{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 39,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08163265306122448,
      "grad_norm": 1.025234580039978,
      "learning_rate": 0.0,
      "logits/chosen": -2.375152349472046,
      "logits/rejected": -2.394904375076294,
      "logps/chosen": -84.37759399414062,
      "logps/rejected": -96.84375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16326530612244897,
      "grad_norm": 1.1001588106155396,
      "learning_rate": 2.5e-06,
      "logits/chosen": -2.3984360694885254,
      "logits/rejected": -2.4267055988311768,
      "logps/chosen": -89.42276000976562,
      "logps/rejected": -96.12063598632812,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 2
    },
    {
      "epoch": 0.24489795918367346,
      "grad_norm": 1.0922303199768066,
      "learning_rate": 5e-06,
      "logits/chosen": -2.426487922668457,
      "logits/rejected": -2.437016487121582,
      "logps/chosen": -90.65304565429688,
      "logps/rejected": -98.02507019042969,
      "loss": 0.6942,
      "rewards/accuracies": 0.48681640625,
      "rewards/chosen": -0.004277409985661507,
      "rewards/margins": -0.0012883469462394714,
      "rewards/rejected": -0.0029890630394220352,
      "step": 3
    },
    {
      "epoch": 0.32653061224489793,
      "grad_norm": 1.0690197944641113,
      "learning_rate": 7.500000000000001e-06,
      "logits/chosen": -2.4016366004943848,
      "logits/rejected": -2.404703378677368,
      "logps/chosen": -88.64875793457031,
      "logps/rejected": -100.15958404541016,
      "loss": 0.6925,
      "rewards/accuracies": 0.48828125,
      "rewards/chosen": 0.0017670390661805868,
      "rewards/margins": 0.0021676155738532543,
      "rewards/rejected": -0.0004005763039458543,
      "step": 4
    },
    {
      "epoch": 0.40816326530612246,
      "grad_norm": 1.0020982027053833,
      "learning_rate": 1e-05,
      "logits/chosen": -2.405047655105591,
      "logits/rejected": -2.4170143604278564,
      "logps/chosen": -86.40444946289062,
      "logps/rejected": -99.32535552978516,
      "loss": 0.692,
      "rewards/accuracies": 0.52099609375,
      "rewards/chosen": 0.002238229615613818,
      "rewards/margins": 0.003263898193836212,
      "rewards/rejected": -0.001025668578222394,
      "step": 5
    },
    {
      "epoch": 0.4897959183673469,
      "grad_norm": 0.9774695038795471,
      "learning_rate": 9.979871469976197e-06,
      "logits/chosen": -2.387697696685791,
      "logits/rejected": -2.40444278717041,
      "logps/chosen": -88.27230072021484,
      "logps/rejected": -96.35655212402344,
      "loss": 0.6868,
      "rewards/accuracies": 0.56005859375,
      "rewards/chosen": 0.026464153081178665,
      "rewards/margins": 0.013977548107504845,
      "rewards/rejected": 0.01248660497367382,
      "step": 6
    },
    {
      "epoch": 0.5714285714285714,
      "grad_norm": 0.9563448429107666,
      "learning_rate": 9.91964794299315e-06,
      "logits/chosen": -2.3706889152526855,
      "logits/rejected": -2.395195722579956,
      "logps/chosen": -85.99250793457031,
      "logps/rejected": -96.60533905029297,
      "loss": 0.6829,
      "rewards/accuracies": 0.59912109375,
      "rewards/chosen": 0.03764727711677551,
      "rewards/margins": 0.02200627699494362,
      "rewards/rejected": 0.015641000121831894,
      "step": 7
    },
    {
      "epoch": 0.6530612244897959,
      "grad_norm": 1.0272670984268188,
      "learning_rate": 9.819814303479268e-06,
      "logits/chosen": -2.3650197982788086,
      "logits/rejected": -2.365727424621582,
      "logps/chosen": -89.17411804199219,
      "logps/rejected": -104.6484146118164,
      "loss": 0.6763,
      "rewards/accuracies": 0.65966796875,
      "rewards/chosen": 0.053641848266124725,
      "rewards/margins": 0.03600878268480301,
      "rewards/rejected": 0.017633063718676567,
      "step": 8
    },
    {
      "epoch": 0.7346938775510204,
      "grad_norm": 0.9614192247390747,
      "learning_rate": 9.681174353198687e-06,
      "logits/chosen": -2.3682937622070312,
      "logits/rejected": -2.3996546268463135,
      "logps/chosen": -84.27474975585938,
      "logps/rejected": -93.97203063964844,
      "loss": 0.6699,
      "rewards/accuracies": 0.6728515625,
      "rewards/chosen": 0.07653943449258804,
      "rewards/margins": 0.05041676014661789,
      "rewards/rejected": 0.026122674345970154,
      "step": 9
    },
    {
      "epoch": 0.8163265306122449,
      "grad_norm": 0.9354036450386047,
      "learning_rate": 9.504844339512096e-06,
      "logits/chosen": -2.3933653831481934,
      "logits/rejected": -2.3978521823883057,
      "logps/chosen": -84.89434051513672,
      "logps/rejected": -97.98092651367188,
      "loss": 0.6657,
      "rewards/accuracies": 0.68896484375,
      "rewards/chosen": 0.10172732174396515,
      "rewards/margins": 0.060384318232536316,
      "rewards/rejected": 0.04134300351142883,
      "step": 10
    },
    {
      "epoch": 0.8979591836734694,
      "grad_norm": 0.9544646143913269,
      "learning_rate": 9.292243968009332e-06,
      "logits/chosen": -2.388801336288452,
      "logits/rejected": -2.4009108543395996,
      "logps/chosen": -87.04521179199219,
      "logps/rejected": -99.57186889648438,
      "loss": 0.6578,
      "rewards/accuracies": 0.71240234375,
      "rewards/chosen": 0.12685298919677734,
      "rewards/margins": 0.07852301746606827,
      "rewards/rejected": 0.048329971730709076,
      "step": 11
    },
    {
      "epoch": 0.9795918367346939,
      "grad_norm": 0.945184051990509,
      "learning_rate": 9.045084971874738e-06,
      "logits/chosen": -2.4180715084075928,
      "logits/rejected": -2.4203011989593506,
      "logps/chosen": -87.21138763427734,
      "logps/rejected": -101.43778991699219,
      "loss": 0.648,
      "rewards/accuracies": 0.7294921875,
      "rewards/chosen": 0.15163584053516388,
      "rewards/margins": 0.10153679549694061,
      "rewards/rejected": 0.05009904503822327,
      "step": 12
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.9313170909881592,
      "learning_rate": 8.765357330018056e-06,
      "logits/chosen": -2.4016616344451904,
      "logits/rejected": -2.461811065673828,
      "logps/chosen": -91.43336486816406,
      "logps/rejected": -93.41850280761719,
      "loss": 0.6428,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.1737888753414154,
      "rewards/margins": 0.11637693643569946,
      "rewards/rejected": 0.05741193890571594,
      "step": 13
    },
    {
      "epoch": 1.0816326530612246,
      "grad_norm": 0.879965603351593,
      "learning_rate": 8.455313244934324e-06,
      "logits/chosen": -2.3858630657196045,
      "logits/rejected": -2.3951520919799805,
      "logps/chosen": -84.2239761352539,
      "logps/rejected": -97.90269470214844,
      "loss": 0.6378,
      "rewards/accuracies": 0.7109375,
      "rewards/chosen": 0.19609522819519043,
      "rewards/margins": 0.13054177165031433,
      "rewards/rejected": 0.0655534565448761,
      "step": 14
    },
    {
      "epoch": 1.163265306122449,
      "grad_norm": 0.8709496855735779,
      "learning_rate": 8.117449009293668e-06,
      "logits/chosen": -2.3995306491851807,
      "logits/rejected": -2.4127840995788574,
      "logps/chosen": -85.27073669433594,
      "logps/rejected": -98.00736236572266,
      "loss": 0.6306,
      "rewards/accuracies": 0.7041015625,
      "rewards/chosen": 0.21614021062850952,
      "rewards/margins": 0.15087103843688965,
      "rewards/rejected": 0.06526917964220047,
      "step": 15
    },
    {
      "epoch": 1.2448979591836735,
      "grad_norm": 0.9011364579200745,
      "learning_rate": 7.754484907260513e-06,
      "logits/chosen": -2.3790643215179443,
      "logits/rejected": -2.3908865451812744,
      "logps/chosen": -93.4906234741211,
      "logps/rejected": -104.75448608398438,
      "loss": 0.6132,
      "rewards/accuracies": 0.7509765625,
      "rewards/chosen": 0.2497057020664215,
      "rewards/margins": 0.19358570873737335,
      "rewards/rejected": 0.05611998587846756,
      "step": 16
    },
    {
      "epoch": 1.3265306122448979,
      "grad_norm": 0.8256746530532837,
      "learning_rate": 7.369343312364994e-06,
      "logits/chosen": -2.4009318351745605,
      "logits/rejected": -2.4279661178588867,
      "logps/chosen": -86.41546630859375,
      "logps/rejected": -96.96250915527344,
      "loss": 0.6066,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.27642762660980225,
      "rewards/margins": 0.2144363969564438,
      "rewards/rejected": 0.06199120730161667,
      "step": 17
    },
    {
      "epoch": 1.4081632653061225,
      "grad_norm": 0.8295791745185852,
      "learning_rate": 6.965125158269619e-06,
      "logits/chosen": -2.3977432250976562,
      "logits/rejected": -2.4250879287719727,
      "logps/chosen": -84.85093688964844,
      "logps/rejected": -95.09344482421875,
      "loss": 0.6024,
      "rewards/accuracies": 0.76025390625,
      "rewards/chosen": 0.27905529737472534,
      "rewards/margins": 0.2258065789937973,
      "rewards/rejected": 0.05324871465563774,
      "step": 18
    },
    {
      "epoch": 1.489795918367347,
      "grad_norm": 0.7997666001319885,
      "learning_rate": 6.545084971874738e-06,
      "logits/chosen": -2.403979778289795,
      "logits/rejected": -2.417026996612549,
      "logps/chosen": -87.35746002197266,
      "logps/rejected": -98.11820220947266,
      "loss": 0.5982,
      "rewards/accuracies": 0.74658203125,
      "rewards/chosen": 0.2821013629436493,
      "rewards/margins": 0.24028602242469788,
      "rewards/rejected": 0.04181533679366112,
      "step": 19
    },
    {
      "epoch": 1.5714285714285714,
      "grad_norm": 0.8002594113349915,
      "learning_rate": 6.112604669781572e-06,
      "logits/chosen": -2.4007248878479004,
      "logits/rejected": -2.4198200702667236,
      "logps/chosen": -84.9500503540039,
      "logps/rejected": -97.81690979003906,
      "loss": 0.595,
      "rewards/accuracies": 0.76416015625,
      "rewards/chosen": 0.27880722284317017,
      "rewards/margins": 0.24656042456626892,
      "rewards/rejected": 0.03224681317806244,
      "step": 20
    },
    {
      "epoch": 1.6530612244897958,
      "grad_norm": 0.8049229383468628,
      "learning_rate": 5.671166329088278e-06,
      "logits/chosen": -2.4079902172088623,
      "logits/rejected": -2.4213085174560547,
      "logps/chosen": -84.05070495605469,
      "logps/rejected": -98.56086730957031,
      "loss": 0.5807,
      "rewards/accuracies": 0.77001953125,
      "rewards/chosen": 0.31120967864990234,
      "rewards/margins": 0.29048869013786316,
      "rewards/rejected": 0.0207209512591362,
      "step": 21
    },
    {
      "epoch": 1.7346938775510203,
      "grad_norm": 0.7420709133148193,
      "learning_rate": 5.224324151752575e-06,
      "logits/chosen": -2.374910831451416,
      "logits/rejected": -2.3909456729888916,
      "logps/chosen": -85.0665283203125,
      "logps/rejected": -94.95793151855469,
      "loss": 0.573,
      "rewards/accuracies": 0.76904296875,
      "rewards/chosen": 0.3096790611743927,
      "rewards/margins": 0.3134555518627167,
      "rewards/rejected": -0.003776476252824068,
      "step": 22
    },
    {
      "epoch": 1.816326530612245,
      "grad_norm": 0.8085404634475708,
      "learning_rate": 4.775675848247427e-06,
      "logits/chosen": -2.3943331241607666,
      "logits/rejected": -2.3917038440704346,
      "logps/chosen": -81.59529876708984,
      "logps/rejected": -101.07594299316406,
      "loss": 0.5658,
      "rewards/accuracies": 0.7763671875,
      "rewards/chosen": 0.31912654638290405,
      "rewards/margins": 0.33659279346466064,
      "rewards/rejected": -0.017466213554143906,
      "step": 23
    },
    {
      "epoch": 1.8979591836734695,
      "grad_norm": 0.7618703246116638,
      "learning_rate": 4.3288336709117246e-06,
      "logits/chosen": -2.415018320083618,
      "logits/rejected": -2.427827835083008,
      "logps/chosen": -77.54719543457031,
      "logps/rejected": -92.66371154785156,
      "loss": 0.5677,
      "rewards/accuracies": 0.76513671875,
      "rewards/chosen": 0.3132694661617279,
      "rewards/margins": 0.33800169825553894,
      "rewards/rejected": -0.024732256308197975,
      "step": 24
    },
    {
      "epoch": 1.9795918367346939,
      "grad_norm": 0.7504754662513733,
      "learning_rate": 3.887395330218429e-06,
      "logits/chosen": -2.4105377197265625,
      "logits/rejected": -2.4160754680633545,
      "logps/chosen": -86.68771362304688,
      "logps/rejected": -102.70646667480469,
      "loss": 0.5621,
      "rewards/accuracies": 0.76171875,
      "rewards/chosen": 0.32979467511177063,
      "rewards/margins": 0.35629236698150635,
      "rewards/rejected": -0.026497643440961838,
      "step": 25
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8305085897445679,
      "learning_rate": 3.4549150281252635e-06,
      "logits/chosen": -2.418914794921875,
      "logits/rejected": -2.3890023231506348,
      "logps/chosen": -82.30009460449219,
      "logps/rejected": -107.8130874633789,
      "loss": 0.5472,
      "rewards/accuracies": 0.775390625,
      "rewards/chosen": 0.33478865027427673,
      "rewards/margins": 0.4053373336791992,
      "rewards/rejected": -0.07054867595434189,
      "step": 26
    },
    {
      "epoch": 2.0816326530612246,
      "grad_norm": 0.6879741549491882,
      "learning_rate": 3.0348748417303826e-06,
      "logits/chosen": -2.4019358158111572,
      "logits/rejected": -2.42034649848938,
      "logps/chosen": -85.02958679199219,
      "logps/rejected": -92.58648681640625,
      "loss": 0.5549,
      "rewards/accuracies": 0.7685546875,
      "rewards/chosen": 0.3364284038543701,
      "rewards/margins": 0.37677001953125,
      "rewards/rejected": -0.04034162312746048,
      "step": 27
    },
    {
      "epoch": 2.163265306122449,
      "grad_norm": 0.6818104982376099,
      "learning_rate": 2.6306566876350072e-06,
      "logits/chosen": -2.426608085632324,
      "logits/rejected": -2.4133074283599854,
      "logps/chosen": -84.07209014892578,
      "logps/rejected": -100.91854858398438,
      "loss": 0.5487,
      "rewards/accuracies": 0.765625,
      "rewards/chosen": 0.3244452476501465,
      "rewards/margins": 0.40365907549858093,
      "rewards/rejected": -0.07921378314495087,
      "step": 28
    },
    {
      "epoch": 2.2448979591836733,
      "grad_norm": 0.6256771087646484,
      "learning_rate": 2.245515092739488e-06,
      "logits/chosen": -2.408965587615967,
      "logits/rejected": -2.4255988597869873,
      "logps/chosen": -84.29199981689453,
      "logps/rejected": -94.82685089111328,
      "loss": 0.5534,
      "rewards/accuracies": 0.76171875,
      "rewards/chosen": 0.31452280282974243,
      "rewards/margins": 0.3928295075893402,
      "rewards/rejected": -0.07830668240785599,
      "step": 29
    },
    {
      "epoch": 2.326530612244898,
      "grad_norm": 0.6441386938095093,
      "learning_rate": 1.8825509907063328e-06,
      "logits/chosen": -2.4270732402801514,
      "logits/rejected": -2.4363362789154053,
      "logps/chosen": -84.19268798828125,
      "logps/rejected": -99.17901611328125,
      "loss": 0.5486,
      "rewards/accuracies": 0.7568359375,
      "rewards/chosen": 0.3075709939002991,
      "rewards/margins": 0.40655040740966797,
      "rewards/rejected": -0.0989793911576271,
      "step": 30
    },
    {
      "epoch": 2.4081632653061225,
      "grad_norm": 0.6837542653083801,
      "learning_rate": 1.544686755065677e-06,
      "logits/chosen": -2.4378089904785156,
      "logits/rejected": -2.447333812713623,
      "logps/chosen": -85.46794128417969,
      "logps/rejected": -102.33061981201172,
      "loss": 0.5394,
      "rewards/accuracies": 0.76904296875,
      "rewards/chosen": 0.34899264574050903,
      "rewards/margins": 0.4352303743362427,
      "rewards/rejected": -0.08623772114515305,
      "step": 31
    },
    {
      "epoch": 2.489795918367347,
      "grad_norm": 0.641403079032898,
      "learning_rate": 1.234642669981946e-06,
      "logits/chosen": -2.4003491401672363,
      "logits/rejected": -2.4253716468811035,
      "logps/chosen": -82.20948791503906,
      "logps/rejected": -100.959228515625,
      "loss": 0.5419,
      "rewards/accuracies": 0.7587890625,
      "rewards/chosen": 0.31559956073760986,
      "rewards/margins": 0.4267823100090027,
      "rewards/rejected": -0.11118277907371521,
      "step": 32
    },
    {
      "epoch": 2.571428571428571,
      "grad_norm": 0.6353703737258911,
      "learning_rate": 9.549150281252633e-07,
      "logits/chosen": -2.4461936950683594,
      "logits/rejected": -2.4458184242248535,
      "logps/chosen": -82.81317138671875,
      "logps/rejected": -100.34042358398438,
      "loss": 0.543,
      "rewards/accuracies": 0.767578125,
      "rewards/chosen": 0.3254584074020386,
      "rewards/margins": 0.4274338185787201,
      "rewards/rejected": -0.1019754409790039,
      "step": 33
    },
    {
      "epoch": 2.6530612244897958,
      "grad_norm": 0.6294788718223572,
      "learning_rate": 7.077560319906696e-07,
      "logits/chosen": -2.4284801483154297,
      "logits/rejected": -2.4406518936157227,
      "logps/chosen": -84.2886734008789,
      "logps/rejected": -101.47149658203125,
      "loss": 0.5405,
      "rewards/accuracies": 0.7626953125,
      "rewards/chosen": 0.3171379864215851,
      "rewards/margins": 0.4382990598678589,
      "rewards/rejected": -0.1211610957980156,
      "step": 34
    },
    {
      "epoch": 2.7346938775510203,
      "grad_norm": 0.6576454043388367,
      "learning_rate": 4.951556604879049e-07,
      "logits/chosen": -2.418337821960449,
      "logits/rejected": -2.4297537803649902,
      "logps/chosen": -86.08992767333984,
      "logps/rejected": -101.49446105957031,
      "loss": 0.5321,
      "rewards/accuracies": 0.779296875,
      "rewards/chosen": 0.343488484621048,
      "rewards/margins": 0.45854640007019043,
      "rewards/rejected": -0.11505790799856186,
      "step": 35
    },
    {
      "epoch": 2.816326530612245,
      "grad_norm": 0.6390217542648315,
      "learning_rate": 3.18825646801314e-07,
      "logits/chosen": -2.4269332885742188,
      "logits/rejected": -2.4400625228881836,
      "logps/chosen": -83.16020965576172,
      "logps/rejected": -98.95231628417969,
      "loss": 0.5411,
      "rewards/accuracies": 0.76708984375,
      "rewards/chosen": 0.32813936471939087,
      "rewards/margins": 0.43710485100746155,
      "rewards/rejected": -0.10896550118923187,
      "step": 36
    },
    {
      "epoch": 2.8979591836734695,
      "grad_norm": 0.6718611121177673,
      "learning_rate": 1.801856965207338e-07,
      "logits/chosen": -2.4263319969177246,
      "logits/rejected": -2.4267172813415527,
      "logps/chosen": -84.72743225097656,
      "logps/rejected": -101.02452850341797,
      "loss": 0.5361,
      "rewards/accuracies": 0.78076171875,
      "rewards/chosen": 0.3393610715866089,
      "rewards/margins": 0.4375483989715576,
      "rewards/rejected": -0.09818730503320694,
      "step": 37
    },
    {
      "epoch": 2.979591836734694,
      "grad_norm": 0.6170840263366699,
      "learning_rate": 8.035205700685167e-08,
      "logits/chosen": -2.3982906341552734,
      "logits/rejected": -2.412749767303467,
      "logps/chosen": -87.18891143798828,
      "logps/rejected": -101.30757141113281,
      "loss": 0.5391,
      "rewards/accuracies": 0.76416015625,
      "rewards/chosen": 0.3484157621860504,
      "rewards/margins": 0.4502694010734558,
      "rewards/rejected": -0.1018536388874054,
      "step": 38
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.7517494559288025,
      "learning_rate": 2.012853002380466e-08,
      "logits/chosen": -2.395568370819092,
      "logits/rejected": -2.403677225112915,
      "logps/chosen": -87.5918960571289,
      "logps/rejected": -93.06141662597656,
      "loss": 0.5331,
      "rewards/accuracies": 0.798828125,
      "rewards/chosen": 0.3457682132720947,
      "rewards/margins": 0.44890642166137695,
      "rewards/rejected": -0.10313821583986282,
      "step": 39
    }
  ],
  "logging_steps": 1,
  "max_steps": 39,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 20,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.297453798761431e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}