{
  "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": 656.0,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 6.25,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 536.0,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.07795715332031,
      "logits/rejected": 80.78132629394531,
      "logps/chosen": -34.14677047729492,
      "logps/rejected": -33.015296936035156,
      "loss": 6.1804,
      "rewards/accuracies": 0.5277777910232544,
      "rewards/chosen": 0.0189919825643301,
      "rewards/margins": 0.030369587242603302,
      "rewards/rejected": -0.011377605609595776,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 592.0,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.67132568359375,
      "logits/rejected": 80.56291198730469,
      "logps/chosen": -33.51841354370117,
      "logps/rejected": -30.76877212524414,
      "loss": 6.6549,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.034299030900001526,
      "rewards/margins": 0.03255577012896538,
      "rewards/rejected": 0.0017432533204555511,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 448.0,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.46919250488281,
      "logits/rejected": 82.5018539428711,
      "logps/chosen": -33.83555603027344,
      "logps/rejected": -31.259449005126953,
      "loss": 7.4075,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": 0.03545312210917473,
      "rewards/margins": 0.0025764047168195248,
      "rewards/rejected": 0.03287670761346817,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 612.0,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 80.91734313964844,
      "logits/rejected": 80.9122314453125,
      "logps/chosen": -32.870399475097656,
      "logps/rejected": -33.192291259765625,
      "loss": 6.9708,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.06073303893208504,
      "rewards/margins": 0.025485849007964134,
      "rewards/rejected": 0.03524719178676605,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 528.0,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.42672729492188,
      "logits/rejected": 78.44120788574219,
      "logps/chosen": -30.90542984008789,
      "logps/rejected": -31.027896881103516,
      "loss": 6.7099,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.059963710606098175,
      "rewards/margins": 0.05088571831583977,
      "rewards/rejected": 0.009077982977032661,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 498.0,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 82.86418151855469,
      "logits/rejected": 82.91536712646484,
      "logps/chosen": -31.078943252563477,
      "logps/rejected": -29.606725692749023,
      "loss": 7.0829,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.0192436333745718,
      "rewards/margins": 0.022484425455331802,
      "rewards/rejected": -0.0032407897524535656,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 960.0,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 83.54774475097656,
      "logits/rejected": 83.57620239257812,
      "logps/chosen": -30.6588077545166,
      "logps/rejected": -33.22206115722656,
      "loss": 6.976,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.0059564015828073025,
      "rewards/margins": 0.0309152714908123,
      "rewards/rejected": -0.03687167167663574,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 556.0,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.12471771240234,
      "logits/rejected": 81.11027526855469,
      "logps/chosen": -31.388225555419922,
      "logps/rejected": -30.88912010192871,
      "loss": 6.8226,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.023343011736869812,
      "rewards/margins": 0.049637455493211746,
      "rewards/rejected": -0.026294440031051636,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 680.0,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.06856536865234,
      "logits/rejected": 78.03953552246094,
      "logps/chosen": -32.45366668701172,
      "logps/rejected": -31.23443603515625,
      "loss": 6.1663,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.03690256550908089,
      "rewards/margins": 0.09349562227725983,
      "rewards/rejected": -0.056593067944049835,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 396.0,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.3121566772461,
      "logits/rejected": 83.32379913330078,
      "logps/chosen": -34.064453125,
      "logps/rejected": -31.79021644592285,
      "loss": 6.7218,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.04810718446969986,
      "rewards/margins": 0.05604150891304016,
      "rewards/rejected": -0.00793431792408228,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.67085266113281,
      "eval_logits/rejected": 98.65489959716797,
      "eval_logps/chosen": -32.457313537597656,
      "eval_logps/rejected": -36.04214859008789,
      "eval_loss": 7.570494651794434,
      "eval_rewards/accuracies": 0.5083056688308716,
      "eval_rewards/chosen": -0.002827130490913987,
      "eval_rewards/margins": 0.0122976154088974,
      "eval_rewards/rejected": -0.015124746598303318,
      "eval_runtime": 104.1006,
      "eval_samples_per_second": 3.295,
      "eval_steps_per_second": 0.413,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 752.0,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.45378112792969,
      "logits/rejected": 83.33932495117188,
      "logps/chosen": -32.50124740600586,
      "logps/rejected": -32.954078674316406,
      "loss": 5.6892,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.104311004281044,
      "rewards/margins": 0.16394877433776855,
      "rewards/rejected": -0.05963778495788574,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 620.0,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 83.64091491699219,
      "logits/rejected": 83.74860382080078,
      "logps/chosen": -28.32021141052246,
      "logps/rejected": -35.66899871826172,
      "loss": 5.3716,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.12187393009662628,
      "rewards/margins": 0.16384465992450714,
      "rewards/rejected": -0.04197072237730026,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 484.0,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 80.95816040039062,
      "logits/rejected": 80.97350311279297,
      "logps/chosen": -30.554763793945312,
      "logps/rejected": -32.3652458190918,
      "loss": 6.2467,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.07051900774240494,
      "rewards/margins": 0.15724298357963562,
      "rewards/rejected": -0.08672399073839188,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 428.0,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 82.39835357666016,
      "logits/rejected": 82.39920806884766,
      "logps/chosen": -27.08688735961914,
      "logps/rejected": -33.00067901611328,
      "loss": 5.3964,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.07208801805973053,
      "rewards/margins": 0.18754826486110687,
      "rewards/rejected": -0.11546025425195694,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 482.0,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 81.19883728027344,
      "logits/rejected": 81.17581939697266,
      "logps/chosen": -28.9885311126709,
      "logps/rejected": -32.907562255859375,
      "loss": 5.3313,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.08011577278375626,
      "rewards/margins": 0.14832790195941925,
      "rewards/rejected": -0.06821214407682419,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 1040.0,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 83.12278747558594,
      "logits/rejected": 83.13607025146484,
      "logps/chosen": -33.499732971191406,
      "logps/rejected": -30.342330932617188,
      "loss": 6.478,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.1274038851261139,
      "rewards/margins": 0.2053629606962204,
      "rewards/rejected": -0.07795912027359009,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 684.0,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 84.02305603027344,
      "logits/rejected": 83.95697784423828,
      "logps/chosen": -30.76401710510254,
      "logps/rejected": -32.3089599609375,
      "loss": 5.8957,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.09905388206243515,
      "rewards/margins": 0.1568334400653839,
      "rewards/rejected": -0.057779550552368164,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 498.0,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 81.92304992675781,
      "logits/rejected": 81.90460205078125,
      "logps/chosen": -30.485610961914062,
      "logps/rejected": -31.443450927734375,
      "loss": 5.0645,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.1423078328371048,
      "rewards/margins": 0.19955609738826752,
      "rewards/rejected": -0.05724828690290451,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 502.0,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 83.73831176757812,
      "logits/rejected": 83.71122741699219,
      "logps/chosen": -30.49334716796875,
      "logps/rejected": -30.74527931213379,
      "loss": 7.3479,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.05461405590176582,
      "rewards/margins": 0.11006426811218262,
      "rewards/rejected": -0.055450212210416794,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 294.0,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 79.29591369628906,
      "logits/rejected": 79.25688171386719,
      "logps/chosen": -34.00603485107422,
      "logps/rejected": -32.55823516845703,
      "loss": 6.2997,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.1493411809206009,
      "rewards/margins": 0.16230711340904236,
      "rewards/rejected": -0.012965935282409191,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 98.92082977294922,
      "eval_logits/rejected": 98.90645599365234,
      "eval_logps/chosen": -32.54380416870117,
      "eval_logps/rejected": -36.24779510498047,
      "eval_loss": 7.145691394805908,
      "eval_rewards/accuracies": 0.5402824282646179,
      "eval_rewards/chosen": -0.02012600749731064,
      "eval_rewards/margins": 0.036127228289842606,
      "eval_rewards/rejected": -0.056253232061862946,
      "eval_runtime": 104.0138,
      "eval_samples_per_second": 3.298,
      "eval_steps_per_second": 0.413,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 1192.0,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 81.78749084472656,
      "logits/rejected": 81.68329620361328,
      "logps/chosen": -33.50835037231445,
      "logps/rejected": -35.296775817871094,
      "loss": 5.7897,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.08747398853302002,
      "rewards/margins": 0.14960959553718567,
      "rewards/rejected": -0.062135614454746246,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 382.0,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 83.9058609008789,
      "logits/rejected": 83.9858627319336,
      "logps/chosen": -31.16668128967285,
      "logps/rejected": -30.917465209960938,
      "loss": 6.2677,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.15645496547222137,
      "rewards/margins": 0.1941571831703186,
      "rewards/rejected": -0.03770221397280693,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 688.0,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 81.14820098876953,
      "logits/rejected": 81.19589233398438,
      "logps/chosen": -32.21388626098633,
      "logps/rejected": -34.05015182495117,
      "loss": 6.1846,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.11837179958820343,
      "rewards/margins": 0.12597264349460602,
      "rewards/rejected": -0.007600849959999323,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 764.0,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 83.51747131347656,
      "logits/rejected": 83.77994537353516,
      "logps/chosen": -30.83194351196289,
      "logps/rejected": -31.766277313232422,
      "loss": 3.8985,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 0.17122383415699005,
      "rewards/margins": 0.21807047724723816,
      "rewards/rejected": -0.0468466654419899,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 528.0,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 82.23722076416016,
      "logits/rejected": 82.29248809814453,
      "logps/chosen": -26.981786727905273,
      "logps/rejected": -29.9254093170166,
      "loss": 5.8549,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.122822105884552,
      "rewards/margins": 0.12775607407093048,
      "rewards/rejected": -0.0049339523538947105,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 744.0,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 79.6116943359375,
      "logits/rejected": 79.72075653076172,
      "logps/chosen": -30.516536712646484,
      "logps/rejected": -36.090003967285156,
      "loss": 4.7707,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.20889195799827576,
      "rewards/margins": 0.20656757056713104,
      "rewards/rejected": 0.0023243515752255917,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 348.0,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 79.08982849121094,
      "logits/rejected": 79.111083984375,
      "logps/chosen": -30.892492294311523,
      "logps/rejected": -31.574138641357422,
      "loss": 4.8727,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.1772860586643219,
      "rewards/margins": 0.1910295933485031,
      "rewards/rejected": -0.013743534684181213,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 700.0,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 81.58210754394531,
      "logits/rejected": 81.3693618774414,
      "logps/chosen": -31.06771469116211,
      "logps/rejected": -29.591297149658203,
      "loss": 7.3884,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.14843526482582092,
      "rewards/margins": 0.13329505920410156,
      "rewards/rejected": 0.015140200033783913,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 428.0,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 81.82354736328125,
      "logits/rejected": 81.73272705078125,
      "logps/chosen": -33.02164077758789,
      "logps/rejected": -32.17236328125,
      "loss": 4.7078,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.21652066707611084,
      "rewards/margins": 0.24762043356895447,
      "rewards/rejected": -0.03109976276755333,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 704.0,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 77.42338562011719,
      "logits/rejected": 77.51200866699219,
      "logps/chosen": -32.24098205566406,
      "logps/rejected": -29.014667510986328,
      "loss": 6.0651,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.2329459935426712,
      "rewards/margins": 0.21547456085681915,
      "rewards/rejected": 0.017471451312303543,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 99.00227355957031,
      "eval_logits/rejected": 98.98667907714844,
      "eval_logps/chosen": -32.43557357788086,
      "eval_logps/rejected": -36.024574279785156,
      "eval_loss": 7.648494243621826,
      "eval_rewards/accuracies": 0.5390365719795227,
      "eval_rewards/chosen": 0.0015206302050501108,
      "eval_rewards/margins": 0.01312990766018629,
      "eval_rewards/rejected": -0.011609279550611973,
      "eval_runtime": 104.0603,
      "eval_samples_per_second": 3.296,
      "eval_steps_per_second": 0.413,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 1200.0,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 84.3228759765625,
      "logits/rejected": 84.34573364257812,
      "logps/chosen": -29.936237335205078,
      "logps/rejected": -32.30895233154297,
      "loss": 5.3743,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.18545907735824585,
      "rewards/margins": 0.2072969228029251,
      "rewards/rejected": -0.02183784916996956,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 406.0,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 81.96125030517578,
      "logits/rejected": 81.95556640625,
      "logps/chosen": -30.652542114257812,
      "logps/rejected": -28.881006240844727,
      "loss": 5.1557,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.18920376896858215,
      "rewards/margins": 0.18150660395622253,
      "rewards/rejected": 0.007697162684053183,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 576.0,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 79.27213287353516,
      "logits/rejected": 79.31598663330078,
      "logps/chosen": -29.210277557373047,
      "logps/rejected": -32.73192596435547,
      "loss": 5.7764,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.22846612334251404,
      "rewards/margins": 0.2478782832622528,
      "rewards/rejected": -0.019412150606513023,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 708.0,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 83.35621643066406,
      "logits/rejected": 83.36161041259766,
      "logps/chosen": -32.25800323486328,
      "logps/rejected": -33.48296356201172,
      "loss": 5.3324,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.19082824885845184,
      "rewards/margins": 0.213682621717453,
      "rewards/rejected": -0.022854382172226906,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 322.0,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 82.49974822998047,
      "logits/rejected": 82.50485229492188,
      "logps/chosen": -32.731300354003906,
      "logps/rejected": -33.191688537597656,
      "loss": 4.9862,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.18248799443244934,
      "rewards/margins": 0.20246508717536926,
      "rewards/rejected": -0.019977085292339325,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 520.0,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 83.8789291381836,
      "logits/rejected": 83.89030456542969,
      "logps/chosen": -28.570327758789062,
      "logps/rejected": -31.54669761657715,
      "loss": 5.4694,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.19637688994407654,
      "rewards/margins": 0.17574702203273773,
      "rewards/rejected": 0.02062985673546791,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 370.0,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 83.33052062988281,
      "logits/rejected": 83.347900390625,
      "logps/chosen": -31.880367279052734,
      "logps/rejected": -35.12144470214844,
      "loss": 7.7502,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.1991589069366455,
      "rewards/margins": 0.20157304406166077,
      "rewards/rejected": -0.0024141333997249603,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 552.0,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 77.32600402832031,
      "logits/rejected": 77.20854187011719,
      "logps/chosen": -29.861562728881836,
      "logps/rejected": -28.112417221069336,
      "loss": 5.5564,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.14291271567344666,
      "rewards/margins": 0.13328143954277039,
      "rewards/rejected": 0.009631268680095673,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 5.985443219271573,
      "train_runtime": 2557.1257,
      "train_samples_per_second": 1.204,
      "train_steps_per_second": 0.151
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}