{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9962264150943396,
  "eval_steps": 500,
  "global_step": 99,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.010062893081761006,
      "grad_norm": 15.443952917965191,
      "learning_rate": 1e-07,
      "logps/chosen": -77.71961212158203,
      "logps/rejected": -77.03057861328125,
      "loss": 1.0499,
      "losses/dpo": 0.6931471824645996,
      "losses/sft": 0.3425563871860504,
      "losses/total": 1.0357035398483276,
      "ref_logps/chosen": -77.71961212158203,
      "ref_logps/rejected": -77.03057861328125,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.02012578616352201,
      "grad_norm": 14.525541091089869,
      "learning_rate": 2e-07,
      "logps/chosen": -81.42757415771484,
      "logps/rejected": -74.36381530761719,
      "loss": 1.0697,
      "losses/dpo": 0.6931471824645996,
      "losses/sft": 0.34544649720191956,
      "losses/total": 1.0385937690734863,
      "ref_logps/chosen": -81.42757415771484,
      "ref_logps/rejected": -74.36381530761719,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 2
    },
    {
      "epoch": 0.03018867924528302,
      "grad_norm": 14.972861178344003,
      "learning_rate": 3e-07,
      "logps/chosen": -77.72991943359375,
      "logps/rejected": -78.03253936767578,
      "loss": 1.0672,
      "losses/dpo": 0.7052053809165955,
      "losses/sft": 0.3688685894012451,
      "losses/total": 1.0740740299224854,
      "ref_logps/chosen": -77.74811553955078,
      "ref_logps/rejected": -78.08845520019531,
      "rewards/accuracies": 0.546875,
      "rewards/chosen": 0.0018183797365054488,
      "rewards/margins": -0.003773280419409275,
      "rewards/rejected": 0.00559166120365262,
      "step": 3
    },
    {
      "epoch": 0.04025157232704402,
      "grad_norm": 14.651655373074027,
      "learning_rate": 4e-07,
      "logps/chosen": -73.76988983154297,
      "logps/rejected": -74.19403839111328,
      "loss": 1.0523,
      "losses/dpo": 0.6988011598587036,
      "losses/sft": 0.3054599463939667,
      "losses/total": 1.0042610168457031,
      "ref_logps/chosen": -73.7587890625,
      "ref_logps/rejected": -74.23301696777344,
      "rewards/accuracies": 0.4609375,
      "rewards/chosen": -0.00110996060539037,
      "rewards/margins": -0.00500828679651022,
      "rewards/rejected": 0.0038983263075351715,
      "step": 4
    },
    {
      "epoch": 0.050314465408805034,
      "grad_norm": 15.110967055469406,
      "learning_rate": 5e-07,
      "logps/chosen": -77.6102523803711,
      "logps/rejected": -72.1615982055664,
      "loss": 1.0563,
      "losses/dpo": 0.6893635988235474,
      "losses/sft": 0.3531861901283264,
      "losses/total": 1.042549729347229,
      "ref_logps/chosen": -77.61317443847656,
      "ref_logps/rejected": -72.10652923583984,
      "rewards/accuracies": 0.5390625,
      "rewards/chosen": 0.0002926290035247803,
      "rewards/margins": 0.005799999460577965,
      "rewards/rejected": -0.005507371388375759,
      "step": 5
    },
    {
      "epoch": 0.06037735849056604,
      "grad_norm": 14.98273253406139,
      "learning_rate": 4.998603909325637e-07,
      "logps/chosen": -82.17599487304688,
      "logps/rejected": -75.54823303222656,
      "loss": 1.0568,
      "losses/dpo": 0.6767082214355469,
      "losses/sft": 0.42767006158828735,
      "losses/total": 1.104378342628479,
      "ref_logps/chosen": -82.23617553710938,
      "ref_logps/rejected": -75.60321807861328,
      "rewards/accuracies": 0.4921875,
      "rewards/chosen": 0.006017628125846386,
      "rewards/margins": 0.0005183331668376923,
      "rewards/rejected": 0.005499294493347406,
      "step": 6
    },
    {
      "epoch": 0.07044025157232704,
      "grad_norm": 15.812032171234971,
      "learning_rate": 4.994417196557883e-07,
      "logps/chosen": -80.02710723876953,
      "logps/rejected": -74.0862045288086,
      "loss": 1.0599,
      "losses/dpo": 0.6971860527992249,
      "losses/sft": 0.36510100960731506,
      "losses/total": 1.0622870922088623,
      "ref_logps/chosen": -80.04718017578125,
      "ref_logps/rejected": -74.11724090576172,
      "rewards/accuracies": 0.5234375,
      "rewards/chosen": 0.002006138674914837,
      "rewards/margins": -0.0010970293078571558,
      "rewards/rejected": 0.0031031689140945673,
      "step": 7
    },
    {
      "epoch": 0.08050314465408805,
      "grad_norm": 16.026806076613063,
      "learning_rate": 4.98744453772126e-07,
      "logps/chosen": -81.10508728027344,
      "logps/rejected": -72.62604522705078,
      "loss": 1.0513,
      "losses/dpo": 0.6706422567367554,
      "losses/sft": 0.32770049571990967,
      "losses/total": 0.998342752456665,
      "ref_logps/chosen": -81.17386627197266,
      "ref_logps/rejected": -72.63655090332031,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.006877122446894646,
      "rewards/margins": 0.005827218294143677,
      "rewards/rejected": 0.0010499038035050035,
      "step": 8
    },
    {
      "epoch": 0.09056603773584905,
      "grad_norm": 14.880720919169196,
      "learning_rate": 4.97769372038695e-07,
      "logps/chosen": -82.00043487548828,
      "logps/rejected": -77.90165710449219,
      "loss": 1.0566,
      "losses/dpo": 0.6756574511528015,
      "losses/sft": 0.3787515163421631,
      "losses/total": 1.0544090270996094,
      "ref_logps/chosen": -82.00239562988281,
      "ref_logps/rejected": -77.88880157470703,
      "rewards/accuracies": 0.5234375,
      "rewards/chosen": 0.0001968789438251406,
      "rewards/margins": 0.001483186730183661,
      "rewards/rejected": -0.0012863080482929945,
      "step": 9
    },
    {
      "epoch": 0.10062893081761007,
      "grad_norm": 15.517516535710937,
      "learning_rate": 4.965175634975071e-07,
      "logps/chosen": -88.80177307128906,
      "logps/rejected": -80.90895080566406,
      "loss": 1.0859,
      "losses/dpo": 0.7250989675521851,
      "losses/sft": 0.3838391602039337,
      "losses/total": 1.108938217163086,
      "ref_logps/chosen": -88.875732421875,
      "ref_logps/rejected": -80.98503112792969,
      "rewards/accuracies": 0.5390625,
      "rewards/chosen": 0.007396275643259287,
      "rewards/margins": -0.00021131080575287342,
      "rewards/rejected": 0.007607587147504091,
      "step": 10
    },
    {
      "epoch": 0.11069182389937107,
      "grad_norm": 14.958071998861861,
      "learning_rate": 4.949904262591467e-07,
      "logps/chosen": -80.21826171875,
      "logps/rejected": -73.97025299072266,
      "loss": 1.0449,
      "losses/dpo": 0.6753466725349426,
      "losses/sft": 0.40270546078681946,
      "losses/total": 1.0780521631240845,
      "ref_logps/chosen": -80.33733367919922,
      "ref_logps/rejected": -74.07726287841797,
      "rewards/accuracies": 0.5078125,
      "rewards/chosen": 0.011908243410289288,
      "rewards/margins": 0.0012067598290741444,
      "rewards/rejected": 0.01070148404687643,
      "step": 11
    },
    {
      "epoch": 0.12075471698113208,
      "grad_norm": 15.554570066549323,
      "learning_rate": 4.931896659412593e-07,
      "logps/chosen": -82.6167221069336,
      "logps/rejected": -79.10201263427734,
      "loss": 1.0498,
      "losses/dpo": 0.6855930685997009,
      "losses/sft": 0.37060120701789856,
      "losses/total": 1.0561943054199219,
      "ref_logps/chosen": -82.68681335449219,
      "ref_logps/rejected": -79.13915252685547,
      "rewards/accuracies": 0.515625,
      "rewards/chosen": 0.007009676657617092,
      "rewards/margins": 0.0032958725932985544,
      "rewards/rejected": 0.003713802434504032,
      "step": 12
    },
    {
      "epoch": 0.13081761006289308,
      "grad_norm": 15.575888738679977,
      "learning_rate": 4.911172937635942e-07,
      "logps/chosen": -78.73028564453125,
      "logps/rejected": -77.23632049560547,
      "loss": 1.0541,
      "losses/dpo": 0.703624963760376,
      "losses/sft": 0.3454435467720032,
      "losses/total": 1.049068570137024,
      "ref_logps/chosen": -78.94412231445312,
      "ref_logps/rejected": -77.30516815185547,
      "rewards/accuracies": 0.5234375,
      "rewards/chosen": 0.021384168416261673,
      "rewards/margins": 0.014499638229608536,
      "rewards/rejected": 0.006884530186653137,
      "step": 13
    },
    {
      "epoch": 0.14088050314465408,
      "grad_norm": 16.866802733593584,
      "learning_rate": 4.88775624301728e-07,
      "logps/chosen": -85.04679870605469,
      "logps/rejected": -71.76718139648438,
      "loss": 1.0699,
      "losses/dpo": 0.7083428502082825,
      "losses/sft": 0.38208234310150146,
      "losses/total": 1.0904252529144287,
      "ref_logps/chosen": -85.30770111083984,
      "ref_logps/rejected": -71.88570404052734,
      "rewards/accuracies": 0.5859375,
      "rewards/chosen": 0.02608950436115265,
      "rewards/margins": 0.014237894676625729,
      "rewards/rejected": 0.011851610615849495,
      "step": 14
    },
    {
      "epoch": 0.1509433962264151,
      "grad_norm": 15.574664831922826,
      "learning_rate": 4.861672729019796e-07,
      "logps/chosen": -80.3941650390625,
      "logps/rejected": -68.88729858398438,
      "loss": 1.0521,
      "losses/dpo": 0.6916662454605103,
      "losses/sft": 0.37930455803871155,
      "losses/total": 1.0709707736968994,
      "ref_logps/chosen": -80.68431091308594,
      "ref_logps/rejected": -68.9327392578125,
      "rewards/accuracies": 0.6328125,
      "rewards/chosen": 0.02901477925479412,
      "rewards/margins": 0.024471886456012726,
      "rewards/rejected": 0.004542896058410406,
      "step": 15
    },
    {
      "epoch": 0.1610062893081761,
      "grad_norm": 14.882165248684807,
      "learning_rate": 4.832951527604006e-07,
      "logps/chosen": -82.48802185058594,
      "logps/rejected": -78.44295501708984,
      "loss": 1.0461,
      "losses/dpo": 0.6916183233261108,
      "losses/sft": 0.36963188648223877,
      "losses/total": 1.0612502098083496,
      "ref_logps/chosen": -82.8018798828125,
      "ref_logps/rejected": -78.61569213867188,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": 0.0313868448138237,
      "rewards/margins": 0.014113493263721466,
      "rewards/rejected": 0.017273351550102234,
      "step": 16
    },
    {
      "epoch": 0.1710691823899371,
      "grad_norm": 14.311327622150413,
      "learning_rate": 4.801624716691072e-07,
      "logps/chosen": -76.1909408569336,
      "logps/rejected": -71.96290588378906,
      "loss": 1.0383,
      "losses/dpo": 0.6969538927078247,
      "losses/sft": 0.3102976977825165,
      "losses/total": 1.0072516202926636,
      "ref_logps/chosen": -76.5750732421875,
      "ref_logps/rejected": -72.12234497070312,
      "rewards/accuracies": 0.6796875,
      "rewards/chosen": 0.038413576781749725,
      "rewards/margins": 0.02246994338929653,
      "rewards/rejected": 0.015943635255098343,
      "step": 17
    },
    {
      "epoch": 0.1811320754716981,
      "grad_norm": 15.165695980080356,
      "learning_rate": 4.767727284335852e-07,
      "logps/chosen": -79.1435317993164,
      "logps/rejected": -74.56543731689453,
      "loss": 1.0555,
      "losses/dpo": 0.7003558278083801,
      "losses/sft": 0.3900759220123291,
      "losses/total": 1.0904316902160645,
      "ref_logps/chosen": -79.60275268554688,
      "ref_logps/rejected": -74.77250671386719,
      "rewards/accuracies": 0.6796875,
      "rewards/chosen": 0.04592147842049599,
      "rewards/margins": 0.02521485835313797,
      "rewards/rejected": 0.020706620067358017,
      "step": 18
    },
    {
      "epoch": 0.19119496855345913,
      "grad_norm": 15.578067919764813,
      "learning_rate": 4.7312970896497027e-07,
      "logps/chosen": -83.0239486694336,
      "logps/rejected": -74.53577423095703,
      "loss": 1.0743,
      "losses/dpo": 0.700573742389679,
      "losses/sft": 0.4360811114311218,
      "losses/total": 1.1366548538208008,
      "ref_logps/chosen": -83.43629455566406,
      "ref_logps/rejected": -74.72197723388672,
      "rewards/accuracies": 0.6015625,
      "rewards/chosen": 0.04123528674244881,
      "rewards/margins": 0.022615134716033936,
      "rewards/rejected": 0.01862015575170517,
      "step": 19
    },
    {
      "epoch": 0.20125786163522014,
      "grad_norm": 16.148080108602343,
      "learning_rate": 4.6923748205166785e-07,
      "logps/chosen": -86.2251205444336,
      "logps/rejected": -78.16014099121094,
      "loss": 1.0688,
      "losses/dpo": 0.6666887998580933,
      "losses/sft": 0.40739819407463074,
      "losses/total": 1.0740869045257568,
      "ref_logps/chosen": -86.82108306884766,
      "ref_logps/rejected": -78.41912078857422,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.05959618091583252,
      "rewards/margins": 0.03369804471731186,
      "rewards/rejected": 0.02589813806116581,
      "step": 20
    },
    {
      "epoch": 0.21132075471698114,
      "grad_norm": 14.878899557347989,
      "learning_rate": 4.6510039481503485e-07,
      "logps/chosen": -75.44450378417969,
      "logps/rejected": -71.4926528930664,
      "loss": 1.0377,
      "losses/dpo": 0.7134190201759338,
      "losses/sft": 0.38239145278930664,
      "losses/total": 1.0958105325698853,
      "ref_logps/chosen": -75.94510650634766,
      "ref_logps/rejected": -71.6917953491211,
      "rewards/accuracies": 0.6796875,
      "rewards/chosen": 0.05006038397550583,
      "rewards/margins": 0.030146455392241478,
      "rewards/rejected": 0.0199139267206192,
      "step": 21
    },
    {
      "epoch": 0.22138364779874214,
      "grad_norm": 15.226230246124961,
      "learning_rate": 4.6072306785419924e-07,
      "logps/chosen": -80.53596496582031,
      "logps/rejected": -80.97723388671875,
      "loss": 1.0284,
      "losses/dpo": 0.6952553391456604,
      "losses/sft": 0.27885523438453674,
      "losses/total": 0.9741106033325195,
      "ref_logps/chosen": -81.25791931152344,
      "ref_logps/rejected": -81.23795318603516,
      "rewards/accuracies": 0.6796875,
      "rewards/chosen": 0.07219455391168594,
      "rewards/margins": 0.046122319996356964,
      "rewards/rejected": 0.026072228327393532,
      "step": 22
    },
    {
      "epoch": 0.23144654088050315,
      "grad_norm": 14.590224847017746,
      "learning_rate": 4.5611039008544007e-07,
      "logps/chosen": -77.15602111816406,
      "logps/rejected": -72.4802017211914,
      "loss": 1.0256,
      "losses/dpo": 0.6954873204231262,
      "losses/sft": 0.30716028809547424,
      "losses/total": 1.0026476383209229,
      "ref_logps/chosen": -77.8787841796875,
      "ref_logps/rejected": -72.68162536621094,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.07227593660354614,
      "rewards/margins": 0.052133046090602875,
      "rewards/rejected": 0.020142890512943268,
      "step": 23
    },
    {
      "epoch": 0.24150943396226415,
      "grad_norm": 14.408995030982073,
      "learning_rate": 4.5126751328189073e-07,
      "logps/chosen": -75.66419982910156,
      "logps/rejected": -73.60345458984375,
      "loss": 1.025,
      "losses/dpo": 0.6652034521102905,
      "losses/sft": 0.353601336479187,
      "losses/total": 1.0188047885894775,
      "ref_logps/chosen": -76.47856903076172,
      "ref_logps/rejected": -73.73396301269531,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.08143681287765503,
      "rewards/margins": 0.06838638335466385,
      "rewards/rejected": 0.013050435110926628,
      "step": 24
    },
    {
      "epoch": 0.25157232704402516,
      "grad_norm": 15.305978491237383,
      "learning_rate": 4.4619984631966524e-07,
      "logps/chosen": -85.6600341796875,
      "logps/rejected": -78.23870849609375,
      "loss": 1.0417,
      "losses/dpo": 0.6948066353797913,
      "losses/sft": 0.4345339834690094,
      "losses/total": 1.1293405294418335,
      "ref_logps/chosen": -86.60327911376953,
      "ref_logps/rejected": -78.51371002197266,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.09432436525821686,
      "rewards/margins": 0.06682416051626205,
      "rewards/rejected": 0.027500208467245102,
      "step": 25
    },
    {
      "epoch": 0.26163522012578616,
      "grad_norm": 14.998262009207098,
      "learning_rate": 4.4091304913683303e-07,
      "logps/chosen": -78.49246215820312,
      "logps/rejected": -73.33612060546875,
      "loss": 1.0172,
      "losses/dpo": 0.6582043766975403,
      "losses/sft": 0.36708328127861023,
      "losses/total": 1.0252876281738281,
      "ref_logps/chosen": -79.48141479492188,
      "ref_logps/rejected": -73.58092498779297,
      "rewards/accuracies": 0.7578125,
      "rewards/chosen": 0.09889513999223709,
      "rewards/margins": 0.07441384345293045,
      "rewards/rejected": 0.024481287226080894,
      "step": 26
    },
    {
      "epoch": 0.27169811320754716,
      "grad_norm": 15.27620011007559,
      "learning_rate": 4.354130264119894e-07,
      "logps/chosen": -84.75664520263672,
      "logps/rejected": -81.69904327392578,
      "loss": 1.0329,
      "losses/dpo": 0.6849141716957092,
      "losses/sft": 0.39731666445732117,
      "losses/total": 1.0822309255599976,
      "ref_logps/chosen": -85.81089782714844,
      "ref_logps/rejected": -82.00923156738281,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 0.10542524605989456,
      "rewards/margins": 0.07440595328807831,
      "rewards/rejected": 0.031019294634461403,
      "step": 27
    },
    {
      "epoch": 0.28176100628930817,
      "grad_norm": 14.743778070925911,
      "learning_rate": 4.2970592096948233e-07,
      "logps/chosen": -80.38286590576172,
      "logps/rejected": -72.54148864746094,
      "loss": 1.0242,
      "losses/dpo": 0.646571695804596,
      "losses/sft": 0.3782905042171478,
      "losses/total": 1.0248621702194214,
      "ref_logps/chosen": -81.38656616210938,
      "ref_logps/rejected": -72.7380142211914,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": 0.10037041455507278,
      "rewards/margins": 0.08071772754192352,
      "rewards/rejected": 0.019652685150504112,
      "step": 28
    },
    {
      "epoch": 0.2918238993710692,
      "grad_norm": 14.816804858263165,
      "learning_rate": 4.237981069186606e-07,
      "logps/chosen": -80.61254119873047,
      "logps/rejected": -68.75757598876953,
      "loss": 1.0156,
      "losses/dpo": 0.6498047709465027,
      "losses/sft": 0.35897865891456604,
      "losses/total": 1.0087834596633911,
      "ref_logps/chosen": -81.66033935546875,
      "ref_logps/rejected": -68.87429809570312,
      "rewards/accuracies": 0.703125,
      "rewards/chosen": 0.10477961599826813,
      "rewards/margins": 0.09310723096132278,
      "rewards/rejected": 0.01167239435017109,
      "step": 29
    },
    {
      "epoch": 0.3018867924528302,
      "grad_norm": 14.461566845563905,
      "learning_rate": 4.1769618253480586e-07,
      "logps/chosen": -82.01522827148438,
      "logps/rejected": -75.83625793457031,
      "loss": 1.0182,
      "losses/dpo": 0.6898115277290344,
      "losses/sft": 0.4069589078426361,
      "losses/total": 1.0967704057693481,
      "ref_logps/chosen": -83.18346405029297,
      "ref_logps/rejected": -76.05616760253906,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.11682344973087311,
      "rewards/margins": 0.09483154118061066,
      "rewards/rejected": 0.021991904824972153,
      "step": 30
    },
    {
      "epoch": 0.3119496855345912,
      "grad_norm": 15.942497361038132,
      "learning_rate": 4.114069628897006e-07,
      "logps/chosen": -85.87028503417969,
      "logps/rejected": -73.85649108886719,
      "loss": 1.0156,
      "losses/dpo": 0.6715642213821411,
      "losses/sft": 0.41779857873916626,
      "losses/total": 1.0893628597259521,
      "ref_logps/chosen": -87.21926879882812,
      "ref_logps/rejected": -74.12530517578125,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": 0.13489778339862823,
      "rewards/margins": 0.10801610350608826,
      "rewards/rejected": 0.02688167802989483,
      "step": 31
    },
    {
      "epoch": 0.3220125786163522,
      "grad_norm": 13.82769050245281,
      "learning_rate": 4.049374722400612e-07,
      "logps/chosen": -74.70889282226562,
      "logps/rejected": -69.65875244140625,
      "loss": 1.0103,
      "losses/dpo": 0.6761119961738586,
      "losses/sft": 0.3233855962753296,
      "losses/total": 0.9994975924491882,
      "ref_logps/chosen": -75.858154296875,
      "ref_logps/rejected": -69.84304809570312,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.11492575705051422,
      "rewards/margins": 0.09649605304002762,
      "rewards/rejected": 0.0184297114610672,
      "step": 32
    },
    {
      "epoch": 0.3320754716981132,
      "grad_norm": 14.038904140570896,
      "learning_rate": 3.982949361823388e-07,
      "logps/chosen": -78.57609558105469,
      "logps/rejected": -74.79994201660156,
      "loss": 1.0256,
      "losses/dpo": 0.6977800130844116,
      "losses/sft": 0.3623635470867157,
      "losses/total": 1.0601435899734497,
      "ref_logps/chosen": -79.52400970458984,
      "ref_logps/rejected": -75.06456756591797,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 0.09479095041751862,
      "rewards/margins": 0.06832793354988098,
      "rewards/rejected": 0.026463022455573082,
      "step": 33
    },
    {
      "epoch": 0.3421383647798742,
      "grad_norm": 15.003139146574807,
      "learning_rate": 3.9148677358264883e-07,
      "logps/chosen": -80.80524444580078,
      "logps/rejected": -75.07316589355469,
      "loss": 1.0322,
      "losses/dpo": 0.6698934435844421,
      "losses/sft": 0.3599715232849121,
      "losses/total": 1.029865026473999,
      "ref_logps/chosen": -82.07255554199219,
      "ref_logps/rejected": -75.4892578125,
      "rewards/accuracies": 0.6796875,
      "rewards/chosen": 0.12673160433769226,
      "rewards/margins": 0.08512222021818161,
      "rewards/rejected": 0.041609376668930054,
      "step": 34
    },
    {
      "epoch": 0.3522012578616352,
      "grad_norm": 14.255000386833563,
      "learning_rate": 3.845205882908432e-07,
      "logps/chosen": -73.87222290039062,
      "logps/rejected": -72.94102478027344,
      "loss": 1.0045,
      "losses/dpo": 0.7203497886657715,
      "losses/sft": 0.36445504426956177,
      "losses/total": 1.0848047733306885,
      "ref_logps/chosen": -75.16232299804688,
      "ref_logps/rejected": -73.02802276611328,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.12901026010513306,
      "rewards/margins": 0.12031099945306778,
      "rewards/rejected": 0.008699269965291023,
      "step": 35
    },
    {
      "epoch": 0.3622641509433962,
      "grad_norm": 12.941056546878112,
      "learning_rate": 3.7740416064797934e-07,
      "logps/chosen": -70.72003173828125,
      "logps/rejected": -68.8385009765625,
      "loss": 0.9812,
      "losses/dpo": 0.6256467700004578,
      "losses/sft": 0.39033231139183044,
      "losses/total": 1.0159790515899658,
      "ref_logps/chosen": -71.86844635009766,
      "ref_logps/rejected": -68.81616973876953,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.11484173685312271,
      "rewards/margins": 0.11707467585802078,
      "rewards/rejected": -0.0022329380735754967,
      "step": 36
    },
    {
      "epoch": 0.3723270440251572,
      "grad_norm": 14.226051241378633,
      "learning_rate": 3.7014543879667095e-07,
      "logps/chosen": -75.77134704589844,
      "logps/rejected": -72.75182342529297,
      "loss": 0.9871,
      "losses/dpo": 0.7132631540298462,
      "losses/sft": 0.3754574656486511,
      "losses/total": 1.0887205600738525,
      "ref_logps/chosen": -77.1048812866211,
      "ref_logps/rejected": -72.76545715332031,
      "rewards/accuracies": 0.7265625,
      "rewards/chosen": 0.133354052901268,
      "rewards/margins": 0.13199105858802795,
      "rewards/rejected": 0.0013630054891109467,
      "step": 37
    },
    {
      "epoch": 0.38238993710691827,
      "grad_norm": 14.196980254865485,
      "learning_rate": 3.6275252980402544e-07,
      "logps/chosen": -85.88675689697266,
      "logps/rejected": -79.93682861328125,
      "loss": 1.0015,
      "losses/dpo": 0.6279609799385071,
      "losses/sft": 0.34624436497688293,
      "losses/total": 0.9742053151130676,
      "ref_logps/chosen": -87.35802459716797,
      "ref_logps/rejected": -79.97398376464844,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.14712581038475037,
      "rewards/margins": 0.14341090619564056,
      "rewards/rejected": 0.0037149174604564905,
      "step": 38
    },
    {
      "epoch": 0.39245283018867927,
      "grad_norm": 14.456015586003081,
      "learning_rate": 3.5523369060708374e-07,
      "logps/chosen": -78.47836303710938,
      "logps/rejected": -79.48917388916016,
      "loss": 1.0232,
      "losses/dpo": 0.7399665117263794,
      "losses/sft": 0.3791757822036743,
      "losses/total": 1.1191422939300537,
      "ref_logps/chosen": -79.63422393798828,
      "ref_logps/rejected": -79.65493774414062,
      "rewards/accuracies": 0.7109375,
      "rewards/chosen": 0.11558625847101212,
      "rewards/margins": 0.09900960326194763,
      "rewards/rejected": 0.016576647758483887,
      "step": 39
    },
    {
      "epoch": 0.4025157232704403,
      "grad_norm": 14.505318885700563,
      "learning_rate": 3.475973187908737e-07,
      "logps/chosen": -88.03107452392578,
      "logps/rejected": -80.5452880859375,
      "loss": 1.0003,
      "losses/dpo": 0.6669642329216003,
      "losses/sft": 0.3931300640106201,
      "losses/total": 1.0600943565368652,
      "ref_logps/chosen": -89.49748992919922,
      "ref_logps/rejected": -80.5037841796875,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": 0.14664065837860107,
      "rewards/margins": 0.15079057216644287,
      "rewards/rejected": -0.004149916581809521,
      "step": 40
    },
    {
      "epoch": 0.4125786163522013,
      "grad_norm": 14.149342282091867,
      "learning_rate": 3.3985194320937815e-07,
      "logps/chosen": -82.236328125,
      "logps/rejected": -77.66332244873047,
      "loss": 1.0057,
      "losses/dpo": 0.7688466310501099,
      "losses/sft": 0.3329574167728424,
      "losses/total": 1.1018040180206299,
      "ref_logps/chosen": -83.41654968261719,
      "ref_logps/rejected": -77.58070373535156,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.11802264302968979,
      "rewards/margins": 0.12628453969955444,
      "rewards/rejected": -0.008261887356638908,
      "step": 41
    },
    {
      "epoch": 0.4226415094339623,
      "grad_norm": 13.618988720784087,
      "learning_rate": 3.3200621445989225e-07,
      "logps/chosen": -80.78431701660156,
      "logps/rejected": -75.85997772216797,
      "loss": 1.0087,
      "losses/dpo": 0.7258040904998779,
      "losses/sft": 0.3291688561439514,
      "losses/total": 1.0549728870391846,
      "ref_logps/chosen": -81.78661346435547,
      "ref_logps/rejected": -75.47698974609375,
      "rewards/accuracies": 0.703125,
      "rewards/chosen": 0.10023011267185211,
      "rewards/margins": 0.1385280042886734,
      "rewards/rejected": -0.038297899067401886,
      "step": 42
    },
    {
      "epoch": 0.4327044025157233,
      "grad_norm": 14.273197027918624,
      "learning_rate": 3.240688952214085e-07,
      "logps/chosen": -78.64404296875,
      "logps/rejected": -75.18115234375,
      "loss": 0.9865,
      "losses/dpo": 0.7115089893341064,
      "losses/sft": 0.3947640061378479,
      "losses/total": 1.1062729358673096,
      "ref_logps/chosen": -80.05891418457031,
      "ref_logps/rejected": -75.02284240722656,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.14148636162281036,
      "rewards/margins": 0.15731626749038696,
      "rewards/rejected": -0.015829918906092644,
      "step": 43
    },
    {
      "epoch": 0.4427672955974843,
      "grad_norm": 13.480343487130659,
      "learning_rate": 3.160488504678215e-07,
      "logps/chosen": -74.60444641113281,
      "logps/rejected": -73.48818969726562,
      "loss": 0.9864,
      "losses/dpo": 0.66825270652771,
      "losses/sft": 0.3295142650604248,
      "losses/total": 0.99776691198349,
      "ref_logps/chosen": -75.63268280029297,
      "ref_logps/rejected": -73.1849365234375,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 0.10282407701015472,
      "rewards/margins": 0.133149191737175,
      "rewards/rejected": -0.030325107276439667,
      "step": 44
    },
    {
      "epoch": 0.4528301886792453,
      "grad_norm": 14.0102985883088,
      "learning_rate": 3.0795503756688205e-07,
      "logps/chosen": -86.0777587890625,
      "logps/rejected": -75.60399627685547,
      "loss": 0.9862,
      "losses/dpo": 0.658064067363739,
      "losses/sft": 0.36845076084136963,
      "losses/total": 1.0265148878097534,
      "ref_logps/chosen": -87.51325988769531,
      "ref_logps/rejected": -74.97911071777344,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.1435496211051941,
      "rewards/margins": 0.20603831112384796,
      "rewards/rejected": -0.06248871609568596,
      "step": 45
    },
    {
      "epoch": 0.4628930817610063,
      "grad_norm": 13.904517806164792,
      "learning_rate": 2.99796496275959e-07,
      "logps/chosen": -85.28681945800781,
      "logps/rejected": -79.6563491821289,
      "loss": 0.9989,
      "losses/dpo": 0.6056103706359863,
      "losses/sft": 0.34353241324424744,
      "losses/total": 0.9491427540779114,
      "ref_logps/chosen": -86.63987731933594,
      "ref_logps/rejected": -79.10623931884766,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.13530582189559937,
      "rewards/margins": 0.19031640887260437,
      "rewards/rejected": -0.0550105907022953,
      "step": 46
    },
    {
      "epoch": 0.4729559748427673,
      "grad_norm": 13.560110623241957,
      "learning_rate": 2.915823386457825e-07,
      "logps/chosen": -77.61348724365234,
      "logps/rejected": -75.27241516113281,
      "loss": 0.9656,
      "losses/dpo": 0.6847131252288818,
      "losses/sft": 0.40810227394104004,
      "losses/total": 1.0928153991699219,
      "ref_logps/chosen": -78.91578674316406,
      "ref_logps/rejected": -74.3929443359375,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.13022957742214203,
      "rewards/margins": 0.2181762158870697,
      "rewards/rejected": -0.08794663846492767,
      "step": 47
    },
    {
      "epoch": 0.4830188679245283,
      "grad_norm": 13.954359522887207,
      "learning_rate": 2.8332173884344473e-07,
      "logps/chosen": -87.6903305053711,
      "logps/rejected": -76.82035064697266,
      "loss": 0.9552,
      "losses/dpo": 0.5738142728805542,
      "losses/sft": 0.3640093207359314,
      "losses/total": 0.9378236532211304,
      "ref_logps/chosen": -89.48294830322266,
      "ref_logps/rejected": -75.78794860839844,
      "rewards/accuracies": 0.828125,
      "rewards/chosen": 0.1792612373828888,
      "rewards/margins": 0.28250131011009216,
      "rewards/rejected": -0.10324011743068695,
      "step": 48
    },
    {
      "epoch": 0.4930817610062893,
      "grad_norm": 14.369413009124699,
      "learning_rate": 2.7502392290602463e-07,
      "logps/chosen": -86.66905212402344,
      "logps/rejected": -76.85792541503906,
      "loss": 0.9481,
      "losses/dpo": 0.6439822316169739,
      "losses/sft": 0.3433207869529724,
      "losses/total": 0.9873029589653015,
      "ref_logps/chosen": -88.44965362548828,
      "ref_logps/rejected": -75.8476333618164,
      "rewards/accuracies": 0.796875,
      "rewards/chosen": 0.17805951833724976,
      "rewards/margins": 0.27908915281295776,
      "rewards/rejected": -0.10102960467338562,
      "step": 49
    },
    {
      "epoch": 0.5031446540880503,
      "grad_norm": 13.398792594698223,
      "learning_rate": 2.6669815843628043e-07,
      "logps/chosen": -80.97222900390625,
      "logps/rejected": -71.44869232177734,
      "loss": 0.9467,
      "losses/dpo": 0.7079877257347107,
      "losses/sft": 0.31868329644203186,
      "losses/total": 1.0266709327697754,
      "ref_logps/chosen": -82.57719421386719,
      "ref_logps/rejected": -70.47540283203125,
      "rewards/accuracies": 0.7578125,
      "rewards/chosen": 0.16049541532993317,
      "rewards/margins": 0.25782501697540283,
      "rewards/rejected": -0.09732960164546967,
      "step": 50
    },
    {
      "epoch": 0.5132075471698113,
      "grad_norm": 14.007896777723824,
      "learning_rate": 2.583537442519186e-07,
      "logps/chosen": -79.37796783447266,
      "logps/rejected": -78.45370483398438,
      "loss": 0.9709,
      "losses/dpo": 0.6879909038543701,
      "losses/sft": 0.4556385278701782,
      "losses/total": 1.1436294317245483,
      "ref_logps/chosen": -80.48783874511719,
      "ref_logps/rejected": -77.22457122802734,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": 0.11098764091730118,
      "rewards/margins": 0.2339007407426834,
      "rewards/rejected": -0.12291311472654343,
      "step": 51
    },
    {
      "epoch": 0.5232704402515723,
      "grad_norm": 13.571929188656334,
      "learning_rate": 2.5e-07,
      "logps/chosen": -76.12169647216797,
      "logps/rejected": -75.76917266845703,
      "loss": 0.9754,
      "losses/dpo": 0.551976203918457,
      "losses/sft": 0.34610363841056824,
      "losses/total": 0.8980798125267029,
      "ref_logps/chosen": -77.15998840332031,
      "ref_logps/rejected": -75.1527328491211,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.10382916778326035,
      "rewards/margins": 0.1654740273952484,
      "rewards/rejected": -0.06164485588669777,
      "step": 52
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 13.78983992632312,
      "learning_rate": 2.416462557480814e-07,
      "logps/chosen": -78.70403289794922,
      "logps/rejected": -74.04467010498047,
      "loss": 0.9668,
      "losses/dpo": 0.7154827117919922,
      "losses/sft": 0.355939120054245,
      "losses/total": 1.0714218616485596,
      "ref_logps/chosen": -79.78579711914062,
      "ref_logps/rejected": -72.72827911376953,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.10817612707614899,
      "rewards/margins": 0.23981451988220215,
      "rewards/rejected": -0.13163839280605316,
      "step": 53
    },
    {
      "epoch": 0.5433962264150943,
      "grad_norm": 13.17030909486059,
      "learning_rate": 2.333018415637196e-07,
      "logps/chosen": -78.02973937988281,
      "logps/rejected": -76.5998764038086,
      "loss": 0.9702,
      "losses/dpo": 0.6811453104019165,
      "losses/sft": 0.3610793948173523,
      "losses/total": 1.0422247648239136,
      "ref_logps/chosen": -79.0947265625,
      "ref_logps/rejected": -75.32565307617188,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.10649853199720383,
      "rewards/margins": 0.23392075300216675,
      "rewards/rejected": -0.12742222845554352,
      "step": 54
    },
    {
      "epoch": 0.5534591194968553,
      "grad_norm": 13.825294877409265,
      "learning_rate": 2.249760770939754e-07,
      "logps/chosen": -82.60687255859375,
      "logps/rejected": -81.40107727050781,
      "loss": 0.9518,
      "losses/dpo": 0.8651823401451111,
      "losses/sft": 0.33423036336898804,
      "losses/total": 1.1994127035140991,
      "ref_logps/chosen": -84.10595703125,
      "ref_logps/rejected": -80.23603057861328,
      "rewards/accuracies": 0.7890625,
      "rewards/chosen": 0.1499093621969223,
      "rewards/margins": 0.2664135694503784,
      "rewards/rejected": -0.11650420725345612,
      "step": 55
    },
    {
      "epoch": 0.5635220125786163,
      "grad_norm": 13.434417359184703,
      "learning_rate": 2.1667826115655535e-07,
      "logps/chosen": -77.57618713378906,
      "logps/rejected": -72.0264892578125,
      "loss": 0.9118,
      "losses/dpo": 0.6483964323997498,
      "losses/sft": 0.4290625751018524,
      "losses/total": 1.0774589776992798,
      "ref_logps/chosen": -79.1749496459961,
      "ref_logps/rejected": -70.45938110351562,
      "rewards/accuracies": 0.796875,
      "rewards/chosen": 0.15987561643123627,
      "rewards/margins": 0.3165854811668396,
      "rewards/rejected": -0.15670984983444214,
      "step": 56
    },
    {
      "epoch": 0.5735849056603773,
      "grad_norm": 13.682845711236109,
      "learning_rate": 2.0841766135421747e-07,
      "logps/chosen": -77.24420928955078,
      "logps/rejected": -74.8659439086914,
      "loss": 0.9385,
      "losses/dpo": 0.5598583221435547,
      "losses/sft": 0.32705366611480713,
      "losses/total": 0.8869119882583618,
      "ref_logps/chosen": -78.34547424316406,
      "ref_logps/rejected": -73.31063842773438,
      "rewards/accuracies": 0.7578125,
      "rewards/chosen": 0.11012652516365051,
      "rewards/margins": 0.26565757393836975,
      "rewards/rejected": -0.15553104877471924,
      "step": 57
    },
    {
      "epoch": 0.5836477987421383,
      "grad_norm": 14.285195042847882,
      "learning_rate": 2.0020350372404102e-07,
      "logps/chosen": -84.25393676757812,
      "logps/rejected": -81.29403686523438,
      "loss": 1.0015,
      "losses/dpo": 0.5723427534103394,
      "losses/sft": 0.350294291973114,
      "losses/total": 0.9226371049880981,
      "ref_logps/chosen": -85.54225158691406,
      "ref_logps/rejected": -80.76034545898438,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.1288313865661621,
      "rewards/margins": 0.18219929933547974,
      "rewards/rejected": -0.05336792394518852,
      "step": 58
    },
    {
      "epoch": 0.5937106918238994,
      "grad_norm": 13.633332779144865,
      "learning_rate": 1.920449624331179e-07,
      "logps/chosen": -81.46337890625,
      "logps/rejected": -78.45436096191406,
      "loss": 0.9685,
      "losses/dpo": 0.5848886370658875,
      "losses/sft": 0.3466652035713196,
      "losses/total": 0.931553840637207,
      "ref_logps/chosen": -82.74446105957031,
      "ref_logps/rejected": -77.16352844238281,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.12810780107975006,
      "rewards/margins": 0.2571909427642822,
      "rewards/rejected": -0.12908315658569336,
      "step": 59
    },
    {
      "epoch": 0.6037735849056604,
      "grad_norm": 13.112336410349611,
      "learning_rate": 1.839511495321785e-07,
      "logps/chosen": -77.9227294921875,
      "logps/rejected": -78.29096221923828,
      "loss": 0.9701,
      "losses/dpo": 0.6548681855201721,
      "losses/sft": 0.34684374928474426,
      "losses/total": 1.0017119646072388,
      "ref_logps/chosen": -78.95404052734375,
      "ref_logps/rejected": -77.03378295898438,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.10313084721565247,
      "rewards/margins": 0.22884881496429443,
      "rewards/rejected": -0.12571796774864197,
      "step": 60
    },
    {
      "epoch": 0.6138364779874214,
      "grad_norm": 13.600151225844877,
      "learning_rate": 1.7593110477859152e-07,
      "logps/chosen": -79.69929504394531,
      "logps/rejected": -81.41608428955078,
      "loss": 0.9724,
      "losses/dpo": 0.7238990664482117,
      "losses/sft": 0.3255448639392853,
      "losses/total": 1.0494439601898193,
      "ref_logps/chosen": -80.66063690185547,
      "ref_logps/rejected": -80.092041015625,
      "rewards/accuracies": 0.6796875,
      "rewards/chosen": 0.09613411873579025,
      "rewards/margins": 0.22853833436965942,
      "rewards/rejected": -0.13240423798561096,
      "step": 61
    },
    {
      "epoch": 0.6238993710691824,
      "grad_norm": 13.73953093508848,
      "learning_rate": 1.679937855401077e-07,
      "logps/chosen": -77.4028091430664,
      "logps/rejected": -76.37467956542969,
      "loss": 0.9901,
      "losses/dpo": 0.8189263939857483,
      "losses/sft": 0.3098299205303192,
      "losses/total": 1.1287564039230347,
      "ref_logps/chosen": -78.21166229248047,
      "ref_logps/rejected": -75.11843872070312,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.0808858722448349,
      "rewards/margins": 0.20650942623615265,
      "rewards/rejected": -0.12562356889247894,
      "step": 62
    },
    {
      "epoch": 0.6339622641509434,
      "grad_norm": 13.434557079661225,
      "learning_rate": 1.6014805679062183e-07,
      "logps/chosen": -82.96167755126953,
      "logps/rejected": -74.5992431640625,
      "loss": 0.9358,
      "losses/dpo": 0.5670045614242554,
      "losses/sft": 0.3809337019920349,
      "losses/total": 0.9479383230209351,
      "ref_logps/chosen": -84.36255645751953,
      "ref_logps/rejected": -72.7698974609375,
      "rewards/accuracies": 0.7578125,
      "rewards/chosen": 0.14008811116218567,
      "rewards/margins": 0.32302191853523254,
      "rewards/rejected": -0.18293380737304688,
      "step": 63
    },
    {
      "epoch": 0.6440251572327044,
      "grad_norm": 13.593026912115091,
      "learning_rate": 1.524026812091263e-07,
      "logps/chosen": -79.88929748535156,
      "logps/rejected": -81.100830078125,
      "loss": 0.9879,
      "losses/dpo": 0.667205274105072,
      "losses/sft": 0.3584054708480835,
      "losses/total": 1.0256106853485107,
      "ref_logps/chosen": -80.77157592773438,
      "ref_logps/rejected": -79.93651580810547,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.0882289931178093,
      "rewards/margins": 0.20465928316116333,
      "rewards/rejected": -0.11643026769161224,
      "step": 64
    },
    {
      "epoch": 0.6540880503144654,
      "grad_norm": 13.167920882870893,
      "learning_rate": 1.447663093929163e-07,
      "logps/chosen": -77.92182922363281,
      "logps/rejected": -75.83228302001953,
      "loss": 0.9598,
      "losses/dpo": 0.7064578533172607,
      "losses/sft": 0.3876456618309021,
      "losses/total": 1.094103455543518,
      "ref_logps/chosen": -78.7850341796875,
      "ref_logps/rejected": -74.05996704101562,
      "rewards/accuracies": 0.7265625,
      "rewards/chosen": 0.08632123470306396,
      "rewards/margins": 0.2635524570941925,
      "rewards/rejected": -0.17723123729228973,
      "step": 65
    },
    {
      "epoch": 0.6641509433962264,
      "grad_norm": 13.6875600213724,
      "learning_rate": 1.372474701959745e-07,
      "logps/chosen": -78.46904754638672,
      "logps/rejected": -74.88655090332031,
      "loss": 0.9566,
      "losses/dpo": 0.8276803493499756,
      "losses/sft": 0.3391786813735962,
      "losses/total": 1.1668590307235718,
      "ref_logps/chosen": -79.26625061035156,
      "ref_logps/rejected": -73.1043701171875,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.07972101867198944,
      "rewards/margins": 0.25793933868408203,
      "rewards/rejected": -0.17821836471557617,
      "step": 66
    },
    {
      "epoch": 0.6742138364779874,
      "grad_norm": 13.714831020339835,
      "learning_rate": 1.2985456120332906e-07,
      "logps/chosen": -83.13707733154297,
      "logps/rejected": -73.50003051757812,
      "loss": 0.9304,
      "losses/dpo": 0.552372932434082,
      "losses/sft": 0.3785245716571808,
      "losses/total": 0.9308975338935852,
      "ref_logps/chosen": -84.77952575683594,
      "ref_logps/rejected": -71.75003814697266,
      "rewards/accuracies": 0.8046875,
      "rewards/chosen": 0.16424375772476196,
      "rewards/margins": 0.3392435610294342,
      "rewards/rejected": -0.17499980330467224,
      "step": 67
    },
    {
      "epoch": 0.6842767295597484,
      "grad_norm": 13.196746131149819,
      "learning_rate": 1.225958393520206e-07,
      "logps/chosen": -78.00275421142578,
      "logps/rejected": -71.03691864013672,
      "loss": 0.9349,
      "losses/dpo": 0.5955305695533752,
      "losses/sft": 0.36189359426498413,
      "losses/total": 0.9574241638183594,
      "ref_logps/chosen": -78.86644744873047,
      "ref_logps/rejected": -68.57377624511719,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 0.0863691195845604,
      "rewards/margins": 0.33268311619758606,
      "rewards/rejected": -0.24631400406360626,
      "step": 68
    },
    {
      "epoch": 0.6943396226415094,
      "grad_norm": 13.32726104884298,
      "learning_rate": 1.1547941170915685e-07,
      "logps/chosen": -85.18400573730469,
      "logps/rejected": -74.97825622558594,
      "loss": 0.9174,
      "losses/dpo": 0.5665487051010132,
      "losses/sft": 0.36319419741630554,
      "losses/total": 0.9297429323196411,
      "ref_logps/chosen": -86.47076416015625,
      "ref_logps/rejected": -72.61034393310547,
      "rewards/accuracies": 0.765625,
      "rewards/chosen": 0.12867560982704163,
      "rewards/margins": 0.36546623706817627,
      "rewards/rejected": -0.23679062724113464,
      "step": 69
    },
    {
      "epoch": 0.7044025157232704,
      "grad_norm": 14.283665942013558,
      "learning_rate": 1.0851322641735117e-07,
      "logps/chosen": -83.07630157470703,
      "logps/rejected": -81.20217895507812,
      "loss": 0.976,
      "losses/dpo": 0.7452372312545776,
      "losses/sft": 0.337566077709198,
      "losses/total": 1.0828032493591309,
      "ref_logps/chosen": -84.08202362060547,
      "ref_logps/rejected": -79.51718139648438,
      "rewards/accuracies": 0.7265625,
      "rewards/chosen": 0.10057278722524643,
      "rewards/margins": 0.2690734267234802,
      "rewards/rejected": -0.1685006469488144,
      "step": 70
    },
    {
      "epoch": 0.7144654088050314,
      "grad_norm": 12.916070035590293,
      "learning_rate": 1.0170506381766119e-07,
      "logps/chosen": -74.7712631225586,
      "logps/rejected": -74.12356567382812,
      "loss": 0.9084,
      "losses/dpo": 0.5794627666473389,
      "losses/sft": 0.40780600905418396,
      "losses/total": 0.9872688055038452,
      "ref_logps/chosen": -75.7598876953125,
      "ref_logps/rejected": -71.68157196044922,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.0988626778125763,
      "rewards/margins": 0.3430623412132263,
      "rewards/rejected": -0.24419964849948883,
      "step": 71
    },
    {
      "epoch": 0.7245283018867924,
      "grad_norm": 13.489560558340342,
      "learning_rate": 9.506252775993881e-08,
      "logps/chosen": -78.59689331054688,
      "logps/rejected": -77.52888488769531,
      "loss": 0.937,
      "losses/dpo": 0.7861602306365967,
      "losses/sft": 0.3728378415107727,
      "losses/total": 1.1589981317520142,
      "ref_logps/chosen": -79.52552032470703,
      "ref_logps/rejected": -75.24673461914062,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.09286268055438995,
      "rewards/margins": 0.3210780918598175,
      "rewards/rejected": -0.22821539640426636,
      "step": 72
    },
    {
      "epoch": 0.7345911949685534,
      "grad_norm": 13.406140514068719,
      "learning_rate": 8.859303711029939e-08,
      "logps/chosen": -79.9459457397461,
      "logps/rejected": -76.30967712402344,
      "loss": 0.9496,
      "losses/dpo": 0.5502427220344543,
      "losses/sft": 0.36247241497039795,
      "losses/total": 0.9127151966094971,
      "ref_logps/chosen": -80.76979064941406,
      "ref_logps/rejected": -73.900390625,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 0.08238394558429718,
      "rewards/margins": 0.323311448097229,
      "rewards/rejected": -0.24092751741409302,
      "step": 73
    },
    {
      "epoch": 0.7446540880503144,
      "grad_norm": 13.255817188222572,
      "learning_rate": 8.230381746519419e-08,
      "logps/chosen": -83.04154968261719,
      "logps/rejected": -78.92507934570312,
      "loss": 0.9526,
      "losses/dpo": 0.5468727946281433,
      "losses/sft": 0.32324424386024475,
      "losses/total": 0.8701170682907104,
      "ref_logps/chosen": -84.06158447265625,
      "ref_logps/rejected": -76.94873046875,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.10200438648462296,
      "rewards/margins": 0.2996390461921692,
      "rewards/rejected": -0.19763466715812683,
      "step": 74
    },
    {
      "epoch": 0.7547169811320755,
      "grad_norm": 13.448309012154779,
      "learning_rate": 7.620189308133943e-08,
      "logps/chosen": -82.13368225097656,
      "logps/rejected": -80.07732391357422,
      "loss": 0.9415,
      "losses/dpo": 0.6907780170440674,
      "losses/sft": 0.3356712758541107,
      "losses/total": 1.0264493227005005,
      "ref_logps/chosen": -82.98661041259766,
      "ref_logps/rejected": -77.53411865234375,
      "rewards/accuracies": 0.765625,
      "rewards/chosen": 0.08529379218816757,
      "rewards/margins": 0.33961349725723267,
      "rewards/rejected": -0.2543197274208069,
      "step": 75
    },
    {
      "epoch": 0.7647798742138365,
      "grad_norm": 13.705859719966018,
      "learning_rate": 7.02940790305177e-08,
      "logps/chosen": -84.20864868164062,
      "logps/rejected": -73.40209197998047,
      "loss": 0.9356,
      "losses/dpo": 0.5647323131561279,
      "losses/sft": 0.3960777521133423,
      "losses/total": 0.960810124874115,
      "ref_logps/chosen": -84.89796447753906,
      "ref_logps/rejected": -70.65544128417969,
      "rewards/accuracies": 0.765625,
      "rewards/chosen": 0.06893075257539749,
      "rewards/margins": 0.3435964584350586,
      "rewards/rejected": -0.2746656835079193,
      "step": 76
    },
    {
      "epoch": 0.7748427672955975,
      "grad_norm": 13.738145113740726,
      "learning_rate": 6.45869735880106e-08,
      "logps/chosen": -78.38431549072266,
      "logps/rejected": -76.76514434814453,
      "loss": 0.9343,
      "losses/dpo": 0.48376235365867615,
      "losses/sft": 0.36545315384864807,
      "losses/total": 0.8492154479026794,
      "ref_logps/chosen": -79.1065673828125,
      "ref_logps/rejected": -74.5013427734375,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.07222675532102585,
      "rewards/margins": 0.29860708117485046,
      "rewards/rejected": -0.2263803333044052,
      "step": 77
    },
    {
      "epoch": 0.7849056603773585,
      "grad_norm": 15.243013296605241,
      "learning_rate": 5.908695086316701e-08,
      "logps/chosen": -86.07107543945312,
      "logps/rejected": -84.10150909423828,
      "loss": 0.9363,
      "losses/dpo": 0.6692842245101929,
      "losses/sft": 0.3787580132484436,
      "losses/total": 1.0480422973632812,
      "ref_logps/chosen": -87.29566955566406,
      "ref_logps/rejected": -81.75081634521484,
      "rewards/accuracies": 0.7578125,
      "rewards/chosen": 0.12246063351631165,
      "rewards/margins": 0.35752907395362854,
      "rewards/rejected": -0.2350684255361557,
      "step": 78
    },
    {
      "epoch": 0.7949685534591195,
      "grad_norm": 13.426200075061494,
      "learning_rate": 5.380015368033475e-08,
      "logps/chosen": -80.97967529296875,
      "logps/rejected": -78.46061706542969,
      "loss": 0.9532,
      "losses/dpo": 0.7194112539291382,
      "losses/sft": 0.3862980306148529,
      "losses/total": 1.1057093143463135,
      "ref_logps/chosen": -81.8296127319336,
      "ref_logps/rejected": -76.18516540527344,
      "rewards/accuracies": 0.6953125,
      "rewards/chosen": 0.08499410003423691,
      "rewards/margins": 0.31253981590270996,
      "rewards/rejected": -0.22754575312137604,
      "step": 79
    },
    {
      "epoch": 0.8050314465408805,
      "grad_norm": 14.18385541473246,
      "learning_rate": 4.8732486718109284e-08,
      "logps/chosen": -80.56340789794922,
      "logps/rejected": -80.10472106933594,
      "loss": 0.956,
      "losses/dpo": 0.7997117042541504,
      "losses/sft": 0.2968393862247467,
      "losses/total": 1.0965510606765747,
      "ref_logps/chosen": -81.14766693115234,
      "ref_logps/rejected": -77.70572662353516,
      "rewards/accuracies": 0.7265625,
      "rewards/chosen": 0.058426402509212494,
      "rewards/margins": 0.2983267903327942,
      "rewards/rejected": -0.2399004101753235,
      "step": 80
    },
    {
      "epoch": 0.8150943396226416,
      "grad_norm": 13.451956561854924,
      "learning_rate": 4.388960991455998e-08,
      "logps/chosen": -83.258544921875,
      "logps/rejected": -73.33403015136719,
      "loss": 0.9075,
      "losses/dpo": 0.6336656808853149,
      "losses/sft": 0.3658246695995331,
      "losses/total": 0.9994903206825256,
      "ref_logps/chosen": -84.42308044433594,
      "ref_logps/rejected": -70.38734436035156,
      "rewards/accuracies": 0.8203125,
      "rewards/chosen": 0.11645462363958359,
      "rewards/margins": 0.4111226797103882,
      "rewards/rejected": -0.294668048620224,
      "step": 81
    },
    {
      "epoch": 0.8251572327044026,
      "grad_norm": 13.420430376295206,
      "learning_rate": 3.927693214580075e-08,
      "logps/chosen": -79.82371520996094,
      "logps/rejected": -76.38064575195312,
      "loss": 0.9636,
      "losses/dpo": 0.7534294128417969,
      "losses/sft": 0.30548882484436035,
      "losses/total": 1.0589182376861572,
      "ref_logps/chosen": -80.31490325927734,
      "ref_logps/rejected": -74.10685729980469,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.04911880940198898,
      "rewards/margins": 0.27649733424186707,
      "rewards/rejected": -0.22737851738929749,
      "step": 82
    },
    {
      "epoch": 0.8352201257861636,
      "grad_norm": 13.178604910224282,
      "learning_rate": 3.4899605184965206e-08,
      "logps/chosen": -79.35425567626953,
      "logps/rejected": -72.01036071777344,
      "loss": 0.9413,
      "losses/dpo": 0.5128786563873291,
      "losses/sft": 0.36703866720199585,
      "losses/total": 0.8799173831939697,
      "ref_logps/chosen": -80.35247039794922,
      "ref_logps/rejected": -69.59880065917969,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": 0.09982186555862427,
      "rewards/margins": 0.3409779965877533,
      "rewards/rejected": -0.24115616083145142,
      "step": 83
    },
    {
      "epoch": 0.8452830188679246,
      "grad_norm": 13.358792610889648,
      "learning_rate": 3.076251794833212e-08,
      "logps/chosen": -82.73087310791016,
      "logps/rejected": -78.44337463378906,
      "loss": 0.9477,
      "losses/dpo": 0.5781779289245605,
      "losses/sft": 0.37605956196784973,
      "losses/total": 0.9542375206947327,
      "ref_logps/chosen": -83.96467590332031,
      "ref_logps/rejected": -76.55040740966797,
      "rewards/accuracies": 0.7578125,
      "rewards/chosen": 0.12338028848171234,
      "rewards/margins": 0.31267744302749634,
      "rewards/rejected": -0.1892971247434616,
      "step": 84
    },
    {
      "epoch": 0.8553459119496856,
      "grad_norm": 13.690907855317144,
      "learning_rate": 2.6870291035029718e-08,
      "logps/chosen": -79.540771484375,
      "logps/rejected": -78.94165802001953,
      "loss": 0.9684,
      "losses/dpo": 0.7159697413444519,
      "losses/sft": 0.3729468584060669,
      "losses/total": 1.0889166593551636,
      "ref_logps/chosen": -80.23945617675781,
      "ref_logps/rejected": -76.9882583618164,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.06986873596906662,
      "rewards/margins": 0.265209436416626,
      "rewards/rejected": -0.19534069299697876,
      "step": 85
    },
    {
      "epoch": 0.8654088050314466,
      "grad_norm": 13.105540805914607,
      "learning_rate": 2.3227271566414823e-08,
      "logps/chosen": -77.78494262695312,
      "logps/rejected": -71.80929565429688,
      "loss": 0.9314,
      "losses/dpo": 0.8130008578300476,
      "losses/sft": 0.33129045367240906,
      "losses/total": 1.1442912817001343,
      "ref_logps/chosen": -78.75529479980469,
      "ref_logps/rejected": -69.36322784423828,
      "rewards/accuracies": 0.8203125,
      "rewards/chosen": 0.09703564643859863,
      "rewards/margins": 0.34164199233055115,
      "rewards/rejected": -0.24460633099079132,
      "step": 86
    },
    {
      "epoch": 0.8754716981132076,
      "grad_norm": 13.612775733069526,
      "learning_rate": 1.9837528330892776e-08,
      "logps/chosen": -82.26605224609375,
      "logps/rejected": -76.89611053466797,
      "loss": 0.9146,
      "losses/dpo": 0.6761611700057983,
      "losses/sft": 0.31893810629844666,
      "losses/total": 0.9950993061065674,
      "ref_logps/chosen": -83.41204071044922,
      "ref_logps/rejected": -74.252197265625,
      "rewards/accuracies": 0.7734375,
      "rewards/chosen": 0.11459847539663315,
      "rewards/margins": 0.3789900541305542,
      "rewards/rejected": -0.26439157128334045,
      "step": 87
    },
    {
      "epoch": 0.8855345911949686,
      "grad_norm": 13.33006314162012,
      "learning_rate": 1.6704847239599363e-08,
      "logps/chosen": -83.5063247680664,
      "logps/rejected": -76.12062072753906,
      "loss": 0.9517,
      "losses/dpo": 0.5405787229537964,
      "losses/sft": 0.33481675386428833,
      "losses/total": 0.8753954768180847,
      "ref_logps/chosen": -84.05892181396484,
      "ref_logps/rejected": -73.51615142822266,
      "rewards/accuracies": 0.7421875,
      "rewards/chosen": 0.05525832623243332,
      "rewards/margins": 0.31570491194725037,
      "rewards/rejected": -0.26044660806655884,
      "step": 88
    },
    {
      "epoch": 0.8955974842767296,
      "grad_norm": 13.537155227659039,
      "learning_rate": 1.383272709802033e-08,
      "logps/chosen": -82.11839294433594,
      "logps/rejected": -79.336669921875,
      "loss": 0.9444,
      "losses/dpo": 0.7212787866592407,
      "losses/sft": 0.3310409486293793,
      "losses/total": 1.0523197650909424,
      "ref_logps/chosen": -83.14412689208984,
      "ref_logps/rejected": -77.21094512939453,
      "rewards/accuracies": 0.8046875,
      "rewards/chosen": 0.10257288068532944,
      "rewards/margins": 0.3151445984840393,
      "rewards/rejected": -0.21257171034812927,
      "step": 89
    },
    {
      "epoch": 0.9056603773584906,
      "grad_norm": 14.311506779786631,
      "learning_rate": 1.1224375698271893e-08,
      "logps/chosen": -80.742919921875,
      "logps/rejected": -80.19947814941406,
      "loss": 0.9524,
      "losses/dpo": 0.808186948299408,
      "losses/sft": 0.3583277463912964,
      "losses/total": 1.1665147542953491,
      "ref_logps/chosen": -81.75139617919922,
      "ref_logps/rejected": -78.16023254394531,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.10084706544876099,
      "rewards/margins": 0.30477190017700195,
      "rewards/rejected": -0.20392483472824097,
      "step": 90
    },
    {
      "epoch": 0.9157232704402516,
      "grad_norm": 13.55537996131118,
      "learning_rate": 8.882706236405885e-09,
      "logps/chosen": -77.98944854736328,
      "logps/rejected": -78.98483276367188,
      "loss": 0.9742,
      "losses/dpo": 0.8716121315956116,
      "losses/sft": 0.3347938060760498,
      "losses/total": 1.2064059972763062,
      "ref_logps/chosen": -78.49334716796875,
      "ref_logps/rejected": -76.73573303222656,
      "rewards/accuracies": 0.671875,
      "rewards/chosen": 0.05039064586162567,
      "rewards/margins": 0.27530133724212646,
      "rewards/rejected": -0.2249106764793396,
      "step": 91
    },
    {
      "epoch": 0.9257861635220126,
      "grad_norm": 13.793163914910798,
      "learning_rate": 6.810334058740736e-09,
      "logps/chosen": -81.20413970947266,
      "logps/rejected": -79.94811248779297,
      "loss": 0.9541,
      "losses/dpo": 0.7405242323875427,
      "losses/sft": 0.37262600660324097,
      "losses/total": 1.1131502389907837,
      "ref_logps/chosen": -82.03514862060547,
      "ref_logps/rejected": -77.8819351196289,
      "rewards/accuracies": 0.7265625,
      "rewards/chosen": 0.08309987932443619,
      "rewards/margins": 0.2897171676158905,
      "rewards/rejected": -0.20661728084087372,
      "step": 92
    },
    {
      "epoch": 0.9358490566037736,
      "grad_norm": 12.979428116887359,
      "learning_rate": 5.009573740853313e-09,
      "logps/chosen": -78.45671844482422,
      "logps/rejected": -72.1712875366211,
      "loss": 0.8749,
      "losses/dpo": 0.6323506832122803,
      "losses/sft": 0.37680134177207947,
      "losses/total": 1.0091519355773926,
      "ref_logps/chosen": -79.63896179199219,
      "ref_logps/rejected": -68.9825210571289,
      "rewards/accuracies": 0.8359375,
      "rewards/chosen": 0.118224136531353,
      "rewards/margins": 0.43710076808929443,
      "rewards/rejected": -0.31887662410736084,
      "step": 93
    },
    {
      "epoch": 0.9459119496855346,
      "grad_norm": 12.91270906410155,
      "learning_rate": 3.482436502492858e-09,
      "logps/chosen": -78.55203247070312,
      "logps/rejected": -76.18038940429688,
      "loss": 0.9144,
      "losses/dpo": 0.6181066036224365,
      "losses/sft": 0.37585684657096863,
      "losses/total": 0.9939634799957275,
      "ref_logps/chosen": -79.37383270263672,
      "ref_logps/rejected": -73.2895278930664,
      "rewards/accuracies": 0.7734375,
      "rewards/chosen": 0.08218016475439072,
      "rewards/margins": 0.3712664246559143,
      "rewards/rejected": -0.2890862822532654,
      "step": 94
    },
    {
      "epoch": 0.9559748427672956,
      "grad_norm": 13.290626630295796,
      "learning_rate": 2.2306279613049926e-09,
      "logps/chosen": -78.57327270507812,
      "logps/rejected": -77.99625396728516,
      "loss": 0.926,
      "losses/dpo": 0.6854672431945801,
      "losses/sft": 0.41287970542907715,
      "losses/total": 1.0983469486236572,
      "ref_logps/chosen": -79.61153411865234,
      "ref_logps/rejected": -75.51338958740234,
      "rewards/accuracies": 0.7734375,
      "rewards/chosen": 0.1038258820772171,
      "rewards/margins": 0.3521122932434082,
      "rewards/rejected": -0.2482864260673523,
      "step": 95
    },
    {
      "epoch": 0.9660377358490566,
      "grad_norm": 13.431385680233705,
      "learning_rate": 1.255546227873966e-09,
      "logps/chosen": -79.0116195678711,
      "logps/rejected": -79.90180969238281,
      "loss": 0.9308,
      "losses/dpo": 0.7681677937507629,
      "losses/sft": 0.41490912437438965,
      "losses/total": 1.1830769777297974,
      "ref_logps/chosen": -79.99933624267578,
      "ref_logps/rejected": -77.70450592041016,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 0.09877175092697144,
      "rewards/margins": 0.31850215792655945,
      "rewards/rejected": -0.21973037719726562,
      "step": 96
    },
    {
      "epoch": 0.9761006289308176,
      "grad_norm": 14.079932869223676,
      "learning_rate": 5.58280344211709e-10,
      "logps/chosen": -84.25556945800781,
      "logps/rejected": -80.85723876953125,
      "loss": 0.9438,
      "losses/dpo": 0.6998444199562073,
      "losses/sft": 0.4053678512573242,
      "losses/total": 1.1052123308181763,
      "ref_logps/chosen": -84.91212463378906,
      "ref_logps/rejected": -78.10282135009766,
      "rewards/accuracies": 0.7109375,
      "rewards/chosen": 0.0656551793217659,
      "rewards/margins": 0.341097354888916,
      "rewards/rejected": -0.2754421830177307,
      "step": 97
    },
    {
      "epoch": 0.9861635220125786,
      "grad_norm": 14.278966959359417,
      "learning_rate": 1.3960906743634703e-10,
      "logps/chosen": -80.89794158935547,
      "logps/rejected": -84.88188171386719,
      "loss": 0.9774,
      "losses/dpo": 0.7035096883773804,
      "losses/sft": 0.31059831380844116,
      "losses/total": 1.0141080617904663,
      "ref_logps/chosen": -81.65719604492188,
      "ref_logps/rejected": -83.00572204589844,
      "rewards/accuracies": 0.671875,
      "rewards/chosen": 0.07592566311359406,
      "rewards/margins": 0.26354214549064636,
      "rewards/rejected": -0.1876164972782135,
      "step": 98
    },
    {
      "epoch": 0.9962264150943396,
      "grad_norm": 12.961129324284881,
      "learning_rate": 0.0,
      "logps/chosen": -77.79227447509766,
      "logps/rejected": -73.44792175292969,
      "loss": 0.935,
      "losses/dpo": 0.673121452331543,
      "losses/sft": 0.3511947691440582,
      "losses/total": 1.0243163108825684,
      "ref_logps/chosen": -78.48436737060547,
      "ref_logps/rejected": -70.95289611816406,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.06920892000198364,
      "rewards/margins": 0.31871092319488525,
      "rewards/rejected": -0.249502032995224,
      "step": 99
    },
    {
      "epoch": 0.9962264150943396,
      "step": 99,
      "total_flos": 0.0,
      "train_loss": 0.9872819769262063,
      "train_runtime": 3540.7882,
      "train_samples_per_second": 3.592,
      "train_steps_per_second": 0.028
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 99,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}