "eval_logits/rejected": 97.68975067138672,
      "eval_logps/chosen": -33.206886291503906,
      "eval_logps/rejected": -37.196205139160156,
      "eval_loss": 0.9533767700195312,
      "eval_rewards/accuracies": 0.5714285373687744,
      "eval_rewards/chosen": -0.07637124508619308,
      "eval_rewards/margins": 0.04659651592373848,
      "eval_rewards/rejected": -0.12296776473522186,
      "eval_runtime": 104.0124,
      "eval_samples_per_second": 3.298,
      "eval_steps_per_second": 0.413,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 13.0,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 80.51168060302734,
      "logits/rejected": 80.51959228515625,
      "logps/chosen": -30.940738677978516,
      "logps/rejected": -33.976158142089844,
      "loss": 0.841,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.00772014120593667,
      "rewards/margins": 0.1699191778898239,
      "rewards/rejected": -0.17763930559158325,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 14.9375,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 77.665771484375,
      "logits/rejected": 77.6807861328125,
      "logps/chosen": -31.073196411132812,
      "logps/rejected": -30.321044921875,
      "loss": 0.8202,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.05253634601831436,
      "rewards/margins": 0.19269177317619324,
      "rewards/rejected": -0.14015543460845947,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 16.375,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 74.7007064819336,
      "logits/rejected": 74.74031829833984,
      "logps/chosen": -29.66641616821289,
      "logps/rejected": -34.42815399169922,
      "loss": 0.7549,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.06861907988786697,
      "rewards/margins": 0.24794825911521912,
      "rewards/rejected": -0.17932915687561035,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 17.125,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 79.32682800292969,
      "logits/rejected": 79.36729431152344,
      "logps/chosen": -33.05027770996094,
      "logps/rejected": -35.377281188964844,
      "loss": 0.7972,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.01618681475520134,
      "rewards/margins": 0.21704569458961487,
      "rewards/rejected": -0.20085887610912323,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 15.5,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 78.2139663696289,
      "logits/rejected": 78.22695922851562,
      "logps/chosen": -33.175575256347656,
      "logps/rejected": -34.91777801513672,
      "loss": 0.7788,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.04681625962257385,
      "rewards/margins": 0.22941403090953827,
      "rewards/rejected": -0.18259775638580322,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 11.6875,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 79.90480041503906,
      "logits/rejected": 79.93946075439453,
      "logps/chosen": -28.85941505432129,
      "logps/rejected": -33.032447814941406,
      "loss": 0.8014,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.06927979737520218,
      "rewards/margins": 0.20753948390483856,
      "rewards/rejected": -0.13825969398021698,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 12.8125,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 79.3287582397461,
      "logits/rejected": 79.33840942382812,
      "logps/chosen": -33.099693298339844,
      "logps/rejected": -37.02233123779297,
      "loss": 0.8425,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.022353025153279305,
      "rewards/margins": 0.16894304752349854,
      "rewards/rejected": -0.1912960708141327,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 13.25,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 72.94795227050781,
      "logits/rejected": 72.81517791748047,
      "logps/chosen": -30.674768447875977,
      "logps/rejected": -29.731273651123047,
      "loss": 0.8564,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.009864235296845436,
      "rewards/margins": 0.14720600843429565,
      "rewards/rejected": -0.15707024931907654,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.88636748450143,
      "train_runtime": 2554.6221,
      "train_samples_per_second": 1.205,
      "train_steps_per_second": 0.151
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}