{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.999244142101285,
  "eval_steps": 300,
  "global_step": 661,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0015117157974300832,
      "grad_norm": 2.8220584392547607,
      "learning_rate": 0.0,
      "logits/chosen": 0.15524622797966003,
      "logits/rejected": 0.17812994122505188,
      "logps/chosen": -72.87983703613281,
      "logps/ref_chosen": -72.91278839111328,
      "logps/ref_rejected": -47.538291931152344,
      "logps/rejected": -47.55305480957031,
      "loss": 1.3863,
      "margin_dpo/margin_mean": 0.04771292209625244,
      "margin_dpo/margin_std": 0.24055811762809753,
      "step": 1
    },
    {
      "epoch": 0.0030234315948601664,
      "grad_norm": 2.786863088607788,
      "learning_rate": 7.462686567164179e-09,
      "logits/chosen": 0.13265366852283478,
      "logits/rejected": 0.06028885394334793,
      "logps/chosen": -47.69560241699219,
      "logps/ref_chosen": -47.742698669433594,
      "logps/ref_rejected": -54.36896896362305,
      "logps/rejected": -54.41385269165039,
      "loss": 1.3859,
      "margin_dpo/margin_mean": 0.09197819232940674,
      "margin_dpo/margin_std": 0.19671888649463654,
      "step": 2
    },
    {
      "epoch": 0.0045351473922902496,
      "grad_norm": 3.1143651008605957,
      "learning_rate": 1.4925373134328357e-08,
      "logits/chosen": 0.12916997075080872,
      "logits/rejected": 0.0313495397567749,
      "logps/chosen": -57.6026611328125,
      "logps/ref_chosen": -57.59052276611328,
      "logps/ref_rejected": -123.58332824707031,
      "logps/rejected": -123.56676483154297,
      "loss": 1.3866,
      "margin_dpo/margin_mean": -0.028696417808532715,
      "margin_dpo/margin_std": 0.30107247829437256,
      "step": 3
    },
    {
      "epoch": 0.006046863189720333,
      "grad_norm": 3.436974287033081,
      "learning_rate": 2.2388059701492534e-08,
      "logits/chosen": 0.11005942523479462,
      "logits/rejected": 0.08743590116500854,
      "logps/chosen": -72.13690185546875,
      "logps/ref_chosen": -72.2152328491211,
      "logps/ref_rejected": -72.17367553710938,
      "logps/rejected": -72.15191650390625,
      "loss": 1.3864,
      "margin_dpo/margin_mean": 0.05657494068145752,
      "margin_dpo/margin_std": 0.35384583473205566,
      "step": 4
    },
    {
      "epoch": 0.007558578987150416,
      "grad_norm": 2.9546422958374023,
      "learning_rate": 2.9850746268656714e-08,
      "logits/chosen": 0.1004139631986618,
      "logits/rejected": 0.015552863478660583,
      "logps/chosen": -56.2198486328125,
      "logps/ref_chosen": -56.174278259277344,
      "logps/ref_rejected": -111.51732635498047,
      "logps/rejected": -111.60870361328125,
      "loss": 1.3863,
      "margin_dpo/margin_mean": 0.04580581188201904,
      "margin_dpo/margin_std": 0.2617151737213135,
      "step": 5
    },
    {
      "epoch": 0.009070294784580499,
      "grad_norm": 2.960731267929077,
      "learning_rate": 3.731343283582089e-08,
      "logits/chosen": 0.09757982194423676,
      "logits/rejected": 0.021553121507167816,
      "logps/chosen": -33.81324005126953,
      "logps/ref_chosen": -33.68452453613281,
      "logps/ref_rejected": -102.40830993652344,
      "logps/rejected": -102.35445404052734,
      "loss": 1.387,
      "margin_dpo/margin_mean": -0.1825714111328125,
      "margin_dpo/margin_std": 0.29179731011390686,
      "step": 6
    },
    {
      "epoch": 0.010582010582010581,
      "grad_norm": 2.716484546661377,
      "learning_rate": 4.477611940298507e-08,
      "logits/chosen": -0.05511504039168358,
      "logits/rejected": -0.10139390826225281,
      "logps/chosen": -36.465389251708984,
      "logps/ref_chosen": -36.51213073730469,
      "logps/ref_rejected": -55.43703842163086,
      "logps/rejected": -55.4085578918457,
      "loss": 1.3864,
      "margin_dpo/margin_mean": 0.01826620101928711,
      "margin_dpo/margin_std": 0.24058791995048523,
      "step": 7
    },
    {
      "epoch": 0.012093726379440665,
      "grad_norm": 2.8784360885620117,
      "learning_rate": 5.223880597014925e-08,
      "logits/chosen": 0.09966567903757095,
      "logits/rejected": 0.06772118806838989,
      "logps/chosen": -91.59811401367188,
      "logps/ref_chosen": -91.63668823242188,
      "logps/ref_rejected": -115.22460174560547,
      "logps/rejected": -115.18936920166016,
      "loss": 1.3864,
      "margin_dpo/margin_mean": 0.0033463239669799805,
      "margin_dpo/margin_std": 0.3163236379623413,
      "step": 8
    },
    {
      "epoch": 0.013605442176870748,
      "grad_norm": 3.1453640460968018,
      "learning_rate": 5.970149253731343e-08,
      "logits/chosen": 0.08192189782857895,
      "logits/rejected": 0.07255452871322632,
      "logps/chosen": -91.2304916381836,
      "logps/ref_chosen": -91.08434295654297,
      "logps/ref_rejected": -78.1643295288086,
      "logps/rejected": -78.20406341552734,
      "loss": 1.3863,
      "margin_dpo/margin_mean": -0.10640597343444824,
      "margin_dpo/margin_std": 0.37841445207595825,
      "step": 9
    },
    {
      "epoch": 0.015117157974300832,
      "grad_norm": 2.9692389965057373,
      "learning_rate": 6.71641791044776e-08,
      "logits/chosen": 0.1450178623199463,
      "logits/rejected": 0.08948463946580887,
      "logps/chosen": -67.3464126586914,
      "logps/ref_chosen": -67.3323974609375,
      "logps/ref_rejected": -115.37638854980469,
      "logps/rejected": -115.22101593017578,
      "loss": 1.3866,
      "margin_dpo/margin_mean": -0.16937530040740967,
      "margin_dpo/margin_std": 0.2835836708545685,
      "step": 10
    },
    {
      "epoch": 0.016628873771730914,
      "grad_norm": 2.9334216117858887,
      "learning_rate": 7.462686567164178e-08,
      "logits/chosen": 0.06959305703639984,
      "logits/rejected": 0.031854622066020966,
      "logps/chosen": -48.15754699707031,
      "logps/ref_chosen": -48.158531188964844,
      "logps/ref_rejected": -66.28083038330078,
      "logps/rejected": -66.29977416992188,
      "loss": 1.3861,
      "margin_dpo/margin_mean": 0.01993393898010254,
      "margin_dpo/margin_std": 0.2706565260887146,
      "step": 11
    },
    {
      "epoch": 0.018140589569160998,
      "grad_norm": 2.821261167526245,
      "learning_rate": 8.208955223880596e-08,
      "logits/chosen": -0.059022121131420135,
      "logits/rejected": -0.055310823023319244,
      "logps/chosen": -68.31194305419922,
      "logps/ref_chosen": -68.27070617675781,
      "logps/ref_rejected": -66.15010070800781,
      "logps/rejected": -66.14984130859375,
      "loss": 1.3862,
      "margin_dpo/margin_mean": -0.04150247573852539,
      "margin_dpo/margin_std": 0.30395328998565674,
      "step": 12
    },
    {
      "epoch": 0.019652305366591082,
      "grad_norm": 2.898451089859009,
      "learning_rate": 8.955223880597014e-08,
      "logits/chosen": 0.07420723140239716,
      "logits/rejected": -0.013622896745800972,
      "logps/chosen": -55.22724151611328,
      "logps/ref_chosen": -55.27293395996094,
      "logps/ref_rejected": -116.98251342773438,
      "logps/rejected": -116.99977111816406,
      "loss": 1.386,
      "margin_dpo/margin_mean": 0.06296110153198242,
      "margin_dpo/margin_std": 0.41959983110427856,
      "step": 13
    },
    {
      "epoch": 0.021164021164021163,
      "grad_norm": 2.963911533355713,
      "learning_rate": 9.701492537313432e-08,
      "logits/chosen": 0.1168874204158783,
      "logits/rejected": 0.0010670926421880722,
      "logps/chosen": -54.75418472290039,
      "logps/ref_chosen": -54.73517608642578,
      "logps/ref_rejected": -79.82144165039062,
      "logps/rejected": -79.80023193359375,
      "loss": 1.3859,
      "margin_dpo/margin_mean": -0.0402069091796875,
      "margin_dpo/margin_std": 0.21686886250972748,
      "step": 14
    },
    {
      "epoch": 0.022675736961451247,
      "grad_norm": 3.3150715827941895,
      "learning_rate": 1.044776119402985e-07,
      "logits/chosen": 0.09902211278676987,
      "logits/rejected": 0.0021891482174396515,
      "logps/chosen": -47.44508361816406,
      "logps/ref_chosen": -47.35077667236328,
      "logps/ref_rejected": -98.93363189697266,
      "logps/rejected": -98.7593994140625,
      "loss": 1.3867,
      "margin_dpo/margin_mean": -0.26853013038635254,
      "margin_dpo/margin_std": 0.3253557085990906,
      "step": 15
    },
    {
      "epoch": 0.02418745275888133,
      "grad_norm": 2.685232639312744,
      "learning_rate": 1.1194029850746268e-07,
      "logits/chosen": 0.13864967226982117,
      "logits/rejected": 0.09606489539146423,
      "logps/chosen": -37.010860443115234,
      "logps/ref_chosen": -36.986331939697266,
      "logps/ref_rejected": -55.56471633911133,
      "logps/rejected": -55.6058464050293,
      "loss": 1.3861,
      "margin_dpo/margin_mean": 0.016601920127868652,
      "margin_dpo/margin_std": 0.2442459762096405,
      "step": 16
    },
    {
      "epoch": 0.025699168556311415,
      "grad_norm": 3.2944958209991455,
      "learning_rate": 1.1940298507462686e-07,
      "logits/chosen": 0.06346956640481949,
      "logits/rejected": 0.028996460139751434,
      "logps/chosen": -53.554481506347656,
      "logps/ref_chosen": -53.60247802734375,
      "logps/ref_rejected": -77.51017761230469,
      "logps/rejected": -77.53253936767578,
      "loss": 1.386,
      "margin_dpo/margin_mean": 0.07034468650817871,
      "margin_dpo/margin_std": 0.30959659814834595,
      "step": 17
    },
    {
      "epoch": 0.027210884353741496,
      "grad_norm": 2.7822513580322266,
      "learning_rate": 1.2686567164179106e-07,
      "logits/chosen": 0.08662936091423035,
      "logits/rejected": 0.05197212100028992,
      "logps/chosen": -43.53667449951172,
      "logps/ref_chosen": -43.651275634765625,
      "logps/ref_rejected": -70.06555938720703,
      "logps/rejected": -70.07490539550781,
      "loss": 1.3857,
      "margin_dpo/margin_mean": 0.12395095825195312,
      "margin_dpo/margin_std": 0.20322147011756897,
      "step": 18
    },
    {
      "epoch": 0.02872260015117158,
      "grad_norm": 3.082880735397339,
      "learning_rate": 1.343283582089552e-07,
      "logits/chosen": 0.09144282341003418,
      "logits/rejected": 0.04711594060063362,
      "logps/chosen": -59.105987548828125,
      "logps/ref_chosen": -59.121559143066406,
      "logps/ref_rejected": -95.91956329345703,
      "logps/rejected": -95.91903686523438,
      "loss": 1.386,
      "margin_dpo/margin_mean": 0.015046358108520508,
      "margin_dpo/margin_std": 0.3638499677181244,
      "step": 19
    },
    {
      "epoch": 0.030234315948601664,
      "grad_norm": 2.926210880279541,
      "learning_rate": 1.4179104477611938e-07,
      "logits/chosen": 0.12322017550468445,
      "logits/rejected": 0.061798982322216034,
      "logps/chosen": -51.08586120605469,
      "logps/ref_chosen": -51.031883239746094,
      "logps/ref_rejected": -96.53536224365234,
      "logps/rejected": -96.57363891601562,
      "loss": 1.3865,
      "margin_dpo/margin_mean": -0.015711426734924316,
      "margin_dpo/margin_std": 0.19928568601608276,
      "step": 20
    },
    {
      "epoch": 0.031746031746031744,
      "grad_norm": 2.9375104904174805,
      "learning_rate": 1.4925373134328355e-07,
      "logits/chosen": -0.0577910915017128,
      "logits/rejected": -0.00928708165884018,
      "logps/chosen": -76.64593505859375,
      "logps/ref_chosen": -76.7261962890625,
      "logps/ref_rejected": -53.63877868652344,
      "logps/rejected": -53.602439880371094,
      "loss": 1.3867,
      "margin_dpo/margin_mean": 0.0439220666885376,
      "margin_dpo/margin_std": 0.23736098408699036,
      "step": 21
    },
    {
      "epoch": 0.03325774754346183,
      "grad_norm": 2.896641969680786,
      "learning_rate": 1.5671641791044775e-07,
      "logits/chosen": -0.03439096361398697,
      "logits/rejected": -0.11815059185028076,
      "logps/chosen": -36.69358444213867,
      "logps/ref_chosen": -36.67463302612305,
      "logps/ref_rejected": -58.37761688232422,
      "logps/rejected": -58.41578674316406,
      "loss": 1.3866,
      "margin_dpo/margin_mean": 0.019217491149902344,
      "margin_dpo/margin_std": 0.21051761507987976,
      "step": 22
    },
    {
      "epoch": 0.03476946334089191,
      "grad_norm": 3.29414963722229,
      "learning_rate": 1.6417910447761193e-07,
      "logits/chosen": 0.2472844123840332,
      "logits/rejected": 0.1739306002855301,
      "logps/chosen": -88.78631591796875,
      "logps/ref_chosen": -88.72676849365234,
      "logps/ref_rejected": -151.3695068359375,
      "logps/rejected": -151.45941162109375,
      "loss": 1.3865,
      "margin_dpo/margin_mean": 0.030363917350769043,
      "margin_dpo/margin_std": 0.18649427592754364,
      "step": 23
    },
    {
      "epoch": 0.036281179138321996,
      "grad_norm": 2.626323938369751,
      "learning_rate": 1.716417910447761e-07,
      "logits/chosen": 0.10805254429578781,
      "logits/rejected": 0.052521951496601105,
      "logps/chosen": -46.81217575073242,
      "logps/ref_chosen": -46.81597137451172,
      "logps/ref_rejected": -77.96964263916016,
      "logps/rejected": -78.01282501220703,
      "loss": 1.3866,
      "margin_dpo/margin_mean": 0.04697418212890625,
      "margin_dpo/margin_std": 0.1974717229604721,
      "step": 24
    },
    {
      "epoch": 0.03779289493575208,
      "grad_norm": 2.9760308265686035,
      "learning_rate": 1.7910447761194027e-07,
      "logits/chosen": 0.027542775496840477,
      "logits/rejected": -0.008236430585384369,
      "logps/chosen": -66.06083679199219,
      "logps/ref_chosen": -66.00652313232422,
      "logps/ref_rejected": -94.2716293334961,
      "logps/rejected": -94.29405212402344,
      "loss": 1.3861,
      "margin_dpo/margin_mean": -0.031876444816589355,
      "margin_dpo/margin_std": 0.2860802412033081,
      "step": 25
    },
    {
      "epoch": 0.039304610733182165,
      "grad_norm": 2.988525152206421,
      "learning_rate": 1.8656716417910447e-07,
      "logits/chosen": 0.16524234414100647,
      "logits/rejected": 0.14268483221530914,
      "logps/chosen": -67.2355728149414,
      "logps/ref_chosen": -67.12877655029297,
      "logps/ref_rejected": -115.98387145996094,
      "logps/rejected": -116.02112579345703,
      "loss": 1.3867,
      "margin_dpo/margin_mean": -0.06954813003540039,
      "margin_dpo/margin_std": 0.4333837032318115,
      "step": 26
    },
    {
      "epoch": 0.04081632653061224,
      "grad_norm": 2.903107166290283,
      "learning_rate": 1.9402985074626865e-07,
      "logits/chosen": 0.1587177813053131,
      "logits/rejected": 0.14436712861061096,
      "logps/chosen": -45.9591178894043,
      "logps/ref_chosen": -45.86199188232422,
      "logps/ref_rejected": -56.50844955444336,
      "logps/rejected": -56.52033615112305,
      "loss": 1.3864,
      "margin_dpo/margin_mean": -0.08524084091186523,
      "margin_dpo/margin_std": 0.2513440251350403,
      "step": 27
    },
    {
      "epoch": 0.042328042328042326,
      "grad_norm": 2.764413833618164,
      "learning_rate": 2.0149253731343282e-07,
      "logits/chosen": -0.02600286155939102,
      "logits/rejected": -0.024466849863529205,
      "logps/chosen": -100.1274642944336,
      "logps/ref_chosen": -100.07420349121094,
      "logps/ref_rejected": -98.97936248779297,
      "logps/rejected": -98.93499755859375,
      "loss": 1.3864,
      "margin_dpo/margin_mean": -0.09762585163116455,
      "margin_dpo/margin_std": 0.4535467326641083,
      "step": 28
    },
    {
      "epoch": 0.04383975812547241,
      "grad_norm": 3.2600700855255127,
      "learning_rate": 2.08955223880597e-07,
      "logits/chosen": 0.09348219633102417,
      "logits/rejected": 0.09450555592775345,
      "logps/chosen": -77.15126037597656,
      "logps/ref_chosen": -77.19854736328125,
      "logps/ref_rejected": -73.5864028930664,
      "logps/rejected": -73.67513275146484,
      "loss": 1.3857,
      "margin_dpo/margin_mean": 0.13602125644683838,
      "margin_dpo/margin_std": 0.40513163805007935,
      "step": 29
    },
    {
      "epoch": 0.045351473922902494,
      "grad_norm": 3.0433361530303955,
      "learning_rate": 2.1641791044776117e-07,
      "logits/chosen": 0.15750174224376678,
      "logits/rejected": 0.1133815348148346,
      "logps/chosen": -59.61478805541992,
      "logps/ref_chosen": -59.58109664916992,
      "logps/ref_rejected": -81.88029479980469,
      "logps/rejected": -81.91694641113281,
      "loss": 1.3854,
      "margin_dpo/margin_mean": 0.002964496612548828,
      "margin_dpo/margin_std": 0.1755043864250183,
      "step": 30
    },
    {
      "epoch": 0.04686318972033258,
      "grad_norm": 3.4319612979888916,
      "learning_rate": 2.2388059701492537e-07,
      "logits/chosen": 0.1796724498271942,
      "logits/rejected": 0.1043396145105362,
      "logps/chosen": -69.11215209960938,
      "logps/ref_chosen": -69.12844848632812,
      "logps/ref_rejected": -135.9136505126953,
      "logps/rejected": -136.04110717773438,
      "loss": 1.3858,
      "margin_dpo/margin_mean": 0.14375102519989014,
      "margin_dpo/margin_std": 0.2275710105895996,
      "step": 31
    },
    {
      "epoch": 0.04837490551776266,
      "grad_norm": 3.0030133724212646,
      "learning_rate": 2.3134328358208954e-07,
      "logits/chosen": 0.08424107730388641,
      "logits/rejected": 0.03786729276180267,
      "logps/chosen": -79.7166748046875,
      "logps/ref_chosen": -79.69361877441406,
      "logps/ref_rejected": -119.91200256347656,
      "logps/rejected": -119.93031311035156,
      "loss": 1.3867,
      "margin_dpo/margin_mean": -0.004748940467834473,
      "margin_dpo/margin_std": 0.19554254412651062,
      "step": 32
    },
    {
      "epoch": 0.049886621315192746,
      "grad_norm": 3.1022226810455322,
      "learning_rate": 2.388059701492537e-07,
      "logits/chosen": 0.05021132156252861,
      "logits/rejected": 0.02479320764541626,
      "logps/chosen": -55.22189712524414,
      "logps/ref_chosen": -55.10992431640625,
      "logps/ref_rejected": -82.52067565917969,
      "logps/rejected": -82.57884216308594,
      "loss": 1.3863,
      "margin_dpo/margin_mean": -0.05379873514175415,
      "margin_dpo/margin_std": 0.37445366382598877,
      "step": 33
    },
    {
      "epoch": 0.05139833711262283,
      "grad_norm": 2.8541648387908936,
      "learning_rate": 2.4626865671641786e-07,
      "logits/chosen": 0.057025909423828125,
      "logits/rejected": -0.0010375156998634338,
      "logps/chosen": -52.59141540527344,
      "logps/ref_chosen": -52.49810028076172,
      "logps/ref_rejected": -74.43905639648438,
      "logps/rejected": -74.50889587402344,
      "loss": 1.3854,
      "margin_dpo/margin_mean": -0.02347743511199951,
      "margin_dpo/margin_std": 0.2468748390674591,
      "step": 34
    },
    {
      "epoch": 0.05291005291005291,
      "grad_norm": 2.7856571674346924,
      "learning_rate": 2.537313432835821e-07,
      "logits/chosen": -0.023874841630458832,
      "logits/rejected": -0.03130623698234558,
      "logps/chosen": -67.71501159667969,
      "logps/ref_chosen": -67.55126953125,
      "logps/ref_rejected": -55.26748275756836,
      "logps/rejected": -55.42961502075195,
      "loss": 1.3864,
      "margin_dpo/margin_mean": -0.0016133785247802734,
      "margin_dpo/margin_std": 0.33188915252685547,
      "step": 35
    },
    {
      "epoch": 0.05442176870748299,
      "grad_norm": 2.8720510005950928,
      "learning_rate": 2.611940298507462e-07,
      "logits/chosen": 0.16390232741832733,
      "logits/rejected": 0.08551906049251556,
      "logps/chosen": -52.384647369384766,
      "logps/ref_chosen": -52.303955078125,
      "logps/ref_rejected": -88.28939819335938,
      "logps/rejected": -88.45569610595703,
      "loss": 1.3856,
      "margin_dpo/margin_mean": 0.085601806640625,
      "margin_dpo/margin_std": 0.35991254448890686,
      "step": 36
    },
    {
      "epoch": 0.055933484504913075,
      "grad_norm": 3.7995986938476562,
      "learning_rate": 2.686567164179104e-07,
      "logits/chosen": 0.06787224858999252,
      "logits/rejected": 0.008850198239088058,
      "logps/chosen": -60.5541877746582,
      "logps/ref_chosen": -60.51344680786133,
      "logps/ref_rejected": -103.91683959960938,
      "logps/rejected": -104.0807113647461,
      "loss": 1.3853,
      "margin_dpo/margin_mean": 0.12312948703765869,
      "margin_dpo/margin_std": 0.36983656883239746,
      "step": 37
    },
    {
      "epoch": 0.05744520030234316,
      "grad_norm": 2.977374792098999,
      "learning_rate": 2.761194029850746e-07,
      "logits/chosen": 0.05500081554055214,
      "logits/rejected": 0.019983187317848206,
      "logps/chosen": -50.67215347290039,
      "logps/ref_chosen": -50.62315368652344,
      "logps/ref_rejected": -60.629703521728516,
      "logps/rejected": -60.82992935180664,
      "loss": 1.3861,
      "margin_dpo/margin_mean": 0.15122318267822266,
      "margin_dpo/margin_std": 0.3122575581073761,
      "step": 38
    },
    {
      "epoch": 0.05895691609977324,
      "grad_norm": 2.931081771850586,
      "learning_rate": 2.8358208955223876e-07,
      "logits/chosen": 0.2395174652338028,
      "logits/rejected": 0.20490288734436035,
      "logps/chosen": -45.89991760253906,
      "logps/ref_chosen": -45.8764533996582,
      "logps/ref_rejected": -69.75498962402344,
      "logps/rejected": -69.87886047363281,
      "loss": 1.3858,
      "margin_dpo/margin_mean": 0.10041189193725586,
      "margin_dpo/margin_std": 0.28835514187812805,
      "step": 39
    },
    {
      "epoch": 0.06046863189720333,
      "grad_norm": 3.1318254470825195,
      "learning_rate": 2.9104477611940296e-07,
      "logits/chosen": 0.18963474035263062,
      "logits/rejected": 0.16616499423980713,
      "logps/chosen": -39.96923828125,
      "logps/ref_chosen": -39.950111389160156,
      "logps/ref_rejected": -62.60026931762695,
      "logps/rejected": -62.79283142089844,
      "loss": 1.3859,
      "margin_dpo/margin_mean": 0.1734316349029541,
      "margin_dpo/margin_std": 0.2163560539484024,
      "step": 40
    },
    {
      "epoch": 0.06198034769463341,
      "grad_norm": 3.0570766925811768,
      "learning_rate": 2.985074626865671e-07,
      "logits/chosen": -0.12535785138607025,
      "logits/rejected": -0.12432264536619186,
      "logps/chosen": -65.45966339111328,
      "logps/ref_chosen": -65.39937591552734,
      "logps/ref_rejected": -83.19218444824219,
      "logps/rejected": -83.37960815429688,
      "loss": 1.3852,
      "margin_dpo/margin_mean": 0.1271369457244873,
      "margin_dpo/margin_std": 0.2651064991950989,
      "step": 41
    },
    {
      "epoch": 0.06349206349206349,
      "grad_norm": 2.934258460998535,
      "learning_rate": 3.059701492537313e-07,
      "logits/chosen": 0.06298904120922089,
      "logits/rejected": 0.05617032200098038,
      "logps/chosen": -61.81056594848633,
      "logps/ref_chosen": -61.71239471435547,
      "logps/ref_rejected": -81.27043151855469,
      "logps/rejected": -81.32711029052734,
      "loss": 1.3864,
      "margin_dpo/margin_mean": -0.04149121046066284,
      "margin_dpo/margin_std": 0.3146214485168457,
      "step": 42
    },
    {
      "epoch": 0.06500377928949358,
      "grad_norm": 3.4698517322540283,
      "learning_rate": 3.134328358208955e-07,
      "logits/chosen": 0.029873725026845932,
      "logits/rejected": -0.042266130447387695,
      "logps/chosen": -68.39386749267578,
      "logps/ref_chosen": -68.25798797607422,
      "logps/ref_rejected": -94.93944549560547,
      "logps/rejected": -95.09349060058594,
      "loss": 1.3861,
      "margin_dpo/margin_mean": 0.018149971961975098,
      "margin_dpo/margin_std": 0.2860589325428009,
      "step": 43
    },
    {
      "epoch": 0.06651549508692366,
      "grad_norm": 3.261988639831543,
      "learning_rate": 3.2089552238805965e-07,
      "logits/chosen": 0.059243109077215195,
      "logits/rejected": 0.07599075883626938,
      "logps/chosen": -76.56920623779297,
      "logps/ref_chosen": -76.42689514160156,
      "logps/ref_rejected": -84.94781494140625,
      "logps/rejected": -84.98534393310547,
      "loss": 1.3858,
      "margin_dpo/margin_mean": -0.10478079319000244,
      "margin_dpo/margin_std": 0.2691563367843628,
      "step": 44
    },
    {
      "epoch": 0.06802721088435375,
      "grad_norm": 3.200819492340088,
      "learning_rate": 3.2835820895522385e-07,
      "logits/chosen": 0.09944377094507217,
      "logits/rejected": 0.05152427405118942,
      "logps/chosen": -74.33448791503906,
      "logps/ref_chosen": -74.13058471679688,
      "logps/ref_rejected": -82.69816589355469,
      "logps/rejected": -82.89010620117188,
      "loss": 1.385,
      "margin_dpo/margin_mean": -0.011955618858337402,
      "margin_dpo/margin_std": 0.3313651978969574,
      "step": 45
    },
    {
      "epoch": 0.06953892668178382,
      "grad_norm": 3.1459238529205322,
      "learning_rate": 3.3582089552238805e-07,
      "logits/chosen": 0.17175917327404022,
      "logits/rejected": 0.15970298647880554,
      "logps/chosen": -96.9361343383789,
      "logps/ref_chosen": -96.75468444824219,
      "logps/ref_rejected": -82.88079833984375,
      "logps/rejected": -83.0982666015625,
      "loss": 1.3848,
      "margin_dpo/margin_mean": 0.03600466251373291,
      "margin_dpo/margin_std": 0.3888518512248993,
      "step": 46
    },
    {
      "epoch": 0.0710506424792139,
      "grad_norm": 2.8698384761810303,
      "learning_rate": 3.432835820895522e-07,
      "logits/chosen": 0.1564871072769165,
      "logits/rejected": 0.1589561551809311,
      "logps/chosen": -53.687896728515625,
      "logps/ref_chosen": -53.5174560546875,
      "logps/ref_rejected": -55.22039794921875,
      "logps/rejected": -55.589107513427734,
      "loss": 1.385,
      "margin_dpo/margin_mean": 0.19826769828796387,
      "margin_dpo/margin_std": 0.3298158049583435,
      "step": 47
    },
    {
      "epoch": 0.07256235827664399,
      "grad_norm": 2.6883156299591064,
      "learning_rate": 3.507462686567164e-07,
      "logits/chosen": 0.17292676866054535,
      "logits/rejected": 0.13669899106025696,
      "logps/chosen": -62.340145111083984,
      "logps/ref_chosen": -62.185054779052734,
      "logps/ref_rejected": -74.84376525878906,
      "logps/rejected": -75.03015899658203,
      "loss": 1.386,
      "margin_dpo/margin_mean": 0.03130638599395752,
      "margin_dpo/margin_std": 0.34968793392181396,
      "step": 48
    },
    {
      "epoch": 0.07407407407407407,
      "grad_norm": 2.750591993331909,
      "learning_rate": 3.5820895522388055e-07,
      "logits/chosen": 0.057202327996492386,
      "logits/rejected": 0.037843670696020126,
      "logps/chosen": -37.986602783203125,
      "logps/ref_chosen": -37.77415084838867,
      "logps/ref_rejected": -51.56128692626953,
      "logps/rejected": -51.90427780151367,
      "loss": 1.3855,
      "margin_dpo/margin_mean": 0.13054275512695312,
      "margin_dpo/margin_std": 0.2379816472530365,
      "step": 49
    },
    {
      "epoch": 0.07558578987150416,
      "grad_norm": 2.820672035217285,
      "learning_rate": 3.6567164179104475e-07,
      "logits/chosen": 0.10407137125730515,
      "logits/rejected": 0.08040489256381989,
      "logps/chosen": -60.96312713623047,
      "logps/ref_chosen": -60.664947509765625,
      "logps/ref_rejected": -89.31661987304688,
      "logps/rejected": -89.6517333984375,
      "loss": 1.3852,
      "margin_dpo/margin_mean": 0.03693675994873047,
      "margin_dpo/margin_std": 0.25188928842544556,
      "step": 50
    },
    {
      "epoch": 0.07709750566893424,
      "grad_norm": 3.1633975505828857,
      "learning_rate": 3.7313432835820895e-07,
      "logits/chosen": 0.11444868892431259,
      "logits/rejected": 0.04659546911716461,
      "logps/chosen": -48.322227478027344,
      "logps/ref_chosen": -48.012168884277344,
      "logps/ref_rejected": -68.38160705566406,
      "logps/rejected": -68.88851928710938,
      "loss": 1.3847,
      "margin_dpo/margin_mean": 0.19685041904449463,
      "margin_dpo/margin_std": 0.40103837847709656,
      "step": 51
    },
    {
      "epoch": 0.07860922146636433,
      "grad_norm": 3.325279951095581,
      "learning_rate": 3.805970149253731e-07,
      "logits/chosen": 0.08999551832675934,
      "logits/rejected": 0.05174541473388672,
      "logps/chosen": -69.64974975585938,
      "logps/ref_chosen": -69.34451293945312,
      "logps/ref_rejected": -115.13761901855469,
      "logps/rejected": -115.90428924560547,
      "loss": 1.3839,
      "margin_dpo/margin_mean": 0.46144330501556396,
      "margin_dpo/margin_std": 0.610642671585083,
      "step": 52
    },
    {
      "epoch": 0.0801209372637944,
      "grad_norm": 2.7831149101257324,
      "learning_rate": 3.880597014925373e-07,
      "logits/chosen": 0.05799449607729912,
      "logits/rejected": 0.057163748890161514,
      "logps/chosen": -61.22389602661133,
      "logps/ref_chosen": -60.885520935058594,
      "logps/ref_rejected": -51.6776237487793,
      "logps/rejected": -52.1979866027832,
      "loss": 1.3842,
      "margin_dpo/margin_mean": 0.18198823928833008,
      "margin_dpo/margin_std": 0.3276767432689667,
      "step": 53
    },
    {
      "epoch": 0.08163265306122448,
      "grad_norm": 2.6667864322662354,
      "learning_rate": 3.9552238805970144e-07,
      "logits/chosen": -0.010790073312819004,
      "logits/rejected": -0.014523019082844257,
      "logps/chosen": -43.61071014404297,
      "logps/ref_chosen": -43.147682189941406,
      "logps/ref_rejected": -48.63517761230469,
      "logps/rejected": -49.14527130126953,
      "loss": 1.3849,
      "margin_dpo/margin_mean": 0.047069668769836426,
      "margin_dpo/margin_std": 0.5451761484146118,
      "step": 54
    },
    {
      "epoch": 0.08314436885865457,
      "grad_norm": 3.2038674354553223,
      "learning_rate": 4.0298507462686564e-07,
      "logits/chosen": 0.09067589789628983,
      "logits/rejected": -0.015168176032602787,
      "logps/chosen": -45.248573303222656,
      "logps/ref_chosen": -44.79475402832031,
      "logps/ref_rejected": -87.80093383789062,
      "logps/rejected": -88.53862762451172,
      "loss": 1.3832,
      "margin_dpo/margin_mean": 0.28387510776519775,
      "margin_dpo/margin_std": 0.6332917809486389,
      "step": 55
    },
    {
      "epoch": 0.08465608465608465,
      "grad_norm": 2.775723934173584,
      "learning_rate": 4.1044776119402984e-07,
      "logits/chosen": 0.09276323020458221,
      "logits/rejected": 0.11390136182308197,
      "logps/chosen": -71.15162658691406,
      "logps/ref_chosen": -70.83395385742188,
      "logps/ref_rejected": -66.29704284667969,
      "logps/rejected": -66.77851867675781,
      "loss": 1.3843,
      "margin_dpo/margin_mean": 0.16381430625915527,
      "margin_dpo/margin_std": 0.6528092622756958,
      "step": 56
    },
    {
      "epoch": 0.08616780045351474,
      "grad_norm": 3.1573398113250732,
      "learning_rate": 4.17910447761194e-07,
      "logits/chosen": 0.13995778560638428,
      "logits/rejected": 0.10124865919351578,
      "logps/chosen": -55.33793640136719,
      "logps/ref_chosen": -54.77841567993164,
      "logps/ref_rejected": -97.11053466796875,
      "logps/rejected": -97.8675537109375,
      "loss": 1.3835,
      "margin_dpo/margin_mean": 0.19750165939331055,
      "margin_dpo/margin_std": 0.8898135423660278,
      "step": 57
    },
    {
      "epoch": 0.08767951625094482,
      "grad_norm": 3.0929791927337646,
      "learning_rate": 4.253731343283582e-07,
      "logits/chosen": -0.024497557431459427,
      "logits/rejected": -0.11392132192850113,
      "logps/chosen": -59.87010955810547,
      "logps/ref_chosen": -59.280887603759766,
      "logps/ref_rejected": -82.80585479736328,
      "logps/rejected": -83.43619537353516,
      "loss": 1.3839,
      "margin_dpo/margin_mean": 0.041118621826171875,
      "margin_dpo/margin_std": 0.6544058322906494,
      "step": 58
    },
    {
      "epoch": 0.08919123204837491,
      "grad_norm": 2.7556982040405273,
      "learning_rate": 4.3283582089552234e-07,
      "logits/chosen": 0.17279072105884552,
      "logits/rejected": 0.04841926693916321,
      "logps/chosen": -38.3438720703125,
      "logps/ref_chosen": -37.76800537109375,
      "logps/ref_rejected": -84.53325653076172,
      "logps/rejected": -85.44366455078125,
      "loss": 1.3833,
      "margin_dpo/margin_mean": 0.33453965187072754,
      "margin_dpo/margin_std": 0.6480045318603516,
      "step": 59
    },
    {
      "epoch": 0.09070294784580499,
      "grad_norm": 2.9187021255493164,
      "learning_rate": 4.4029850746268654e-07,
      "logits/chosen": 0.2775009274482727,
      "logits/rejected": 0.2738117575645447,
      "logps/chosen": -67.29905700683594,
      "logps/ref_chosen": -66.45524597167969,
      "logps/ref_rejected": -67.77998352050781,
      "logps/rejected": -68.42279052734375,
      "loss": 1.3855,
      "margin_dpo/margin_mean": -0.20100653171539307,
      "margin_dpo/margin_std": 1.0322511196136475,
      "step": 60
    },
    {
      "epoch": 0.09221466364323508,
      "grad_norm": 2.7786881923675537,
      "learning_rate": 4.4776119402985074e-07,
      "logits/chosen": 0.02796616032719612,
      "logits/rejected": -0.045561283826828,
      "logps/chosen": -41.18583679199219,
      "logps/ref_chosen": -40.7296257019043,
      "logps/ref_rejected": -82.87712097167969,
      "logps/rejected": -83.8257827758789,
      "loss": 1.3858,
      "margin_dpo/margin_mean": 0.4924490451812744,
      "margin_dpo/margin_std": 0.6613912582397461,
      "step": 61
    },
    {
      "epoch": 0.09372637944066516,
      "grad_norm": 2.8831284046173096,
      "learning_rate": 4.552238805970149e-07,
      "logits/chosen": 0.16551688313484192,
      "logits/rejected": 0.05868356674909592,
      "logps/chosen": -47.318138122558594,
      "logps/ref_chosen": -46.39446258544922,
      "logps/ref_rejected": -79.28895568847656,
      "logps/rejected": -80.39089965820312,
      "loss": 1.3838,
      "margin_dpo/margin_mean": 0.17827367782592773,
      "margin_dpo/margin_std": 1.0472949743270874,
      "step": 62
    },
    {
      "epoch": 0.09523809523809523,
      "grad_norm": 3.3160440921783447,
      "learning_rate": 4.626865671641791e-07,
      "logits/chosen": 0.2583298087120056,
      "logits/rejected": 0.2626573443412781,
      "logps/chosen": -75.17607116699219,
      "logps/ref_chosen": -74.44918823242188,
      "logps/ref_rejected": -115.08148193359375,
      "logps/rejected": -116.2177734375,
      "loss": 1.3806,
      "margin_dpo/margin_mean": 0.4094170331954956,
      "margin_dpo/margin_std": 1.2669267654418945,
      "step": 63
    },
    {
      "epoch": 0.09674981103552532,
      "grad_norm": 2.5979647636413574,
      "learning_rate": 4.701492537313433e-07,
      "logits/chosen": 0.0404229536652565,
      "logits/rejected": -0.03692961111664772,
      "logps/chosen": -42.32487869262695,
      "logps/ref_chosen": -41.524444580078125,
      "logps/ref_rejected": -76.30519104003906,
      "logps/rejected": -77.62030029296875,
      "loss": 1.3855,
      "margin_dpo/margin_mean": 0.5146619081497192,
      "margin_dpo/margin_std": 0.8373426198959351,
      "step": 64
    },
    {
      "epoch": 0.0982615268329554,
      "grad_norm": 3.180335760116577,
      "learning_rate": 4.776119402985074e-07,
      "logits/chosen": 0.22089830040931702,
      "logits/rejected": 0.2287295162677765,
      "logps/chosen": -79.51335144042969,
      "logps/ref_chosen": -78.54818725585938,
      "logps/ref_rejected": -58.596473693847656,
      "logps/rejected": -59.96930694580078,
      "loss": 1.3814,
      "margin_dpo/margin_mean": 0.4076697826385498,
      "margin_dpo/margin_std": 0.8789803981781006,
      "step": 65
    },
    {
      "epoch": 0.09977324263038549,
      "grad_norm": 3.0080127716064453,
      "learning_rate": 4.850746268656717e-07,
      "logits/chosen": 0.15983106195926666,
      "logits/rejected": 0.14227060973644257,
      "logps/chosen": -69.20197296142578,
      "logps/ref_chosen": -68.18994903564453,
      "logps/ref_rejected": -83.489013671875,
      "logps/rejected": -85.03985595703125,
      "loss": 1.3826,
      "margin_dpo/margin_mean": 0.5388225317001343,
      "margin_dpo/margin_std": 1.0032914876937866,
      "step": 66
    },
    {
      "epoch": 0.10128495842781557,
      "grad_norm": 2.978635311126709,
      "learning_rate": 4.925373134328357e-07,
      "logits/chosen": 0.13758046925067902,
      "logits/rejected": 0.084464430809021,
      "logps/chosen": -63.52684020996094,
      "logps/ref_chosen": -62.41529083251953,
      "logps/ref_rejected": -53.900169372558594,
      "logps/rejected": -54.98922348022461,
      "loss": 1.3834,
      "margin_dpo/margin_mean": -0.02249324321746826,
      "margin_dpo/margin_std": 1.1306034326553345,
      "step": 67
    },
    {
      "epoch": 0.10279667422524566,
      "grad_norm": 3.2404768466949463,
      "learning_rate": 5e-07,
      "logits/chosen": 0.06684955954551697,
      "logits/rejected": -0.005514336749911308,
      "logps/chosen": -81.26565551757812,
      "logps/ref_chosen": -80.14068603515625,
      "logps/ref_rejected": -116.50318145751953,
      "logps/rejected": -118.35859680175781,
      "loss": 1.3832,
      "margin_dpo/margin_mean": 0.7304507493972778,
      "margin_dpo/margin_std": 1.5464614629745483,
      "step": 68
    },
    {
      "epoch": 0.10430839002267574,
      "grad_norm": 3.3028857707977295,
      "learning_rate": 4.999965034812934e-07,
      "logits/chosen": 0.16428935527801514,
      "logits/rejected": 0.11686080694198608,
      "logps/chosen": -54.37907409667969,
      "logps/ref_chosen": -53.415428161621094,
      "logps/ref_rejected": -71.89765930175781,
      "logps/rejected": -73.68708801269531,
      "loss": 1.3792,
      "margin_dpo/margin_mean": 0.8257811069488525,
      "margin_dpo/margin_std": 0.9166597127914429,
      "step": 69
    },
    {
      "epoch": 0.10582010582010581,
      "grad_norm": 2.9263834953308105,
      "learning_rate": 4.999860140229787e-07,
      "logits/chosen": 0.006922289729118347,
      "logits/rejected": 0.01605862006545067,
      "logps/chosen": -72.73168182373047,
      "logps/ref_chosen": -71.148193359375,
      "logps/ref_rejected": -67.78597259521484,
      "logps/rejected": -69.44383239746094,
      "loss": 1.383,
      "margin_dpo/margin_mean": 0.07435917854309082,
      "margin_dpo/margin_std": 1.4182093143463135,
      "step": 70
    },
    {
      "epoch": 0.1073318216175359,
      "grad_norm": 2.8662917613983154,
      "learning_rate": 4.999685319184688e-07,
      "logits/chosen": 0.1492251604795456,
      "logits/rejected": 0.13701248168945312,
      "logps/chosen": -46.276390075683594,
      "logps/ref_chosen": -44.75934982299805,
      "logps/ref_rejected": -65.2507553100586,
      "logps/rejected": -66.79752349853516,
      "loss": 1.3832,
      "margin_dpo/margin_mean": 0.029730796813964844,
      "margin_dpo/margin_std": 1.3828539848327637,
      "step": 71
    },
    {
      "epoch": 0.10884353741496598,
      "grad_norm": 3.228018045425415,
      "learning_rate": 4.999440576567755e-07,
      "logits/chosen": 0.09348595142364502,
      "logits/rejected": 0.025297988206148148,
      "logps/chosen": -48.47857666015625,
      "logps/ref_chosen": -47.3697395324707,
      "logps/ref_rejected": -52.99060821533203,
      "logps/rejected": -54.79874801635742,
      "loss": 1.3786,
      "margin_dpo/margin_mean": 0.699303388595581,
      "margin_dpo/margin_std": 1.9094823598861694,
      "step": 72
    },
    {
      "epoch": 0.11035525321239607,
      "grad_norm": 3.229674816131592,
      "learning_rate": 4.999125919224965e-07,
      "logits/chosen": 0.10412274301052094,
      "logits/rejected": 0.11492769420146942,
      "logps/chosen": -58.23365783691406,
      "logps/ref_chosen": -56.49576187133789,
      "logps/ref_rejected": -49.6622200012207,
      "logps/rejected": -50.660560607910156,
      "loss": 1.3857,
      "margin_dpo/margin_mean": -0.7395575046539307,
      "margin_dpo/margin_std": 1.3145250082015991,
      "step": 73
    },
    {
      "epoch": 0.11186696900982615,
      "grad_norm": 2.963764190673828,
      "learning_rate": 4.998741355957963e-07,
      "logits/chosen": 0.2109803855419159,
      "logits/rejected": 0.14343787729740143,
      "logps/chosen": -58.918479919433594,
      "logps/ref_chosen": -57.7161750793457,
      "logps/ref_rejected": -114.44114685058594,
      "logps/rejected": -117.02381896972656,
      "loss": 1.3789,
      "margin_dpo/margin_mean": 1.3803461790084839,
      "margin_dpo/margin_std": 1.7276947498321533,
      "step": 74
    },
    {
      "epoch": 0.11337868480725624,
      "grad_norm": 2.831622362136841,
      "learning_rate": 4.998286897523808e-07,
      "logits/chosen": 0.19602981209754944,
      "logits/rejected": 0.17468248307704926,
      "logps/chosen": -41.735137939453125,
      "logps/ref_chosen": -39.99764633178711,
      "logps/ref_rejected": -56.469932556152344,
      "logps/rejected": -58.855743408203125,
      "loss": 1.3781,
      "margin_dpo/margin_mean": 0.6483190059661865,
      "margin_dpo/margin_std": 1.347118616104126,
      "step": 75
    },
    {
      "epoch": 0.11489040060468632,
      "grad_norm": 2.9740262031555176,
      "learning_rate": 4.997762556634679e-07,
      "logits/chosen": -0.004973419010639191,
      "logits/rejected": -0.06459110975265503,
      "logps/chosen": -60.27781677246094,
      "logps/ref_chosen": -57.66736602783203,
      "logps/ref_rejected": -80.56336212158203,
      "logps/rejected": -83.19686889648438,
      "loss": 1.3828,
      "margin_dpo/margin_mean": 0.023052692413330078,
      "margin_dpo/margin_std": 2.44602632522583,
      "step": 76
    },
    {
      "epoch": 0.1164021164021164,
      "grad_norm": 3.190723419189453,
      "learning_rate": 4.99716834795752e-07,
      "logits/chosen": -0.021925870329141617,
      "logits/rejected": -0.0063882917165756226,
      "logps/chosen": -55.694488525390625,
      "logps/ref_chosen": -53.56401824951172,
      "logps/ref_rejected": -48.90995407104492,
      "logps/rejected": -51.173919677734375,
      "loss": 1.3771,
      "margin_dpo/margin_mean": 0.13348984718322754,
      "margin_dpo/margin_std": 1.8197870254516602,
      "step": 77
    },
    {
      "epoch": 0.11791383219954649,
      "grad_norm": 2.9408459663391113,
      "learning_rate": 4.996504288113623e-07,
      "logits/chosen": 0.07129096984863281,
      "logits/rejected": 0.04697857052087784,
      "logps/chosen": -74.45344543457031,
      "logps/ref_chosen": -72.49877166748047,
      "logps/ref_rejected": -85.70948791503906,
      "logps/rejected": -88.24615478515625,
      "loss": 1.3815,
      "margin_dpo/margin_mean": 0.5819820165634155,
      "margin_dpo/margin_std": 1.5051074028015137,
      "step": 78
    },
    {
      "epoch": 0.11942554799697656,
      "grad_norm": 3.75710391998291,
      "learning_rate": 4.995770395678171e-07,
      "logits/chosen": 0.2161446511745453,
      "logits/rejected": 0.2010088711977005,
      "logps/chosen": -57.58828353881836,
      "logps/ref_chosen": -55.347450256347656,
      "logps/ref_rejected": -58.748321533203125,
      "logps/rejected": -61.51164245605469,
      "loss": 1.3723,
      "margin_dpo/margin_mean": 0.522484540939331,
      "margin_dpo/margin_std": 2.684553384780884,
      "step": 79
    },
    {
      "epoch": 0.12093726379440665,
      "grad_norm": 2.9566855430603027,
      "learning_rate": 4.994966691179711e-07,
      "logits/chosen": 0.16256621479988098,
      "logits/rejected": 0.08070458471775055,
      "logps/chosen": -61.3144416809082,
      "logps/ref_chosen": -58.95909118652344,
      "logps/ref_rejected": -62.06755065917969,
      "logps/rejected": -65.10287475585938,
      "loss": 1.3796,
      "margin_dpo/margin_mean": 0.6799747943878174,
      "margin_dpo/margin_std": 2.2472503185272217,
      "step": 80
    },
    {
      "epoch": 0.12244897959183673,
      "grad_norm": 3.1515696048736572,
      "learning_rate": 4.994093197099587e-07,
      "logits/chosen": 0.10985112190246582,
      "logits/rejected": 0.07357762008905411,
      "logps/chosen": -60.51456069946289,
      "logps/ref_chosen": -57.94086456298828,
      "logps/ref_rejected": -64.90560913085938,
      "logps/rejected": -68.10665893554688,
      "loss": 1.3767,
      "margin_dpo/margin_mean": 0.6273548603057861,
      "margin_dpo/margin_std": 2.312195301055908,
      "step": 81
    },
    {
      "epoch": 0.12396069538926682,
      "grad_norm": 2.9304919242858887,
      "learning_rate": 4.993149937871306e-07,
      "logits/chosen": 0.1599178910255432,
      "logits/rejected": 0.06249549984931946,
      "logps/chosen": -27.858848571777344,
      "logps/ref_chosen": -26.338790893554688,
      "logps/ref_rejected": -56.80085754394531,
      "logps/rejected": -59.76110076904297,
      "loss": 1.374,
      "margin_dpo/margin_mean": 1.4401792287826538,
      "margin_dpo/margin_std": 1.2950589656829834,
      "step": 82
    },
    {
      "epoch": 0.1254724111866969,
      "grad_norm": 3.3920578956604004,
      "learning_rate": 4.992136939879856e-07,
      "logits/chosen": 0.07286302000284195,
      "logits/rejected": -0.008843222633004189,
      "logps/chosen": -58.036651611328125,
      "logps/ref_chosen": -55.43024444580078,
      "logps/ref_rejected": -90.0384521484375,
      "logps/rejected": -94.02841186523438,
      "loss": 1.3722,
      "margin_dpo/margin_mean": 1.3835418224334717,
      "margin_dpo/margin_std": 2.0829572677612305,
      "step": 83
    },
    {
      "epoch": 0.12698412698412698,
      "grad_norm": 3.293107271194458,
      "learning_rate": 4.991054231460969e-07,
      "logits/chosen": 0.14062954485416412,
      "logits/rejected": 0.09686338156461716,
      "logps/chosen": -64.60865020751953,
      "logps/ref_chosen": -60.98677444458008,
      "logps/ref_rejected": -61.2043342590332,
      "logps/rejected": -66.28961944580078,
      "loss": 1.3749,
      "margin_dpo/margin_mean": 1.463415503501892,
      "margin_dpo/margin_std": 2.3290019035339355,
      "step": 84
    },
    {
      "epoch": 0.12849584278155707,
      "grad_norm": 3.0449564456939697,
      "learning_rate": 4.989901842900325e-07,
      "logits/chosen": 0.2767505645751953,
      "logits/rejected": 0.2392296940088272,
      "logps/chosen": -58.43796157836914,
      "logps/ref_chosen": -55.2398681640625,
      "logps/ref_rejected": -69.06980895996094,
      "logps/rejected": -73.96517181396484,
      "loss": 1.3711,
      "margin_dpo/margin_mean": 1.6972711086273193,
      "margin_dpo/margin_std": 1.7987269163131714,
      "step": 85
    },
    {
      "epoch": 0.13000755857898716,
      "grad_norm": 3.007859468460083,
      "learning_rate": 4.988679806432711e-07,
      "logits/chosen": 0.16140137612819672,
      "logits/rejected": 0.1533210277557373,
      "logps/chosen": -64.31861114501953,
      "logps/ref_chosen": -60.13792419433594,
      "logps/ref_rejected": -66.80750274658203,
      "logps/rejected": -72.00105285644531,
      "loss": 1.3752,
      "margin_dpo/margin_mean": 1.0128706693649292,
      "margin_dpo/margin_std": 3.6858959197998047,
      "step": 86
    },
    {
      "epoch": 0.13151927437641722,
      "grad_norm": 3.371891975402832,
      "learning_rate": 4.987388156241114e-07,
      "logits/chosen": 0.1272026002407074,
      "logits/rejected": 0.08599858731031418,
      "logps/chosen": -95.75196075439453,
      "logps/ref_chosen": -90.86170959472656,
      "logps/ref_rejected": -81.98894500732422,
      "logps/rejected": -88.12923431396484,
      "loss": 1.3721,
      "margin_dpo/margin_mean": 1.2500503063201904,
      "margin_dpo/margin_std": 4.517343997955322,
      "step": 87
    },
    {
      "epoch": 0.1330309901738473,
      "grad_norm": 3.1865837574005127,
      "learning_rate": 4.986026928455767e-07,
      "logits/chosen": 0.10692833364009857,
      "logits/rejected": 0.08406249433755875,
      "logps/chosen": -47.30534362792969,
      "logps/ref_chosen": -44.642494201660156,
      "logps/ref_rejected": -57.57598114013672,
      "logps/rejected": -62.85344314575195,
      "loss": 1.3767,
      "margin_dpo/margin_mean": 2.6146082878112793,
      "margin_dpo/margin_std": 3.4296021461486816,
      "step": 88
    },
    {
      "epoch": 0.1345427059712774,
      "grad_norm": 3.5698328018188477,
      "learning_rate": 4.984596161153135e-07,
      "logits/chosen": 0.23458774387836456,
      "logits/rejected": 0.11213146895170212,
      "logps/chosen": -47.105751037597656,
      "logps/ref_chosen": -43.498695373535156,
      "logps/ref_rejected": -72.5770034790039,
      "logps/rejected": -78.28662109375,
      "loss": 1.3671,
      "margin_dpo/margin_mean": 2.10256290435791,
      "margin_dpo/margin_std": 2.2028491497039795,
      "step": 89
    },
    {
      "epoch": 0.1360544217687075,
      "grad_norm": 3.3276965618133545,
      "learning_rate": 4.983095894354857e-07,
      "logits/chosen": 0.24859435856342316,
      "logits/rejected": 0.14004702866077423,
      "logps/chosen": -35.86720275878906,
      "logps/ref_chosen": -32.05683517456055,
      "logps/ref_rejected": -76.27911376953125,
      "logps/rejected": -82.3067626953125,
      "loss": 1.3688,
      "margin_dpo/margin_mean": 2.2172858715057373,
      "margin_dpo/margin_std": 4.022238731384277,
      "step": 90
    },
    {
      "epoch": 0.13756613756613756,
      "grad_norm": 2.8259410858154297,
      "learning_rate": 4.98152617002662e-07,
      "logits/chosen": 0.2829027473926544,
      "logits/rejected": 0.1933433711528778,
      "logps/chosen": -42.66006088256836,
      "logps/ref_chosen": -38.95655822753906,
      "logps/ref_rejected": -85.30648803710938,
      "logps/rejected": -92.27479553222656,
      "loss": 1.3726,
      "margin_dpo/margin_mean": 3.264805793762207,
      "margin_dpo/margin_std": 5.691169738769531,
      "step": 91
    },
    {
      "epoch": 0.13907785336356765,
      "grad_norm": 3.3671655654907227,
      "learning_rate": 4.979887032076988e-07,
      "logits/chosen": 0.1655869483947754,
      "logits/rejected": 0.14279474318027496,
      "logps/chosen": -50.55265808105469,
      "logps/ref_chosen": -46.43939208984375,
      "logps/ref_rejected": -56.59052276611328,
      "logps/rejected": -61.2132568359375,
      "loss": 1.3639,
      "margin_dpo/margin_mean": 0.5094633102416992,
      "margin_dpo/margin_std": 2.5982000827789307,
      "step": 92
    },
    {
      "epoch": 0.14058956916099774,
      "grad_norm": 2.7980592250823975,
      "learning_rate": 4.978178526356172e-07,
      "logits/chosen": 0.23437252640724182,
      "logits/rejected": 0.18408940732479095,
      "logps/chosen": -47.498390197753906,
      "logps/ref_chosen": -42.59188461303711,
      "logps/ref_rejected": -66.06549072265625,
      "logps/rejected": -73.67166900634766,
      "loss": 1.3733,
      "margin_dpo/margin_mean": 2.699659824371338,
      "margin_dpo/margin_std": 6.147436141967773,
      "step": 93
    },
    {
      "epoch": 0.1421012849584278,
      "grad_norm": 3.8317432403564453,
      "learning_rate": 4.976400700654751e-07,
      "logits/chosen": 0.23454198241233826,
      "logits/rejected": 0.24940252304077148,
      "logps/chosen": -82.16134643554688,
      "logps/ref_chosen": -77.08745574951172,
      "logps/ref_rejected": -72.75309753417969,
      "logps/rejected": -78.77753448486328,
      "loss": 1.3598,
      "margin_dpo/margin_mean": 0.9505494832992554,
      "margin_dpo/margin_std": 5.345444679260254,
      "step": 94
    },
    {
      "epoch": 0.1436130007558579,
      "grad_norm": 3.494490146636963,
      "learning_rate": 4.974553604702332e-07,
      "logits/chosen": 0.13477879762649536,
      "logits/rejected": 0.0803680494427681,
      "logps/chosen": -63.49205017089844,
      "logps/ref_chosen": -57.205665588378906,
      "logps/ref_rejected": -101.0076904296875,
      "logps/rejected": -108.45451354980469,
      "loss": 1.3646,
      "margin_dpo/margin_mean": 1.1604300737380981,
      "margin_dpo/margin_std": 4.039283752441406,
      "step": 95
    },
    {
      "epoch": 0.14512471655328799,
      "grad_norm": 3.4159934520721436,
      "learning_rate": 4.972637290166157e-07,
      "logits/chosen": 0.27271753549575806,
      "logits/rejected": 0.200592041015625,
      "logps/chosen": -78.8208999633789,
      "logps/ref_chosen": -71.87321472167969,
      "logps/ref_rejected": -108.4333267211914,
      "logps/rejected": -119.43417358398438,
      "loss": 1.3618,
      "margin_dpo/margin_mean": 4.053152084350586,
      "margin_dpo/margin_std": 8.04190731048584,
      "step": 96
    },
    {
      "epoch": 0.14663643235071808,
      "grad_norm": 3.486067056655884,
      "learning_rate": 4.970651810649666e-07,
      "logits/chosen": 0.15996894240379333,
      "logits/rejected": 0.2043858766555786,
      "logps/chosen": -104.0246810913086,
      "logps/ref_chosen": -93.0615463256836,
      "logps/ref_rejected": -54.65837097167969,
      "logps/rejected": -61.65705871582031,
      "loss": 1.3753,
      "margin_dpo/margin_mean": -3.9644370079040527,
      "margin_dpo/margin_std": 11.941303253173828,
      "step": 97
    },
    {
      "epoch": 0.14814814814814814,
      "grad_norm": 2.9447083473205566,
      "learning_rate": 4.968597221690985e-07,
      "logits/chosen": 0.19196490943431854,
      "logits/rejected": 0.1588142365217209,
      "logps/chosen": -72.50187683105469,
      "logps/ref_chosen": -64.01171875,
      "logps/ref_rejected": -64.54795837402344,
      "logps/rejected": -73.04808044433594,
      "loss": 1.3791,
      "margin_dpo/margin_mean": 0.009970307350158691,
      "margin_dpo/margin_std": 4.951154708862305,
      "step": 98
    },
    {
      "epoch": 0.14965986394557823,
      "grad_norm": 3.1145100593566895,
      "learning_rate": 4.966473580761389e-07,
      "logits/chosen": 0.3419758081436157,
      "logits/rejected": 0.2835521697998047,
      "logps/chosen": -59.71137619018555,
      "logps/ref_chosen": -53.621009826660156,
      "logps/ref_rejected": -93.46940612792969,
      "logps/rejected": -103.66671752929688,
      "loss": 1.3786,
      "margin_dpo/margin_mean": 4.106945991516113,
      "margin_dpo/margin_std": 5.849977016448975,
      "step": 99
    },
    {
      "epoch": 0.15117157974300832,
      "grad_norm": 3.7920093536376953,
      "learning_rate": 4.964280947263676e-07,
      "logits/chosen": 0.24154677987098694,
      "logits/rejected": 0.23493698239326477,
      "logps/chosen": -102.99555969238281,
      "logps/ref_chosen": -93.66744995117188,
      "logps/ref_rejected": -110.8492431640625,
      "logps/rejected": -124.81897735595703,
      "loss": 1.3656,
      "margin_dpo/margin_mean": 4.6416401863098145,
      "margin_dpo/margin_std": 7.766916275024414,
      "step": 100
    },
    {
      "epoch": 0.15268329554043839,
      "grad_norm": 3.0823545455932617,
      "learning_rate": 4.96201938253052e-07,
      "logits/chosen": 0.09690912812948227,
      "logits/rejected": 0.05974133312702179,
      "logps/chosen": -71.51644897460938,
      "logps/ref_chosen": -63.71492004394531,
      "logps/ref_rejected": -83.56236267089844,
      "logps/rejected": -93.69319152832031,
      "loss": 1.3574,
      "margin_dpo/margin_mean": 2.32930326461792,
      "margin_dpo/margin_std": 8.121504783630371,
      "step": 101
    },
    {
      "epoch": 0.15419501133786848,
      "grad_norm": 3.003969430923462,
      "learning_rate": 4.959688949822748e-07,
      "logits/chosen": 0.11170051246881485,
      "logits/rejected": 0.12498050928115845,
      "logps/chosen": -73.76181030273438,
      "logps/ref_chosen": -65.03767395019531,
      "logps/ref_rejected": -65.62431335449219,
      "logps/rejected": -73.43655395507812,
      "loss": 1.381,
      "margin_dpo/margin_mean": -0.9118953943252563,
      "margin_dpo/margin_std": 7.858981132507324,
      "step": 102
    },
    {
      "epoch": 0.15570672713529857,
      "grad_norm": 3.363142251968384,
      "learning_rate": 4.957289714327572e-07,
      "logits/chosen": 0.28459155559539795,
      "logits/rejected": 0.2856888771057129,
      "logps/chosen": -62.49891662597656,
      "logps/ref_chosen": -54.93040466308594,
      "logps/ref_rejected": -56.902076721191406,
      "logps/rejected": -65.07192993164062,
      "loss": 1.3643,
      "margin_dpo/margin_mean": 0.6013284921646118,
      "margin_dpo/margin_std": 7.1304521560668945,
      "step": 103
    },
    {
      "epoch": 0.15721844293272866,
      "grad_norm": 4.440778732299805,
      "learning_rate": 4.954821743156767e-07,
      "logits/chosen": 0.28602921962738037,
      "logits/rejected": 0.20815476775169373,
      "logps/chosen": -76.05664825439453,
      "logps/ref_chosen": -66.83149719238281,
      "logps/ref_rejected": -114.09371185302734,
      "logps/rejected": -128.50567626953125,
      "loss": 1.3449,
      "margin_dpo/margin_mean": 5.186819553375244,
      "margin_dpo/margin_std": 18.05282974243164,
      "step": 104
    },
    {
      "epoch": 0.15873015873015872,
      "grad_norm": 3.9823782444000244,
      "learning_rate": 4.952285105344791e-07,
      "logits/chosen": 0.2622066140174866,
      "logits/rejected": 0.18628236651420593,
      "logps/chosen": -52.09508514404297,
      "logps/ref_chosen": -46.05514144897461,
      "logps/ref_rejected": -78.72990417480469,
      "logps/rejected": -89.09690856933594,
      "loss": 1.3504,
      "margin_dpo/margin_mean": 4.327066898345947,
      "margin_dpo/margin_std": 8.973100662231445,
      "step": 105
    },
    {
      "epoch": 0.1602418745275888,
      "grad_norm": 3.0525312423706055,
      "learning_rate": 4.949679871846857e-07,
      "logits/chosen": 0.2560195326805115,
      "logits/rejected": 0.21072283387184143,
      "logps/chosen": -62.558570861816406,
      "logps/ref_chosen": -53.9271240234375,
      "logps/ref_rejected": -66.69943237304688,
      "logps/rejected": -78.46056365966797,
      "loss": 1.3676,
      "margin_dpo/margin_mean": 3.129687786102295,
      "margin_dpo/margin_std": 10.326120376586914,
      "step": 106
    },
    {
      "epoch": 0.1617535903250189,
      "grad_norm": 3.6129863262176514,
      "learning_rate": 4.947006115536947e-07,
      "logits/chosen": 0.20093579590320587,
      "logits/rejected": 0.26784881949424744,
      "logps/chosen": -103.26329040527344,
      "logps/ref_chosen": -87.38455200195312,
      "logps/ref_rejected": -76.76750183105469,
      "logps/rejected": -89.45541381835938,
      "loss": 1.3811,
      "margin_dpo/margin_mean": -3.1908369064331055,
      "margin_dpo/margin_std": 17.2308292388916,
      "step": 107
    },
    {
      "epoch": 0.16326530612244897,
      "grad_norm": 3.4911181926727295,
      "learning_rate": 4.944263911205772e-07,
      "logits/chosen": 0.2522239685058594,
      "logits/rejected": 0.1837356686592102,
      "logps/chosen": -72.13520812988281,
      "logps/ref_chosen": -62.109657287597656,
      "logps/ref_rejected": -91.77952575683594,
      "logps/rejected": -107.07212829589844,
      "loss": 1.3568,
      "margin_dpo/margin_mean": 5.267055034637451,
      "margin_dpo/margin_std": 8.78154182434082,
      "step": 108
    },
    {
      "epoch": 0.16477702191987906,
      "grad_norm": 4.055303573608398,
      "learning_rate": 4.941453335558681e-07,
      "logits/chosen": 0.3957793116569519,
      "logits/rejected": 0.35673829913139343,
      "logps/chosen": -84.13023376464844,
      "logps/ref_chosen": -70.22321319580078,
      "logps/ref_rejected": -103.89926147460938,
      "logps/rejected": -124.2558364868164,
      "loss": 1.3295,
      "margin_dpo/margin_mean": 6.449550151824951,
      "margin_dpo/margin_std": 12.290249824523926,
      "step": 109
    },
    {
      "epoch": 0.16628873771730915,
      "grad_norm": 4.451986312866211,
      "learning_rate": 4.938574467213517e-07,
      "logits/chosen": 0.2378203570842743,
      "logits/rejected": 0.25508666038513184,
      "logps/chosen": -85.72879028320312,
      "logps/ref_chosen": -72.66490173339844,
      "logps/ref_rejected": -62.63153076171875,
      "logps/rejected": -74.57170867919922,
      "loss": 1.4061,
      "margin_dpo/margin_mean": -1.1237014532089233,
      "margin_dpo/margin_std": 12.696800231933594,
      "step": 110
    },
    {
      "epoch": 0.16780045351473924,
      "grad_norm": 3.417356014251709,
      "learning_rate": 4.935627386698418e-07,
      "logits/chosen": 0.26538121700286865,
      "logits/rejected": 0.27719810605049133,
      "logps/chosen": -81.70519256591797,
      "logps/ref_chosen": -67.52632141113281,
      "logps/ref_rejected": -73.84911346435547,
      "logps/rejected": -89.34820556640625,
      "loss": 1.3481,
      "margin_dpo/margin_mean": 1.3202223777770996,
      "margin_dpo/margin_std": 9.906134605407715,
      "step": 111
    },
    {
      "epoch": 0.1693121693121693,
      "grad_norm": 4.943251132965088,
      "learning_rate": 4.932612176449559e-07,
      "logits/chosen": 0.19840675592422485,
      "logits/rejected": 0.1570388376712799,
      "logps/chosen": -72.82176971435547,
      "logps/ref_chosen": -61.945491790771484,
      "logps/ref_rejected": -113.05886840820312,
      "logps/rejected": -132.51878356933594,
      "loss": 1.3318,
      "margin_dpo/margin_mean": 8.58362865447998,
      "margin_dpo/margin_std": 13.595785140991211,
      "step": 112
    },
    {
      "epoch": 0.1708238851095994,
      "grad_norm": 3.653705358505249,
      "learning_rate": 4.929528920808854e-07,
      "logits/chosen": 0.18068620562553406,
      "logits/rejected": 0.22447502613067627,
      "logps/chosen": -83.29192352294922,
      "logps/ref_chosen": -68.68717193603516,
      "logps/ref_rejected": -49.824913024902344,
      "logps/rejected": -59.991519927978516,
      "loss": 1.3697,
      "margin_dpo/margin_mean": -4.438149452209473,
      "margin_dpo/margin_std": 8.201181411743164,
      "step": 113
    },
    {
      "epoch": 0.17233560090702948,
      "grad_norm": 3.5117268562316895,
      "learning_rate": 4.92637770602159e-07,
      "logits/chosen": 0.37020614743232727,
      "logits/rejected": 0.3371107578277588,
      "logps/chosen": -83.77205657958984,
      "logps/ref_chosen": -68.70469665527344,
      "logps/ref_rejected": -66.02621459960938,
      "logps/rejected": -80.38362121582031,
      "loss": 1.3549,
      "margin_dpo/margin_mean": -0.7099564075469971,
      "margin_dpo/margin_std": 13.317527770996094,
      "step": 114
    },
    {
      "epoch": 0.17384731670445955,
      "grad_norm": 3.4344635009765625,
      "learning_rate": 4.923158620234019e-07,
      "logits/chosen": 0.5059663653373718,
      "logits/rejected": 0.3934090733528137,
      "logps/chosen": -78.51239776611328,
      "logps/ref_chosen": -64.35462951660156,
      "logps/ref_rejected": -91.54823303222656,
      "logps/rejected": -109.61679077148438,
      "loss": 1.3504,
      "margin_dpo/margin_mean": 3.910794496536255,
      "margin_dpo/margin_std": 20.801626205444336,
      "step": 115
    },
    {
      "epoch": 0.17535903250188964,
      "grad_norm": 4.414751052856445,
      "learning_rate": 4.91987175349089e-07,
      "logits/chosen": 0.35310861468315125,
      "logits/rejected": 0.28566908836364746,
      "logps/chosen": -62.23815155029297,
      "logps/ref_chosen": -49.481990814208984,
      "logps/ref_rejected": -81.90135192871094,
      "logps/rejected": -104.6424560546875,
      "loss": 1.3141,
      "margin_dpo/margin_mean": 9.984952926635742,
      "margin_dpo/margin_std": 16.990009307861328,
      "step": 116
    },
    {
      "epoch": 0.17687074829931973,
      "grad_norm": 3.1478612422943115,
      "learning_rate": 4.916517197732933e-07,
      "logits/chosen": 0.4462072551250458,
      "logits/rejected": 0.3813575208187103,
      "logps/chosen": -64.82511901855469,
      "logps/ref_chosen": -52.36582946777344,
      "logps/ref_rejected": -76.98123931884766,
      "logps/rejected": -93.8857192993164,
      "loss": 1.3637,
      "margin_dpo/margin_mean": 4.445184707641602,
      "margin_dpo/margin_std": 15.84908676147461,
      "step": 117
    },
    {
      "epoch": 0.17838246409674982,
      "grad_norm": 4.316784381866455,
      "learning_rate": 4.913095046794281e-07,
      "logits/chosen": 0.33763909339904785,
      "logits/rejected": 0.33866265416145325,
      "logps/chosen": -76.38975524902344,
      "logps/ref_chosen": -59.73781204223633,
      "logps/ref_rejected": -71.17877197265625,
      "logps/rejected": -88.17243194580078,
      "loss": 1.3327,
      "margin_dpo/margin_mean": 0.3417208194732666,
      "margin_dpo/margin_std": 10.12466049194336,
      "step": 118
    },
    {
      "epoch": 0.17989417989417988,
      "grad_norm": 3.382061719894409,
      "learning_rate": 4.909605396399855e-07,
      "logits/chosen": 0.42505258321762085,
      "logits/rejected": 0.4622851014137268,
      "logps/chosen": -75.50498962402344,
      "logps/ref_chosen": -57.297969818115234,
      "logps/ref_rejected": -43.269874572753906,
      "logps/rejected": -58.09909439086914,
      "loss": 1.3609,
      "margin_dpo/margin_mean": -3.377796173095703,
      "margin_dpo/margin_std": 17.354841232299805,
      "step": 119
    },
    {
      "epoch": 0.18140589569160998,
      "grad_norm": 4.64827823638916,
      "learning_rate": 4.906048344162676e-07,
      "logits/chosen": 0.3838917016983032,
      "logits/rejected": 0.2518424391746521,
      "logps/chosen": -73.36843872070312,
      "logps/ref_chosen": -59.898441314697266,
      "logps/ref_rejected": -103.18621826171875,
      "logps/rejected": -131.10989379882812,
      "loss": 1.3166,
      "margin_dpo/margin_mean": 14.453689575195312,
      "margin_dpo/margin_std": 17.748037338256836,
      "step": 120
    },
    {
      "epoch": 0.18291761148904007,
      "grad_norm": 5.509910583496094,
      "learning_rate": 4.902423989581143e-07,
      "logits/chosen": 0.4782486855983734,
      "logits/rejected": 0.3805123567581177,
      "logps/chosen": -88.04644775390625,
      "logps/ref_chosen": -67.99308776855469,
      "logps/ref_rejected": -128.11260986328125,
      "logps/rejected": -163.5806427001953,
      "loss": 1.2951,
      "margin_dpo/margin_mean": 15.414678573608398,
      "margin_dpo/margin_std": 29.943111419677734,
      "step": 121
    },
    {
      "epoch": 0.18442932728647016,
      "grad_norm": 3.7154958248138428,
      "learning_rate": 4.898732434036243e-07,
      "logits/chosen": 0.4762170612812042,
      "logits/rejected": 0.41412898898124695,
      "logps/chosen": -74.78926086425781,
      "logps/ref_chosen": -57.45248031616211,
      "logps/ref_rejected": -94.9098892211914,
      "logps/rejected": -132.783447265625,
      "loss": 1.3373,
      "margin_dpo/margin_mean": 20.53677749633789,
      "margin_dpo/margin_std": 39.82550811767578,
      "step": 122
    },
    {
      "epoch": 0.18594104308390022,
      "grad_norm": 3.9479894638061523,
      "learning_rate": 4.894973780788722e-07,
      "logits/chosen": 0.4904557466506958,
      "logits/rejected": 0.3759152293205261,
      "logps/chosen": -68.86778259277344,
      "logps/ref_chosen": -51.397483825683594,
      "logps/ref_rejected": -106.05813598632812,
      "logps/rejected": -142.84266662597656,
      "loss": 1.3303,
      "margin_dpo/margin_mean": 19.314233779907227,
      "margin_dpo/margin_std": 27.63502311706543,
      "step": 123
    },
    {
      "epoch": 0.1874527588813303,
      "grad_norm": 4.315618515014648,
      "learning_rate": 4.89114813497619e-07,
      "logits/chosen": 0.5175679922103882,
      "logits/rejected": 0.3383052349090576,
      "logps/chosen": -46.36642837524414,
      "logps/ref_chosen": -35.62065887451172,
      "logps/ref_rejected": -100.73350524902344,
      "logps/rejected": -135.20120239257812,
      "loss": 1.3198,
      "margin_dpo/margin_mean": 23.721927642822266,
      "margin_dpo/margin_std": 20.40395736694336,
      "step": 124
    },
    {
      "epoch": 0.1889644746787604,
      "grad_norm": 4.568764686584473,
      "learning_rate": 4.887255603610184e-07,
      "logits/chosen": 0.5989629030227661,
      "logits/rejected": 0.4955187439918518,
      "logps/chosen": -93.56169891357422,
      "logps/ref_chosen": -66.91831970214844,
      "logps/ref_rejected": -116.68222045898438,
      "logps/rejected": -157.46566772460938,
      "loss": 1.313,
      "margin_dpo/margin_mean": 14.140068054199219,
      "margin_dpo/margin_std": 47.2589225769043,
      "step": 125
    },
    {
      "epoch": 0.19047619047619047,
      "grad_norm": 4.645914077758789,
      "learning_rate": 4.883296295573176e-07,
      "logits/chosen": 0.3819458484649658,
      "logits/rejected": 0.3799610137939453,
      "logps/chosen": -78.845703125,
      "logps/ref_chosen": -60.64362335205078,
      "logps/ref_rejected": -54.69144821166992,
      "logps/rejected": -76.06550598144531,
      "loss": 1.3929,
      "margin_dpo/margin_mean": 3.171980381011963,
      "margin_dpo/margin_std": 13.540816307067871,
      "step": 126
    },
    {
      "epoch": 0.19198790627362056,
      "grad_norm": 3.9225544929504395,
      "learning_rate": 4.87927032161552e-07,
      "logits/chosen": 0.5223321914672852,
      "logits/rejected": 0.4658096730709076,
      "logps/chosen": -100.17329406738281,
      "logps/ref_chosen": -73.2366943359375,
      "logps/ref_rejected": -83.54694366455078,
      "logps/rejected": -119.79974365234375,
      "loss": 1.3401,
      "margin_dpo/margin_mean": 9.31619930267334,
      "margin_dpo/margin_std": 25.015779495239258,
      "step": 127
    },
    {
      "epoch": 0.19349962207105065,
      "grad_norm": 4.997095584869385,
      "learning_rate": 4.875177794352363e-07,
      "logits/chosen": 0.4383442997932434,
      "logits/rejected": 0.4075517952442169,
      "logps/chosen": -108.09526062011719,
      "logps/ref_chosen": -81.03706359863281,
      "logps/ref_rejected": -95.60237121582031,
      "logps/rejected": -140.10910034179688,
      "loss": 1.2741,
      "margin_dpo/margin_mean": 17.448543548583984,
      "margin_dpo/margin_std": 43.10020446777344,
      "step": 128
    },
    {
      "epoch": 0.19501133786848074,
      "grad_norm": 5.618246078491211,
      "learning_rate": 4.871018828260491e-07,
      "logits/chosen": 0.44885939359664917,
      "logits/rejected": 0.41066890954971313,
      "logps/chosen": -70.06056213378906,
      "logps/ref_chosen": -46.31350326538086,
      "logps/ref_rejected": -68.42625427246094,
      "logps/rejected": -98.83761596679688,
      "loss": 1.3926,
      "margin_dpo/margin_mean": 6.664305686950684,
      "margin_dpo/margin_std": 23.976375579833984,
      "step": 129
    },
    {
      "epoch": 0.1965230536659108,
      "grad_norm": 5.429229259490967,
      "learning_rate": 4.866793539675126e-07,
      "logits/chosen": 0.4905567169189453,
      "logits/rejected": 0.4603291153907776,
      "logps/chosen": -101.65276336669922,
      "logps/ref_chosen": -65.72907257080078,
      "logps/ref_rejected": -97.77519226074219,
      "logps/rejected": -144.22227478027344,
      "loss": 1.2914,
      "margin_dpo/margin_mean": 10.523391723632812,
      "margin_dpo/margin_std": 27.65914535522461,
      "step": 130
    },
    {
      "epoch": 0.1980347694633409,
      "grad_norm": 7.263116359710693,
      "learning_rate": 4.86250204678667e-07,
      "logits/chosen": 0.46883371472358704,
      "logits/rejected": 0.37517213821411133,
      "logps/chosen": -45.65996551513672,
      "logps/ref_chosen": -30.02720069885254,
      "logps/ref_rejected": -52.36793518066406,
      "logps/rejected": -78.58646392822266,
      "loss": 1.2509,
      "margin_dpo/margin_mean": 10.585772514343262,
      "margin_dpo/margin_std": 22.57959747314453,
      "step": 131
    },
    {
      "epoch": 0.19954648526077098,
      "grad_norm": 4.200234413146973,
      "learning_rate": 4.858144469637408e-07,
      "logits/chosen": 0.5417824387550354,
      "logits/rejected": 0.5621850490570068,
      "logps/chosen": -76.25199127197266,
      "logps/ref_chosen": -51.91057586669922,
      "logps/ref_rejected": -42.87281799316406,
      "logps/rejected": -67.77296447753906,
      "loss": 1.3117,
      "margin_dpo/margin_mean": 0.5587306022644043,
      "margin_dpo/margin_std": 15.402095794677734,
      "step": 132
    },
    {
      "epoch": 0.20105820105820105,
      "grad_norm": 6.736537933349609,
      "learning_rate": 4.853720930118138e-07,
      "logits/chosen": 0.5202792882919312,
      "logits/rejected": 0.49684813618659973,
      "logps/chosen": -79.01284790039062,
      "logps/ref_chosen": -51.23572540283203,
      "logps/ref_rejected": -57.740684509277344,
      "logps/rejected": -93.11215209960938,
      "loss": 1.3884,
      "margin_dpo/margin_mean": 7.594354152679443,
      "margin_dpo/margin_std": 35.9656867980957,
      "step": 133
    },
    {
      "epoch": 0.20256991685563114,
      "grad_norm": 4.357104301452637,
      "learning_rate": 4.849231551964771e-07,
      "logits/chosen": 0.6357393264770508,
      "logits/rejected": 0.6053987741470337,
      "logps/chosen": -93.45689392089844,
      "logps/ref_chosen": -60.42839813232422,
      "logps/ref_rejected": -74.99949645996094,
      "logps/rejected": -121.59634399414062,
      "loss": 1.281,
      "margin_dpo/margin_mean": 13.568346977233887,
      "margin_dpo/margin_std": 36.36805725097656,
      "step": 134
    },
    {
      "epoch": 0.20408163265306123,
      "grad_norm": 4.28125524520874,
      "learning_rate": 4.844676460754862e-07,
      "logits/chosen": 0.5447328090667725,
      "logits/rejected": 0.5454561710357666,
      "logps/chosen": -93.66383361816406,
      "logps/ref_chosen": -57.496219635009766,
      "logps/ref_rejected": -65.18019104003906,
      "logps/rejected": -101.57806396484375,
      "loss": 1.324,
      "margin_dpo/margin_mean": 0.23026132583618164,
      "margin_dpo/margin_std": 23.302963256835938,
      "step": 135
    },
    {
      "epoch": 0.20559334845049132,
      "grad_norm": 5.157009124755859,
      "learning_rate": 4.840055783904106e-07,
      "logits/chosen": 0.4646771550178528,
      "logits/rejected": 0.49990448355674744,
      "logps/chosen": -127.57366180419922,
      "logps/ref_chosen": -86.13214111328125,
      "logps/ref_rejected": -57.09363555908203,
      "logps/rejected": -86.46631622314453,
      "loss": 1.2754,
      "margin_dpo/margin_mean": -12.068828582763672,
      "margin_dpo/margin_std": 32.699989318847656,
      "step": 136
    },
    {
      "epoch": 0.20710506424792138,
      "grad_norm": 5.023794174194336,
      "learning_rate": 4.835369650662767e-07,
      "logits/chosen": 0.4657016396522522,
      "logits/rejected": 0.41886240243911743,
      "logps/chosen": -117.60987091064453,
      "logps/ref_chosen": -85.38725280761719,
      "logps/ref_rejected": -74.20018005371094,
      "logps/rejected": -122.04615020751953,
      "loss": 1.321,
      "margin_dpo/margin_mean": 15.623347282409668,
      "margin_dpo/margin_std": 53.613494873046875,
      "step": 137
    },
    {
      "epoch": 0.20861678004535147,
      "grad_norm": 6.945674896240234,
      "learning_rate": 4.830618192112065e-07,
      "logits/chosen": 0.47126084566116333,
      "logits/rejected": 0.48690515756607056,
      "logps/chosen": -163.68258666992188,
      "logps/ref_chosen": -91.57633972167969,
      "logps/ref_rejected": -72.2829360961914,
      "logps/rejected": -130.02001953125,
      "loss": 1.3501,
      "margin_dpo/margin_mean": -14.36915397644043,
      "margin_dpo/margin_std": 60.06218338012695,
      "step": 138
    },
    {
      "epoch": 0.21012849584278157,
      "grad_norm": 6.725894927978516,
      "learning_rate": 4.825801541160509e-07,
      "logits/chosen": 0.4866417348384857,
      "logits/rejected": 0.49573272466659546,
      "logps/chosen": -97.94691467285156,
      "logps/ref_chosen": -61.32575225830078,
      "logps/ref_rejected": -58.40146255493164,
      "logps/rejected": -101.65697479248047,
      "loss": 1.3529,
      "margin_dpo/margin_mean": 6.634352684020996,
      "margin_dpo/margin_std": 25.386695861816406,
      "step": 139
    },
    {
      "epoch": 0.21164021164021163,
      "grad_norm": 6.632240295410156,
      "learning_rate": 4.820919832540181e-07,
      "logits/chosen": 0.4285508096218109,
      "logits/rejected": 0.4244239926338196,
      "logps/chosen": -144.7003631591797,
      "logps/ref_chosen": -90.57447814941406,
      "logps/ref_rejected": -99.98605346679688,
      "logps/rejected": -158.67343139648438,
      "loss": 1.3232,
      "margin_dpo/margin_mean": 4.56149959564209,
      "margin_dpo/margin_std": 71.53721618652344,
      "step": 140
    },
    {
      "epoch": 0.21315192743764172,
      "grad_norm": 5.187021255493164,
      "learning_rate": 4.815973202802966e-07,
      "logits/chosen": 0.5863088965415955,
      "logits/rejected": 0.5247952938079834,
      "logps/chosen": -98.68556213378906,
      "logps/ref_chosen": -48.589813232421875,
      "logps/ref_rejected": -86.63417053222656,
      "logps/rejected": -149.1506805419922,
      "loss": 1.299,
      "margin_dpo/margin_mean": 12.420760154724121,
      "margin_dpo/margin_std": 37.57579040527344,
      "step": 141
    },
    {
      "epoch": 0.2146636432350718,
      "grad_norm": 4.731565475463867,
      "learning_rate": 4.810961790316729e-07,
      "logits/chosen": 0.5422487258911133,
      "logits/rejected": 0.5291632413864136,
      "logps/chosen": -82.97702026367188,
      "logps/ref_chosen": -52.91022491455078,
      "logps/ref_rejected": -69.94438934326172,
      "logps/rejected": -105.55659484863281,
      "loss": 1.3097,
      "margin_dpo/margin_mean": 5.545400142669678,
      "margin_dpo/margin_std": 24.14703941345215,
      "step": 142
    },
    {
      "epoch": 0.2161753590325019,
      "grad_norm": 7.777275085449219,
      "learning_rate": 4.805885735261454e-07,
      "logits/chosen": 0.6047331690788269,
      "logits/rejected": 0.5495315194129944,
      "logps/chosen": -65.96603393554688,
      "logps/ref_chosen": -41.020355224609375,
      "logps/ref_rejected": -76.39324951171875,
      "logps/rejected": -121.12960815429688,
      "loss": 1.399,
      "margin_dpo/margin_mean": 19.79067611694336,
      "margin_dpo/margin_std": 37.09278106689453,
      "step": 143
    },
    {
      "epoch": 0.21768707482993196,
      "grad_norm": 7.344333171844482,
      "learning_rate": 4.800745179625307e-07,
      "logits/chosen": 0.5492057204246521,
      "logits/rejected": 0.5295698046684265,
      "logps/chosen": -72.70046997070312,
      "logps/ref_chosen": -42.882530212402344,
      "logps/ref_rejected": -55.65336608886719,
      "logps/rejected": -92.50035095214844,
      "loss": 1.3536,
      "margin_dpo/margin_mean": 7.029045104980469,
      "margin_dpo/margin_std": 13.494336128234863,
      "step": 144
    },
    {
      "epoch": 0.21919879062736206,
      "grad_norm": 13.654488563537598,
      "learning_rate": 4.795540267200686e-07,
      "logits/chosen": 0.3647320866584778,
      "logits/rejected": 0.3844793438911438,
      "logps/chosen": -107.22215270996094,
      "logps/ref_chosen": -64.46841430664062,
      "logps/ref_rejected": -74.22611236572266,
      "logps/rejected": -127.66377258300781,
      "loss": 1.4044,
      "margin_dpo/margin_mean": 10.683923721313477,
      "margin_dpo/margin_std": 56.14466857910156,
      "step": 145
    },
    {
      "epoch": 0.22071050642479215,
      "grad_norm": 5.76055908203125,
      "learning_rate": 4.790271143580173e-07,
      "logits/chosen": 0.5389979481697083,
      "logits/rejected": 0.5553423166275024,
      "logps/chosen": -108.01651000976562,
      "logps/ref_chosen": -66.7105712890625,
      "logps/ref_rejected": -65.54419708251953,
      "logps/rejected": -108.33381652832031,
      "loss": 1.3497,
      "margin_dpo/margin_mean": 1.4836831092834473,
      "margin_dpo/margin_std": 34.92258071899414,
      "step": 146
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": 5.631329536437988,
      "learning_rate": 4.784937956152489e-07,
      "logits/chosen": 0.42997854948043823,
      "logits/rejected": 0.3938245177268982,
      "logps/chosen": -113.09949493408203,
      "logps/ref_chosen": -69.16227722167969,
      "logps/ref_rejected": -93.64544677734375,
      "logps/rejected": -146.63816833496094,
      "loss": 1.3586,
      "margin_dpo/margin_mean": 9.055498123168945,
      "margin_dpo/margin_std": 28.972103118896484,
      "step": 147
    },
    {
      "epoch": 0.2237339380196523,
      "grad_norm": 5.31839656829834,
      "learning_rate": 4.779540854098347e-07,
      "logits/chosen": 0.7581343650817871,
      "logits/rejected": 0.6984580755233765,
      "logps/chosen": -70.00361633300781,
      "logps/ref_chosen": -43.368194580078125,
      "logps/ref_rejected": -76.12153625488281,
      "logps/rejected": -122.95445251464844,
      "loss": 1.2599,
      "margin_dpo/margin_mean": 20.197490692138672,
      "margin_dpo/margin_std": 35.63082504272461,
      "step": 148
    },
    {
      "epoch": 0.2252456538170824,
      "grad_norm": 4.855349063873291,
      "learning_rate": 4.774079988386296e-07,
      "logits/chosen": 0.4739559292793274,
      "logits/rejected": 0.5105680227279663,
      "logps/chosen": -98.31448364257812,
      "logps/ref_chosen": -59.394657135009766,
      "logps/ref_rejected": -50.62220001220703,
      "logps/rejected": -90.08949279785156,
      "loss": 1.3099,
      "margin_dpo/margin_mean": 0.5474551916122437,
      "margin_dpo/margin_std": 29.59663963317871,
      "step": 149
    },
    {
      "epoch": 0.22675736961451248,
      "grad_norm": 4.553208827972412,
      "learning_rate": 4.768555511768486e-07,
      "logits/chosen": 0.41769298911094666,
      "logits/rejected": 0.42729195952415466,
      "logps/chosen": -122.25434875488281,
      "logps/ref_chosen": -75.73005676269531,
      "logps/ref_rejected": -88.11376190185547,
      "logps/rejected": -137.2162322998047,
      "loss": 1.277,
      "margin_dpo/margin_mean": 2.578186511993408,
      "margin_dpo/margin_std": 40.36921691894531,
      "step": 150
    },
    {
      "epoch": 0.22826908541194255,
      "grad_norm": 4.969564437866211,
      "learning_rate": 4.762967578776406e-07,
      "logits/chosen": 0.528440535068512,
      "logits/rejected": 0.4810951352119446,
      "logps/chosen": -94.2465591430664,
      "logps/ref_chosen": -64.3418960571289,
      "logps/ref_rejected": -81.62193298339844,
      "logps/rejected": -135.70932006835938,
      "loss": 1.2305,
      "margin_dpo/margin_mean": 24.182723999023438,
      "margin_dpo/margin_std": 51.8789176940918,
      "step": 151
    },
    {
      "epoch": 0.22978080120937264,
      "grad_norm": 4.437036037445068,
      "learning_rate": 4.757316345716553e-07,
      "logits/chosen": 0.4535418152809143,
      "logits/rejected": 0.4450215995311737,
      "logps/chosen": -111.42601013183594,
      "logps/ref_chosen": -71.10409545898438,
      "logps/ref_rejected": -85.16281127929688,
      "logps/rejected": -132.8240203857422,
      "loss": 1.2759,
      "margin_dpo/margin_mean": 7.339301586151123,
      "margin_dpo/margin_std": 32.49983596801758,
      "step": 152
    },
    {
      "epoch": 0.23129251700680273,
      "grad_norm": 4.642518043518066,
      "learning_rate": 4.751601970666064e-07,
      "logits/chosen": 0.30414754152297974,
      "logits/rejected": 0.25842684507369995,
      "logps/chosen": -119.28223419189453,
      "logps/ref_chosen": -72.91175842285156,
      "logps/ref_rejected": -77.55793762207031,
      "logps/rejected": -136.66567993164062,
      "loss": 1.2979,
      "margin_dpo/margin_mean": 12.737245559692383,
      "margin_dpo/margin_std": 24.552635192871094,
      "step": 153
    },
    {
      "epoch": 0.2328042328042328,
      "grad_norm": 4.9511237144470215,
      "learning_rate": 4.745824613468292e-07,
      "logits/chosen": 0.35246241092681885,
      "logits/rejected": 0.5255336761474609,
      "logps/chosen": -121.34557342529297,
      "logps/ref_chosen": -75.31983947753906,
      "logps/ref_rejected": -54.22064208984375,
      "logps/rejected": -90.056884765625,
      "loss": 1.3369,
      "margin_dpo/margin_mean": -10.189491271972656,
      "margin_dpo/margin_std": 41.34107971191406,
      "step": 154
    },
    {
      "epoch": 0.23431594860166288,
      "grad_norm": 5.262056350708008,
      "learning_rate": 4.7399844357283393e-07,
      "logits/chosen": 0.5434817671775818,
      "logits/rejected": 0.46925586462020874,
      "logps/chosen": -76.32972717285156,
      "logps/ref_chosen": -44.81412124633789,
      "logps/ref_rejected": -76.29104614257812,
      "logps/rejected": -128.21963500976562,
      "loss": 1.3033,
      "margin_dpo/margin_mean": 20.412994384765625,
      "margin_dpo/margin_std": 27.084177017211914,
      "step": 155
    },
    {
      "epoch": 0.23582766439909297,
      "grad_norm": 5.0457353591918945,
      "learning_rate": 4.7340816008085305e-07,
      "logits/chosen": 0.44192785024642944,
      "logits/rejected": 0.4855366349220276,
      "logps/chosen": -137.42698669433594,
      "logps/ref_chosen": -94.79901123046875,
      "logps/ref_rejected": -81.31282806396484,
      "logps/rejected": -137.68701171875,
      "loss": 1.2281,
      "margin_dpo/margin_mean": 13.746213912963867,
      "margin_dpo/margin_std": 34.27767562866211,
      "step": 156
    },
    {
      "epoch": 0.23733938019652306,
      "grad_norm": 4.6686692237854,
      "learning_rate": 4.728116273823847e-07,
      "logits/chosen": 0.4652218222618103,
      "logits/rejected": 0.45060834288597107,
      "logps/chosen": -98.19709777832031,
      "logps/ref_chosen": -58.356666564941406,
      "logps/ref_rejected": -77.88996887207031,
      "logps/rejected": -127.36456298828125,
      "loss": 1.2929,
      "margin_dpo/margin_mean": 9.634150505065918,
      "margin_dpo/margin_std": 30.75609588623047,
      "step": 157
    },
    {
      "epoch": 0.23885109599395313,
      "grad_norm": 4.9093017578125,
      "learning_rate": 4.7220886216373085e-07,
      "logits/chosen": 0.4573588967323303,
      "logits/rejected": 0.458060622215271,
      "logps/chosen": -108.16549682617188,
      "logps/ref_chosen": -66.18313598632812,
      "logps/ref_rejected": -68.98072814941406,
      "logps/rejected": -120.69749450683594,
      "loss": 1.2881,
      "margin_dpo/margin_mean": 9.734394073486328,
      "margin_dpo/margin_std": 32.37115478515625,
      "step": 158
    },
    {
      "epoch": 0.24036281179138322,
      "grad_norm": 4.744980812072754,
      "learning_rate": 4.715998812855304e-07,
      "logits/chosen": 0.6077166199684143,
      "logits/rejected": 0.550294816493988,
      "logps/chosen": -97.12281799316406,
      "logps/ref_chosen": -63.03964614868164,
      "logps/ref_rejected": -94.90765380859375,
      "logps/rejected": -150.78564453125,
      "loss": 1.2671,
      "margin_dpo/margin_mean": 21.7947998046875,
      "margin_dpo/margin_std": 43.32719039916992,
      "step": 159
    },
    {
      "epoch": 0.2418745275888133,
      "grad_norm": 5.374088287353516,
      "learning_rate": 4.7098470178228755e-07,
      "logits/chosen": 0.3768664002418518,
      "logits/rejected": 0.35334479808807373,
      "logps/chosen": -82.2614517211914,
      "logps/ref_chosen": -53.58171081542969,
      "logps/ref_rejected": -70.01301574707031,
      "logps/rejected": -118.45684814453125,
      "loss": 1.2652,
      "margin_dpo/margin_mean": 19.764087677001953,
      "margin_dpo/margin_std": 29.50685691833496,
      "step": 160
    },
    {
      "epoch": 0.24338624338624337,
      "grad_norm": 5.152878761291504,
      "learning_rate": 4.703633408618955e-07,
      "logits/chosen": 0.5939978957176208,
      "logits/rejected": 0.5104175806045532,
      "logps/chosen": -85.92366790771484,
      "logps/ref_chosen": -51.761775970458984,
      "logps/ref_rejected": -89.95056915283203,
      "logps/rejected": -145.84454345703125,
      "loss": 1.2587,
      "margin_dpo/margin_mean": 21.732070922851562,
      "margin_dpo/margin_std": 34.75413513183594,
      "step": 161
    },
    {
      "epoch": 0.24489795918367346,
      "grad_norm": 6.838240146636963,
      "learning_rate": 4.697358159051549e-07,
      "logits/chosen": 0.5239239931106567,
      "logits/rejected": 0.5126291513442993,
      "logps/chosen": -148.87701416015625,
      "logps/ref_chosen": -93.13358306884766,
      "logps/ref_rejected": -96.1287841796875,
      "logps/rejected": -170.743408203125,
      "loss": 1.2403,
      "margin_dpo/margin_mean": 18.871191024780273,
      "margin_dpo/margin_std": 61.78876495361328,
      "step": 162
    },
    {
      "epoch": 0.24640967498110355,
      "grad_norm": 5.787431240081787,
      "learning_rate": 4.691021444652876e-07,
      "logits/chosen": 0.4375625252723694,
      "logits/rejected": 0.2801172137260437,
      "logps/chosen": -79.89445495605469,
      "logps/ref_chosen": -46.83258056640625,
      "logps/ref_rejected": -108.36699676513672,
      "logps/rejected": -170.7933807373047,
      "loss": 1.2107,
      "margin_dpo/margin_mean": 29.364521026611328,
      "margin_dpo/margin_std": 36.74135208129883,
      "step": 163
    },
    {
      "epoch": 0.24792139077853365,
      "grad_norm": 5.994203090667725,
      "learning_rate": 4.6846234426744624e-07,
      "logits/chosen": 0.490617573261261,
      "logits/rejected": 0.4238327443599701,
      "logps/chosen": -90.10145568847656,
      "logps/ref_chosen": -54.36710739135742,
      "logps/ref_rejected": -96.78005981445312,
      "logps/rejected": -158.1606903076172,
      "loss": 1.1904,
      "margin_dpo/margin_mean": 25.646286010742188,
      "margin_dpo/margin_std": 32.140342712402344,
      "step": 164
    },
    {
      "epoch": 0.2494331065759637,
      "grad_norm": 5.857151508331299,
      "learning_rate": 4.678164332082175e-07,
      "logits/chosen": 0.6752983331680298,
      "logits/rejected": 0.6979560852050781,
      "logps/chosen": -97.69610595703125,
      "logps/ref_chosen": -57.447242736816406,
      "logps/ref_rejected": -57.388519287109375,
      "logps/rejected": -110.6009750366211,
      "loss": 1.2174,
      "margin_dpo/margin_mean": 12.963600158691406,
      "margin_dpo/margin_std": 32.28080749511719,
      "step": 165
    },
    {
      "epoch": 0.2509448223733938,
      "grad_norm": 6.883800029754639,
      "learning_rate": 4.6716442935512214e-07,
      "logits/chosen": 0.6320142149925232,
      "logits/rejected": 0.524321436882019,
      "logps/chosen": -117.2408447265625,
      "logps/ref_chosen": -75.21217346191406,
      "logps/ref_rejected": -97.00743103027344,
      "logps/rejected": -162.87738037109375,
      "loss": 1.2198,
      "margin_dpo/margin_mean": 23.84128189086914,
      "margin_dpo/margin_std": 30.363832473754883,
      "step": 166
    },
    {
      "epoch": 0.25245653817082386,
      "grad_norm": 7.713088512420654,
      "learning_rate": 4.6650635094610966e-07,
      "logits/chosen": 0.6184214949607849,
      "logits/rejected": 0.5350244641304016,
      "logps/chosen": -72.7265853881836,
      "logps/ref_chosen": -39.998783111572266,
      "logps/ref_rejected": -68.48226928710938,
      "logps/rejected": -122.67631530761719,
      "loss": 1.2041,
      "margin_dpo/margin_mean": 21.466243743896484,
      "margin_dpo/margin_std": 22.175495147705078,
      "step": 167
    },
    {
      "epoch": 0.25396825396825395,
      "grad_norm": 6.5206685066223145,
      "learning_rate": 4.6584221638904767e-07,
      "logits/chosen": 0.46801936626434326,
      "logits/rejected": 0.5133869647979736,
      "logps/chosen": -126.92230224609375,
      "logps/ref_chosen": -78.39299774169922,
      "logps/ref_rejected": -56.804256439208984,
      "logps/rejected": -111.1657485961914,
      "loss": 1.2315,
      "margin_dpo/margin_mean": 5.832178592681885,
      "margin_dpo/margin_std": 34.10231399536133,
      "step": 168
    },
    {
      "epoch": 0.25547996976568405,
      "grad_norm": 6.418181419372559,
      "learning_rate": 4.651720442612075e-07,
      "logits/chosen": 0.5367448329925537,
      "logits/rejected": 0.49534183740615845,
      "logps/chosen": -96.13229370117188,
      "logps/ref_chosen": -61.65520477294922,
      "logps/ref_rejected": -74.08130645751953,
      "logps/rejected": -126.73211669921875,
      "loss": 1.2526,
      "margin_dpo/margin_mean": 18.173721313476562,
      "margin_dpo/margin_std": 33.40047073364258,
      "step": 169
    },
    {
      "epoch": 0.25699168556311414,
      "grad_norm": 6.973052501678467,
      "learning_rate": 4.6449585330874425e-07,
      "logits/chosen": 0.6434294581413269,
      "logits/rejected": 0.6778239011764526,
      "logps/chosen": -76.18331146240234,
      "logps/ref_chosen": -48.08928680419922,
      "logps/ref_rejected": -49.94983673095703,
      "logps/rejected": -97.5601577758789,
      "loss": 1.2333,
      "margin_dpo/margin_mean": 19.516292572021484,
      "margin_dpo/margin_std": 36.30491638183594,
      "step": 170
    },
    {
      "epoch": 0.2585034013605442,
      "grad_norm": 7.348653316497803,
      "learning_rate": 4.6381366244617224e-07,
      "logits/chosen": 0.7886656522750854,
      "logits/rejected": 0.7550907135009766,
      "logps/chosen": -70.39274597167969,
      "logps/ref_chosen": -49.932838439941406,
      "logps/ref_rejected": -69.74531555175781,
      "logps/rejected": -124.51280212402344,
      "loss": 1.2094,
      "margin_dpo/margin_mean": 34.307586669921875,
      "margin_dpo/margin_std": 35.918819427490234,
      "step": 171
    },
    {
      "epoch": 0.2600151171579743,
      "grad_norm": 8.536656379699707,
      "learning_rate": 4.631254907558365e-07,
      "logits/chosen": 0.8007753491401672,
      "logits/rejected": 0.7138994932174683,
      "logps/chosen": -96.61810302734375,
      "logps/ref_chosen": -49.946144104003906,
      "logps/ref_rejected": -93.07025146484375,
      "logps/rejected": -180.4694366455078,
      "loss": 1.1897,
      "margin_dpo/margin_mean": 40.72722244262695,
      "margin_dpo/margin_std": 49.622161865234375,
      "step": 172
    },
    {
      "epoch": 0.2615268329554044,
      "grad_norm": 7.331446647644043,
      "learning_rate": 4.624313574873786e-07,
      "logits/chosen": 0.6109236478805542,
      "logits/rejected": 0.6213595867156982,
      "logps/chosen": -75.77210235595703,
      "logps/ref_chosen": -49.288230895996094,
      "logps/ref_rejected": -48.86689758300781,
      "logps/rejected": -101.83216094970703,
      "loss": 1.1658,
      "margin_dpo/margin_mean": 26.481386184692383,
      "margin_dpo/margin_std": 57.468788146972656,
      "step": 173
    },
    {
      "epoch": 0.26303854875283444,
      "grad_norm": 8.760350227355957,
      "learning_rate": 4.61731282057198e-07,
      "logits/chosen": 0.6046967506408691,
      "logits/rejected": 0.5037060379981995,
      "logps/chosen": -112.89518737792969,
      "logps/ref_chosen": -64.05992126464844,
      "logps/ref_rejected": -93.19732666015625,
      "logps/rejected": -186.00070190429688,
      "loss": 1.1071,
      "margin_dpo/margin_mean": 43.968109130859375,
      "margin_dpo/margin_std": 35.23388671875,
      "step": 174
    },
    {
      "epoch": 0.26455026455026454,
      "grad_norm": 9.087226867675781,
      "learning_rate": 4.6102528404790965e-07,
      "logits/chosen": 0.8991174697875977,
      "logits/rejected": 0.8246195316314697,
      "logps/chosen": -83.77995300292969,
      "logps/ref_chosen": -40.91709899902344,
      "logps/ref_rejected": -84.71610260009766,
      "logps/rejected": -157.31451416015625,
      "loss": 1.1565,
      "margin_dpo/margin_mean": 29.735551834106445,
      "margin_dpo/margin_std": 53.80352783203125,
      "step": 175
    },
    {
      "epoch": 0.2660619803476946,
      "grad_norm": 13.09632396697998,
      "learning_rate": 4.603133832077953e-07,
      "logits/chosen": 0.6280794143676758,
      "logits/rejected": 0.6072982549667358,
      "logps/chosen": -153.69549560546875,
      "logps/ref_chosen": -83.32914733886719,
      "logps/ref_rejected": -69.77914428710938,
      "logps/rejected": -140.9823455810547,
      "loss": 1.2784,
      "margin_dpo/margin_mean": 0.8368606567382812,
      "margin_dpo/margin_std": 59.72896957397461,
      "step": 176
    },
    {
      "epoch": 0.2675736961451247,
      "grad_norm": 11.76092529296875,
      "learning_rate": 4.5959559945025183e-07,
      "logits/chosen": 0.843967080116272,
      "logits/rejected": 0.8128781318664551,
      "logps/chosen": -91.85818481445312,
      "logps/ref_chosen": -56.50011444091797,
      "logps/ref_rejected": -76.53157806396484,
      "logps/rejected": -166.71331787109375,
      "loss": 1.008,
      "margin_dpo/margin_mean": 54.82364273071289,
      "margin_dpo/margin_std": 54.522438049316406,
      "step": 177
    },
    {
      "epoch": 0.2690854119425548,
      "grad_norm": 12.708898544311523,
      "learning_rate": 4.588719528532341e-07,
      "logits/chosen": 0.7774707078933716,
      "logits/rejected": 0.6817446351051331,
      "logps/chosen": -82.52491760253906,
      "logps/ref_chosen": -35.268131256103516,
      "logps/ref_rejected": -70.45591735839844,
      "logps/rejected": -159.62277221679688,
      "loss": 1.1395,
      "margin_dpo/margin_mean": 41.910072326660156,
      "margin_dpo/margin_std": 58.84284973144531,
      "step": 178
    },
    {
      "epoch": 0.2705971277399849,
      "grad_norm": 11.324702262878418,
      "learning_rate": 4.581424636586928e-07,
      "logits/chosen": 0.724172830581665,
      "logits/rejected": 0.7666547298431396,
      "logps/chosen": -149.95443725585938,
      "logps/ref_chosen": -82.57086181640625,
      "logps/ref_rejected": -75.7454605102539,
      "logps/rejected": -170.7916717529297,
      "loss": 1.1648,
      "margin_dpo/margin_mean": 27.66265106201172,
      "margin_dpo/margin_std": 52.755733489990234,
      "step": 179
    },
    {
      "epoch": 0.272108843537415,
      "grad_norm": 10.527196884155273,
      "learning_rate": 4.5740715227200897e-07,
      "logits/chosen": 0.6827090978622437,
      "logits/rejected": 0.591106116771698,
      "logps/chosen": -93.48646545410156,
      "logps/ref_chosen": -49.27946472167969,
      "logps/ref_rejected": -70.7477035522461,
      "logps/rejected": -130.7356719970703,
      "loss": 1.2205,
      "margin_dpo/margin_mean": 15.780977249145508,
      "margin_dpo/margin_std": 23.79660415649414,
      "step": 180
    },
    {
      "epoch": 0.273620559334845,
      "grad_norm": 12.358736991882324,
      "learning_rate": 4.566660392614228e-07,
      "logits/chosen": 0.7833654880523682,
      "logits/rejected": 0.7754448652267456,
      "logps/chosen": -116.57708740234375,
      "logps/ref_chosen": -68.44458770751953,
      "logps/ref_rejected": -75.00827026367188,
      "logps/rejected": -140.25169372558594,
      "loss": 1.0862,
      "margin_dpo/margin_mean": 17.11092185974121,
      "margin_dpo/margin_std": 26.67954444885254,
      "step": 181
    },
    {
      "epoch": 0.2751322751322751,
      "grad_norm": 11.587776184082031,
      "learning_rate": 4.5591914535745817e-07,
      "logits/chosen": 0.7390822768211365,
      "logits/rejected": 0.6326345205307007,
      "logps/chosen": -106.85360717773438,
      "logps/ref_chosen": -53.30392074584961,
      "logps/ref_rejected": -106.17438507080078,
      "logps/rejected": -206.2800750732422,
      "loss": 1.0938,
      "margin_dpo/margin_mean": 46.55598449707031,
      "margin_dpo/margin_std": 62.35004425048828,
      "step": 182
    },
    {
      "epoch": 0.2766439909297052,
      "grad_norm": 14.335674285888672,
      "learning_rate": 4.551664914523433e-07,
      "logits/chosen": 0.6751140356063843,
      "logits/rejected": 0.6843445301055908,
      "logps/chosen": -129.42478942871094,
      "logps/ref_chosen": -61.0320930480957,
      "logps/ref_rejected": -64.14570617675781,
      "logps/rejected": -138.19187927246094,
      "loss": 1.3218,
      "margin_dpo/margin_mean": 5.653476238250732,
      "margin_dpo/margin_std": 64.97230529785156,
      "step": 183
    },
    {
      "epoch": 0.2781557067271353,
      "grad_norm": 14.724100112915039,
      "learning_rate": 4.544080985994258e-07,
      "logits/chosen": 0.8014061450958252,
      "logits/rejected": 0.7991318702697754,
      "logps/chosen": -102.4467544555664,
      "logps/ref_chosen": -58.40687561035156,
      "logps/ref_rejected": -65.13507080078125,
      "logps/rejected": -135.5673828125,
      "loss": 1.1217,
      "margin_dpo/margin_mean": 26.392431259155273,
      "margin_dpo/margin_std": 42.3356819152832,
      "step": 184
    },
    {
      "epoch": 0.2796674225245654,
      "grad_norm": 13.151495933532715,
      "learning_rate": 4.5364398801258394e-07,
      "logits/chosen": 0.8872311115264893,
      "logits/rejected": 0.8844795227050781,
      "logps/chosen": -117.208984375,
      "logps/ref_chosen": -68.01717376708984,
      "logps/ref_rejected": -68.56169128417969,
      "logps/rejected": -155.34658813476562,
      "loss": 1.1561,
      "margin_dpo/margin_mean": 37.59309387207031,
      "margin_dpo/margin_std": 63.59172058105469,
      "step": 185
    },
    {
      "epoch": 0.2811791383219955,
      "grad_norm": 18.820205688476562,
      "learning_rate": 4.5287418106563354e-07,
      "logits/chosen": 0.7388228178024292,
      "logits/rejected": 0.6680508255958557,
      "logps/chosen": -118.04601287841797,
      "logps/ref_chosen": -66.14515686035156,
      "logps/ref_rejected": -101.60063934326172,
      "logps/rejected": -203.31298828125,
      "loss": 1.1578,
      "margin_dpo/margin_mean": 49.811492919921875,
      "margin_dpo/margin_std": 85.83422088623047,
      "step": 186
    },
    {
      "epoch": 0.28269085411942557,
      "grad_norm": 13.48037052154541,
      "learning_rate": 4.520986992917297e-07,
      "logits/chosen": 0.6655247211456299,
      "logits/rejected": 0.5620222687721252,
      "logps/chosen": -162.19210815429688,
      "logps/ref_chosen": -80.47019958496094,
      "logps/ref_rejected": -118.81498718261719,
      "logps/rejected": -255.6341094970703,
      "loss": 1.1797,
      "margin_dpo/margin_mean": 55.09719467163086,
      "margin_dpo/margin_std": 82.15184020996094,
      "step": 187
    },
    {
      "epoch": 0.2842025699168556,
      "grad_norm": 12.772809982299805,
      "learning_rate": 4.5131756438276466e-07,
      "logits/chosen": 0.8540096282958984,
      "logits/rejected": 0.7868589758872986,
      "logps/chosen": -118.0152587890625,
      "logps/ref_chosen": -66.75248718261719,
      "logps/ref_rejected": -97.99404907226562,
      "logps/rejected": -214.93295288085938,
      "loss": 1.1456,
      "margin_dpo/margin_mean": 65.6761474609375,
      "margin_dpo/margin_std": 84.41012573242188,
      "step": 188
    },
    {
      "epoch": 0.2857142857142857,
      "grad_norm": 19.746803283691406,
      "learning_rate": 4.5053079818876096e-07,
      "logits/chosen": 0.742944598197937,
      "logits/rejected": 0.8294093012809753,
      "logps/chosen": -144.683349609375,
      "logps/ref_chosen": -75.2247314453125,
      "logps/ref_rejected": -72.38041687011719,
      "logps/rejected": -162.834228515625,
      "loss": 1.3007,
      "margin_dpo/margin_mean": 20.995182037353516,
      "margin_dpo/margin_std": 69.88140869140625,
      "step": 189
    },
    {
      "epoch": 0.2872260015117158,
      "grad_norm": 15.231517791748047,
      "learning_rate": 4.4973842271726024e-07,
      "logits/chosen": 0.8356347680091858,
      "logits/rejected": 0.7419763803482056,
      "logps/chosen": -101.99494934082031,
      "logps/ref_chosen": -47.77008056640625,
      "logps/ref_rejected": -84.10618591308594,
      "logps/rejected": -201.740966796875,
      "loss": 0.9599,
      "margin_dpo/margin_mean": 63.4099235534668,
      "margin_dpo/margin_std": 64.51853942871094,
      "step": 190
    },
    {
      "epoch": 0.2887377173091459,
      "grad_norm": 20.671388626098633,
      "learning_rate": 4.48940460132708e-07,
      "logits/chosen": 0.8784452676773071,
      "logits/rejected": 0.8466886281967163,
      "logps/chosen": -134.3704833984375,
      "logps/ref_chosen": -66.10474395751953,
      "logps/ref_rejected": -79.35490417480469,
      "logps/rejected": -206.15818786621094,
      "loss": 1.2401,
      "margin_dpo/margin_mean": 58.53753662109375,
      "margin_dpo/margin_std": 76.90192413330078,
      "step": 191
    },
    {
      "epoch": 0.29024943310657597,
      "grad_norm": 21.7989559173584,
      "learning_rate": 4.481369327558329e-07,
      "logits/chosen": 0.8773584961891174,
      "logits/rejected": 0.8835272789001465,
      "logps/chosen": -154.4402618408203,
      "logps/ref_chosen": -61.76245880126953,
      "logps/ref_rejected": -60.038848876953125,
      "logps/rejected": -150.0125732421875,
      "loss": 1.279,
      "margin_dpo/margin_mean": -2.704071521759033,
      "margin_dpo/margin_std": 65.4942626953125,
      "step": 192
    },
    {
      "epoch": 0.29176114890400606,
      "grad_norm": 13.274942398071289,
      "learning_rate": 4.47327863063023e-07,
      "logits/chosen": 0.7359751462936401,
      "logits/rejected": 0.6782011985778809,
      "logps/chosen": -131.29852294921875,
      "logps/ref_chosen": -50.086849212646484,
      "logps/ref_rejected": -69.65550231933594,
      "logps/rejected": -187.8815460205078,
      "loss": 1.1063,
      "margin_dpo/margin_mean": 37.01435089111328,
      "margin_dpo/margin_std": 56.66736602783203,
      "step": 193
    },
    {
      "epoch": 0.29327286470143615,
      "grad_norm": 21.689014434814453,
      "learning_rate": 4.4651327368569684e-07,
      "logits/chosen": 0.7673474550247192,
      "logits/rejected": 0.8195520639419556,
      "logps/chosen": -152.31005859375,
      "logps/ref_chosen": -72.05107116699219,
      "logps/ref_rejected": -71.8668212890625,
      "logps/rejected": -168.17160034179688,
      "loss": 1.2219,
      "margin_dpo/margin_mean": 16.045808792114258,
      "margin_dpo/margin_std": 65.90227508544922,
      "step": 194
    },
    {
      "epoch": 0.2947845804988662,
      "grad_norm": 19.105045318603516,
      "learning_rate": 4.4569318740967043e-07,
      "logits/chosen": 0.6917105317115784,
      "logits/rejected": 0.761000394821167,
      "logps/chosen": -192.57481384277344,
      "logps/ref_chosen": -83.29867553710938,
      "logps/ref_rejected": -66.07734680175781,
      "logps/rejected": -193.638671875,
      "loss": 1.179,
      "margin_dpo/margin_mean": 18.285186767578125,
      "margin_dpo/margin_std": 70.0030288696289,
      "step": 195
    },
    {
      "epoch": 0.2962962962962963,
      "grad_norm": 14.678657531738281,
      "learning_rate": 4.448676271745197e-07,
      "logits/chosen": 0.8080000877380371,
      "logits/rejected": 0.8198258876800537,
      "logps/chosen": -177.1346435546875,
      "logps/ref_chosen": -81.20787048339844,
      "logps/ref_rejected": -79.35533142089844,
      "logps/rejected": -194.04977416992188,
      "loss": 1.1227,
      "margin_dpo/margin_mean": 18.76766014099121,
      "margin_dpo/margin_std": 53.9957160949707,
      "step": 196
    },
    {
      "epoch": 0.29780801209372637,
      "grad_norm": 20.904054641723633,
      "learning_rate": 4.440366160729392e-07,
      "logits/chosen": 1.08381986618042,
      "logits/rejected": 0.9524755477905273,
      "logps/chosen": -121.47453308105469,
      "logps/ref_chosen": -38.60869598388672,
      "logps/ref_rejected": -74.34709167480469,
      "logps/rejected": -196.74826049804688,
      "loss": 1.1447,
      "margin_dpo/margin_mean": 39.53533935546875,
      "margin_dpo/margin_std": 104.43431091308594,
      "step": 197
    },
    {
      "epoch": 0.29931972789115646,
      "grad_norm": 18.375619888305664,
      "learning_rate": 4.432001773500957e-07,
      "logits/chosen": 1.0280174016952515,
      "logits/rejected": 0.9963433742523193,
      "logps/chosen": -113.64097595214844,
      "logps/ref_chosen": -48.90552520751953,
      "logps/ref_rejected": -63.93467712402344,
      "logps/rejected": -166.58355712890625,
      "loss": 1.1247,
      "margin_dpo/margin_mean": 37.91343307495117,
      "margin_dpo/margin_std": 68.63130187988281,
      "step": 198
    },
    {
      "epoch": 0.30083144368858655,
      "grad_norm": 13.685064315795898,
      "learning_rate": 4.4235833440297856e-07,
      "logits/chosen": 0.8831381797790527,
      "logits/rejected": 0.8222548961639404,
      "logps/chosen": -129.5958709716797,
      "logps/ref_chosen": -57.552146911621094,
      "logps/ref_rejected": -76.73622131347656,
      "logps/rejected": -196.77474975585938,
      "loss": 1.1776,
      "margin_dpo/margin_mean": 47.994808197021484,
      "margin_dpo/margin_std": 66.5357437133789,
      "step": 199
    },
    {
      "epoch": 0.30234315948601664,
      "grad_norm": 13.301373481750488,
      "learning_rate": 4.415111107797445e-07,
      "logits/chosen": 0.9643712043762207,
      "logits/rejected": 0.7956128120422363,
      "logps/chosen": -104.0108413696289,
      "logps/ref_chosen": -49.7380485534668,
      "logps/ref_rejected": -119.85460662841797,
      "logps/rejected": -247.071533203125,
      "loss": 1.0529,
      "margin_dpo/margin_mean": 72.94413757324219,
      "margin_dpo/margin_std": 100.05728149414062,
      "step": 200
    },
    {
      "epoch": 0.30385487528344673,
      "grad_norm": 14.938904762268066,
      "learning_rate": 4.4065853017905953e-07,
      "logits/chosen": 0.9494801759719849,
      "logits/rejected": 0.963019847869873,
      "logps/chosen": -168.0663604736328,
      "logps/ref_chosen": -76.25155639648438,
      "logps/ref_rejected": -75.89337921142578,
      "logps/rejected": -206.00071716308594,
      "loss": 1.1877,
      "margin_dpo/margin_mean": 38.292537689208984,
      "margin_dpo/margin_std": 74.59246063232422,
      "step": 201
    },
    {
      "epoch": 0.30536659108087677,
      "grad_norm": 19.734487533569336,
      "learning_rate": 4.3980061644943575e-07,
      "logits/chosen": 0.8432959318161011,
      "logits/rejected": 0.6981616020202637,
      "logps/chosen": -75.70082092285156,
      "logps/ref_chosen": -31.3531551361084,
      "logps/ref_rejected": -70.96551513671875,
      "logps/rejected": -184.80126953125,
      "loss": 1.0936,
      "margin_dpo/margin_mean": 69.48809051513672,
      "margin_dpo/margin_std": 59.054290771484375,
      "step": 202
    },
    {
      "epoch": 0.30687830687830686,
      "grad_norm": 17.070323944091797,
      "learning_rate": 4.3893739358856455e-07,
      "logits/chosen": 0.9048564434051514,
      "logits/rejected": 0.7673661708831787,
      "logps/chosen": -144.82659912109375,
      "logps/ref_chosen": -63.048912048339844,
      "logps/ref_rejected": -137.18235778808594,
      "logps/rejected": -288.850341796875,
      "loss": 1.0973,
      "margin_dpo/margin_mean": 69.89026641845703,
      "margin_dpo/margin_std": 88.3589096069336,
      "step": 203
    },
    {
      "epoch": 0.30839002267573695,
      "grad_norm": 14.029024124145508,
      "learning_rate": 4.380688857426449e-07,
      "logits/chosen": 0.8778685331344604,
      "logits/rejected": 0.7175527811050415,
      "logps/chosen": -111.14114379882812,
      "logps/ref_chosen": -45.349220275878906,
      "logps/ref_rejected": -89.06627655029297,
      "logps/rejected": -215.51800537109375,
      "loss": 1.0848,
      "margin_dpo/margin_mean": 60.6597900390625,
      "margin_dpo/margin_std": 76.72471618652344,
      "step": 204
    },
    {
      "epoch": 0.30990173847316704,
      "grad_norm": 15.573190689086914,
      "learning_rate": 4.3719511720570814e-07,
      "logits/chosen": 1.0194439888000488,
      "logits/rejected": 0.9591749310493469,
      "logps/chosen": -138.189697265625,
      "logps/ref_chosen": -68.1277847290039,
      "logps/ref_rejected": -94.7017593383789,
      "logps/rejected": -237.84596252441406,
      "loss": 1.1396,
      "margin_dpo/margin_mean": 73.08229064941406,
      "margin_dpo/margin_std": 103.01658630371094,
      "step": 205
    },
    {
      "epoch": 0.31141345427059713,
      "grad_norm": 21.639747619628906,
      "learning_rate": 4.363161124189387e-07,
      "logits/chosen": 1.0190932750701904,
      "logits/rejected": 0.9483546614646912,
      "logps/chosen": -123.68317413330078,
      "logps/ref_chosen": -49.561851501464844,
      "logps/ref_rejected": -91.47630310058594,
      "logps/rejected": -198.70071411132812,
      "loss": 1.2179,
      "margin_dpo/margin_mean": 33.103092193603516,
      "margin_dpo/margin_std": 57.71031951904297,
      "step": 206
    },
    {
      "epoch": 0.3129251700680272,
      "grad_norm": 17.420026779174805,
      "learning_rate": 4.3543189596998986e-07,
      "logits/chosen": 0.9188834428787231,
      "logits/rejected": 0.8394033908843994,
      "logps/chosen": -133.870361328125,
      "logps/ref_chosen": -39.89246368408203,
      "logps/ref_rejected": -70.57868194580078,
      "logps/rejected": -204.41519165039062,
      "loss": 1.14,
      "margin_dpo/margin_mean": 39.858604431152344,
      "margin_dpo/margin_std": 71.88990783691406,
      "step": 207
    },
    {
      "epoch": 0.3144368858654573,
      "grad_norm": 16.912046432495117,
      "learning_rate": 4.3454249259229664e-07,
      "logits/chosen": 0.8252842426300049,
      "logits/rejected": 0.7790793776512146,
      "logps/chosen": -161.83831787109375,
      "logps/ref_chosen": -72.9476547241211,
      "logps/ref_rejected": -94.71766662597656,
      "logps/rejected": -216.20074462890625,
      "loss": 1.31,
      "margin_dpo/margin_mean": 32.592403411865234,
      "margin_dpo/margin_std": 57.00579071044922,
      "step": 208
    },
    {
      "epoch": 0.31594860166288735,
      "grad_norm": 25.769882202148438,
      "learning_rate": 4.336479271643833e-07,
      "logits/chosen": 0.9054467678070068,
      "logits/rejected": 0.831355631351471,
      "logps/chosen": -154.79385375976562,
      "logps/ref_chosen": -65.3952865600586,
      "logps/ref_rejected": -94.79216003417969,
      "logps/rejected": -261.7325439453125,
      "loss": 1.166,
      "margin_dpo/margin_mean": 77.54179382324219,
      "margin_dpo/margin_std": 97.3060073852539,
      "step": 209
    },
    {
      "epoch": 0.31746031746031744,
      "grad_norm": 12.30976676940918,
      "learning_rate": 4.327482247091679e-07,
      "logits/chosen": 0.8562976121902466,
      "logits/rejected": 0.6786376237869263,
      "logps/chosen": -197.23080444335938,
      "logps/ref_chosen": -88.16167449951172,
      "logps/ref_rejected": -143.62713623046875,
      "logps/rejected": -364.16387939453125,
      "loss": 1.0647,
      "margin_dpo/margin_mean": 111.46759033203125,
      "margin_dpo/margin_std": 77.64178466796875,
      "step": 210
    },
    {
      "epoch": 0.31897203325774753,
      "grad_norm": 16.888795852661133,
      "learning_rate": 4.3184341039326217e-07,
      "logits/chosen": 1.019036054611206,
      "logits/rejected": 0.9224546551704407,
      "logps/chosen": -116.31884765625,
      "logps/ref_chosen": -43.544952392578125,
      "logps/ref_rejected": -96.58848571777344,
      "logps/rejected": -221.36495971679688,
      "loss": 1.0403,
      "margin_dpo/margin_mean": 52.002593994140625,
      "margin_dpo/margin_std": 71.8609390258789,
      "step": 211
    },
    {
      "epoch": 0.3204837490551776,
      "grad_norm": 14.35758113861084,
      "learning_rate": 4.309335095262675e-07,
      "logits/chosen": 0.9997249841690063,
      "logits/rejected": 1.0045267343521118,
      "logps/chosen": -140.09591674804688,
      "logps/ref_chosen": -57.794212341308594,
      "logps/ref_rejected": -72.26956176757812,
      "logps/rejected": -212.42254638671875,
      "loss": 1.0802,
      "margin_dpo/margin_mean": 57.851280212402344,
      "margin_dpo/margin_std": 110.18971252441406,
      "step": 212
    },
    {
      "epoch": 0.3219954648526077,
      "grad_norm": 15.879948616027832,
      "learning_rate": 4.3001854756006724e-07,
      "logits/chosen": 0.8429861664772034,
      "logits/rejected": 0.8865307569503784,
      "logps/chosen": -176.1222381591797,
      "logps/ref_chosen": -81.40860748291016,
      "logps/ref_rejected": -64.03448486328125,
      "logps/rejected": -180.094970703125,
      "loss": 1.1033,
      "margin_dpo/margin_mean": 21.346853256225586,
      "margin_dpo/margin_std": 100.63861083984375,
      "step": 213
    },
    {
      "epoch": 0.3235071806500378,
      "grad_norm": 18.610511779785156,
      "learning_rate": 4.290985500881143e-07,
      "logits/chosen": 0.7390300035476685,
      "logits/rejected": 0.8046228885650635,
      "logps/chosen": -145.5377655029297,
      "logps/ref_chosen": -63.65519714355469,
      "logps/ref_rejected": -61.175392150878906,
      "logps/rejected": -205.8516845703125,
      "loss": 1.1949,
      "margin_dpo/margin_mean": 62.793731689453125,
      "margin_dpo/margin_std": 71.61257934570312,
      "step": 214
    },
    {
      "epoch": 0.3250188964474679,
      "grad_norm": 15.646477699279785,
      "learning_rate": 4.281735428447157e-07,
      "logits/chosen": 0.8204779624938965,
      "logits/rejected": 0.7469308376312256,
      "logps/chosen": -167.19033813476562,
      "logps/ref_chosen": -59.09471893310547,
      "logps/ref_rejected": -93.95791625976562,
      "logps/rejected": -225.2965087890625,
      "loss": 1.0742,
      "margin_dpo/margin_mean": 23.242979049682617,
      "margin_dpo/margin_std": 72.47611999511719,
      "step": 215
    },
    {
      "epoch": 0.32653061224489793,
      "grad_norm": 25.667827606201172,
      "learning_rate": 4.2724355170431247e-07,
      "logits/chosen": 0.8592058420181274,
      "logits/rejected": 0.8133991360664368,
      "logps/chosen": -183.30780029296875,
      "logps/ref_chosen": -88.48361206054688,
      "logps/ref_rejected": -95.00409698486328,
      "logps/rejected": -257.6964111328125,
      "loss": 0.9971,
      "margin_dpo/margin_mean": 67.86811065673828,
      "margin_dpo/margin_std": 59.22956466674805,
      "step": 216
    },
    {
      "epoch": 0.328042328042328,
      "grad_norm": 13.31557559967041,
      "learning_rate": 4.26308602680756e-07,
      "logits/chosen": 0.9996564388275146,
      "logits/rejected": 0.8627007007598877,
      "logps/chosen": -174.81881713867188,
      "logps/ref_chosen": -62.363914489746094,
      "logps/ref_rejected": -104.9856948852539,
      "logps/rejected": -284.2158203125,
      "loss": 1.0242,
      "margin_dpo/margin_mean": 66.77523803710938,
      "margin_dpo/margin_std": 126.6890869140625,
      "step": 217
    },
    {
      "epoch": 0.3295540438397581,
      "grad_norm": 18.151487350463867,
      "learning_rate": 4.253687219265803e-07,
      "logits/chosen": 0.7707165479660034,
      "logits/rejected": 0.7527076005935669,
      "logps/chosen": -201.19351196289062,
      "logps/ref_chosen": -87.87580871582031,
      "logps/ref_rejected": -102.97655487060547,
      "logps/rejected": -231.63278198242188,
      "loss": 1.2726,
      "margin_dpo/margin_mean": 15.338525772094727,
      "margin_dpo/margin_std": 67.2188720703125,
      "step": 218
    },
    {
      "epoch": 0.3310657596371882,
      "grad_norm": 14.18816089630127,
      "learning_rate": 4.2442393573227043e-07,
      "logits/chosen": 0.9285779595375061,
      "logits/rejected": 0.9085969924926758,
      "logps/chosen": -155.8358154296875,
      "logps/ref_chosen": -58.567657470703125,
      "logps/ref_rejected": -94.29285430908203,
      "logps/rejected": -231.346923828125,
      "loss": 1.1618,
      "margin_dpo/margin_mean": 39.78589630126953,
      "margin_dpo/margin_std": 81.78353881835938,
      "step": 219
    },
    {
      "epoch": 0.3325774754346183,
      "grad_norm": 13.471064567565918,
      "learning_rate": 4.234742705255272e-07,
      "logits/chosen": 0.8746180534362793,
      "logits/rejected": 0.7802847623825073,
      "logps/chosen": -130.3991241455078,
      "logps/ref_chosen": -45.8538703918457,
      "logps/ref_rejected": -76.9227066040039,
      "logps/rejected": -201.6156005859375,
      "loss": 1.0699,
      "margin_dpo/margin_mean": 40.14763641357422,
      "margin_dpo/margin_std": 70.33467102050781,
      "step": 220
    },
    {
      "epoch": 0.3340891912320484,
      "grad_norm": 13.25949478149414,
      "learning_rate": 4.22519752870528e-07,
      "logits/chosen": 0.8579016923904419,
      "logits/rejected": 0.7474366426467896,
      "logps/chosen": -150.62062072753906,
      "logps/ref_chosen": -55.79085922241211,
      "logps/ref_rejected": -81.47898864746094,
      "logps/rejected": -204.39466857910156,
      "loss": 1.0478,
      "margin_dpo/margin_mean": 28.085922241210938,
      "margin_dpo/margin_std": 60.90304946899414,
      "step": 221
    },
    {
      "epoch": 0.3356009070294785,
      "grad_norm": 11.687209129333496,
      "learning_rate": 4.2156040946718343e-07,
      "logits/chosen": 0.9172873497009277,
      "logits/rejected": 0.8860268592834473,
      "logps/chosen": -135.50936889648438,
      "logps/ref_chosen": -51.75988006591797,
      "logps/ref_rejected": -97.38671875,
      "logps/rejected": -218.99237060546875,
      "loss": 0.9846,
      "margin_dpo/margin_mean": 37.856170654296875,
      "margin_dpo/margin_std": 90.37283325195312,
      "step": 222
    },
    {
      "epoch": 0.3371126228269085,
      "grad_norm": 12.334152221679688,
      "learning_rate": 4.2059626715039065e-07,
      "logits/chosen": 0.8898248672485352,
      "logits/rejected": 0.8581580519676208,
      "logps/chosen": -131.7398681640625,
      "logps/ref_chosen": -60.334190368652344,
      "logps/ref_rejected": -67.8521728515625,
      "logps/rejected": -187.02847290039062,
      "loss": 1.0231,
      "margin_dpo/margin_mean": 47.77062225341797,
      "margin_dpo/margin_std": 67.18444061279297,
      "step": 223
    },
    {
      "epoch": 0.3386243386243386,
      "grad_norm": 13.227733612060547,
      "learning_rate": 4.1962735288928304e-07,
      "logits/chosen": 1.0880929231643677,
      "logits/rejected": 1.0430279970169067,
      "logps/chosen": -98.79350280761719,
      "logps/ref_chosen": -36.07902908325195,
      "logps/ref_rejected": -63.69470977783203,
      "logps/rejected": -191.50698852539062,
      "loss": 1.3037,
      "margin_dpo/margin_mean": 65.09780883789062,
      "margin_dpo/margin_std": 61.339786529541016,
      "step": 224
    },
    {
      "epoch": 0.3401360544217687,
      "grad_norm": 13.988600730895996,
      "learning_rate": 4.186536937864752e-07,
      "logits/chosen": 0.8999257683753967,
      "logits/rejected": 0.7167999744415283,
      "logps/chosen": -125.11752319335938,
      "logps/ref_chosen": -46.84956359863281,
      "logps/ref_rejected": -112.8951416015625,
      "logps/rejected": -255.221435546875,
      "loss": 1.0097,
      "margin_dpo/margin_mean": 64.05833435058594,
      "margin_dpo/margin_std": 75.08706665039062,
      "step": 225
    },
    {
      "epoch": 0.3416477702191988,
      "grad_norm": 11.088345527648926,
      "learning_rate": 4.176753170773052e-07,
      "logits/chosen": 0.9160268306732178,
      "logits/rejected": 0.8631049990653992,
      "logps/chosen": -130.91664123535156,
      "logps/ref_chosen": -44.67559814453125,
      "logps/ref_rejected": -63.792022705078125,
      "logps/rejected": -193.21331787109375,
      "loss": 1.1595,
      "margin_dpo/margin_mean": 43.18025207519531,
      "margin_dpo/margin_std": 66.46514892578125,
      "step": 226
    },
    {
      "epoch": 0.3431594860166289,
      "grad_norm": 13.128989219665527,
      "learning_rate": 4.166922501290729e-07,
      "logits/chosen": 0.8024705648422241,
      "logits/rejected": 0.7597426772117615,
      "logps/chosen": -116.00537109375,
      "logps/ref_chosen": -47.16712188720703,
      "logps/ref_rejected": -63.74501419067383,
      "logps/rejected": -188.3765106201172,
      "loss": 1.1549,
      "margin_dpo/margin_mean": 55.79324722290039,
      "margin_dpo/margin_std": 75.62820434570312,
      "step": 227
    },
    {
      "epoch": 0.34467120181405897,
      "grad_norm": 18.30823516845703,
      "learning_rate": 4.1570452044027405e-07,
      "logits/chosen": 0.9423525333404541,
      "logits/rejected": 0.9120630621910095,
      "logps/chosen": -164.05587768554688,
      "logps/ref_chosen": -63.36100387573242,
      "logps/ref_rejected": -80.64863586425781,
      "logps/rejected": -217.6392364501953,
      "loss": 1.0638,
      "margin_dpo/margin_mean": 36.29573059082031,
      "margin_dpo/margin_std": 62.555362701416016,
      "step": 228
    },
    {
      "epoch": 0.34618291761148906,
      "grad_norm": 15.304150581359863,
      "learning_rate": 4.147121556398312e-07,
      "logits/chosen": 1.0203680992126465,
      "logits/rejected": 0.9438357353210449,
      "logps/chosen": -110.92491149902344,
      "logps/ref_chosen": -46.663169860839844,
      "logps/ref_rejected": -88.81582641601562,
      "logps/rejected": -221.1804962158203,
      "loss": 1.1233,
      "margin_dpo/margin_mean": 68.10292053222656,
      "margin_dpo/margin_std": 46.62117004394531,
      "step": 229
    },
    {
      "epoch": 0.3476946334089191,
      "grad_norm": 17.100791931152344,
      "learning_rate": 4.137151834863213e-07,
      "logits/chosen": 0.8101639151573181,
      "logits/rejected": 0.9272513389587402,
      "logps/chosen": -166.52886962890625,
      "logps/ref_chosen": -74.076171875,
      "logps/ref_rejected": -60.03583526611328,
      "logps/rejected": -188.06703186035156,
      "loss": 1.1756,
      "margin_dpo/margin_mean": 35.5784912109375,
      "margin_dpo/margin_std": 85.3873519897461,
      "step": 230
    },
    {
      "epoch": 0.3492063492063492,
      "grad_norm": 15.612849235534668,
      "learning_rate": 4.1271363186719835e-07,
      "logits/chosen": 0.7588058710098267,
      "logits/rejected": 0.6893700957298279,
      "logps/chosen": -156.947998046875,
      "logps/ref_chosen": -60.65083312988281,
      "logps/ref_rejected": -88.65960693359375,
      "logps/rejected": -248.28106689453125,
      "loss": 0.9731,
      "margin_dpo/margin_mean": 63.32429504394531,
      "margin_dpo/margin_std": 63.744873046875,
      "step": 231
    },
    {
      "epoch": 0.3507180650037793,
      "grad_norm": 16.46584701538086,
      "learning_rate": 4.1170752879801436e-07,
      "logits/chosen": 0.8945074081420898,
      "logits/rejected": 0.9332794547080994,
      "logps/chosen": -147.59791564941406,
      "logps/ref_chosen": -69.09077453613281,
      "logps/ref_rejected": -57.635154724121094,
      "logps/rejected": -174.9073486328125,
      "loss": 1.1969,
      "margin_dpo/margin_mean": 38.76505661010742,
      "margin_dpo/margin_std": 74.07946014404297,
      "step": 232
    },
    {
      "epoch": 0.35222978080120937,
      "grad_norm": 17.694772720336914,
      "learning_rate": 4.106969024216348e-07,
      "logits/chosen": 1.0592719316482544,
      "logits/rejected": 0.9124851226806641,
      "logps/chosen": -154.83291625976562,
      "logps/ref_chosen": -47.50149917602539,
      "logps/ref_rejected": -65.11629486083984,
      "logps/rejected": -212.47425842285156,
      "loss": 1.1982,
      "margin_dpo/margin_mean": 40.02655792236328,
      "margin_dpo/margin_std": 79.98373413085938,
      "step": 233
    },
    {
      "epoch": 0.35374149659863946,
      "grad_norm": 15.55416488647461,
      "learning_rate": 4.09681781007452e-07,
      "logits/chosen": 0.6289401650428772,
      "logits/rejected": 0.6006139516830444,
      "logps/chosen": -184.160888671875,
      "logps/ref_chosen": -71.29449462890625,
      "logps/ref_rejected": -75.5903091430664,
      "logps/rejected": -222.56118774414062,
      "loss": 1.1917,
      "margin_dpo/margin_mean": 34.104488372802734,
      "margin_dpo/margin_std": 52.7165412902832,
      "step": 234
    },
    {
      "epoch": 0.35525321239606955,
      "grad_norm": 12.542901039123535,
      "learning_rate": 4.08662192950594e-07,
      "logits/chosen": 0.8517894744873047,
      "logits/rejected": 0.8705936670303345,
      "logps/chosen": -171.2367401123047,
      "logps/ref_chosen": -85.16849517822266,
      "logps/ref_rejected": -84.92215728759766,
      "logps/rejected": -236.8095703125,
      "loss": 1.0258,
      "margin_dpo/margin_mean": 65.81916809082031,
      "margin_dpo/margin_std": 52.77665710449219,
      "step": 235
    },
    {
      "epoch": 0.35676492819349964,
      "grad_norm": 14.79189682006836,
      "learning_rate": 4.076381667711306e-07,
      "logits/chosen": 0.9688406586647034,
      "logits/rejected": 0.9256927967071533,
      "logps/chosen": -192.56378173828125,
      "logps/ref_chosen": -79.2076416015625,
      "logps/ref_rejected": -100.40895080566406,
      "logps/rejected": -259.85693359375,
      "loss": 1.1711,
      "margin_dpo/margin_mean": 46.09184265136719,
      "margin_dpo/margin_std": 55.658546447753906,
      "step": 236
    },
    {
      "epoch": 0.35827664399092973,
      "grad_norm": 18.867671966552734,
      "learning_rate": 4.066097311132753e-07,
      "logits/chosen": 0.8245556354522705,
      "logits/rejected": 0.794967770576477,
      "logps/chosen": -178.4556121826172,
      "logps/ref_chosen": -75.55789184570312,
      "logps/ref_rejected": -86.2087631225586,
      "logps/rejected": -248.25030517578125,
      "loss": 1.0857,
      "margin_dpo/margin_mean": 59.14380645751953,
      "margin_dpo/margin_std": 55.71824645996094,
      "step": 237
    },
    {
      "epoch": 0.35978835978835977,
      "grad_norm": 16.57659339904785,
      "learning_rate": 4.0557691474458414e-07,
      "logits/chosen": 0.8863713145256042,
      "logits/rejected": 0.8904386758804321,
      "logps/chosen": -154.5315399169922,
      "logps/ref_chosen": -59.983848571777344,
      "logps/ref_rejected": -59.57722854614258,
      "logps/rejected": -193.907958984375,
      "loss": 1.1075,
      "margin_dpo/margin_mean": 39.78303527832031,
      "margin_dpo/margin_std": 73.26469421386719,
      "step": 238
    },
    {
      "epoch": 0.36130007558578986,
      "grad_norm": 21.753231048583984,
      "learning_rate": 4.045397465551513e-07,
      "logits/chosen": 0.9004707336425781,
      "logits/rejected": 0.8397661447525024,
      "logps/chosen": -204.30355834960938,
      "logps/ref_chosen": -71.07061767578125,
      "logps/ref_rejected": -99.01707458496094,
      "logps/rejected": -268.3192443847656,
      "loss": 1.0612,
      "margin_dpo/margin_mean": 36.0692138671875,
      "margin_dpo/margin_std": 65.23834228515625,
      "step": 239
    },
    {
      "epoch": 0.36281179138321995,
      "grad_norm": 14.38486099243164,
      "learning_rate": 4.0349825555680045e-07,
      "logits/chosen": 0.9007062911987305,
      "logits/rejected": 0.8683948516845703,
      "logps/chosen": -195.65853881835938,
      "logps/ref_chosen": -62.13431930541992,
      "logps/ref_rejected": -94.44198608398438,
      "logps/rejected": -284.7762451171875,
      "loss": 1.0184,
      "margin_dpo/margin_mean": 56.8100471496582,
      "margin_dpo/margin_std": 98.01765441894531,
      "step": 240
    },
    {
      "epoch": 0.36432350718065004,
      "grad_norm": 13.766716003417969,
      "learning_rate": 4.0245247088227377e-07,
      "logits/chosen": 0.8529886603355408,
      "logits/rejected": 0.8130964040756226,
      "logps/chosen": -183.49896240234375,
      "logps/ref_chosen": -64.64911651611328,
      "logps/ref_rejected": -76.52389526367188,
      "logps/rejected": -236.65924072265625,
      "loss": 1.1746,
      "margin_dpo/margin_mean": 41.285484313964844,
      "margin_dpo/margin_std": 58.063209533691406,
      "step": 241
    },
    {
      "epoch": 0.36583522297808013,
      "grad_norm": 12.187637329101562,
      "learning_rate": 4.0140242178441665e-07,
      "logits/chosen": 0.8713346719741821,
      "logits/rejected": 0.8219842910766602,
      "logps/chosen": -147.36550903320312,
      "logps/ref_chosen": -50.54002380371094,
      "logps/ref_rejected": -72.02078247070312,
      "logps/rejected": -221.0113067626953,
      "loss": 1.0658,
      "margin_dpo/margin_mean": 52.1650505065918,
      "margin_dpo/margin_std": 68.18669891357422,
      "step": 242
    },
    {
      "epoch": 0.3673469387755102,
      "grad_norm": 15.1566801071167,
      "learning_rate": 4.003481376353596e-07,
      "logits/chosen": 0.9196364879608154,
      "logits/rejected": 0.9491422176361084,
      "logps/chosen": -203.58029174804688,
      "logps/ref_chosen": -89.10255432128906,
      "logps/ref_rejected": -77.90412902832031,
      "logps/rejected": -224.44149780273438,
      "loss": 1.1345,
      "margin_dpo/margin_mean": 32.059627532958984,
      "margin_dpo/margin_std": 67.19066619873047,
      "step": 243
    },
    {
      "epoch": 0.3688586545729403,
      "grad_norm": 16.03683090209961,
      "learning_rate": 3.9928964792569654e-07,
      "logits/chosen": 0.8897314071655273,
      "logits/rejected": 0.8393882513046265,
      "logps/chosen": -187.6811981201172,
      "logps/ref_chosen": -66.64555358886719,
      "logps/ref_rejected": -92.27547454833984,
      "logps/rejected": -261.2939453125,
      "loss": 0.9875,
      "margin_dpo/margin_mean": 47.982826232910156,
      "margin_dpo/margin_std": 67.17561340332031,
      "step": 244
    },
    {
      "epoch": 0.37037037037037035,
      "grad_norm": 12.031768798828125,
      "learning_rate": 3.982269822636601e-07,
      "logits/chosen": 0.9159705638885498,
      "logits/rejected": 0.9142988324165344,
      "logps/chosen": -185.25064086914062,
      "logps/ref_chosen": -79.13615417480469,
      "logps/ref_rejected": -79.21699523925781,
      "logps/rejected": -262.3095397949219,
      "loss": 0.9243,
      "margin_dpo/margin_mean": 76.97804260253906,
      "margin_dpo/margin_std": 56.90997314453125,
      "step": 245
    },
    {
      "epoch": 0.37188208616780044,
      "grad_norm": 19.480737686157227,
      "learning_rate": 3.971601703742932e-07,
      "logits/chosen": 0.9996259212493896,
      "logits/rejected": 1.0093920230865479,
      "logps/chosen": -149.17922973632812,
      "logps/ref_chosen": -57.22200012207031,
      "logps/ref_rejected": -77.80888366699219,
      "logps/rejected": -234.3947296142578,
      "loss": 1.1167,
      "margin_dpo/margin_mean": 64.62861633300781,
      "margin_dpo/margin_std": 74.76765441894531,
      "step": 246
    },
    {
      "epoch": 0.37339380196523053,
      "grad_norm": 19.866626739501953,
      "learning_rate": 3.960892420986177e-07,
      "logits/chosen": 0.7821018695831299,
      "logits/rejected": 0.7258821725845337,
      "logps/chosen": -219.67767333984375,
      "logps/ref_chosen": -94.68956756591797,
      "logps/ref_rejected": -122.22605895996094,
      "logps/rejected": -318.41412353515625,
      "loss": 1.257,
      "margin_dpo/margin_mean": 71.19993591308594,
      "margin_dpo/margin_std": 81.91409301757812,
      "step": 247
    },
    {
      "epoch": 0.3749055177626606,
      "grad_norm": 22.328622817993164,
      "learning_rate": 3.9501422739279953e-07,
      "logits/chosen": 0.7986111640930176,
      "logits/rejected": 0.8373353481292725,
      "logps/chosen": -175.7537384033203,
      "logps/ref_chosen": -60.833953857421875,
      "logps/ref_rejected": -60.90985107421875,
      "logps/rejected": -218.660888671875,
      "loss": 1.1383,
      "margin_dpo/margin_mean": 42.83125305175781,
      "margin_dpo/margin_std": 89.69184112548828,
      "step": 248
    },
    {
      "epoch": 0.3764172335600907,
      "grad_norm": 19.794172286987305,
      "learning_rate": 3.9393515632731094e-07,
      "logits/chosen": 1.0464762449264526,
      "logits/rejected": 1.056666374206543,
      "logps/chosen": -196.64761352539062,
      "logps/ref_chosen": -74.77812194824219,
      "logps/ref_rejected": -75.08592987060547,
      "logps/rejected": -216.56761169433594,
      "loss": 1.4198,
      "margin_dpo/margin_mean": 19.612201690673828,
      "margin_dpo/margin_std": 99.21858215332031,
      "step": 249
    },
    {
      "epoch": 0.3779289493575208,
      "grad_norm": 19.19304847717285,
      "learning_rate": 3.9285205908608934e-07,
      "logits/chosen": 1.015788197517395,
      "logits/rejected": 0.939686119556427,
      "logps/chosen": -132.5025634765625,
      "logps/ref_chosen": -47.93787384033203,
      "logps/ref_rejected": -61.55204772949219,
      "logps/rejected": -208.47003173828125,
      "loss": 1.0567,
      "margin_dpo/margin_mean": 62.353302001953125,
      "margin_dpo/margin_std": 72.25045776367188,
      "step": 250
    },
    {
      "epoch": 0.3794406651549509,
      "grad_norm": 18.333349227905273,
      "learning_rate": 3.9176496596569265e-07,
      "logits/chosen": 0.7942206263542175,
      "logits/rejected": 0.8282942771911621,
      "logps/chosen": -186.79269409179688,
      "logps/ref_chosen": -77.54549407958984,
      "logps/ref_rejected": -68.58549499511719,
      "logps/rejected": -215.02297973632812,
      "loss": 1.1574,
      "margin_dpo/margin_mean": 37.190284729003906,
      "margin_dpo/margin_std": 79.62153625488281,
      "step": 251
    },
    {
      "epoch": 0.38095238095238093,
      "grad_norm": 26.73542594909668,
      "learning_rate": 3.9067390737445254e-07,
      "logits/chosen": 0.9108133316040039,
      "logits/rejected": 0.8635146021842957,
      "logps/chosen": -196.31478881835938,
      "logps/ref_chosen": -73.16785430908203,
      "logps/ref_rejected": -70.33341979980469,
      "logps/rejected": -239.6460723876953,
      "loss": 1.2713,
      "margin_dpo/margin_mean": 46.16571044921875,
      "margin_dpo/margin_std": 107.81159973144531,
      "step": 252
    },
    {
      "epoch": 0.382464096749811,
      "grad_norm": 21.395902633666992,
      "learning_rate": 3.8957891383162304e-07,
      "logits/chosen": 0.9631872177124023,
      "logits/rejected": 0.9409669637680054,
      "logps/chosen": -125.77433776855469,
      "logps/ref_chosen": -33.16981506347656,
      "logps/ref_rejected": -42.722469329833984,
      "logps/rejected": -179.7202911376953,
      "loss": 1.1379,
      "margin_dpo/margin_mean": 44.39329528808594,
      "margin_dpo/margin_std": 63.10321044921875,
      "step": 253
    },
    {
      "epoch": 0.3839758125472411,
      "grad_norm": 13.39516544342041,
      "learning_rate": 3.884800159665276e-07,
      "logits/chosen": 0.9127005934715271,
      "logits/rejected": 0.8767493367195129,
      "logps/chosen": -138.35330200195312,
      "logps/ref_chosen": -50.45591735839844,
      "logps/ref_rejected": -71.39493560791016,
      "logps/rejected": -223.3489532470703,
      "loss": 1.1077,
      "margin_dpo/margin_mean": 64.056640625,
      "margin_dpo/margin_std": 61.15374755859375,
      "step": 254
    },
    {
      "epoch": 0.3854875283446712,
      "grad_norm": 21.80673599243164,
      "learning_rate": 3.873772445177015e-07,
      "logits/chosen": 0.929595410823822,
      "logits/rejected": 0.8490023612976074,
      "logps/chosen": -134.93621826171875,
      "logps/ref_chosen": -57.36243438720703,
      "logps/ref_rejected": -83.85030364990234,
      "logps/rejected": -219.91397094726562,
      "loss": 1.0894,
      "margin_dpo/margin_mean": 58.489891052246094,
      "margin_dpo/margin_std": 79.15504455566406,
      "step": 255
    },
    {
      "epoch": 0.3869992441421013,
      "grad_norm": 17.780683517456055,
      "learning_rate": 3.862706303320329e-07,
      "logits/chosen": 0.8780696988105774,
      "logits/rejected": 0.8686283230781555,
      "logps/chosen": -191.927734375,
      "logps/ref_chosen": -72.70278930664062,
      "logps/ref_rejected": -87.58718872070312,
      "logps/rejected": -250.40631103515625,
      "loss": 1.1458,
      "margin_dpo/margin_mean": 43.59416580200195,
      "margin_dpo/margin_std": 86.399169921875,
      "step": 256
    },
    {
      "epoch": 0.3885109599395314,
      "grad_norm": 13.700689315795898,
      "learning_rate": 3.851602043638994e-07,
      "logits/chosen": 0.9361928701400757,
      "logits/rejected": 0.8859877586364746,
      "logps/chosen": -137.43472290039062,
      "logps/ref_chosen": -56.16447448730469,
      "logps/ref_rejected": -79.54229736328125,
      "logps/rejected": -244.7681427001953,
      "loss": 0.9962,
      "margin_dpo/margin_mean": 83.9555892944336,
      "margin_dpo/margin_std": 90.04832458496094,
      "step": 257
    },
    {
      "epoch": 0.3900226757369615,
      "grad_norm": 15.653840065002441,
      "learning_rate": 3.840459976743023e-07,
      "logits/chosen": 0.857273280620575,
      "logits/rejected": 0.8150414228439331,
      "logps/chosen": -171.31089782714844,
      "logps/ref_chosen": -68.2958755493164,
      "logps/ref_rejected": -88.37379455566406,
      "logps/rejected": -259.4873046875,
      "loss": 1.0813,
      "margin_dpo/margin_mean": 68.0985107421875,
      "margin_dpo/margin_std": 73.86476135253906,
      "step": 258
    },
    {
      "epoch": 0.3915343915343915,
      "grad_norm": 16.806671142578125,
      "learning_rate": 3.8292804142999796e-07,
      "logits/chosen": 0.8986009955406189,
      "logits/rejected": 0.8750613927841187,
      "logps/chosen": -197.89768981933594,
      "logps/ref_chosen": -86.00422668457031,
      "logps/ref_rejected": -106.7618408203125,
      "logps/rejected": -272.7347106933594,
      "loss": 0.9853,
      "margin_dpo/margin_mean": 54.07939910888672,
      "margin_dpo/margin_std": 119.56117248535156,
      "step": 259
    },
    {
      "epoch": 0.3930461073318216,
      "grad_norm": 22.537059783935547,
      "learning_rate": 3.818063669026256e-07,
      "logits/chosen": 0.9305140972137451,
      "logits/rejected": 0.8760488629341125,
      "logps/chosen": -117.2443618774414,
      "logps/ref_chosen": -53.69929504394531,
      "logps/ref_rejected": -80.42738342285156,
      "logps/rejected": -225.81988525390625,
      "loss": 1.0383,
      "margin_dpo/margin_mean": 81.84744262695312,
      "margin_dpo/margin_std": 77.02603149414062,
      "step": 260
    },
    {
      "epoch": 0.3945578231292517,
      "grad_norm": 16.964977264404297,
      "learning_rate": 3.806810054678331e-07,
      "logits/chosen": 0.8644047975540161,
      "logits/rejected": 1.017871379852295,
      "logps/chosen": -177.18267822265625,
      "logps/ref_chosen": -87.48385620117188,
      "logps/ref_rejected": -45.14533233642578,
      "logps/rejected": -152.22369384765625,
      "loss": 1.1648,
      "margin_dpo/margin_mean": 17.37952995300293,
      "margin_dpo/margin_std": 70.14508056640625,
      "step": 261
    },
    {
      "epoch": 0.3960695389266818,
      "grad_norm": 15.852773666381836,
      "learning_rate": 3.7955198860439887e-07,
      "logits/chosen": 1.01748526096344,
      "logits/rejected": 0.9842813014984131,
      "logps/chosen": -153.70370483398438,
      "logps/ref_chosen": -52.95308303833008,
      "logps/ref_rejected": -64.26335906982422,
      "logps/rejected": -199.62222290039062,
      "loss": 1.0934,
      "margin_dpo/margin_mean": 34.60824203491211,
      "margin_dpo/margin_std": 63.7818489074707,
      "step": 262
    },
    {
      "epoch": 0.3975812547241119,
      "grad_norm": 18.748830795288086,
      "learning_rate": 3.784193478933516e-07,
      "logits/chosen": 0.9187393188476562,
      "logits/rejected": 0.8768531084060669,
      "logps/chosen": -145.96719360351562,
      "logps/ref_chosen": -59.93905258178711,
      "logps/ref_rejected": -70.00927734375,
      "logps/rejected": -191.09591674804688,
      "loss": 1.0738,
      "margin_dpo/margin_mean": 35.05849838256836,
      "margin_dpo/margin_std": 68.49369812011719,
      "step": 263
    },
    {
      "epoch": 0.39909297052154197,
      "grad_norm": 16.476381301879883,
      "learning_rate": 3.7728311501708674e-07,
      "logits/chosen": 0.8161172866821289,
      "logits/rejected": 0.7861305475234985,
      "logps/chosen": -201.89556884765625,
      "logps/ref_chosen": -77.3072509765625,
      "logps/ref_rejected": -83.79466247558594,
      "logps/rejected": -217.04876708984375,
      "loss": 1.0982,
      "margin_dpo/margin_mean": 8.665780067443848,
      "margin_dpo/margin_std": 75.14938354492188,
      "step": 264
    },
    {
      "epoch": 0.40060468631897206,
      "grad_norm": 14.398598670959473,
      "learning_rate": 3.7614332175848027e-07,
      "logits/chosen": 0.8012948036193848,
      "logits/rejected": 0.7619481086730957,
      "logps/chosen": -183.10635375976562,
      "logps/ref_chosen": -62.340206146240234,
      "logps/ref_rejected": -57.182029724121094,
      "logps/rejected": -195.21893310546875,
      "loss": 1.0655,
      "margin_dpo/margin_mean": 17.27075958251953,
      "margin_dpo/margin_std": 70.34654235839844,
      "step": 265
    },
    {
      "epoch": 0.4021164021164021,
      "grad_norm": 19.04807472229004,
      "learning_rate": 3.75e-07,
      "logits/chosen": 0.9468996524810791,
      "logits/rejected": 0.8705604076385498,
      "logps/chosen": -128.9197235107422,
      "logps/ref_chosen": -47.8638916015625,
      "logps/ref_rejected": -65.85595703125,
      "logps/rejected": -201.5211181640625,
      "loss": 1.1209,
      "margin_dpo/margin_mean": 54.60932922363281,
      "margin_dpo/margin_std": 57.88256072998047,
      "step": 266
    },
    {
      "epoch": 0.4036281179138322,
      "grad_norm": 22.434011459350586,
      "learning_rate": 3.738531817228131e-07,
      "logits/chosen": 0.8454642295837402,
      "logits/rejected": 0.786316990852356,
      "logps/chosen": -106.54922485351562,
      "logps/ref_chosen": -43.666568756103516,
      "logps/ref_rejected": -68.1474380493164,
      "logps/rejected": -161.97854614257812,
      "loss": 1.2043,
      "margin_dpo/margin_mean": 30.94845199584961,
      "margin_dpo/margin_std": 81.37274169921875,
      "step": 267
    },
    {
      "epoch": 0.4051398337112623,
      "grad_norm": 18.147586822509766,
      "learning_rate": 3.7270289900589204e-07,
      "logits/chosen": 0.7394974231719971,
      "logits/rejected": 0.7207775115966797,
      "logps/chosen": -133.07296752929688,
      "logps/ref_chosen": -67.96279907226562,
      "logps/ref_rejected": -72.69281005859375,
      "logps/rejected": -186.4730224609375,
      "loss": 1.1032,
      "margin_dpo/margin_mean": 48.67003631591797,
      "margin_dpo/margin_std": 64.53730010986328,
      "step": 268
    },
    {
      "epoch": 0.40665154950869237,
      "grad_norm": 12.17220687866211,
      "learning_rate": 3.7154918402511714e-07,
      "logits/chosen": 0.8191482424736023,
      "logits/rejected": 0.8883626461029053,
      "logps/chosen": -182.84963989257812,
      "logps/ref_chosen": -80.52581787109375,
      "logps/ref_rejected": -65.83181762695312,
      "logps/rejected": -205.61041259765625,
      "loss": 1.0584,
      "margin_dpo/margin_mean": 37.45478820800781,
      "margin_dpo/margin_std": 63.13644027709961,
      "step": 269
    },
    {
      "epoch": 0.40816326530612246,
      "grad_norm": 18.165111541748047,
      "learning_rate": 3.7039206905237656e-07,
      "logits/chosen": 0.8216699957847595,
      "logits/rejected": 0.8250892758369446,
      "logps/chosen": -129.14108276367188,
      "logps/ref_chosen": -64.29264831542969,
      "logps/ref_rejected": -65.4633560180664,
      "logps/rejected": -201.03909301757812,
      "loss": 1.0999,
      "margin_dpo/margin_mean": 70.72732543945312,
      "margin_dpo/margin_std": 76.4561767578125,
      "step": 270
    },
    {
      "epoch": 0.40967498110355255,
      "grad_norm": 21.99321174621582,
      "learning_rate": 3.692315864546635e-07,
      "logits/chosen": 0.9689754247665405,
      "logits/rejected": 0.8169035911560059,
      "logps/chosen": -90.29707336425781,
      "logps/ref_chosen": -39.26963424682617,
      "logps/ref_rejected": -87.15721130371094,
      "logps/rejected": -214.21963500976562,
      "loss": 1.2105,
      "margin_dpo/margin_mean": 76.03497314453125,
      "margin_dpo/margin_std": 88.13627624511719,
      "step": 271
    },
    {
      "epoch": 0.41118669690098264,
      "grad_norm": 12.809111595153809,
      "learning_rate": 3.6806776869317067e-07,
      "logits/chosen": 0.9838389158248901,
      "logits/rejected": 1.008858323097229,
      "logps/chosen": -140.18511962890625,
      "logps/ref_chosen": -55.88648223876953,
      "logps/ref_rejected": -57.95124816894531,
      "logps/rejected": -215.2462158203125,
      "loss": 0.9634,
      "margin_dpo/margin_mean": 72.996337890625,
      "margin_dpo/margin_std": 73.1607666015625,
      "step": 272
    },
    {
      "epoch": 0.4126984126984127,
      "grad_norm": 17.283329010009766,
      "learning_rate": 3.669006483223828e-07,
      "logits/chosen": 0.9234431982040405,
      "logits/rejected": 0.8033227324485779,
      "logps/chosen": -175.26605224609375,
      "logps/ref_chosen": -59.144004821777344,
      "logps/ref_rejected": -123.4438247680664,
      "logps/rejected": -278.2921447753906,
      "loss": 1.14,
      "margin_dpo/margin_mean": 38.72627258300781,
      "margin_dpo/margin_std": 74.85566711425781,
      "step": 273
    },
    {
      "epoch": 0.41421012849584277,
      "grad_norm": 18.224842071533203,
      "learning_rate": 3.657302579891656e-07,
      "logits/chosen": 0.7041195631027222,
      "logits/rejected": 0.8130519986152649,
      "logps/chosen": -175.5207977294922,
      "logps/ref_chosen": -77.74801635742188,
      "logps/ref_rejected": -63.99616241455078,
      "logps/rejected": -199.80519104003906,
      "loss": 1.1015,
      "margin_dpo/margin_mean": 38.036251068115234,
      "margin_dpo/margin_std": 89.42131042480469,
      "step": 274
    },
    {
      "epoch": 0.41572184429327286,
      "grad_norm": 21.59168815612793,
      "learning_rate": 3.645566304318526e-07,
      "logits/chosen": 0.9048452377319336,
      "logits/rejected": 0.8862916231155396,
      "logps/chosen": -169.61819458007812,
      "logps/ref_chosen": -56.68327331542969,
      "logps/ref_rejected": -65.53984832763672,
      "logps/rejected": -225.29954528808594,
      "loss": 1.0305,
      "margin_dpo/margin_mean": 46.82477569580078,
      "margin_dpo/margin_std": 65.18940734863281,
      "step": 275
    },
    {
      "epoch": 0.41723356009070295,
      "grad_norm": 20.635337829589844,
      "learning_rate": 3.633797984793294e-07,
      "logits/chosen": 0.8774361610412598,
      "logits/rejected": 0.8986802101135254,
      "logps/chosen": -137.88626098632812,
      "logps/ref_chosen": -43.50504684448242,
      "logps/ref_rejected": -43.088871002197266,
      "logps/rejected": -173.4464111328125,
      "loss": 1.123,
      "margin_dpo/margin_mean": 35.97632598876953,
      "margin_dpo/margin_std": 63.101436614990234,
      "step": 276
    },
    {
      "epoch": 0.41874527588813304,
      "grad_norm": 18.922773361206055,
      "learning_rate": 3.6219979505011555e-07,
      "logits/chosen": 0.7700687646865845,
      "logits/rejected": 0.7433135509490967,
      "logps/chosen": -165.19195556640625,
      "logps/ref_chosen": -52.701934814453125,
      "logps/ref_rejected": -62.01823425292969,
      "logps/rejected": -202.63037109375,
      "loss": 1.3038,
      "margin_dpo/margin_mean": 28.122108459472656,
      "margin_dpo/margin_std": 74.96601867675781,
      "step": 277
    },
    {
      "epoch": 0.42025699168556313,
      "grad_norm": 24.745040893554688,
      "learning_rate": 3.6101665315144353e-07,
      "logits/chosen": 0.9205065965652466,
      "logits/rejected": 0.8737306594848633,
      "logps/chosen": -186.35955810546875,
      "logps/ref_chosen": -66.36759948730469,
      "logps/ref_rejected": -71.12834930419922,
      "logps/rejected": -234.69064331054688,
      "loss": 1.0695,
      "margin_dpo/margin_mean": 43.57034683227539,
      "margin_dpo/margin_std": 91.35447692871094,
      "step": 278
    },
    {
      "epoch": 0.4217687074829932,
      "grad_norm": 19.72016143798828,
      "learning_rate": 3.5983040587833563e-07,
      "logits/chosen": 0.8044095039367676,
      "logits/rejected": 0.7835577726364136,
      "logps/chosen": -125.09947967529297,
      "logps/ref_chosen": -57.34808349609375,
      "logps/ref_rejected": -60.212989807128906,
      "logps/rejected": -194.96324157714844,
      "loss": 0.942,
      "margin_dpo/margin_mean": 66.99885559082031,
      "margin_dpo/margin_std": 86.55064392089844,
      "step": 279
    },
    {
      "epoch": 0.42328042328042326,
      "grad_norm": 13.24905014038086,
      "learning_rate": 3.586410864126781e-07,
      "logits/chosen": 0.8917016983032227,
      "logits/rejected": 0.8699558973312378,
      "logps/chosen": -111.32423400878906,
      "logps/ref_chosen": -58.24922561645508,
      "logps/ref_rejected": -59.01625061035156,
      "logps/rejected": -200.49822998046875,
      "loss": 0.8917,
      "margin_dpo/margin_mean": 88.40696716308594,
      "margin_dpo/margin_std": 55.23664474487305,
      "step": 280
    },
    {
      "epoch": 0.42479213907785335,
      "grad_norm": 14.725621223449707,
      "learning_rate": 3.574487280222929e-07,
      "logits/chosen": 0.8508050441741943,
      "logits/rejected": 0.8704952597618103,
      "logps/chosen": -189.38882446289062,
      "logps/ref_chosen": -78.43274688720703,
      "logps/ref_rejected": -91.09056091308594,
      "logps/rejected": -267.9164733886719,
      "loss": 1.081,
      "margin_dpo/margin_mean": 65.86985778808594,
      "margin_dpo/margin_std": 70.21514892578125,
      "step": 281
    },
    {
      "epoch": 0.42630385487528344,
      "grad_norm": 16.900634765625,
      "learning_rate": 3.562533640600075e-07,
      "logits/chosen": 0.8684970140457153,
      "logits/rejected": 0.7901803255081177,
      "logps/chosen": -150.54010009765625,
      "logps/ref_chosen": -53.83773422241211,
      "logps/ref_rejected": -75.1729507446289,
      "logps/rejected": -234.26873779296875,
      "loss": 1.1197,
      "margin_dpo/margin_mean": 62.39341354370117,
      "margin_dpo/margin_std": 64.95208740234375,
      "step": 282
    },
    {
      "epoch": 0.42781557067271353,
      "grad_norm": 15.944957733154297,
      "learning_rate": 3.550550279627215e-07,
      "logits/chosen": 0.8715301156044006,
      "logits/rejected": 0.8303920030593872,
      "logps/chosen": -142.8111572265625,
      "logps/ref_chosen": -53.5611457824707,
      "logps/ref_rejected": -73.18958282470703,
      "logps/rejected": -201.0227813720703,
      "loss": 1.0702,
      "margin_dpo/margin_mean": 38.58319091796875,
      "margin_dpo/margin_std": 71.02609252929688,
      "step": 283
    },
    {
      "epoch": 0.4293272864701436,
      "grad_norm": 19.076061248779297,
      "learning_rate": 3.5385375325047163e-07,
      "logits/chosen": 0.9148997068405151,
      "logits/rejected": 0.8821941614151001,
      "logps/chosen": -152.02041625976562,
      "logps/ref_chosen": -55.81263732910156,
      "logps/ref_rejected": -90.23190307617188,
      "logps/rejected": -226.96029663085938,
      "loss": 1.0726,
      "margin_dpo/margin_mean": 40.52062225341797,
      "margin_dpo/margin_std": 77.91551971435547,
      "step": 284
    },
    {
      "epoch": 0.4308390022675737,
      "grad_norm": 24.047950744628906,
      "learning_rate": 3.5264957352549375e-07,
      "logits/chosen": 0.9095529913902283,
      "logits/rejected": 0.9729138016700745,
      "logps/chosen": -185.89060974121094,
      "logps/ref_chosen": -71.53235626220703,
      "logps/ref_rejected": -46.31084060668945,
      "logps/rejected": -196.288818359375,
      "loss": 1.1315,
      "margin_dpo/margin_mean": 35.61972427368164,
      "margin_dpo/margin_std": 69.28761291503906,
      "step": 285
    },
    {
      "epoch": 0.4323507180650038,
      "grad_norm": 16.382957458496094,
      "learning_rate": 3.514425224712835e-07,
      "logits/chosen": 0.9263152480125427,
      "logits/rejected": 0.9703081846237183,
      "logps/chosen": -178.46542358398438,
      "logps/ref_chosen": -56.29132080078125,
      "logps/ref_rejected": -54.583534240722656,
      "logps/rejected": -236.7250518798828,
      "loss": 0.9443,
      "margin_dpo/margin_mean": 59.96741485595703,
      "margin_dpo/margin_std": 58.34821319580078,
      "step": 286
    },
    {
      "epoch": 0.43386243386243384,
      "grad_norm": 18.149490356445312,
      "learning_rate": 3.502326338516534e-07,
      "logits/chosen": 0.9496700763702393,
      "logits/rejected": 0.798512876033783,
      "logps/chosen": -130.92161560058594,
      "logps/ref_chosen": -39.158355712890625,
      "logps/ref_rejected": -74.88023376464844,
      "logps/rejected": -248.06649780273438,
      "loss": 1.0283,
      "margin_dpo/margin_mean": 81.4229965209961,
      "margin_dpo/margin_std": 76.41511535644531,
      "step": 287
    },
    {
      "epoch": 0.43537414965986393,
      "grad_norm": 17.73630714416504,
      "learning_rate": 3.490199415097892e-07,
      "logits/chosen": 0.7672666311264038,
      "logits/rejected": 0.7976804971694946,
      "logps/chosen": -178.63949584960938,
      "logps/ref_chosen": -69.44332885742188,
      "logps/ref_rejected": -54.98228454589844,
      "logps/rejected": -200.60418701171875,
      "loss": 1.1142,
      "margin_dpo/margin_mean": 36.425743103027344,
      "margin_dpo/margin_std": 56.845184326171875,
      "step": 288
    },
    {
      "epoch": 0.436885865457294,
      "grad_norm": 16.403818130493164,
      "learning_rate": 3.4780447936730247e-07,
      "logits/chosen": 0.8484159111976624,
      "logits/rejected": 0.8917375802993774,
      "logps/chosen": -221.443359375,
      "logps/ref_chosen": -64.24480438232422,
      "logps/ref_rejected": -67.39839172363281,
      "logps/rejected": -246.82095336914062,
      "loss": 1.1333,
      "margin_dpo/margin_mean": 22.224018096923828,
      "margin_dpo/margin_std": 46.693260192871094,
      "step": 289
    },
    {
      "epoch": 0.4383975812547241,
      "grad_norm": 17.30687141418457,
      "learning_rate": 3.465862814232821e-07,
      "logits/chosen": 0.9498525261878967,
      "logits/rejected": 0.9050639867782593,
      "logps/chosen": -192.14419555664062,
      "logps/ref_chosen": -65.73394012451172,
      "logps/ref_rejected": -74.00738525390625,
      "logps/rejected": -251.295654296875,
      "loss": 1.0606,
      "margin_dpo/margin_mean": 50.87800598144531,
      "margin_dpo/margin_std": 90.56031799316406,
      "step": 290
    },
    {
      "epoch": 0.4399092970521542,
      "grad_norm": 19.28776741027832,
      "learning_rate": 3.4536538175334343e-07,
      "logits/chosen": 0.9647561311721802,
      "logits/rejected": 0.9699366092681885,
      "logps/chosen": -183.28567504882812,
      "logps/ref_chosen": -67.45711517333984,
      "logps/ref_rejected": -75.27851867675781,
      "logps/rejected": -262.3265075683594,
      "loss": 1.0179,
      "margin_dpo/margin_mean": 71.21941375732422,
      "margin_dpo/margin_std": 88.33670806884766,
      "step": 291
    },
    {
      "epoch": 0.4414210128495843,
      "grad_norm": 15.975472450256348,
      "learning_rate": 3.4414181450867465e-07,
      "logits/chosen": 0.9394838809967041,
      "logits/rejected": 0.9308391809463501,
      "logps/chosen": -167.02105712890625,
      "logps/ref_chosen": -58.86817932128906,
      "logps/ref_rejected": -64.91166687011719,
      "logps/rejected": -216.97520446777344,
      "loss": 1.1114,
      "margin_dpo/margin_mean": 43.910648345947266,
      "margin_dpo/margin_std": 96.6522216796875,
      "step": 292
    },
    {
      "epoch": 0.4429327286470144,
      "grad_norm": 14.782191276550293,
      "learning_rate": 3.4291561391508185e-07,
      "logits/chosen": 0.9456629753112793,
      "logits/rejected": 0.9435967803001404,
      "logps/chosen": -138.83485412597656,
      "logps/ref_chosen": -48.30006408691406,
      "logps/ref_rejected": -58.28700256347656,
      "logps/rejected": -214.57583618164062,
      "loss": 0.9865,
      "margin_dpo/margin_mean": 65.7540512084961,
      "margin_dpo/margin_std": 53.61219024658203,
      "step": 293
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 13.40251350402832,
      "learning_rate": 3.4168681427203153e-07,
      "logits/chosen": 0.8696566224098206,
      "logits/rejected": 0.7992103099822998,
      "logps/chosen": -201.7752685546875,
      "logps/ref_chosen": -73.32835388183594,
      "logps/ref_rejected": -100.44032287597656,
      "logps/rejected": -282.5135192871094,
      "loss": 1.0812,
      "margin_dpo/margin_mean": 53.62628936767578,
      "margin_dpo/margin_std": 67.26288604736328,
      "step": 294
    },
    {
      "epoch": 0.4459561602418745,
      "grad_norm": 20.00554656982422,
      "learning_rate": 3.4045544995169125e-07,
      "logits/chosen": 1.0336451530456543,
      "logits/rejected": 0.9124792814254761,
      "logps/chosen": -154.9568328857422,
      "logps/ref_chosen": -36.71764373779297,
      "logps/ref_rejected": -77.01786804199219,
      "logps/rejected": -266.3442687988281,
      "loss": 1.0952,
      "margin_dpo/margin_mean": 71.08719635009766,
      "margin_dpo/margin_std": 75.32640838623047,
      "step": 295
    },
    {
      "epoch": 0.4474678760393046,
      "grad_norm": 17.829896926879883,
      "learning_rate": 3.392215553979679e-07,
      "logits/chosen": 0.8829468488693237,
      "logits/rejected": 0.7915094494819641,
      "logps/chosen": -156.69998168945312,
      "logps/ref_chosen": -47.72833251953125,
      "logps/ref_rejected": -85.0424575805664,
      "logps/rejected": -267.91741943359375,
      "loss": 1.0734,
      "margin_dpo/margin_mean": 73.9033203125,
      "margin_dpo/margin_std": 85.67912292480469,
      "step": 296
    },
    {
      "epoch": 0.4489795918367347,
      "grad_norm": 13.065276145935059,
      "learning_rate": 3.3798516512554485e-07,
      "logits/chosen": 1.0190558433532715,
      "logits/rejected": 0.9399404525756836,
      "logps/chosen": -171.317626953125,
      "logps/ref_chosen": -45.453880310058594,
      "logps/ref_rejected": -60.58012390136719,
      "logps/rejected": -240.07229614257812,
      "loss": 0.9618,
      "margin_dpo/margin_mean": 53.62842559814453,
      "margin_dpo/margin_std": 68.6543197631836,
      "step": 297
    },
    {
      "epoch": 0.4504913076341648,
      "grad_norm": 14.449824333190918,
      "learning_rate": 3.367463137189156e-07,
      "logits/chosen": 1.0604900121688843,
      "logits/rejected": 1.045938491821289,
      "logps/chosen": -196.72889709472656,
      "logps/ref_chosen": -65.93342590332031,
      "logps/ref_rejected": -81.31886291503906,
      "logps/rejected": -258.4837341308594,
      "loss": 1.1329,
      "margin_dpo/margin_mean": 46.3693962097168,
      "margin_dpo/margin_std": 84.15235900878906,
      "step": 298
    },
    {
      "epoch": 0.4520030234315949,
      "grad_norm": 18.29178237915039,
      "learning_rate": 3.355050358314172e-07,
      "logits/chosen": 0.9493238925933838,
      "logits/rejected": 0.9166898727416992,
      "logps/chosen": -167.92984008789062,
      "logps/ref_chosen": -41.45861053466797,
      "logps/ref_rejected": -55.44845199584961,
      "logps/rejected": -203.72457885742188,
      "loss": 1.2108,
      "margin_dpo/margin_mean": 21.804901123046875,
      "margin_dpo/margin_std": 75.58480834960938,
      "step": 299
    },
    {
      "epoch": 0.45351473922902497,
      "grad_norm": 15.63215446472168,
      "learning_rate": 3.3426136618426043e-07,
      "logits/chosen": 0.95320063829422,
      "logits/rejected": 1.0048692226409912,
      "logps/chosen": -177.46788024902344,
      "logps/ref_chosen": -61.02867889404297,
      "logps/ref_rejected": -52.653968811035156,
      "logps/rejected": -235.43170166015625,
      "loss": 1.1015,
      "margin_dpo/margin_mean": 66.33853149414062,
      "margin_dpo/margin_std": 77.45952606201172,
      "step": 300
    },
    {
      "epoch": 0.45351473922902497,
      "eval_logits/chosen": 0.910781741142273,
      "eval_logits/rejected": 0.8756802678108215,
      "eval_logps/chosen": -200.87997436523438,
      "eval_logps/ref_chosen": -75.30646514892578,
      "eval_logps/ref_rejected": -77.75511932373047,
      "eval_logps/rejected": -251.94520568847656,
      "eval_loss": 0.557418942451477,
      "eval_margin_dpo/margin_mean": 48.61654281616211,
      "eval_margin_dpo/margin_std": 83.31050872802734,
      "eval_runtime": 37.4536,
      "eval_samples_per_second": 61.489,
      "eval_steps_per_second": 1.922,
      "step": 300
    },
    {
      "epoch": 0.455026455026455,
      "grad_norm": 14.323222160339355,
      "learning_rate": 3.3301533956555885e-07,
      "logits/chosen": 1.0346510410308838,
      "logits/rejected": 0.985994815826416,
      "logps/chosen": -206.44583129882812,
      "logps/ref_chosen": -49.377403259277344,
      "logps/ref_rejected": -72.45796966552734,
      "logps/rejected": -270.4298095703125,
      "loss": 1.1779,
      "margin_dpo/margin_mean": 40.90338897705078,
      "margin_dpo/margin_std": 61.039005279541016,
      "step": 301
    },
    {
      "epoch": 0.4565381708238851,
      "grad_norm": 18.329959869384766,
      "learning_rate": 3.317669908293554e-07,
      "logits/chosen": 0.8981478810310364,
      "logits/rejected": 0.8204025030136108,
      "logps/chosen": -166.45431518554688,
      "logps/ref_chosen": -42.884490966796875,
      "logps/ref_rejected": -72.04084777832031,
      "logps/rejected": -267.53070068359375,
      "loss": 1.2516,
      "margin_dpo/margin_mean": 71.92005920410156,
      "margin_dpo/margin_std": 84.18696594238281,
      "step": 302
    },
    {
      "epoch": 0.4580498866213152,
      "grad_norm": 13.802994728088379,
      "learning_rate": 3.3051635489464793e-07,
      "logits/chosen": 0.9232026934623718,
      "logits/rejected": 0.8720027804374695,
      "logps/chosen": -161.49925231933594,
      "logps/ref_chosen": -50.44966506958008,
      "logps/ref_rejected": -77.09220886230469,
      "logps/rejected": -234.8109893798828,
      "loss": 1.0738,
      "margin_dpo/margin_mean": 46.669189453125,
      "margin_dpo/margin_std": 89.64495849609375,
      "step": 303
    },
    {
      "epoch": 0.4595616024187453,
      "grad_norm": 13.76878833770752,
      "learning_rate": 3.292634667444117e-07,
      "logits/chosen": 1.0514378547668457,
      "logits/rejected": 0.9998120069503784,
      "logps/chosen": -169.84353637695312,
      "logps/ref_chosen": -54.7811279296875,
      "logps/ref_rejected": -74.56997680664062,
      "logps/rejected": -264.3577880859375,
      "loss": 0.9872,
      "margin_dpo/margin_mean": 74.72541809082031,
      "margin_dpo/margin_std": 65.64923858642578,
      "step": 304
    },
    {
      "epoch": 0.46107331821617537,
      "grad_norm": 18.287494659423828,
      "learning_rate": 3.280083614246217e-07,
      "logits/chosen": 0.8871467113494873,
      "logits/rejected": 0.917915940284729,
      "logps/chosen": -238.80645751953125,
      "logps/ref_chosen": -81.34001159667969,
      "logps/ref_rejected": -77.41258239746094,
      "logps/rejected": -272.7764892578125,
      "loss": 1.1496,
      "margin_dpo/margin_mean": 37.89744567871094,
      "margin_dpo/margin_std": 99.84722900390625,
      "step": 305
    },
    {
      "epoch": 0.46258503401360546,
      "grad_norm": 13.584040641784668,
      "learning_rate": 3.267510740432719e-07,
      "logits/chosen": 1.079813003540039,
      "logits/rejected": 1.0361995697021484,
      "logps/chosen": -143.89230346679688,
      "logps/ref_chosen": -49.74858474731445,
      "logps/ref_rejected": -54.940982818603516,
      "logps/rejected": -209.38763427734375,
      "loss": 1.0814,
      "margin_dpo/margin_mean": 60.302940368652344,
      "margin_dpo/margin_std": 79.73137664794922,
      "step": 306
    },
    {
      "epoch": 0.46409674981103555,
      "grad_norm": 19.44350814819336,
      "learning_rate": 3.2549163976939285e-07,
      "logits/chosen": 0.8657441139221191,
      "logits/rejected": 0.8813103437423706,
      "logps/chosen": -183.44241333007812,
      "logps/ref_chosen": -65.90791320800781,
      "logps/ref_rejected": -65.25321960449219,
      "logps/rejected": -221.45703125,
      "loss": 1.2949,
      "margin_dpo/margin_mean": 38.6693229675293,
      "margin_dpo/margin_std": 113.20942687988281,
      "step": 307
    },
    {
      "epoch": 0.4656084656084656,
      "grad_norm": 17.951297760009766,
      "learning_rate": 3.2423009383206874e-07,
      "logits/chosen": 0.9202988147735596,
      "logits/rejected": 0.9259182214736938,
      "logps/chosen": -203.8446807861328,
      "logps/ref_chosen": -71.3767318725586,
      "logps/ref_rejected": -82.10542297363281,
      "logps/rejected": -271.75042724609375,
      "loss": 1.1168,
      "margin_dpo/margin_mean": 57.17706298828125,
      "margin_dpo/margin_std": 100.58518981933594,
      "step": 308
    },
    {
      "epoch": 0.4671201814058957,
      "grad_norm": 14.301473617553711,
      "learning_rate": 3.229664715194511e-07,
      "logits/chosen": 0.9366397857666016,
      "logits/rejected": 0.913209080696106,
      "logps/chosen": -179.5953369140625,
      "logps/ref_chosen": -41.846153259277344,
      "logps/ref_rejected": -61.37134552001953,
      "logps/rejected": -237.18511962890625,
      "loss": 1.0857,
      "margin_dpo/margin_mean": 38.064598083496094,
      "margin_dpo/margin_std": 66.49103546142578,
      "step": 309
    },
    {
      "epoch": 0.46863189720332576,
      "grad_norm": 16.19812774658203,
      "learning_rate": 3.2170080817777257e-07,
      "logits/chosen": 0.969029426574707,
      "logits/rejected": 0.9913873672485352,
      "logps/chosen": -214.78329467773438,
      "logps/ref_chosen": -70.55810546875,
      "logps/ref_rejected": -64.62115478515625,
      "logps/rejected": -227.16543579101562,
      "loss": 1.2835,
      "margin_dpo/margin_mean": 18.319095611572266,
      "margin_dpo/margin_std": 73.50588989257812,
      "step": 310
    },
    {
      "epoch": 0.47014361300075586,
      "grad_norm": 16.054964065551758,
      "learning_rate": 3.204331392103574e-07,
      "logits/chosen": 0.8452152013778687,
      "logits/rejected": 0.8800424337387085,
      "logps/chosen": -181.9866485595703,
      "logps/ref_chosen": -60.32414245605469,
      "logps/ref_rejected": -68.1629638671875,
      "logps/rejected": -208.29612731933594,
      "loss": 1.0987,
      "margin_dpo/margin_mean": 18.47066307067871,
      "margin_dpo/margin_std": 71.49610137939453,
      "step": 311
    },
    {
      "epoch": 0.47165532879818595,
      "grad_norm": 13.902632713317871,
      "learning_rate": 3.1916350007663176e-07,
      "logits/chosen": 1.0027267932891846,
      "logits/rejected": 0.952292799949646,
      "logps/chosen": -169.0614013671875,
      "logps/ref_chosen": -55.58141326904297,
      "logps/ref_rejected": -71.82810974121094,
      "logps/rejected": -234.40756225585938,
      "loss": 0.9795,
      "margin_dpo/margin_mean": 49.09947967529297,
      "margin_dpo/margin_std": 47.16817855834961,
      "step": 312
    },
    {
      "epoch": 0.47316704459561604,
      "grad_norm": 15.096352577209473,
      "learning_rate": 3.178919262911314e-07,
      "logits/chosen": 0.9660900831222534,
      "logits/rejected": 1.0016883611679077,
      "logps/chosen": -156.38772583007812,
      "logps/ref_chosen": -52.92902374267578,
      "logps/ref_rejected": -47.91901779174805,
      "logps/rejected": -186.65386962890625,
      "loss": 1.2484,
      "margin_dpo/margin_mean": 35.27616500854492,
      "margin_dpo/margin_std": 60.62500762939453,
      "step": 313
    },
    {
      "epoch": 0.47467876039304613,
      "grad_norm": 15.742197036743164,
      "learning_rate": 3.166184534225087e-07,
      "logits/chosen": 0.9835371971130371,
      "logits/rejected": 0.9550716876983643,
      "logps/chosen": -169.02850341796875,
      "logps/ref_chosen": -64.4450454711914,
      "logps/ref_rejected": -76.83822631835938,
      "logps/rejected": -248.9261474609375,
      "loss": 1.0354,
      "margin_dpo/margin_mean": 67.50444793701172,
      "margin_dpo/margin_std": 74.06520080566406,
      "step": 314
    },
    {
      "epoch": 0.47619047619047616,
      "grad_norm": 13.899375915527344,
      "learning_rate": 3.1534311709253723e-07,
      "logits/chosen": 0.9853957295417786,
      "logits/rejected": 0.9791759252548218,
      "logps/chosen": -192.0851593017578,
      "logps/ref_chosen": -56.93284225463867,
      "logps/ref_rejected": -50.28406524658203,
      "logps/rejected": -199.18292236328125,
      "loss": 1.0917,
      "margin_dpo/margin_mean": 13.746543884277344,
      "margin_dpo/margin_std": 78.6737060546875,
      "step": 315
    },
    {
      "epoch": 0.47770219198790626,
      "grad_norm": 16.974472045898438,
      "learning_rate": 3.1406595297511564e-07,
      "logits/chosen": 0.7342613935470581,
      "logits/rejected": 0.5312126874923706,
      "logps/chosen": -182.5743408203125,
      "logps/ref_chosen": -66.3306884765625,
      "logps/ref_rejected": -137.49655151367188,
      "logps/rejected": -309.92919921875,
      "loss": 1.0034,
      "margin_dpo/margin_mean": 56.18898391723633,
      "margin_dpo/margin_std": 85.84481048583984,
      "step": 316
    },
    {
      "epoch": 0.47921390778533635,
      "grad_norm": 14.526897430419922,
      "learning_rate": 3.1278699679526975e-07,
      "logits/chosen": 0.9011315107345581,
      "logits/rejected": 0.8500291705131531,
      "logps/chosen": -154.86817932128906,
      "logps/ref_chosen": -42.494422912597656,
      "logps/ref_rejected": -66.68242645263672,
      "logps/rejected": -229.81800842285156,
      "loss": 0.9879,
      "margin_dpo/margin_mean": 50.761817932128906,
      "margin_dpo/margin_std": 63.989234924316406,
      "step": 317
    },
    {
      "epoch": 0.48072562358276644,
      "grad_norm": 17.438730239868164,
      "learning_rate": 3.1150628432815336e-07,
      "logits/chosen": 0.8117409944534302,
      "logits/rejected": 0.8378602266311646,
      "logps/chosen": -200.1689453125,
      "logps/ref_chosen": -80.13600158691406,
      "logps/ref_rejected": -83.40070343017578,
      "logps/rejected": -256.28057861328125,
      "loss": 1.1962,
      "margin_dpo/margin_mean": 52.8469352722168,
      "margin_dpo/margin_std": 89.18522644042969,
      "step": 318
    },
    {
      "epoch": 0.48223733938019653,
      "grad_norm": 13.24758529663086,
      "learning_rate": 3.1022385139804707e-07,
      "logits/chosen": 0.8036596179008484,
      "logits/rejected": 0.758333683013916,
      "logps/chosen": -186.25949096679688,
      "logps/ref_chosen": -83.42949676513672,
      "logps/ref_rejected": -113.88960266113281,
      "logps/rejected": -307.47576904296875,
      "loss": 1.0176,
      "margin_dpo/margin_mean": 90.75616455078125,
      "margin_dpo/margin_std": 84.7353515625,
      "step": 319
    },
    {
      "epoch": 0.4837490551776266,
      "grad_norm": 16.077190399169922,
      "learning_rate": 3.0893973387735683e-07,
      "logits/chosen": 0.8179005980491638,
      "logits/rejected": 0.7942554950714111,
      "logps/chosen": -119.91881561279297,
      "logps/ref_chosen": -34.690284729003906,
      "logps/ref_rejected": -57.00449752807617,
      "logps/rejected": -183.24380493164062,
      "loss": 1.1051,
      "margin_dpo/margin_mean": 41.01078796386719,
      "margin_dpo/margin_std": 86.33978271484375,
      "step": 320
    },
    {
      "epoch": 0.4852607709750567,
      "grad_norm": 18.350908279418945,
      "learning_rate": 3.0765396768561004e-07,
      "logits/chosen": 0.8898118734359741,
      "logits/rejected": 0.9107556343078613,
      "logps/chosen": -122.81785583496094,
      "logps/ref_chosen": -43.647361755371094,
      "logps/ref_rejected": -49.380775451660156,
      "logps/rejected": -199.43606567382812,
      "loss": 1.0686,
      "margin_dpo/margin_mean": 70.88478088378906,
      "margin_dpo/margin_std": 72.49562072753906,
      "step": 321
    },
    {
      "epoch": 0.48677248677248675,
      "grad_norm": 15.368215560913086,
      "learning_rate": 3.063665887884511e-07,
      "logits/chosen": 1.0382732152938843,
      "logits/rejected": 0.9146447777748108,
      "logps/chosen": -97.49690246582031,
      "logps/ref_chosen": -30.90003204345703,
      "logps/ref_rejected": -70.10389709472656,
      "logps/rejected": -234.061767578125,
      "loss": 0.9306,
      "margin_dpo/margin_mean": 97.36099243164062,
      "margin_dpo/margin_std": 62.040199279785156,
      "step": 322
    },
    {
      "epoch": 0.48828420256991684,
      "grad_norm": 14.468438148498535,
      "learning_rate": 3.0507763319663517e-07,
      "logits/chosen": 0.8593270778656006,
      "logits/rejected": 0.8456133008003235,
      "logps/chosen": -175.40945434570312,
      "logps/ref_chosen": -65.93765258789062,
      "logps/ref_rejected": -73.23563385009766,
      "logps/rejected": -233.8284912109375,
      "loss": 1.1484,
      "margin_dpo/margin_mean": 51.121055603027344,
      "margin_dpo/margin_std": 80.7939453125,
      "step": 323
    },
    {
      "epoch": 0.4897959183673469,
      "grad_norm": 15.986069679260254,
      "learning_rate": 3.0378713696502097e-07,
      "logits/chosen": 0.8656540513038635,
      "logits/rejected": 0.8433347940444946,
      "logps/chosen": -124.39891052246094,
      "logps/ref_chosen": -40.30308532714844,
      "logps/ref_rejected": -56.531700134277344,
      "logps/rejected": -194.01043701171875,
      "loss": 1.0246,
      "margin_dpo/margin_mean": 53.382911682128906,
      "margin_dpo/margin_std": 58.651451110839844,
      "step": 324
    },
    {
      "epoch": 0.491307634164777,
      "grad_norm": 14.863397598266602,
      "learning_rate": 3.0249513619156206e-07,
      "logits/chosen": 1.018980860710144,
      "logits/rejected": 1.0166943073272705,
      "logps/chosen": -157.23825073242188,
      "logps/ref_chosen": -52.309547424316406,
      "logps/ref_rejected": -56.243019104003906,
      "logps/rejected": -230.12759399414062,
      "loss": 1.0454,
      "margin_dpo/margin_mean": 68.95586395263672,
      "margin_dpo/margin_std": 71.14509582519531,
      "step": 325
    },
    {
      "epoch": 0.4928193499622071,
      "grad_norm": 18.127382278442383,
      "learning_rate": 3.012016670162977e-07,
      "logits/chosen": 0.9008328914642334,
      "logits/rejected": 0.8836438655853271,
      "logps/chosen": -221.15011596679688,
      "logps/ref_chosen": -74.94476318359375,
      "logps/ref_rejected": -84.40521240234375,
      "logps/rejected": -264.2186584472656,
      "loss": 1.2972,
      "margin_dpo/margin_mean": 33.60809326171875,
      "margin_dpo/margin_std": 81.20984649658203,
      "step": 326
    },
    {
      "epoch": 0.4943310657596372,
      "grad_norm": 15.47046947479248,
      "learning_rate": 2.99906765620341e-07,
      "logits/chosen": 0.7541144490242004,
      "logits/rejected": 0.7670720815658569,
      "logps/chosen": -199.9362030029297,
      "logps/ref_chosen": -75.78781127929688,
      "logps/ref_rejected": -58.949928283691406,
      "logps/rejected": -254.4146728515625,
      "loss": 1.1629,
      "margin_dpo/margin_mean": 71.31634521484375,
      "margin_dpo/margin_std": 152.7041473388672,
      "step": 327
    },
    {
      "epoch": 0.4958427815570673,
      "grad_norm": 14.062821388244629,
      "learning_rate": 2.9861046822486766e-07,
      "logits/chosen": 0.7631938457489014,
      "logits/rejected": 0.7273193597793579,
      "logps/chosen": -151.08351135253906,
      "logps/ref_chosen": -64.362060546875,
      "logps/ref_rejected": -83.52467346191406,
      "logps/rejected": -229.90032958984375,
      "loss": 1.0252,
      "margin_dpo/margin_mean": 59.65420150756836,
      "margin_dpo/margin_std": 63.522666931152344,
      "step": 328
    },
    {
      "epoch": 0.4973544973544973,
      "grad_norm": 16.59518814086914,
      "learning_rate": 2.9731281109010253e-07,
      "logits/chosen": 0.9434309005737305,
      "logits/rejected": 0.8342186212539673,
      "logps/chosen": -146.32696533203125,
      "logps/ref_chosen": -49.1827278137207,
      "logps/ref_rejected": -84.71371459960938,
      "logps/rejected": -222.47955322265625,
      "loss": 1.0562,
      "margin_dpo/margin_mean": 40.62158966064453,
      "margin_dpo/margin_std": 68.19479370117188,
      "step": 329
    },
    {
      "epoch": 0.4988662131519274,
      "grad_norm": 14.8185396194458,
      "learning_rate": 2.9601383051430505e-07,
      "logits/chosen": 0.9047882556915283,
      "logits/rejected": 0.8687087297439575,
      "logps/chosen": -178.20071411132812,
      "logps/ref_chosen": -55.316497802734375,
      "logps/ref_rejected": -65.46820831298828,
      "logps/rejected": -196.5502166748047,
      "loss": 1.1321,
      "margin_dpo/margin_mean": 8.197792053222656,
      "margin_dpo/margin_std": 85.20492553710938,
      "step": 330
    },
    {
      "epoch": 0.5003779289493575,
      "grad_norm": 15.615235328674316,
      "learning_rate": 2.947135628327544e-07,
      "logits/chosen": 0.8331134915351868,
      "logits/rejected": 0.7597118616104126,
      "logps/chosen": -159.10202026367188,
      "logps/ref_chosen": -59.31645965576172,
      "logps/ref_rejected": -85.34983825683594,
      "logps/rejected": -289.3092346191406,
      "loss": 0.957,
      "margin_dpo/margin_mean": 104.17383575439453,
      "margin_dpo/margin_std": 75.5068588256836,
      "step": 331
    },
    {
      "epoch": 0.5018896447467877,
      "grad_norm": 16.04841423034668,
      "learning_rate": 2.934120444167326e-07,
      "logits/chosen": 0.8443226218223572,
      "logits/rejected": 0.8055863976478577,
      "logps/chosen": -169.5580291748047,
      "logps/ref_chosen": -58.70336151123047,
      "logps/ref_rejected": -75.91543579101562,
      "logps/rejected": -240.45140075683594,
      "loss": 1.0368,
      "margin_dpo/margin_mean": 53.68130874633789,
      "margin_dpo/margin_std": 67.17088317871094,
      "step": 332
    },
    {
      "epoch": 0.5034013605442177,
      "grad_norm": 17.165664672851562,
      "learning_rate": 2.921093116725076e-07,
      "logits/chosen": 0.9496264457702637,
      "logits/rejected": 0.8828880190849304,
      "logps/chosen": -196.00656127929688,
      "logps/ref_chosen": -68.11222839355469,
      "logps/ref_rejected": -106.30081939697266,
      "logps/rejected": -314.76043701171875,
      "loss": 0.9578,
      "margin_dpo/margin_mean": 80.5653076171875,
      "margin_dpo/margin_std": 64.04613494873047,
      "step": 333
    },
    {
      "epoch": 0.5049130763416477,
      "grad_norm": 13.934477806091309,
      "learning_rate": 2.9080540104031484e-07,
      "logits/chosen": 0.9245538711547852,
      "logits/rejected": 0.8373509645462036,
      "logps/chosen": -169.71932983398438,
      "logps/ref_chosen": -59.10272216796875,
      "logps/ref_rejected": -81.27894592285156,
      "logps/rejected": -235.17379760742188,
      "loss": 1.1124,
      "margin_dpo/margin_mean": 43.278263092041016,
      "margin_dpo/margin_std": 132.07708740234375,
      "step": 334
    },
    {
      "epoch": 0.5064247921390779,
      "grad_norm": 18.533241271972656,
      "learning_rate": 2.895003489933375e-07,
      "logits/chosen": 1.05397367477417,
      "logits/rejected": 0.9336830973625183,
      "logps/chosen": -147.48800659179688,
      "logps/ref_chosen": -59.12438201904297,
      "logps/ref_rejected": -121.7302017211914,
      "logps/rejected": -331.6442565917969,
      "loss": 1.0951,
      "margin_dpo/margin_mean": 121.5504150390625,
      "margin_dpo/margin_std": 88.70282745361328,
      "step": 335
    },
    {
      "epoch": 0.5079365079365079,
      "grad_norm": 16.828493118286133,
      "learning_rate": 2.8819419203668675e-07,
      "logits/chosen": 0.8980951309204102,
      "logits/rejected": 0.8811999559402466,
      "logps/chosen": -182.00933837890625,
      "logps/ref_chosen": -58.688018798828125,
      "logps/ref_rejected": -89.30653381347656,
      "logps/rejected": -294.16778564453125,
      "loss": 1.0804,
      "margin_dpo/margin_mean": 81.53993225097656,
      "margin_dpo/margin_std": 102.6664047241211,
      "step": 336
    },
    {
      "epoch": 0.509448223733938,
      "grad_norm": 13.274103164672852,
      "learning_rate": 2.8688696670638053e-07,
      "logits/chosen": 0.6885286569595337,
      "logits/rejected": 0.6868454813957214,
      "logps/chosen": -265.0749206542969,
      "logps/ref_chosen": -97.68962860107422,
      "logps/ref_rejected": -91.40831756591797,
      "logps/rejected": -269.9544677734375,
      "loss": 1.1635,
      "margin_dpo/margin_mean": 11.160860061645508,
      "margin_dpo/margin_std": 96.92953491210938,
      "step": 337
    },
    {
      "epoch": 0.5109599395313681,
      "grad_norm": 14.477042198181152,
      "learning_rate": 2.8557870956832133e-07,
      "logits/chosen": 0.8425217866897583,
      "logits/rejected": 0.7798970937728882,
      "logps/chosen": -184.1267852783203,
      "logps/ref_chosen": -73.13327026367188,
      "logps/ref_rejected": -104.0283432006836,
      "logps/rejected": -278.3711853027344,
      "loss": 1.1303,
      "margin_dpo/margin_mean": 63.349342346191406,
      "margin_dpo/margin_std": 92.2508544921875,
      "step": 338
    },
    {
      "epoch": 0.5124716553287982,
      "grad_norm": 17.419946670532227,
      "learning_rate": 2.842694572172736e-07,
      "logits/chosen": 0.9565955400466919,
      "logits/rejected": 0.8105146884918213,
      "logps/chosen": -125.23301696777344,
      "logps/ref_chosen": -27.726638793945312,
      "logps/ref_rejected": -54.045658111572266,
      "logps/rejected": -183.23825073242188,
      "loss": 1.0861,
      "margin_dpo/margin_mean": 31.686233520507812,
      "margin_dpo/margin_std": 49.92768859863281,
      "step": 339
    },
    {
      "epoch": 0.5139833711262283,
      "grad_norm": 18.45969009399414,
      "learning_rate": 2.8295924627584004e-07,
      "logits/chosen": 0.9575048685073853,
      "logits/rejected": 0.8758723735809326,
      "logps/chosen": -149.28018188476562,
      "logps/ref_chosen": -37.378753662109375,
      "logps/ref_rejected": -67.66883850097656,
      "logps/rejected": -262.060546875,
      "loss": 1.1941,
      "margin_dpo/margin_mean": 82.4902572631836,
      "margin_dpo/margin_std": 83.99423217773438,
      "step": 340
    },
    {
      "epoch": 0.5154950869236583,
      "grad_norm": 17.296634674072266,
      "learning_rate": 2.816481133934373e-07,
      "logits/chosen": 1.0063215494155884,
      "logits/rejected": 0.9997066259384155,
      "logps/chosen": -138.48793029785156,
      "logps/ref_chosen": -42.783775329589844,
      "logps/ref_rejected": -59.344329833984375,
      "logps/rejected": -233.50625610351562,
      "loss": 1.0223,
      "margin_dpo/margin_mean": 78.457763671875,
      "margin_dpo/margin_std": 88.61229705810547,
      "step": 341
    },
    {
      "epoch": 0.5170068027210885,
      "grad_norm": 14.958573341369629,
      "learning_rate": 2.8033609524527046e-07,
      "logits/chosen": 0.9206830859184265,
      "logits/rejected": 0.9482388496398926,
      "logps/chosen": -197.25350952148438,
      "logps/ref_chosen": -72.35289764404297,
      "logps/ref_rejected": -63.26990509033203,
      "logps/rejected": -223.66114807128906,
      "loss": 1.0141,
      "margin_dpo/margin_mean": 35.490623474121094,
      "margin_dpo/margin_std": 68.44136810302734,
      "step": 342
    },
    {
      "epoch": 0.5185185185185185,
      "grad_norm": 14.569660186767578,
      "learning_rate": 2.7902322853130753e-07,
      "logits/chosen": 0.8026296496391296,
      "logits/rejected": 0.8457683324813843,
      "logps/chosen": -219.88909912109375,
      "logps/ref_chosen": -83.87641906738281,
      "logps/ref_rejected": -75.55497741699219,
      "logps/rejected": -238.33535766601562,
      "loss": 1.1457,
      "margin_dpo/margin_mean": 26.76772689819336,
      "margin_dpo/margin_std": 85.9823226928711,
      "step": 343
    },
    {
      "epoch": 0.5200302343159486,
      "grad_norm": 16.628511428833008,
      "learning_rate": 2.7770954997525274e-07,
      "logits/chosen": 1.025362491607666,
      "logits/rejected": 0.8970023393630981,
      "logps/chosen": -157.8984832763672,
      "logps/ref_chosen": -35.154476165771484,
      "logps/ref_rejected": -77.97383880615234,
      "logps/rejected": -260.02105712890625,
      "loss": 1.0538,
      "margin_dpo/margin_mean": 59.303184509277344,
      "margin_dpo/margin_std": 57.50261306762695,
      "step": 344
    },
    {
      "epoch": 0.5215419501133787,
      "grad_norm": 18.29877471923828,
      "learning_rate": 2.7639509632351927e-07,
      "logits/chosen": 0.9883379936218262,
      "logits/rejected": 0.9794098734855652,
      "logps/chosen": -154.01934814453125,
      "logps/ref_chosen": -39.99463653564453,
      "logps/ref_rejected": -52.60383224487305,
      "logps/rejected": -210.22317504882812,
      "loss": 1.0774,
      "margin_dpo/margin_mean": 43.594635009765625,
      "margin_dpo/margin_std": 90.1194076538086,
      "step": 345
    },
    {
      "epoch": 0.5230536659108088,
      "grad_norm": 14.863842964172363,
      "learning_rate": 2.7507990434420123e-07,
      "logits/chosen": 0.9803563356399536,
      "logits/rejected": 0.9136096239089966,
      "logps/chosen": -155.381103515625,
      "logps/ref_chosen": -61.2567024230957,
      "logps/ref_rejected": -102.48171997070312,
      "logps/rejected": -321.823974609375,
      "loss": 0.9885,
      "margin_dpo/margin_mean": 125.21788787841797,
      "margin_dpo/margin_std": 68.31744384765625,
      "step": 346
    },
    {
      "epoch": 0.5245653817082389,
      "grad_norm": 16.280048370361328,
      "learning_rate": 2.737640108260456e-07,
      "logits/chosen": 1.080568790435791,
      "logits/rejected": 1.0347011089324951,
      "logps/chosen": -183.25091552734375,
      "logps/ref_chosen": -58.63034439086914,
      "logps/ref_rejected": -79.94859313964844,
      "logps/rejected": -263.5897216796875,
      "loss": 1.058,
      "margin_dpo/margin_mean": 59.020530700683594,
      "margin_dpo/margin_std": 88.26313781738281,
      "step": 347
    },
    {
      "epoch": 0.5260770975056689,
      "grad_norm": 18.49781036376953,
      "learning_rate": 2.724474525774229e-07,
      "logits/chosen": 0.9393756985664368,
      "logits/rejected": 0.9025633335113525,
      "logps/chosen": -187.76321411132812,
      "logps/ref_chosen": -72.25175476074219,
      "logps/ref_rejected": -95.0661392211914,
      "logps/rejected": -275.28118896484375,
      "loss": 1.0541,
      "margin_dpo/margin_mean": 64.70359802246094,
      "margin_dpo/margin_std": 109.8628158569336,
      "step": 348
    },
    {
      "epoch": 0.527588813303099,
      "grad_norm": 15.981618881225586,
      "learning_rate": 2.711302664252973e-07,
      "logits/chosen": 0.9622572660446167,
      "logits/rejected": 0.8644802570343018,
      "logps/chosen": -134.60650634765625,
      "logps/ref_chosen": -34.93451690673828,
      "logps/ref_rejected": -71.41903686523438,
      "logps/rejected": -250.6934814453125,
      "loss": 1.0037,
      "margin_dpo/margin_mean": 79.60247039794922,
      "margin_dpo/margin_std": 61.48447799682617,
      "step": 349
    },
    {
      "epoch": 0.5291005291005291,
      "grad_norm": 17.303754806518555,
      "learning_rate": 2.698124892141971e-07,
      "logits/chosen": 0.9259968996047974,
      "logits/rejected": 0.8843666911125183,
      "logps/chosen": -206.68112182617188,
      "logps/ref_chosen": -71.93693542480469,
      "logps/ref_rejected": -97.71165466308594,
      "logps/rejected": -320.2801513671875,
      "loss": 0.8634,
      "margin_dpo/margin_mean": 87.82434844970703,
      "margin_dpo/margin_std": 68.40359497070312,
      "step": 350
    },
    {
      "epoch": 0.5306122448979592,
      "grad_norm": 15.037851333618164,
      "learning_rate": 2.6849415780518357e-07,
      "logits/chosen": 0.8401812314987183,
      "logits/rejected": 0.6766891479492188,
      "logps/chosen": -163.47976684570312,
      "logps/ref_chosen": -55.08075714111328,
      "logps/ref_rejected": -107.10870361328125,
      "logps/rejected": -288.9677734375,
      "loss": 1.092,
      "margin_dpo/margin_mean": 73.46005249023438,
      "margin_dpo/margin_std": 82.07776641845703,
      "step": 351
    },
    {
      "epoch": 0.5321239606953893,
      "grad_norm": 19.045122146606445,
      "learning_rate": 2.6717530907482024e-07,
      "logits/chosen": 0.8240054845809937,
      "logits/rejected": 0.8280425071716309,
      "logps/chosen": -186.6539306640625,
      "logps/ref_chosen": -75.67378234863281,
      "logps/ref_rejected": -88.66374206542969,
      "logps/rejected": -267.3773193359375,
      "loss": 1.0297,
      "margin_dpo/margin_mean": 67.73342895507812,
      "margin_dpo/margin_std": 94.85681915283203,
      "step": 352
    },
    {
      "epoch": 0.5336356764928194,
      "grad_norm": 16.103891372680664,
      "learning_rate": 2.658559799141411e-07,
      "logits/chosen": 0.9016702175140381,
      "logits/rejected": 0.7808640003204346,
      "logps/chosen": -195.90447998046875,
      "logps/ref_chosen": -62.94065856933594,
      "logps/ref_rejected": -104.08489990234375,
      "logps/rejected": -294.2989807128906,
      "loss": 1.082,
      "margin_dpo/margin_mean": 57.25025177001953,
      "margin_dpo/margin_std": 92.04182434082031,
      "step": 353
    },
    {
      "epoch": 0.5351473922902494,
      "grad_norm": 18.149988174438477,
      "learning_rate": 2.6453620722761895e-07,
      "logits/chosen": 1.0865049362182617,
      "logits/rejected": 1.0457913875579834,
      "logps/chosen": -135.3533172607422,
      "logps/ref_chosen": -28.847824096679688,
      "logps/ref_rejected": -53.78091812133789,
      "logps/rejected": -229.83432006835938,
      "loss": 1.0416,
      "margin_dpo/margin_mean": 69.54791259765625,
      "margin_dpo/margin_std": 84.38679504394531,
      "step": 354
    },
    {
      "epoch": 0.5366591080876795,
      "grad_norm": 17.005126953125,
      "learning_rate": 2.632160279321328e-07,
      "logits/chosen": 1.035976767539978,
      "logits/rejected": 0.9110543727874756,
      "logps/chosen": -159.94003295898438,
      "logps/ref_chosen": -53.094722747802734,
      "logps/ref_rejected": -91.13424682617188,
      "logps/rejected": -290.590087890625,
      "loss": 0.9458,
      "margin_dpo/margin_mean": 92.61053466796875,
      "margin_dpo/margin_std": 103.92430114746094,
      "step": 355
    },
    {
      "epoch": 0.5381708238851096,
      "grad_norm": 15.574033737182617,
      "learning_rate": 2.618954789559356e-07,
      "logits/chosen": 0.9952447414398193,
      "logits/rejected": 0.85740727186203,
      "logps/chosen": -137.01632690429688,
      "logps/ref_chosen": -34.362483978271484,
      "logps/ref_rejected": -77.31940460205078,
      "logps/rejected": -252.238037109375,
      "loss": 1.0505,
      "margin_dpo/margin_mean": 72.2647933959961,
      "margin_dpo/margin_std": 116.42106628417969,
      "step": 356
    },
    {
      "epoch": 0.5396825396825397,
      "grad_norm": 18.58148765563965,
      "learning_rate": 2.6057459723762076e-07,
      "logits/chosen": 0.9232524633407593,
      "logits/rejected": 0.8515598773956299,
      "logps/chosen": -193.28883361816406,
      "logps/ref_chosen": -64.16845703125,
      "logps/ref_rejected": -78.76988983154297,
      "logps/rejected": -273.12249755859375,
      "loss": 1.1607,
      "margin_dpo/margin_mean": 65.23223114013672,
      "margin_dpo/margin_std": 106.97843933105469,
      "step": 357
    },
    {
      "epoch": 0.5411942554799698,
      "grad_norm": 25.09504508972168,
      "learning_rate": 2.5925341972508954e-07,
      "logits/chosen": 0.8110201954841614,
      "logits/rejected": 0.862175464630127,
      "logps/chosen": -185.81414794921875,
      "logps/ref_chosen": -64.39706420898438,
      "logps/ref_rejected": -56.678443908691406,
      "logps/rejected": -252.69879150390625,
      "loss": 1.107,
      "margin_dpo/margin_mean": 74.60326385498047,
      "margin_dpo/margin_std": 79.77681732177734,
      "step": 358
    },
    {
      "epoch": 0.5427059712773998,
      "grad_norm": 27.411643981933594,
      "learning_rate": 2.579319833745169e-07,
      "logits/chosen": 0.9938050508499146,
      "logits/rejected": 1.0131170749664307,
      "logps/chosen": -222.49644470214844,
      "logps/ref_chosen": -71.20832824707031,
      "logps/ref_rejected": -75.58880615234375,
      "logps/rejected": -236.30592346191406,
      "loss": 1.3531,
      "margin_dpo/margin_mean": 9.428986549377441,
      "margin_dpo/margin_std": 102.51341247558594,
      "step": 359
    },
    {
      "epoch": 0.54421768707483,
      "grad_norm": 16.536500930786133,
      "learning_rate": 2.5661032514931834e-07,
      "logits/chosen": 0.8350504636764526,
      "logits/rejected": 0.687712550163269,
      "logps/chosen": -199.9302520751953,
      "logps/ref_chosen": -65.89573669433594,
      "logps/ref_rejected": -97.73664855957031,
      "logps/rejected": -276.3427734375,
      "loss": 1.0164,
      "margin_dpo/margin_mean": 44.57158660888672,
      "margin_dpo/margin_std": 97.77774047851562,
      "step": 360
    },
    {
      "epoch": 0.54572940287226,
      "grad_norm": 15.011344909667969,
      "learning_rate": 2.552884820191154e-07,
      "logits/chosen": 0.9806392192840576,
      "logits/rejected": 0.9534279108047485,
      "logps/chosen": -175.77911376953125,
      "logps/ref_chosen": -50.53264617919922,
      "logps/ref_rejected": -59.25585174560547,
      "logps/rejected": -233.24551391601562,
      "loss": 1.0024,
      "margin_dpo/margin_mean": 48.743194580078125,
      "margin_dpo/margin_std": 77.80059051513672,
      "step": 361
    },
    {
      "epoch": 0.54724111866969,
      "grad_norm": 16.17755699157715,
      "learning_rate": 2.53966490958702e-07,
      "logits/chosen": 1.0447218418121338,
      "logits/rejected": 0.9513689279556274,
      "logps/chosen": -221.77896118164062,
      "logps/ref_chosen": -58.92408752441406,
      "logps/ref_rejected": -104.06151580810547,
      "logps/rejected": -342.5570373535156,
      "loss": 0.971,
      "margin_dpo/margin_mean": 75.64065551757812,
      "margin_dpo/margin_std": 177.78872680664062,
      "step": 362
    },
    {
      "epoch": 0.5487528344671202,
      "grad_norm": 21.469377517700195,
      "learning_rate": 2.526443889470099e-07,
      "logits/chosen": 0.9666943550109863,
      "logits/rejected": 0.7734654545783997,
      "logps/chosen": -171.7001495361328,
      "logps/ref_chosen": -46.72846984863281,
      "logps/ref_rejected": -140.4446258544922,
      "logps/rejected": -366.104736328125,
      "loss": 0.9585,
      "margin_dpo/margin_mean": 100.68843078613281,
      "margin_dpo/margin_std": 130.09930419921875,
      "step": 363
    },
    {
      "epoch": 0.5502645502645502,
      "grad_norm": 16.017593383789062,
      "learning_rate": 2.513222129660744e-07,
      "logits/chosen": 0.9620314240455627,
      "logits/rejected": 0.8384478688240051,
      "logps/chosen": -144.81683349609375,
      "logps/ref_chosen": -47.71454620361328,
      "logps/ref_rejected": -85.33769226074219,
      "logps/rejected": -307.9134521484375,
      "loss": 0.9883,
      "margin_dpo/margin_mean": 125.4734878540039,
      "margin_dpo/margin_std": 125.29339599609375,
      "step": 364
    },
    {
      "epoch": 0.5517762660619804,
      "grad_norm": 16.882993698120117,
      "learning_rate": 2.5e-07,
      "logits/chosen": 0.9969468116760254,
      "logits/rejected": 1.0294103622436523,
      "logps/chosen": -152.65603637695312,
      "logps/ref_chosen": -53.76380157470703,
      "logps/ref_rejected": -46.24406433105469,
      "logps/rejected": -212.58950805664062,
      "loss": 0.9774,
      "margin_dpo/margin_mean": 67.45320892333984,
      "margin_dpo/margin_std": 76.16273498535156,
      "step": 365
    },
    {
      "epoch": 0.5532879818594104,
      "grad_norm": 19.378551483154297,
      "learning_rate": 2.486777870339255e-07,
      "logits/chosen": 0.9522498846054077,
      "logits/rejected": 0.9834457635879517,
      "logps/chosen": -179.36004638671875,
      "logps/ref_chosen": -67.52264404296875,
      "logps/ref_rejected": -70.28094482421875,
      "logps/rejected": -228.3491973876953,
      "loss": 1.0973,
      "margin_dpo/margin_mean": 46.230857849121094,
      "margin_dpo/margin_std": 88.74593353271484,
      "step": 366
    },
    {
      "epoch": 0.5547996976568406,
      "grad_norm": 18.30280876159668,
      "learning_rate": 2.4735561105299014e-07,
      "logits/chosen": 0.9707492589950562,
      "logits/rejected": 0.8614064455032349,
      "logps/chosen": -161.2140655517578,
      "logps/ref_chosen": -55.156681060791016,
      "logps/ref_rejected": -82.34903717041016,
      "logps/rejected": -260.894775390625,
      "loss": 1.0604,
      "margin_dpo/margin_mean": 72.48836517333984,
      "margin_dpo/margin_std": 68.52483367919922,
      "step": 367
    },
    {
      "epoch": 0.5563114134542706,
      "grad_norm": 19.369901657104492,
      "learning_rate": 2.46033509041298e-07,
      "logits/chosen": 0.7760097980499268,
      "logits/rejected": 0.8570929765701294,
      "logps/chosen": -175.87571716308594,
      "logps/ref_chosen": -74.56654357910156,
      "logps/ref_rejected": -55.081199645996094,
      "logps/rejected": -217.58401489257812,
      "loss": 1.137,
      "margin_dpo/margin_mean": 61.19365692138672,
      "margin_dpo/margin_std": 94.93489074707031,
      "step": 368
    },
    {
      "epoch": 0.5578231292517006,
      "grad_norm": 24.349010467529297,
      "learning_rate": 2.447115179808846e-07,
      "logits/chosen": 0.9876595735549927,
      "logits/rejected": 0.9113900661468506,
      "logps/chosen": -190.04141235351562,
      "logps/ref_chosen": -61.67764663696289,
      "logps/ref_rejected": -88.59959411621094,
      "logps/rejected": -285.03338623046875,
      "loss": 1.1874,
      "margin_dpo/margin_mean": 68.07002258300781,
      "margin_dpo/margin_std": 85.15034484863281,
      "step": 369
    },
    {
      "epoch": 0.5593348450491308,
      "grad_norm": 16.110780715942383,
      "learning_rate": 2.4338967485068164e-07,
      "logits/chosen": 1.0228030681610107,
      "logits/rejected": 0.9811384677886963,
      "logps/chosen": -119.4067153930664,
      "logps/ref_chosen": -44.50119400024414,
      "logps/ref_rejected": -81.18331909179688,
      "logps/rejected": -273.24188232421875,
      "loss": 0.9948,
      "margin_dpo/margin_mean": 117.15304565429688,
      "margin_dpo/margin_std": 106.01179504394531,
      "step": 370
    },
    {
      "epoch": 0.5608465608465608,
      "grad_norm": 21.480369567871094,
      "learning_rate": 2.420680166254831e-07,
      "logits/chosen": 1.0966333150863647,
      "logits/rejected": 1.0227328538894653,
      "logps/chosen": -148.44422912597656,
      "logps/ref_chosen": -39.17439651489258,
      "logps/ref_rejected": -76.12638092041016,
      "logps/rejected": -258.49078369140625,
      "loss": 1.0422,
      "margin_dpo/margin_mean": 73.09457397460938,
      "margin_dpo/margin_std": 105.94315338134766,
      "step": 371
    },
    {
      "epoch": 0.562358276643991,
      "grad_norm": 20.347423553466797,
      "learning_rate": 2.4074658027491044e-07,
      "logits/chosen": 0.9855044484138489,
      "logits/rejected": 0.8605490922927856,
      "logps/chosen": -152.38363647460938,
      "logps/ref_chosen": -52.27345275878906,
      "logps/ref_rejected": -87.656494140625,
      "logps/rejected": -296.9883117675781,
      "loss": 1.2809,
      "margin_dpo/margin_mean": 109.22164916992188,
      "margin_dpo/margin_std": 77.62582397460938,
      "step": 372
    },
    {
      "epoch": 0.563869992441421,
      "grad_norm": 17.717870712280273,
      "learning_rate": 2.394254027623792e-07,
      "logits/chosen": 1.0428986549377441,
      "logits/rejected": 1.0207685232162476,
      "logps/chosen": -223.8621368408203,
      "logps/ref_chosen": -68.01244354248047,
      "logps/ref_rejected": -64.4259033203125,
      "logps/rejected": -273.6564636230469,
      "loss": 1.1361,
      "margin_dpo/margin_mean": 53.3808708190918,
      "margin_dpo/margin_std": 147.9482421875,
      "step": 373
    },
    {
      "epoch": 0.5653817082388511,
      "grad_norm": 27.212177276611328,
      "learning_rate": 2.381045210440644e-07,
      "logits/chosen": 0.8525890111923218,
      "logits/rejected": 0.7748229503631592,
      "logps/chosen": -152.43893432617188,
      "logps/ref_chosen": -56.639495849609375,
      "logps/ref_rejected": -83.10781860351562,
      "logps/rejected": -317.96673583984375,
      "loss": 1.0262,
      "margin_dpo/margin_mean": 139.05946350097656,
      "margin_dpo/margin_std": 114.1051025390625,
      "step": 374
    },
    {
      "epoch": 0.5668934240362812,
      "grad_norm": 21.642242431640625,
      "learning_rate": 2.3678397206786715e-07,
      "logits/chosen": 1.0544872283935547,
      "logits/rejected": 0.9103308320045471,
      "logps/chosen": -140.70907592773438,
      "logps/ref_chosen": -31.620290756225586,
      "logps/ref_rejected": -70.079345703125,
      "logps/rejected": -234.93954467773438,
      "loss": 1.1789,
      "margin_dpo/margin_mean": 55.77141189575195,
      "margin_dpo/margin_std": 87.87168884277344,
      "step": 375
    },
    {
      "epoch": 0.5684051398337112,
      "grad_norm": 18.192161560058594,
      "learning_rate": 2.3546379277238103e-07,
      "logits/chosen": 0.788690447807312,
      "logits/rejected": 0.8902658224105835,
      "logps/chosen": -237.19358825683594,
      "logps/ref_chosen": -78.47712707519531,
      "logps/ref_rejected": -57.03622055053711,
      "logps/rejected": -222.91226196289062,
      "loss": 1.0841,
      "margin_dpo/margin_mean": 7.1595916748046875,
      "margin_dpo/margin_std": 78.5771255493164,
      "step": 376
    },
    {
      "epoch": 0.5699168556311414,
      "grad_norm": 16.722244262695312,
      "learning_rate": 2.3414402008585886e-07,
      "logits/chosen": 0.9365607500076294,
      "logits/rejected": 0.937558650970459,
      "logps/chosen": -185.59255981445312,
      "logps/ref_chosen": -60.769771575927734,
      "logps/ref_rejected": -65.29888916015625,
      "logps/rejected": -243.30319213867188,
      "loss": 1.1232,
      "margin_dpo/margin_mean": 53.181495666503906,
      "margin_dpo/margin_std": 76.733642578125,
      "step": 377
    },
    {
      "epoch": 0.5714285714285714,
      "grad_norm": 18.274368286132812,
      "learning_rate": 2.3282469092517977e-07,
      "logits/chosen": 1.0017985105514526,
      "logits/rejected": 0.944943904876709,
      "logps/chosen": -159.78944396972656,
      "logps/ref_chosen": -50.79759979248047,
      "logps/ref_rejected": -75.34347534179688,
      "logps/rejected": -254.2080841064453,
      "loss": 1.1378,
      "margin_dpo/margin_mean": 69.87276458740234,
      "margin_dpo/margin_std": 101.12820434570312,
      "step": 378
    },
    {
      "epoch": 0.5729402872260015,
      "grad_norm": 17.872270584106445,
      "learning_rate": 2.3150584219481643e-07,
      "logits/chosen": 0.9760261178016663,
      "logits/rejected": 0.9050667881965637,
      "logps/chosen": -185.8768310546875,
      "logps/ref_chosen": -61.738677978515625,
      "logps/ref_rejected": -97.11418151855469,
      "logps/rejected": -275.95050048828125,
      "loss": 1.0379,
      "margin_dpo/margin_mean": 54.698150634765625,
      "margin_dpo/margin_std": 97.39860534667969,
      "step": 379
    },
    {
      "epoch": 0.5744520030234316,
      "grad_norm": 15.110595703125,
      "learning_rate": 2.3018751078580283e-07,
      "logits/chosen": 0.8375756740570068,
      "logits/rejected": 0.9170357584953308,
      "logps/chosen": -169.6956329345703,
      "logps/ref_chosen": -44.782066345214844,
      "logps/ref_rejected": -41.68242263793945,
      "logps/rejected": -221.0014190673828,
      "loss": 0.9764,
      "margin_dpo/margin_mean": 54.40543746948242,
      "margin_dpo/margin_std": 82.51588439941406,
      "step": 380
    },
    {
      "epoch": 0.5759637188208617,
      "grad_norm": 19.764453887939453,
      "learning_rate": 2.288697335747027e-07,
      "logits/chosen": 0.975003719329834,
      "logits/rejected": 0.9306457042694092,
      "logps/chosen": -211.8162384033203,
      "logps/ref_chosen": -59.876434326171875,
      "logps/ref_rejected": -56.535682678222656,
      "logps/rejected": -277.5439147949219,
      "loss": 1.3201,
      "margin_dpo/margin_mean": 69.06842041015625,
      "margin_dpo/margin_std": 162.8582763671875,
      "step": 381
    },
    {
      "epoch": 0.5774754346182918,
      "grad_norm": 14.49204158782959,
      "learning_rate": 2.2755254742257706e-07,
      "logits/chosen": 1.0008131265640259,
      "logits/rejected": 0.9001951217651367,
      "logps/chosen": -203.72560119628906,
      "logps/ref_chosen": -63.76511764526367,
      "logps/ref_rejected": -100.84956359863281,
      "logps/rejected": -320.87860107421875,
      "loss": 1.0773,
      "margin_dpo/margin_mean": 80.06857299804688,
      "margin_dpo/margin_std": 91.30355072021484,
      "step": 382
    },
    {
      "epoch": 0.5789871504157218,
      "grad_norm": 20.305614471435547,
      "learning_rate": 2.2623598917395436e-07,
      "logits/chosen": 0.9180213809013367,
      "logits/rejected": 0.8855539560317993,
      "logps/chosen": -171.7388458251953,
      "logps/ref_chosen": -56.75093078613281,
      "logps/ref_rejected": -63.029056549072266,
      "logps/rejected": -220.02169799804688,
      "loss": 1.1631,
      "margin_dpo/margin_mean": 42.00471115112305,
      "margin_dpo/margin_std": 75.70281982421875,
      "step": 383
    },
    {
      "epoch": 0.5804988662131519,
      "grad_norm": 19.855918884277344,
      "learning_rate": 2.2492009565579875e-07,
      "logits/chosen": 0.8351778984069824,
      "logits/rejected": 0.8428291082382202,
      "logps/chosen": -205.7308807373047,
      "logps/ref_chosen": -79.68529510498047,
      "logps/ref_rejected": -81.70601654052734,
      "logps/rejected": -268.79345703125,
      "loss": 1.0003,
      "margin_dpo/margin_mean": 61.041873931884766,
      "margin_dpo/margin_std": 60.530391693115234,
      "step": 384
    },
    {
      "epoch": 0.582010582010582,
      "grad_norm": 15.318973541259766,
      "learning_rate": 2.2360490367648084e-07,
      "logits/chosen": 0.8146077394485474,
      "logits/rejected": 0.808496356010437,
      "logps/chosen": -191.92584228515625,
      "logps/ref_chosen": -55.45124053955078,
      "logps/ref_rejected": -74.28287506103516,
      "logps/rejected": -266.0337219238281,
      "loss": 0.9743,
      "margin_dpo/margin_mean": 55.27625274658203,
      "margin_dpo/margin_std": 67.60725402832031,
      "step": 385
    },
    {
      "epoch": 0.5835222978080121,
      "grad_norm": 16.764909744262695,
      "learning_rate": 2.2229045002474724e-07,
      "logits/chosen": 0.8237930536270142,
      "logits/rejected": 0.7379618287086487,
      "logps/chosen": -165.93934631347656,
      "logps/ref_chosen": -49.65403747558594,
      "logps/ref_rejected": -81.07264709472656,
      "logps/rejected": -254.802978515625,
      "loss": 1.1167,
      "margin_dpo/margin_mean": 57.445030212402344,
      "margin_dpo/margin_std": 128.3809814453125,
      "step": 386
    },
    {
      "epoch": 0.5850340136054422,
      "grad_norm": 20.964162826538086,
      "learning_rate": 2.209767714686924e-07,
      "logits/chosen": 1.0351340770721436,
      "logits/rejected": 0.93593430519104,
      "logps/chosen": -142.808349609375,
      "logps/ref_chosen": -30.506126403808594,
      "logps/ref_rejected": -75.85283660888672,
      "logps/rejected": -261.2908935546875,
      "loss": 0.9562,
      "margin_dpo/margin_mean": 73.13583374023438,
      "margin_dpo/margin_std": 91.49317932128906,
      "step": 387
    },
    {
      "epoch": 0.5865457294028723,
      "grad_norm": 16.960439682006836,
      "learning_rate": 2.1966390475472954e-07,
      "logits/chosen": 0.8540966510772705,
      "logits/rejected": 0.7713406085968018,
      "logps/chosen": -190.65914916992188,
      "logps/ref_chosen": -74.6607437133789,
      "logps/ref_rejected": -112.83131408691406,
      "logps/rejected": -320.3790588378906,
      "loss": 1.1545,
      "margin_dpo/margin_mean": 91.54933166503906,
      "margin_dpo/margin_std": 84.64852142333984,
      "step": 388
    },
    {
      "epoch": 0.5880574452003023,
      "grad_norm": 28.818973541259766,
      "learning_rate": 2.1835188660656265e-07,
      "logits/chosen": 0.9836728572845459,
      "logits/rejected": 0.8946930170059204,
      "logps/chosen": -191.19837951660156,
      "logps/ref_chosen": -57.50859069824219,
      "logps/ref_rejected": -99.1073989868164,
      "logps/rejected": -297.79901123046875,
      "loss": 1.0733,
      "margin_dpo/margin_mean": 65.00183868408203,
      "margin_dpo/margin_std": 77.60054016113281,
      "step": 389
    },
    {
      "epoch": 0.5895691609977324,
      "grad_norm": 14.67129135131836,
      "learning_rate": 2.170407537241599e-07,
      "logits/chosen": 0.8287357687950134,
      "logits/rejected": 0.8107761144638062,
      "logps/chosen": -167.33096313476562,
      "logps/ref_chosen": -48.60906982421875,
      "logps/ref_rejected": -54.462955474853516,
      "logps/rejected": -222.6549530029297,
      "loss": 1.0588,
      "margin_dpo/margin_mean": 49.4700927734375,
      "margin_dpo/margin_std": 65.86190795898438,
      "step": 390
    },
    {
      "epoch": 0.5910808767951625,
      "grad_norm": 15.129668235778809,
      "learning_rate": 2.1573054278272636e-07,
      "logits/chosen": 0.8232520818710327,
      "logits/rejected": 0.7771520018577576,
      "logps/chosen": -169.54945373535156,
      "logps/ref_chosen": -64.54489135742188,
      "logps/ref_rejected": -90.6060791015625,
      "logps/rejected": -272.054931640625,
      "loss": 1.0709,
      "margin_dpo/margin_mean": 76.44432067871094,
      "margin_dpo/margin_std": 110.20875549316406,
      "step": 391
    },
    {
      "epoch": 0.5925925925925926,
      "grad_norm": 15.302267074584961,
      "learning_rate": 2.1442129043167873e-07,
      "logits/chosen": 0.946514904499054,
      "logits/rejected": 0.9243011474609375,
      "logps/chosen": -161.78128051757812,
      "logps/ref_chosen": -62.38185119628906,
      "logps/ref_rejected": -71.40414428710938,
      "logps/rejected": -220.90402221679688,
      "loss": 1.0166,
      "margin_dpo/margin_mean": 50.100440979003906,
      "margin_dpo/margin_std": 103.29415130615234,
      "step": 392
    },
    {
      "epoch": 0.5941043083900227,
      "grad_norm": 15.310603141784668,
      "learning_rate": 2.131130332936195e-07,
      "logits/chosen": 0.8303643465042114,
      "logits/rejected": 0.830100953578949,
      "logps/chosen": -150.3733367919922,
      "logps/ref_chosen": -40.24174499511719,
      "logps/ref_rejected": -50.49744415283203,
      "logps/rejected": -233.74623107910156,
      "loss": 0.9411,
      "margin_dpo/margin_mean": 73.11720275878906,
      "margin_dpo/margin_std": 63.29140853881836,
      "step": 393
    },
    {
      "epoch": 0.5956160241874527,
      "grad_norm": 15.765230178833008,
      "learning_rate": 2.1180580796331323e-07,
      "logits/chosen": 0.971742570400238,
      "logits/rejected": 0.847053050994873,
      "logps/chosen": -156.80836486816406,
      "logps/ref_chosen": -46.0341911315918,
      "logps/ref_rejected": -90.94654846191406,
      "logps/rejected": -264.73736572265625,
      "loss": 1.076,
      "margin_dpo/margin_mean": 63.01665496826172,
      "margin_dpo/margin_std": 79.52315521240234,
      "step": 394
    },
    {
      "epoch": 0.5971277399848829,
      "grad_norm": 14.962550163269043,
      "learning_rate": 2.104996510066625e-07,
      "logits/chosen": 1.010396122932434,
      "logits/rejected": 0.888831615447998,
      "logps/chosen": -136.7878875732422,
      "logps/ref_chosen": -45.55821990966797,
      "logps/ref_rejected": -93.29295349121094,
      "logps/rejected": -282.3095397949219,
      "loss": 1.0567,
      "margin_dpo/margin_mean": 97.78689575195312,
      "margin_dpo/margin_std": 64.2834701538086,
      "step": 395
    },
    {
      "epoch": 0.5986394557823129,
      "grad_norm": 15.773469924926758,
      "learning_rate": 2.0919459895968517e-07,
      "logits/chosen": 0.9865584373474121,
      "logits/rejected": 0.9610118865966797,
      "logps/chosen": -135.16543579101562,
      "logps/ref_chosen": -47.607505798339844,
      "logps/ref_rejected": -52.5338020324707,
      "logps/rejected": -236.95713806152344,
      "loss": 1.0154,
      "margin_dpo/margin_mean": 96.86540222167969,
      "margin_dpo/margin_std": 54.51676940917969,
      "step": 396
    },
    {
      "epoch": 0.600151171579743,
      "grad_norm": 17.012723922729492,
      "learning_rate": 2.078906883274924e-07,
      "logits/chosen": 0.8290398120880127,
      "logits/rejected": 0.7684098482131958,
      "logps/chosen": -193.76319885253906,
      "logps/ref_chosen": -61.47978973388672,
      "logps/ref_rejected": -80.81649780273438,
      "logps/rejected": -270.89862060546875,
      "loss": 1.2961,
      "margin_dpo/margin_mean": 57.79869842529297,
      "margin_dpo/margin_std": 66.68041229248047,
      "step": 397
    },
    {
      "epoch": 0.6016628873771731,
      "grad_norm": 14.493592262268066,
      "learning_rate": 2.065879555832674e-07,
      "logits/chosen": 0.8568699359893799,
      "logits/rejected": 0.7129446268081665,
      "logps/chosen": -158.29476928710938,
      "logps/ref_chosen": -47.49082946777344,
      "logps/ref_rejected": -101.38699340820312,
      "logps/rejected": -306.26104736328125,
      "loss": 0.9401,
      "margin_dpo/margin_mean": 94.07012939453125,
      "margin_dpo/margin_std": 80.53518676757812,
      "step": 398
    },
    {
      "epoch": 0.6031746031746031,
      "grad_norm": 16.002872467041016,
      "learning_rate": 2.052864371672457e-07,
      "logits/chosen": 0.9500819444656372,
      "logits/rejected": 0.8516980409622192,
      "logps/chosen": -241.56881713867188,
      "logps/ref_chosen": -76.00422668457031,
      "logps/ref_rejected": -139.26205444335938,
      "logps/rejected": -371.9350280761719,
      "loss": 0.9017,
      "margin_dpo/margin_mean": 67.10838317871094,
      "margin_dpo/margin_std": 61.81909942626953,
      "step": 399
    },
    {
      "epoch": 0.6046863189720333,
      "grad_norm": 22.764698028564453,
      "learning_rate": 2.0398616948569493e-07,
      "logits/chosen": 0.8686426281929016,
      "logits/rejected": 0.9016132354736328,
      "logps/chosen": -222.953857421875,
      "logps/ref_chosen": -95.81818389892578,
      "logps/ref_rejected": -84.06385803222656,
      "logps/rejected": -276.75958251953125,
      "loss": 1.1296,
      "margin_dpo/margin_mean": 65.56008911132812,
      "margin_dpo/margin_std": 57.78480529785156,
      "step": 400
    },
    {
      "epoch": 0.6061980347694633,
      "grad_norm": 11.519392013549805,
      "learning_rate": 2.0268718890989752e-07,
      "logits/chosen": 0.8911803960800171,
      "logits/rejected": 0.8166502714157104,
      "logps/chosen": -165.61497497558594,
      "logps/ref_chosen": -53.86456298828125,
      "logps/ref_rejected": -53.25059509277344,
      "logps/rejected": -205.1787567138672,
      "loss": 0.9321,
      "margin_dpo/margin_mean": 40.17775344848633,
      "margin_dpo/margin_std": 83.69491577148438,
      "step": 401
    },
    {
      "epoch": 0.6077097505668935,
      "grad_norm": 16.777515411376953,
      "learning_rate": 2.013895317751323e-07,
      "logits/chosen": 0.9933120012283325,
      "logits/rejected": 0.814152717590332,
      "logps/chosen": -142.3320770263672,
      "logps/ref_chosen": -37.891700744628906,
      "logps/ref_rejected": -85.77980041503906,
      "logps/rejected": -285.2149963378906,
      "loss": 1.0406,
      "margin_dpo/margin_mean": 94.99481201171875,
      "margin_dpo/margin_std": 68.64460754394531,
      "step": 402
    },
    {
      "epoch": 0.6092214663643235,
      "grad_norm": 13.203496932983398,
      "learning_rate": 2.0009323437965898e-07,
      "logits/chosen": 1.0703718662261963,
      "logits/rejected": 0.9925715923309326,
      "logps/chosen": -180.50936889648438,
      "logps/ref_chosen": -63.857696533203125,
      "logps/ref_rejected": -93.38938903808594,
      "logps/rejected": -301.53302001953125,
      "loss": 0.9559,
      "margin_dpo/margin_mean": 91.49195861816406,
      "margin_dpo/margin_std": 91.93600463867188,
      "step": 403
    },
    {
      "epoch": 0.6107331821617535,
      "grad_norm": 15.767083168029785,
      "learning_rate": 1.9879833298370237e-07,
      "logits/chosen": 0.9310320019721985,
      "logits/rejected": 0.9439609050750732,
      "logps/chosen": -240.16302490234375,
      "logps/ref_chosen": -84.43171691894531,
      "logps/ref_rejected": -89.65742492675781,
      "logps/rejected": -274.97003173828125,
      "loss": 0.9804,
      "margin_dpo/margin_mean": 29.5813045501709,
      "margin_dpo/margin_std": 90.8002700805664,
      "step": 404
    },
    {
      "epoch": 0.6122448979591837,
      "grad_norm": 14.391467094421387,
      "learning_rate": 1.975048638084379e-07,
      "logits/chosen": 1.0961458683013916,
      "logits/rejected": 1.0419206619262695,
      "logps/chosen": -144.2384033203125,
      "logps/ref_chosen": -43.280792236328125,
      "logps/ref_rejected": -58.38227844238281,
      "logps/rejected": -217.02481079101562,
      "loss": 1.0274,
      "margin_dpo/margin_mean": 57.6849365234375,
      "margin_dpo/margin_std": 87.62693786621094,
      "step": 405
    },
    {
      "epoch": 0.6137566137566137,
      "grad_norm": 17.21456527709961,
      "learning_rate": 1.9621286303497914e-07,
      "logits/chosen": 1.0429320335388184,
      "logits/rejected": 0.8181569576263428,
      "logps/chosen": -143.40634155273438,
      "logps/ref_chosen": -38.76139831542969,
      "logps/ref_rejected": -95.4449462890625,
      "logps/rejected": -312.931396484375,
      "loss": 0.9634,
      "margin_dpo/margin_mean": 112.84149169921875,
      "margin_dpo/margin_std": 117.30062866210938,
      "step": 406
    },
    {
      "epoch": 0.6152683295540439,
      "grad_norm": 17.83815574645996,
      "learning_rate": 1.9492236680336483e-07,
      "logits/chosen": 0.921512246131897,
      "logits/rejected": 0.9170160293579102,
      "logps/chosen": -201.85311889648438,
      "logps/ref_chosen": -70.94854736328125,
      "logps/ref_rejected": -83.073486328125,
      "logps/rejected": -236.95440673828125,
      "loss": 1.065,
      "margin_dpo/margin_mean": 22.976346969604492,
      "margin_dpo/margin_std": 84.53071594238281,
      "step": 407
    },
    {
      "epoch": 0.6167800453514739,
      "grad_norm": 14.140363693237305,
      "learning_rate": 1.9363341121154895e-07,
      "logits/chosen": 0.9865923523902893,
      "logits/rejected": 0.8969758749008179,
      "logps/chosen": -151.7926788330078,
      "logps/ref_chosen": -57.809539794921875,
      "logps/ref_rejected": -81.67845153808594,
      "logps/rejected": -253.80831909179688,
      "loss": 0.8738,
      "margin_dpo/margin_mean": 78.146728515625,
      "margin_dpo/margin_std": 85.02195739746094,
      "step": 408
    },
    {
      "epoch": 0.618291761148904,
      "grad_norm": 18.92486000061035,
      "learning_rate": 1.9234603231438994e-07,
      "logits/chosen": 0.9477880001068115,
      "logits/rejected": 1.024693250656128,
      "logps/chosen": -208.4686279296875,
      "logps/ref_chosen": -77.07215118408203,
      "logps/ref_rejected": -52.54692459106445,
      "logps/rejected": -207.8866729736328,
      "loss": 1.1856,
      "margin_dpo/margin_mean": 23.943286895751953,
      "margin_dpo/margin_std": 57.939476013183594,
      "step": 409
    },
    {
      "epoch": 0.6198034769463341,
      "grad_norm": 16.562345504760742,
      "learning_rate": 1.9106026612264315e-07,
      "logits/chosen": 1.0837414264678955,
      "logits/rejected": 1.0848373174667358,
      "logps/chosen": -157.5943603515625,
      "logps/ref_chosen": -44.102970123291016,
      "logps/ref_rejected": -48.312713623046875,
      "logps/rejected": -213.13218688964844,
      "loss": 1.0328,
      "margin_dpo/margin_mean": 51.32807159423828,
      "margin_dpo/margin_std": 73.68001556396484,
      "step": 410
    },
    {
      "epoch": 0.6213151927437641,
      "grad_norm": 15.483366012573242,
      "learning_rate": 1.8977614860195296e-07,
      "logits/chosen": 0.9602583646774292,
      "logits/rejected": 0.9861100316047668,
      "logps/chosen": -207.64120483398438,
      "logps/ref_chosen": -68.72139739990234,
      "logps/ref_rejected": -60.70808792114258,
      "logps/rejected": -261.77276611328125,
      "loss": 1.0486,
      "margin_dpo/margin_mean": 62.14488983154297,
      "margin_dpo/margin_std": 83.70924377441406,
      "step": 411
    },
    {
      "epoch": 0.6228269085411943,
      "grad_norm": 18.00762176513672,
      "learning_rate": 1.8849371567184662e-07,
      "logits/chosen": 0.9532754421234131,
      "logits/rejected": 0.9993252754211426,
      "logps/chosen": -164.32028198242188,
      "logps/ref_chosen": -48.907501220703125,
      "logps/ref_rejected": -40.223628997802734,
      "logps/rejected": -204.09124755859375,
      "loss": 1.0641,
      "margin_dpo/margin_mean": 48.454830169677734,
      "margin_dpo/margin_std": 77.01388549804688,
      "step": 412
    },
    {
      "epoch": 0.6243386243386243,
      "grad_norm": 18.54705810546875,
      "learning_rate": 1.872130032047302e-07,
      "logits/chosen": 0.7836633324623108,
      "logits/rejected": 0.6924010515213013,
      "logps/chosen": -224.32937622070312,
      "logps/ref_chosen": -66.48075103759766,
      "logps/ref_rejected": -88.64950561523438,
      "logps/rejected": -281.3914794921875,
      "loss": 1.1014,
      "margin_dpo/margin_mean": 34.89335250854492,
      "margin_dpo/margin_std": 108.35519409179688,
      "step": 413
    },
    {
      "epoch": 0.6258503401360545,
      "grad_norm": 17.233596801757812,
      "learning_rate": 1.8593404702488436e-07,
      "logits/chosen": 0.8586117029190063,
      "logits/rejected": 0.8090198040008545,
      "logps/chosen": -217.32992553710938,
      "logps/ref_chosen": -64.2005386352539,
      "logps/ref_rejected": -92.14444732666016,
      "logps/rejected": -309.4424743652344,
      "loss": 0.9987,
      "margin_dpo/margin_mean": 64.16864013671875,
      "margin_dpo/margin_std": 81.55545806884766,
      "step": 414
    },
    {
      "epoch": 0.6273620559334845,
      "grad_norm": 15.781834602355957,
      "learning_rate": 1.846568829074628e-07,
      "logits/chosen": 0.8974360227584839,
      "logits/rejected": 1.0069502592086792,
      "logps/chosen": -171.6870880126953,
      "logps/ref_chosen": -58.819007873535156,
      "logps/ref_rejected": -41.336639404296875,
      "logps/rejected": -227.85916137695312,
      "loss": 1.1027,
      "margin_dpo/margin_mean": 73.65443420410156,
      "margin_dpo/margin_std": 78.06534576416016,
      "step": 415
    },
    {
      "epoch": 0.6288737717309146,
      "grad_norm": 18.85218620300293,
      "learning_rate": 1.8338154657749128e-07,
      "logits/chosen": 0.892058253288269,
      "logits/rejected": 0.7859885692596436,
      "logps/chosen": -194.0150146484375,
      "logps/ref_chosen": -53.452110290527344,
      "logps/ref_rejected": -97.50613403320312,
      "logps/rejected": -326.833984375,
      "loss": 1.1854,
      "margin_dpo/margin_mean": 88.76496124267578,
      "margin_dpo/margin_std": 134.0952911376953,
      "step": 416
    },
    {
      "epoch": 0.6303854875283447,
      "grad_norm": 16.071107864379883,
      "learning_rate": 1.8210807370886849e-07,
      "logits/chosen": 0.9685580730438232,
      "logits/rejected": 0.9870299100875854,
      "logps/chosen": -228.75973510742188,
      "logps/ref_chosen": -75.47906494140625,
      "logps/ref_rejected": -67.37366485595703,
      "logps/rejected": -281.687255859375,
      "loss": 1.0758,
      "margin_dpo/margin_mean": 61.032928466796875,
      "margin_dpo/margin_std": 164.45205688476562,
      "step": 417
    },
    {
      "epoch": 0.6318972033257747,
      "grad_norm": 31.492109298706055,
      "learning_rate": 1.8083649992336825e-07,
      "logits/chosen": 1.0411242246627808,
      "logits/rejected": 0.9932562112808228,
      "logps/chosen": -163.39306640625,
      "logps/ref_chosen": -51.03925323486328,
      "logps/ref_rejected": -80.96292877197266,
      "logps/rejected": -269.3048095703125,
      "loss": 1.2937,
      "margin_dpo/margin_mean": 75.98806762695312,
      "margin_dpo/margin_std": 79.07701110839844,
      "step": 418
    },
    {
      "epoch": 0.6334089191232048,
      "grad_norm": 15.3997802734375,
      "learning_rate": 1.7956686078964255e-07,
      "logits/chosen": 0.9788599014282227,
      "logits/rejected": 0.8751944303512573,
      "logps/chosen": -196.85659790039062,
      "logps/ref_chosen": -60.755767822265625,
      "logps/ref_rejected": -77.95507049560547,
      "logps/rejected": -270.27056884765625,
      "loss": 0.9464,
      "margin_dpo/margin_mean": 56.214637756347656,
      "margin_dpo/margin_std": 98.2620849609375,
      "step": 419
    },
    {
      "epoch": 0.6349206349206349,
      "grad_norm": 16.48794937133789,
      "learning_rate": 1.782991918222275e-07,
      "logits/chosen": 0.9061227440834045,
      "logits/rejected": 0.9082077741622925,
      "logps/chosen": -246.45130920410156,
      "logps/ref_chosen": -64.6197738647461,
      "logps/ref_rejected": -65.47144317626953,
      "logps/rejected": -271.0919494628906,
      "loss": 1.2391,
      "margin_dpo/margin_mean": 23.788970947265625,
      "margin_dpo/margin_std": 90.23184204101562,
      "step": 420
    },
    {
      "epoch": 0.636432350718065,
      "grad_norm": 20.588409423828125,
      "learning_rate": 1.7703352848054887e-07,
      "logits/chosen": 1.1465280055999756,
      "logits/rejected": 0.9681707620620728,
      "logps/chosen": -153.81906127929688,
      "logps/ref_chosen": -37.7196159362793,
      "logps/ref_rejected": -102.12132263183594,
      "logps/rejected": -318.0316162109375,
      "loss": 1.1831,
      "margin_dpo/margin_mean": 99.81085205078125,
      "margin_dpo/margin_std": 135.99864196777344,
      "step": 421
    },
    {
      "epoch": 0.6379440665154951,
      "grad_norm": 15.120969772338867,
      "learning_rate": 1.7576990616793137e-07,
      "logits/chosen": 0.988216757774353,
      "logits/rejected": 0.943016529083252,
      "logps/chosen": -174.78521728515625,
      "logps/ref_chosen": -70.57130432128906,
      "logps/ref_rejected": -81.15480041503906,
      "logps/rejected": -301.62823486328125,
      "loss": 0.9731,
      "margin_dpo/margin_mean": 116.259521484375,
      "margin_dpo/margin_std": 165.31149291992188,
      "step": 422
    },
    {
      "epoch": 0.6394557823129252,
      "grad_norm": 18.961877822875977,
      "learning_rate": 1.745083602306071e-07,
      "logits/chosen": 1.0410032272338867,
      "logits/rejected": 1.0402342081069946,
      "logps/chosen": -181.06886291503906,
      "logps/ref_chosen": -65.94102478027344,
      "logps/ref_rejected": -58.68115234375,
      "logps/rejected": -218.81155395507812,
      "loss": 0.968,
      "margin_dpo/margin_mean": 45.00257873535156,
      "margin_dpo/margin_std": 87.43341827392578,
      "step": 423
    },
    {
      "epoch": 0.6409674981103552,
      "grad_norm": 16.557573318481445,
      "learning_rate": 1.7324892595672804e-07,
      "logits/chosen": 0.966809868812561,
      "logits/rejected": 0.9489647150039673,
      "logps/chosen": -159.09210205078125,
      "logps/ref_chosen": -47.5775032043457,
      "logps/ref_rejected": -81.55694580078125,
      "logps/rejected": -263.04534912109375,
      "loss": 0.9385,
      "margin_dpo/margin_mean": 69.97380828857422,
      "margin_dpo/margin_std": 96.92681884765625,
      "step": 424
    },
    {
      "epoch": 0.6424792139077853,
      "grad_norm": 18.266592025756836,
      "learning_rate": 1.7199163857537824e-07,
      "logits/chosen": 0.9221489429473877,
      "logits/rejected": 0.9283965826034546,
      "logps/chosen": -210.3824462890625,
      "logps/ref_chosen": -73.816650390625,
      "logps/ref_rejected": -68.2657470703125,
      "logps/rejected": -270.66265869140625,
      "loss": 1.0809,
      "margin_dpo/margin_mean": 65.83113861083984,
      "margin_dpo/margin_std": 87.40162658691406,
      "step": 425
    },
    {
      "epoch": 0.6439909297052154,
      "grad_norm": 23.678329467773438,
      "learning_rate": 1.7073653325558828e-07,
      "logits/chosen": 0.7612646818161011,
      "logits/rejected": 0.8192156553268433,
      "logps/chosen": -212.634033203125,
      "logps/ref_chosen": -73.34886169433594,
      "logps/ref_rejected": -49.84626007080078,
      "logps/rejected": -236.68417358398438,
      "loss": 1.28,
      "margin_dpo/margin_mean": 47.552730560302734,
      "margin_dpo/margin_std": 78.64155578613281,
      "step": 426
    },
    {
      "epoch": 0.6455026455026455,
      "grad_norm": 18.935203552246094,
      "learning_rate": 1.6948364510535218e-07,
      "logits/chosen": 0.8741757869720459,
      "logits/rejected": 0.8858860731124878,
      "logps/chosen": -192.26693725585938,
      "logps/ref_chosen": -59.81298828125,
      "logps/ref_rejected": -72.67082214355469,
      "logps/rejected": -229.54574584960938,
      "loss": 1.1213,
      "margin_dpo/margin_mean": 24.420982360839844,
      "margin_dpo/margin_std": 100.69892883300781,
      "step": 427
    },
    {
      "epoch": 0.6470143613000756,
      "grad_norm": 14.1466646194458,
      "learning_rate": 1.6823300917064458e-07,
      "logits/chosen": 0.810982346534729,
      "logits/rejected": 0.8677164316177368,
      "logps/chosen": -231.75439453125,
      "logps/ref_chosen": -75.25834655761719,
      "logps/ref_rejected": -71.87213134765625,
      "logps/rejected": -261.0834655761719,
      "loss": 1.0756,
      "margin_dpo/margin_mean": 32.71527099609375,
      "margin_dpo/margin_std": 101.5775146484375,
      "step": 428
    },
    {
      "epoch": 0.6485260770975056,
      "grad_norm": 18.27950668334961,
      "learning_rate": 1.669846604344412e-07,
      "logits/chosen": 0.8632928133010864,
      "logits/rejected": 0.9631503820419312,
      "logps/chosen": -232.76438903808594,
      "logps/ref_chosen": -85.73371887207031,
      "logps/ref_rejected": -54.903968811035156,
      "logps/rejected": -241.41387939453125,
      "loss": 1.0955,
      "margin_dpo/margin_mean": 39.47923278808594,
      "margin_dpo/margin_std": 66.2182846069336,
      "step": 429
    },
    {
      "epoch": 0.6500377928949358,
      "grad_norm": 17.305450439453125,
      "learning_rate": 1.6573863381573954e-07,
      "logits/chosen": 0.8608442544937134,
      "logits/rejected": 0.8532319068908691,
      "logps/chosen": -167.2602081298828,
      "logps/ref_chosen": -54.592891693115234,
      "logps/ref_rejected": -57.52851867675781,
      "logps/rejected": -240.46214294433594,
      "loss": 0.9472,
      "margin_dpo/margin_mean": 70.26631164550781,
      "margin_dpo/margin_std": 83.55381774902344,
      "step": 430
    },
    {
      "epoch": 0.6515495086923658,
      "grad_norm": 15.106849670410156,
      "learning_rate": 1.6449496416858282e-07,
      "logits/chosen": 0.9315043091773987,
      "logits/rejected": 0.8595200777053833,
      "logps/chosen": -139.00140380859375,
      "logps/ref_chosen": -28.757225036621094,
      "logps/ref_rejected": -60.048500061035156,
      "logps/rejected": -241.73719787597656,
      "loss": 1.0949,
      "margin_dpo/margin_mean": 71.44451141357422,
      "margin_dpo/margin_std": 95.85964965820312,
      "step": 431
    },
    {
      "epoch": 0.6530612244897959,
      "grad_norm": 15.847650527954102,
      "learning_rate": 1.632536862810844e-07,
      "logits/chosen": 0.9007859230041504,
      "logits/rejected": 0.9773823022842407,
      "logps/chosen": -222.88815307617188,
      "logps/ref_chosen": -72.49076080322266,
      "logps/ref_rejected": -61.208106994628906,
      "logps/rejected": -219.5262451171875,
      "loss": 1.0237,
      "margin_dpo/margin_mean": 7.920762062072754,
      "margin_dpo/margin_std": 85.05551147460938,
      "step": 432
    },
    {
      "epoch": 0.654572940287226,
      "grad_norm": 18.91417121887207,
      "learning_rate": 1.6201483487445515e-07,
      "logits/chosen": 1.0263733863830566,
      "logits/rejected": 1.010549545288086,
      "logps/chosen": -214.4171905517578,
      "logps/ref_chosen": -74.03857421875,
      "logps/ref_rejected": -75.64851379394531,
      "logps/rejected": -267.9232177734375,
      "loss": 0.9721,
      "margin_dpo/margin_mean": 51.896095275878906,
      "margin_dpo/margin_std": 69.40410614013672,
      "step": 433
    },
    {
      "epoch": 0.656084656084656,
      "grad_norm": 14.838190078735352,
      "learning_rate": 1.6077844460203204e-07,
      "logits/chosen": 0.9536264538764954,
      "logits/rejected": 0.9389366507530212,
      "logps/chosen": -183.00430297851562,
      "logps/ref_chosen": -56.56264114379883,
      "logps/ref_rejected": -75.03836822509766,
      "logps/rejected": -236.72100830078125,
      "loss": 1.0005,
      "margin_dpo/margin_mean": 35.24098587036133,
      "margin_dpo/margin_std": 116.1950912475586,
      "step": 434
    },
    {
      "epoch": 0.6575963718820862,
      "grad_norm": 14.722707748413086,
      "learning_rate": 1.5954455004830878e-07,
      "logits/chosen": 1.0074257850646973,
      "logits/rejected": 0.9908655881881714,
      "logps/chosen": -178.95724487304688,
      "logps/ref_chosen": -52.70317840576172,
      "logps/ref_rejected": -59.57474899291992,
      "logps/rejected": -245.07876586914062,
      "loss": 1.1085,
      "margin_dpo/margin_mean": 59.24994659423828,
      "margin_dpo/margin_std": 70.93247985839844,
      "step": 435
    },
    {
      "epoch": 0.6591080876795162,
      "grad_norm": 21.451398849487305,
      "learning_rate": 1.5831318572796847e-07,
      "logits/chosen": 0.8413320183753967,
      "logits/rejected": 0.8829286098480225,
      "logps/chosen": -164.8529815673828,
      "logps/ref_chosen": -54.026947021484375,
      "logps/ref_rejected": -50.91650390625,
      "logps/rejected": -223.85400390625,
      "loss": 1.1831,
      "margin_dpo/margin_mean": 62.111473083496094,
      "margin_dpo/margin_std": 77.27197265625,
      "step": 436
    },
    {
      "epoch": 0.6606198034769464,
      "grad_norm": 15.607953071594238,
      "learning_rate": 1.5708438608491815e-07,
      "logits/chosen": 0.8712018728256226,
      "logits/rejected": 0.796289324760437,
      "logps/chosen": -210.6629180908203,
      "logps/ref_chosen": -65.94082641601562,
      "logps/ref_rejected": -101.48641967773438,
      "logps/rejected": -329.0594482421875,
      "loss": 1.1607,
      "margin_dpo/margin_mean": 82.8509292602539,
      "margin_dpo/margin_std": 156.46224975585938,
      "step": 437
    },
    {
      "epoch": 0.6621315192743764,
      "grad_norm": 18.58790397644043,
      "learning_rate": 1.558581854913253e-07,
      "logits/chosen": 0.9901245832443237,
      "logits/rejected": 0.8835450410842896,
      "logps/chosen": -163.1690673828125,
      "logps/ref_chosen": -37.30860137939453,
      "logps/ref_rejected": -83.07009887695312,
      "logps/rejected": -284.5603942871094,
      "loss": 1.0057,
      "margin_dpo/margin_mean": 75.62983703613281,
      "margin_dpo/margin_std": 144.90997314453125,
      "step": 438
    },
    {
      "epoch": 0.6636432350718064,
      "grad_norm": 14.376500129699707,
      "learning_rate": 1.5463461824665658e-07,
      "logits/chosen": 0.9146884679794312,
      "logits/rejected": 0.8561520576477051,
      "logps/chosen": -174.28402709960938,
      "logps/ref_chosen": -58.679443359375,
      "logps/ref_rejected": -86.585693359375,
      "logps/rejected": -287.9344177246094,
      "loss": 0.9811,
      "margin_dpo/margin_mean": 85.74412536621094,
      "margin_dpo/margin_std": 75.79908752441406,
      "step": 439
    },
    {
      "epoch": 0.6651549508692366,
      "grad_norm": 23.994163513183594,
      "learning_rate": 1.534137185767178e-07,
      "logits/chosen": 0.90851229429245,
      "logits/rejected": 0.7904536724090576,
      "logps/chosen": -171.59495544433594,
      "logps/ref_chosen": -52.94178009033203,
      "logps/ref_rejected": -108.83475494384766,
      "logps/rejected": -340.7264404296875,
      "loss": 1.0222,
      "margin_dpo/margin_mean": 113.23851013183594,
      "margin_dpo/margin_std": 85.38230895996094,
      "step": 440
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 16.344532012939453,
      "learning_rate": 1.521955206326976e-07,
      "logits/chosen": 0.9741239547729492,
      "logits/rejected": 0.7976517081260681,
      "logps/chosen": -146.74644470214844,
      "logps/ref_chosen": -43.878997802734375,
      "logps/ref_rejected": -85.12787628173828,
      "logps/rejected": -268.0721130371094,
      "loss": 0.9318,
      "margin_dpo/margin_mean": 80.07679748535156,
      "margin_dpo/margin_std": 80.79681396484375,
      "step": 441
    },
    {
      "epoch": 0.6681783824640968,
      "grad_norm": 16.937883377075195,
      "learning_rate": 1.5098005849021078e-07,
      "logits/chosen": 0.9499423503875732,
      "logits/rejected": 0.8875406980514526,
      "logps/chosen": -202.962158203125,
      "logps/ref_chosen": -82.47845458984375,
      "logps/ref_rejected": -119.25947570800781,
      "logps/rejected": -351.79766845703125,
      "loss": 1.0069,
      "margin_dpo/margin_mean": 112.05450439453125,
      "margin_dpo/margin_std": 77.75383758544922,
      "step": 442
    },
    {
      "epoch": 0.6696900982615268,
      "grad_norm": 15.244129180908203,
      "learning_rate": 1.4976736614834662e-07,
      "logits/chosen": 0.7210106253623962,
      "logits/rejected": 0.6923055648803711,
      "logps/chosen": -197.3117218017578,
      "logps/ref_chosen": -68.7590560913086,
      "logps/ref_rejected": -75.88587951660156,
      "logps/rejected": -294.58660888671875,
      "loss": 0.8711,
      "margin_dpo/margin_mean": 90.1480712890625,
      "margin_dpo/margin_std": 78.41068267822266,
      "step": 443
    },
    {
      "epoch": 0.671201814058957,
      "grad_norm": 21.625139236450195,
      "learning_rate": 1.4855747752871654e-07,
      "logits/chosen": 0.8997494578361511,
      "logits/rejected": 0.8529607057571411,
      "logps/chosen": -243.5794677734375,
      "logps/ref_chosen": -69.02406311035156,
      "logps/ref_rejected": -82.6009750366211,
      "logps/rejected": -275.14569091796875,
      "loss": 1.2998,
      "margin_dpo/margin_mean": 17.989328384399414,
      "margin_dpo/margin_std": 90.46211242675781,
      "step": 444
    },
    {
      "epoch": 0.672713529856387,
      "grad_norm": 18.841861724853516,
      "learning_rate": 1.473504264745062e-07,
      "logits/chosen": 0.9031407833099365,
      "logits/rejected": 0.9619897603988647,
      "logps/chosen": -269.9713439941406,
      "logps/ref_chosen": -90.35289001464844,
      "logps/ref_rejected": -69.4399185180664,
      "logps/rejected": -283.7539367675781,
      "loss": 0.9734,
      "margin_dpo/margin_mean": 34.695556640625,
      "margin_dpo/margin_std": 91.03128051757812,
      "step": 445
    },
    {
      "epoch": 0.674225245653817,
      "grad_norm": 14.573994636535645,
      "learning_rate": 1.461462467495284e-07,
      "logits/chosen": 0.8910402655601501,
      "logits/rejected": 0.8468337655067444,
      "logps/chosen": -196.3031768798828,
      "logps/ref_chosen": -68.73054504394531,
      "logps/ref_rejected": -94.55728149414062,
      "logps/rejected": -304.6093444824219,
      "loss": 0.8754,
      "margin_dpo/margin_mean": 82.47945404052734,
      "margin_dpo/margin_std": 77.50213623046875,
      "step": 446
    },
    {
      "epoch": 0.6757369614512472,
      "grad_norm": 19.80596351623535,
      "learning_rate": 1.4494497203727843e-07,
      "logits/chosen": 0.9224183559417725,
      "logits/rejected": 0.8308309316635132,
      "logps/chosen": -180.90097045898438,
      "logps/ref_chosen": -63.873809814453125,
      "logps/ref_rejected": -93.55643463134766,
      "logps/rejected": -295.10546875,
      "loss": 0.9013,
      "margin_dpo/margin_mean": 84.52186584472656,
      "margin_dpo/margin_std": 88.04441833496094,
      "step": 447
    },
    {
      "epoch": 0.6772486772486772,
      "grad_norm": 15.010350227355957,
      "learning_rate": 1.4374663593999256e-07,
      "logits/chosen": 1.0064988136291504,
      "logits/rejected": 0.9316452145576477,
      "logps/chosen": -173.94631958007812,
      "logps/ref_chosen": -48.06145477294922,
      "logps/ref_rejected": -74.88455963134766,
      "logps/rejected": -297.13531494140625,
      "loss": 1.015,
      "margin_dpo/margin_mean": 96.36589050292969,
      "margin_dpo/margin_std": 80.901611328125,
      "step": 448
    },
    {
      "epoch": 0.6787603930461074,
      "grad_norm": 21.837177276611328,
      "learning_rate": 1.4255127197770707e-07,
      "logits/chosen": 0.7422596216201782,
      "logits/rejected": 0.7190842628479004,
      "logps/chosen": -238.9716796875,
      "logps/ref_chosen": -68.59271240234375,
      "logps/ref_rejected": -88.01607513427734,
      "logps/rejected": -317.0701599121094,
      "loss": 1.233,
      "margin_dpo/margin_mean": 58.67515563964844,
      "margin_dpo/margin_std": 102.90806579589844,
      "step": 449
    },
    {
      "epoch": 0.6802721088435374,
      "grad_norm": 16.762248992919922,
      "learning_rate": 1.4135891358732205e-07,
      "logits/chosen": 1.0215989351272583,
      "logits/rejected": 0.838610053062439,
      "logps/chosen": -167.00927734375,
      "logps/ref_chosen": -35.716896057128906,
      "logps/ref_rejected": -95.92900085449219,
      "logps/rejected": -310.751953125,
      "loss": 1.049,
      "margin_dpo/margin_mean": 83.53057861328125,
      "margin_dpo/margin_std": 82.19971466064453,
      "step": 450
    },
    {
      "epoch": 0.6817838246409675,
      "grad_norm": 15.963050842285156,
      "learning_rate": 1.4016959412166437e-07,
      "logits/chosen": 0.8930540084838867,
      "logits/rejected": 0.7441238760948181,
      "logps/chosen": -221.75592041015625,
      "logps/ref_chosen": -73.0178451538086,
      "logps/ref_rejected": -109.17976379394531,
      "logps/rejected": -333.75726318359375,
      "loss": 1.0905,
      "margin_dpo/margin_mean": 75.83943939208984,
      "margin_dpo/margin_std": 87.21731567382812,
      "step": 451
    },
    {
      "epoch": 0.6832955404383976,
      "grad_norm": 17.484153747558594,
      "learning_rate": 1.3898334684855645e-07,
      "logits/chosen": 0.7930479645729065,
      "logits/rejected": 0.7267175316810608,
      "logps/chosen": -191.09988403320312,
      "logps/ref_chosen": -68.31144714355469,
      "logps/ref_rejected": -97.96285247802734,
      "logps/rejected": -300.38970947265625,
      "loss": 1.096,
      "margin_dpo/margin_mean": 79.63841247558594,
      "margin_dpo/margin_std": 85.86831665039062,
      "step": 452
    },
    {
      "epoch": 0.6848072562358276,
      "grad_norm": 24.5980167388916,
      "learning_rate": 1.3780020494988445e-07,
      "logits/chosen": 0.8875958919525146,
      "logits/rejected": 0.8231840133666992,
      "logps/chosen": -145.66799926757812,
      "logps/ref_chosen": -51.74473190307617,
      "logps/ref_rejected": -60.13861083984375,
      "logps/rejected": -229.53916931152344,
      "loss": 1.0671,
      "margin_dpo/margin_mean": 75.477294921875,
      "margin_dpo/margin_std": 93.73162841796875,
      "step": 453
    },
    {
      "epoch": 0.6863189720332578,
      "grad_norm": 12.73465633392334,
      "learning_rate": 1.366202015206706e-07,
      "logits/chosen": 0.9996371269226074,
      "logits/rejected": 1.0085588693618774,
      "logps/chosen": -156.74002075195312,
      "logps/ref_chosen": -44.98827362060547,
      "logps/ref_rejected": -50.99993133544922,
      "logps/rejected": -196.87745666503906,
      "loss": 1.0144,
      "margin_dpo/margin_mean": 34.12577819824219,
      "margin_dpo/margin_std": 120.2644271850586,
      "step": 454
    },
    {
      "epoch": 0.6878306878306878,
      "grad_norm": 16.96663475036621,
      "learning_rate": 1.354433695681474e-07,
      "logits/chosen": 0.8430243134498596,
      "logits/rejected": 0.8127152919769287,
      "logps/chosen": -215.993896484375,
      "logps/ref_chosen": -71.83073425292969,
      "logps/ref_rejected": -84.22274780273438,
      "logps/rejected": -321.873046875,
      "loss": 1.0001,
      "margin_dpo/margin_mean": 93.48712921142578,
      "margin_dpo/margin_std": 97.56796264648438,
      "step": 455
    },
    {
      "epoch": 0.6893424036281179,
      "grad_norm": 21.244468688964844,
      "learning_rate": 1.3426974201083439e-07,
      "logits/chosen": 0.8168896436691284,
      "logits/rejected": 0.7225136756896973,
      "logps/chosen": -166.49514770507812,
      "logps/ref_chosen": -37.67304229736328,
      "logps/ref_rejected": -87.09375,
      "logps/rejected": -294.4206848144531,
      "loss": 1.0769,
      "margin_dpo/margin_mean": 78.50482177734375,
      "margin_dpo/margin_std": 95.56376647949219,
      "step": 456
    },
    {
      "epoch": 0.690854119425548,
      "grad_norm": 20.784381866455078,
      "learning_rate": 1.3309935167761717e-07,
      "logits/chosen": 0.9731137752532959,
      "logits/rejected": 0.890478789806366,
      "logps/chosen": -193.4537353515625,
      "logps/ref_chosen": -36.889923095703125,
      "logps/ref_rejected": -71.26411437988281,
      "logps/rejected": -267.1976623535156,
      "loss": 1.0515,
      "margin_dpo/margin_mean": 39.369747161865234,
      "margin_dpo/margin_std": 71.71172332763672,
      "step": 457
    },
    {
      "epoch": 0.6923658352229781,
      "grad_norm": 17.790769577026367,
      "learning_rate": 1.3193223130682936e-07,
      "logits/chosen": 0.9141269326210022,
      "logits/rejected": 0.8772815465927124,
      "logps/chosen": -211.3562774658203,
      "logps/ref_chosen": -61.25056076049805,
      "logps/ref_rejected": -70.772216796875,
      "logps/rejected": -278.8810119628906,
      "loss": 0.9925,
      "margin_dpo/margin_mean": 58.00306701660156,
      "margin_dpo/margin_std": 117.56779479980469,
      "step": 458
    },
    {
      "epoch": 0.6938775510204082,
      "grad_norm": 15.180387496948242,
      "learning_rate": 1.3076841354533658e-07,
      "logits/chosen": 1.0155048370361328,
      "logits/rejected": 0.9921817779541016,
      "logps/chosen": -206.2635498046875,
      "logps/ref_chosen": -73.28994750976562,
      "logps/ref_rejected": -108.29696655273438,
      "logps/rejected": -306.605712890625,
      "loss": 0.9325,
      "margin_dpo/margin_mean": 65.33513641357422,
      "margin_dpo/margin_std": 104.4261474609375,
      "step": 459
    },
    {
      "epoch": 0.6953892668178382,
      "grad_norm": 15.735591888427734,
      "learning_rate": 1.2960793094762345e-07,
      "logits/chosen": 0.8665391802787781,
      "logits/rejected": 0.744087815284729,
      "logps/chosen": -220.26666259765625,
      "logps/ref_chosen": -76.68836975097656,
      "logps/ref_rejected": -99.02154541015625,
      "logps/rejected": -351.54107666015625,
      "loss": 0.8928,
      "margin_dpo/margin_mean": 108.94125366210938,
      "margin_dpo/margin_std": 133.6810302734375,
      "step": 460
    },
    {
      "epoch": 0.6969009826152683,
      "grad_norm": 14.483108520507812,
      "learning_rate": 1.2845081597488286e-07,
      "logits/chosen": 0.9958152770996094,
      "logits/rejected": 0.9232292175292969,
      "logps/chosen": -167.3024139404297,
      "logps/ref_chosen": -63.83565902709961,
      "logps/ref_rejected": -87.00081634521484,
      "logps/rejected": -267.07049560546875,
      "loss": 0.883,
      "margin_dpo/margin_mean": 76.60294342041016,
      "margin_dpo/margin_std": 62.0404167175293,
      "step": 461
    },
    {
      "epoch": 0.6984126984126984,
      "grad_norm": 18.1173095703125,
      "learning_rate": 1.27297100994108e-07,
      "logits/chosen": 0.9541089534759521,
      "logits/rejected": 0.9522888660430908,
      "logps/chosen": -175.2286376953125,
      "logps/ref_chosen": -64.87322998046875,
      "logps/ref_rejected": -68.76766967773438,
      "logps/rejected": -291.7859191894531,
      "loss": 0.9019,
      "margin_dpo/margin_mean": 112.66285705566406,
      "margin_dpo/margin_std": 76.85243225097656,
      "step": 462
    },
    {
      "epoch": 0.6999244142101285,
      "grad_norm": 21.672767639160156,
      "learning_rate": 1.2614681827718695e-07,
      "logits/chosen": 0.9698790907859802,
      "logits/rejected": 1.0222316980361938,
      "logps/chosen": -219.83706665039062,
      "logps/ref_chosen": -66.8828125,
      "logps/ref_rejected": -47.84074783325195,
      "logps/rejected": -255.33847045898438,
      "loss": 1.0292,
      "margin_dpo/margin_mean": 54.54346466064453,
      "margin_dpo/margin_std": 97.27693176269531,
      "step": 463
    },
    {
      "epoch": 0.7014361300075586,
      "grad_norm": 16.856752395629883,
      "learning_rate": 1.2500000000000005e-07,
      "logits/chosen": 1.0088590383529663,
      "logits/rejected": 1.0314637422561646,
      "logps/chosen": -243.63446044921875,
      "logps/ref_chosen": -90.68864440917969,
      "logps/ref_rejected": -78.24504852294922,
      "logps/rejected": -304.3553161621094,
      "loss": 0.9942,
      "margin_dpo/margin_mean": 73.16445922851562,
      "margin_dpo/margin_std": 115.3930435180664,
      "step": 464
    },
    {
      "epoch": 0.7029478458049887,
      "grad_norm": 18.76186752319336,
      "learning_rate": 1.238566782415197e-07,
      "logits/chosen": 0.9639427661895752,
      "logits/rejected": 0.8763701915740967,
      "logps/chosen": -196.27841186523438,
      "logps/ref_chosen": -68.18783569335938,
      "logps/ref_rejected": -107.69645690917969,
      "logps/rejected": -325.6939697265625,
      "loss": 1.0667,
      "margin_dpo/margin_mean": 89.90691375732422,
      "margin_dpo/margin_std": 111.15441131591797,
      "step": 465
    },
    {
      "epoch": 0.7044595616024187,
      "grad_norm": 25.630605697631836,
      "learning_rate": 1.2271688498291334e-07,
      "logits/chosen": 0.7273076772689819,
      "logits/rejected": 0.7946709394454956,
      "logps/chosen": -287.0624084472656,
      "logps/ref_chosen": -109.91573333740234,
      "logps/ref_rejected": -71.43077087402344,
      "logps/rejected": -251.70777893066406,
      "loss": 1.2282,
      "margin_dpo/margin_mean": 3.1303582191467285,
      "margin_dpo/margin_std": 107.84750366210938,
      "step": 466
    },
    {
      "epoch": 0.7059712773998488,
      "grad_norm": 19.084739685058594,
      "learning_rate": 1.2158065210664848e-07,
      "logits/chosen": 0.861819326877594,
      "logits/rejected": 0.7350021600723267,
      "logps/chosen": -200.34225463867188,
      "logps/ref_chosen": -44.50347900390625,
      "logps/ref_rejected": -68.12409973144531,
      "logps/rejected": -296.81658935546875,
      "loss": 0.9962,
      "margin_dpo/margin_mean": 72.85372924804688,
      "margin_dpo/margin_std": 124.93696594238281,
      "step": 467
    },
    {
      "epoch": 0.7074829931972789,
      "grad_norm": 17.534069061279297,
      "learning_rate": 1.204480113956011e-07,
      "logits/chosen": 0.9389829635620117,
      "logits/rejected": 0.8962525129318237,
      "logps/chosen": -154.1627197265625,
      "logps/ref_chosen": -46.26074981689453,
      "logps/ref_rejected": -77.6624755859375,
      "logps/rejected": -269.74609375,
      "loss": 0.9274,
      "margin_dpo/margin_mean": 84.1816635131836,
      "margin_dpo/margin_std": 95.72941589355469,
      "step": 468
    },
    {
      "epoch": 0.708994708994709,
      "grad_norm": 18.879037857055664,
      "learning_rate": 1.1931899453216697e-07,
      "logits/chosen": 0.98770672082901,
      "logits/rejected": 0.9201364517211914,
      "logps/chosen": -188.12139892578125,
      "logps/ref_chosen": -62.30226516723633,
      "logps/ref_rejected": -88.74978637695312,
      "logps/rejected": -278.533203125,
      "loss": 0.9651,
      "margin_dpo/margin_mean": 63.96430206298828,
      "margin_dpo/margin_std": 68.25845336914062,
      "step": 469
    },
    {
      "epoch": 0.7105064247921391,
      "grad_norm": 17.858858108520508,
      "learning_rate": 1.1819363309737438e-07,
      "logits/chosen": 1.0070170164108276,
      "logits/rejected": 0.9005523920059204,
      "logps/chosen": -177.07325744628906,
      "logps/ref_chosen": -53.556175231933594,
      "logps/ref_rejected": -83.76905059814453,
      "logps/rejected": -305.9222717285156,
      "loss": 1.0157,
      "margin_dpo/margin_mean": 98.63612365722656,
      "margin_dpo/margin_std": 116.62651062011719,
      "step": 470
    },
    {
      "epoch": 0.7120181405895691,
      "grad_norm": 22.893625259399414,
      "learning_rate": 1.1707195857000215e-07,
      "logits/chosen": 0.9534423351287842,
      "logits/rejected": 0.9319720268249512,
      "logps/chosen": -163.5334014892578,
      "logps/ref_chosen": -51.11562728881836,
      "logps/ref_rejected": -63.00703048706055,
      "logps/rejected": -267.5377197265625,
      "loss": 0.9278,
      "margin_dpo/margin_mean": 92.11293029785156,
      "margin_dpo/margin_std": 100.85304260253906,
      "step": 471
    },
    {
      "epoch": 0.7135298563869993,
      "grad_norm": 16.294567108154297,
      "learning_rate": 1.1595400232569768e-07,
      "logits/chosen": 0.7927904725074768,
      "logits/rejected": 0.8178097009658813,
      "logps/chosen": -197.45468139648438,
      "logps/ref_chosen": -69.28681182861328,
      "logps/ref_rejected": -85.08757019042969,
      "logps/rejected": -298.28179931640625,
      "loss": 1.103,
      "margin_dpo/margin_mean": 85.0263900756836,
      "margin_dpo/margin_std": 92.62736511230469,
      "step": 472
    },
    {
      "epoch": 0.7150415721844293,
      "grad_norm": 19.770648956298828,
      "learning_rate": 1.1483979563610069e-07,
      "logits/chosen": 0.9858765602111816,
      "logits/rejected": 0.7901525497436523,
      "logps/chosen": -140.59815979003906,
      "logps/ref_chosen": -34.896080017089844,
      "logps/ref_rejected": -81.10395812988281,
      "logps/rejected": -261.3387145996094,
      "loss": 1.0053,
      "margin_dpo/margin_mean": 74.53266906738281,
      "margin_dpo/margin_std": 110.72260284423828,
      "step": 473
    },
    {
      "epoch": 0.7165532879818595,
      "grad_norm": 23.826879501342773,
      "learning_rate": 1.1372936966796709e-07,
      "logits/chosen": 1.108849287033081,
      "logits/rejected": 1.0479471683502197,
      "logps/chosen": -195.81671142578125,
      "logps/ref_chosen": -45.78113555908203,
      "logps/ref_rejected": -68.88629150390625,
      "logps/rejected": -293.0316162109375,
      "loss": 1.1275,
      "margin_dpo/margin_mean": 74.10977172851562,
      "margin_dpo/margin_std": 110.27363586425781,
      "step": 474
    },
    {
      "epoch": 0.7180650037792895,
      "grad_norm": 13.597807884216309,
      "learning_rate": 1.126227554822985e-07,
      "logits/chosen": 0.8909091949462891,
      "logits/rejected": 0.9018880724906921,
      "logps/chosen": -191.0635223388672,
      "logps/ref_chosen": -58.526344299316406,
      "logps/ref_rejected": -76.01811981201172,
      "logps/rejected": -290.8351745605469,
      "loss": 0.8632,
      "margin_dpo/margin_mean": 82.27987670898438,
      "margin_dpo/margin_std": 109.62013244628906,
      "step": 475
    },
    {
      "epoch": 0.7195767195767195,
      "grad_norm": 22.143871307373047,
      "learning_rate": 1.1151998403347243e-07,
      "logits/chosen": 0.8504942059516907,
      "logits/rejected": 0.801427960395813,
      "logps/chosen": -185.2305145263672,
      "logps/ref_chosen": -54.09751510620117,
      "logps/ref_rejected": -84.35127258300781,
      "logps/rejected": -310.95294189453125,
      "loss": 1.0874,
      "margin_dpo/margin_mean": 95.46868133544922,
      "margin_dpo/margin_std": 103.3380126953125,
      "step": 476
    },
    {
      "epoch": 0.7210884353741497,
      "grad_norm": 20.90325355529785,
      "learning_rate": 1.1042108616837692e-07,
      "logits/chosen": 0.9148606657981873,
      "logits/rejected": 0.8745808601379395,
      "logps/chosen": -211.44253540039062,
      "logps/ref_chosen": -67.8685531616211,
      "logps/ref_rejected": -72.77481842041016,
      "logps/rejected": -246.86721801757812,
      "loss": 1.1216,
      "margin_dpo/margin_mean": 30.518428802490234,
      "margin_dpo/margin_std": 85.03158569335938,
      "step": 477
    },
    {
      "epoch": 0.7226001511715797,
      "grad_norm": 19.139371871948242,
      "learning_rate": 1.0932609262554746e-07,
      "logits/chosen": 0.8678562641143799,
      "logits/rejected": 0.8669169545173645,
      "logps/chosen": -157.50787353515625,
      "logps/ref_chosen": -56.62529754638672,
      "logps/ref_rejected": -58.720787048339844,
      "logps/rejected": -254.533203125,
      "loss": 1.1308,
      "margin_dpo/margin_mean": 94.92986297607422,
      "margin_dpo/margin_std": 100.27252197265625,
      "step": 478
    },
    {
      "epoch": 0.7241118669690099,
      "grad_norm": 20.085996627807617,
      "learning_rate": 1.0823503403430734e-07,
      "logits/chosen": 0.8876814842224121,
      "logits/rejected": 0.7483628988265991,
      "logps/chosen": -176.6420135498047,
      "logps/ref_chosen": -45.32330322265625,
      "logps/ref_rejected": -55.02246856689453,
      "logps/rejected": -235.74169921875,
      "loss": 1.1586,
      "margin_dpo/margin_mean": 49.40052795410156,
      "margin_dpo/margin_std": 96.12824249267578,
      "step": 479
    },
    {
      "epoch": 0.7256235827664399,
      "grad_norm": 21.663545608520508,
      "learning_rate": 1.0714794091391072e-07,
      "logits/chosen": 0.8385661244392395,
      "logits/rejected": 0.8454256057739258,
      "logps/chosen": -211.0986328125,
      "logps/ref_chosen": -63.22759246826172,
      "logps/ref_rejected": -74.44642639160156,
      "logps/rejected": -270.2560729980469,
      "loss": 1.0536,
      "margin_dpo/margin_mean": 47.93859100341797,
      "margin_dpo/margin_std": 117.68538665771484,
      "step": 480
    },
    {
      "epoch": 0.72713529856387,
      "grad_norm": 17.779111862182617,
      "learning_rate": 1.0606484367268906e-07,
      "logits/chosen": 0.8755944967269897,
      "logits/rejected": 0.9314774870872498,
      "logps/chosen": -223.0485076904297,
      "logps/ref_chosen": -79.15220642089844,
      "logps/ref_rejected": -68.31973266601562,
      "logps/rejected": -270.32794189453125,
      "loss": 1.049,
      "margin_dpo/margin_mean": 58.11188507080078,
      "margin_dpo/margin_std": 97.23744201660156,
      "step": 481
    },
    {
      "epoch": 0.7286470143613001,
      "grad_norm": 24.795852661132812,
      "learning_rate": 1.0498577260720048e-07,
      "logits/chosen": 0.8218737840652466,
      "logits/rejected": 0.7339863777160645,
      "logps/chosen": -193.96629333496094,
      "logps/ref_chosen": -54.06950378417969,
      "logps/ref_rejected": -87.615234375,
      "logps/rejected": -284.41217041015625,
      "loss": 1.1686,
      "margin_dpo/margin_mean": 56.900150299072266,
      "margin_dpo/margin_std": 103.31169128417969,
      "step": 482
    },
    {
      "epoch": 0.7301587301587301,
      "grad_norm": 15.781610488891602,
      "learning_rate": 1.0391075790138232e-07,
      "logits/chosen": 0.894173800945282,
      "logits/rejected": 0.9251315593719482,
      "logps/chosen": -174.34373474121094,
      "logps/ref_chosen": -54.128578186035156,
      "logps/ref_rejected": -67.60116577148438,
      "logps/rejected": -278.50946044921875,
      "loss": 0.986,
      "margin_dpo/margin_mean": 90.693115234375,
      "margin_dpo/margin_std": 83.16683197021484,
      "step": 483
    },
    {
      "epoch": 0.7316704459561603,
      "grad_norm": 22.09979820251465,
      "learning_rate": 1.0283982962570681e-07,
      "logits/chosen": 0.8971969485282898,
      "logits/rejected": 0.8970204591751099,
      "logps/chosen": -179.95797729492188,
      "logps/ref_chosen": -50.43122100830078,
      "logps/ref_rejected": -66.30445861816406,
      "logps/rejected": -237.34353637695312,
      "loss": 1.0404,
      "margin_dpo/margin_mean": 41.512306213378906,
      "margin_dpo/margin_std": 72.52426147460938,
      "step": 484
    },
    {
      "epoch": 0.7331821617535903,
      "grad_norm": 17.71631622314453,
      "learning_rate": 1.0177301773633992e-07,
      "logits/chosen": 1.0137141942977905,
      "logits/rejected": 0.9617197513580322,
      "logps/chosen": -193.55841064453125,
      "logps/ref_chosen": -51.1461067199707,
      "logps/ref_rejected": -61.148231506347656,
      "logps/rejected": -281.61785888671875,
      "loss": 1.0243,
      "margin_dpo/margin_mean": 78.0573501586914,
      "margin_dpo/margin_std": 87.72028350830078,
      "step": 485
    },
    {
      "epoch": 0.7346938775510204,
      "grad_norm": 18.199512481689453,
      "learning_rate": 1.007103520743035e-07,
      "logits/chosen": 0.9664397239685059,
      "logits/rejected": 0.8338623046875,
      "logps/chosen": -239.65943908691406,
      "logps/ref_chosen": -70.02229309082031,
      "logps/ref_rejected": -114.05645751953125,
      "logps/rejected": -355.57293701171875,
      "loss": 1.0931,
      "margin_dpo/margin_mean": 71.87936401367188,
      "margin_dpo/margin_std": 91.45281982421875,
      "step": 486
    },
    {
      "epoch": 0.7362055933484505,
      "grad_norm": 21.441146850585938,
      "learning_rate": 9.965186236464046e-08,
      "logits/chosen": 0.8630800843238831,
      "logits/rejected": 0.9823142290115356,
      "logps/chosen": -227.5937042236328,
      "logps/ref_chosen": -81.48747253417969,
      "logps/ref_rejected": -57.63191604614258,
      "logps/rejected": -252.54348754882812,
      "loss": 1.0282,
      "margin_dpo/margin_mean": 48.80535125732422,
      "margin_dpo/margin_std": 73.48953247070312,
      "step": 487
    },
    {
      "epoch": 0.7377173091458806,
      "grad_norm": 16.673267364501953,
      "learning_rate": 9.859757821558337e-08,
      "logits/chosen": 1.1053991317749023,
      "logits/rejected": 1.0277700424194336,
      "logps/chosen": -159.07391357421875,
      "logps/ref_chosen": -40.96950912475586,
      "logps/ref_rejected": -58.97332000732422,
      "logps/rejected": -254.75424194335938,
      "loss": 0.9396,
      "margin_dpo/margin_mean": 77.67652130126953,
      "margin_dpo/margin_std": 103.35063934326172,
      "step": 488
    },
    {
      "epoch": 0.7392290249433107,
      "grad_norm": 17.88896942138672,
      "learning_rate": 9.754752911772615e-08,
      "logits/chosen": 0.9406849145889282,
      "logits/rejected": 0.9168277978897095,
      "logps/chosen": -206.83840942382812,
      "logps/ref_chosen": -61.101173400878906,
      "logps/ref_rejected": -98.35012817382812,
      "logps/rejected": -302.9677734375,
      "loss": 1.2049,
      "margin_dpo/margin_mean": 58.880401611328125,
      "margin_dpo/margin_std": 115.15332794189453,
      "step": 489
    },
    {
      "epoch": 0.7407407407407407,
      "grad_norm": 17.050439834594727,
      "learning_rate": 9.650174444319956e-08,
      "logits/chosen": 1.056211233139038,
      "logits/rejected": 1.0357109308242798,
      "logps/chosen": -221.88174438476562,
      "logps/ref_chosen": -62.3327751159668,
      "logps/ref_rejected": -79.33484649658203,
      "logps/rejected": -289.1636962890625,
      "loss": 1.1471,
      "margin_dpo/margin_mean": 50.279884338378906,
      "margin_dpo/margin_std": 150.81320190429688,
      "step": 490
    },
    {
      "epoch": 0.7422524565381708,
      "grad_norm": 15.308135032653809,
      "learning_rate": 9.546025344484868e-08,
      "logits/chosen": 0.9274444580078125,
      "logits/rejected": 0.9464068412780762,
      "logps/chosen": -178.4617919921875,
      "logps/ref_chosen": -56.26042175292969,
      "logps/ref_rejected": -60.113643646240234,
      "logps/rejected": -242.59280395507812,
      "loss": 1.0716,
      "margin_dpo/margin_mean": 60.27778625488281,
      "margin_dpo/margin_std": 79.55590057373047,
      "step": 491
    },
    {
      "epoch": 0.7437641723356009,
      "grad_norm": 18.683191299438477,
      "learning_rate": 9.442308525541589e-08,
      "logits/chosen": 1.0775320529937744,
      "logits/rejected": 0.8653636574745178,
      "logps/chosen": -203.3031463623047,
      "logps/ref_chosen": -45.579383850097656,
      "logps/ref_rejected": -113.18993377685547,
      "logps/rejected": -362.9856872558594,
      "loss": 1.1972,
      "margin_dpo/margin_mean": 92.07199096679688,
      "margin_dpo/margin_std": 111.55503845214844,
      "step": 492
    },
    {
      "epoch": 0.745275888133031,
      "grad_norm": 16.538679122924805,
      "learning_rate": 9.339026888672468e-08,
      "logits/chosen": 0.9095848798751831,
      "logits/rejected": 0.843256950378418,
      "logps/chosen": -190.7752685546875,
      "logps/ref_chosen": -59.649986267089844,
      "logps/ref_rejected": -87.15731811523438,
      "logps/rejected": -255.50759887695312,
      "loss": 0.9946,
      "margin_dpo/margin_mean": 37.224998474121094,
      "margin_dpo/margin_std": 83.05178833007812,
      "step": 493
    },
    {
      "epoch": 0.7467876039304611,
      "grad_norm": 22.777530670166016,
      "learning_rate": 9.236183322886945e-08,
      "logits/chosen": 0.7308402061462402,
      "logits/rejected": 0.6717466711997986,
      "logps/chosen": -180.57354736328125,
      "logps/ref_chosen": -52.495460510253906,
      "logps/ref_rejected": -76.69441223144531,
      "logps/rejected": -251.54196166992188,
      "loss": 1.1615,
      "margin_dpo/margin_mean": 46.76948547363281,
      "margin_dpo/margin_std": 64.22396850585938,
      "step": 494
    },
    {
      "epoch": 0.7482993197278912,
      "grad_norm": 20.922956466674805,
      "learning_rate": 9.133780704940594e-08,
      "logits/chosen": 0.8490627408027649,
      "logits/rejected": 0.8381332755088806,
      "logps/chosen": -224.40399169921875,
      "logps/ref_chosen": -61.861454010009766,
      "logps/ref_rejected": -66.13208770751953,
      "logps/rejected": -220.16583251953125,
      "loss": 1.1718,
      "margin_dpo/margin_mean": -8.508773803710938,
      "margin_dpo/margin_std": 103.41836547851562,
      "step": 495
    },
    {
      "epoch": 0.7498110355253212,
      "grad_norm": 17.408597946166992,
      "learning_rate": 9.031821899254797e-08,
      "logits/chosen": 0.9441404938697815,
      "logits/rejected": 0.7932374477386475,
      "logps/chosen": -185.04629516601562,
      "logps/ref_chosen": -45.550537109375,
      "logps/ref_rejected": -100.87998962402344,
      "logps/rejected": -331.1201477050781,
      "loss": 1.0355,
      "margin_dpo/margin_mean": 90.74439239501953,
      "margin_dpo/margin_std": 99.05509948730469,
      "step": 496
    },
    {
      "epoch": 0.7513227513227513,
      "grad_norm": 17.972551345825195,
      "learning_rate": 8.930309757836516e-08,
      "logits/chosen": 0.9454480409622192,
      "logits/rejected": 0.8560576438903809,
      "logps/chosen": -160.0005645751953,
      "logps/ref_chosen": -59.57224655151367,
      "logps/ref_rejected": -98.0445556640625,
      "logps/rejected": -283.73138427734375,
      "loss": 1.0466,
      "margin_dpo/margin_mean": 85.25851440429688,
      "margin_dpo/margin_std": 80.23634338378906,
      "step": 497
    },
    {
      "epoch": 0.7528344671201814,
      "grad_norm": 15.286055564880371,
      "learning_rate": 8.829247120198563e-08,
      "logits/chosen": 0.9226865768432617,
      "logits/rejected": 0.8130264282226562,
      "logps/chosen": -209.95169067382812,
      "logps/ref_chosen": -62.02671432495117,
      "logps/ref_rejected": -102.47027587890625,
      "logps/rejected": -339.2867431640625,
      "loss": 0.9486,
      "margin_dpo/margin_mean": 88.89149475097656,
      "margin_dpo/margin_std": 98.82290649414062,
      "step": 498
    },
    {
      "epoch": 0.7543461829176115,
      "grad_norm": 19.15370750427246,
      "learning_rate": 8.728636813280163e-08,
      "logits/chosen": 0.9353519082069397,
      "logits/rejected": 0.817765474319458,
      "logps/chosen": -201.40264892578125,
      "logps/ref_chosen": -59.132904052734375,
      "logps/ref_rejected": -109.99728393554688,
      "logps/rejected": -319.89678955078125,
      "loss": 1.0559,
      "margin_dpo/margin_mean": 67.6297607421875,
      "margin_dpo/margin_std": 111.14754486083984,
      "step": 499
    },
    {
      "epoch": 0.7558578987150416,
      "grad_norm": 19.469440460205078,
      "learning_rate": 8.628481651367875e-08,
      "logits/chosen": 0.7339733839035034,
      "logits/rejected": 0.7467474341392517,
      "logps/chosen": -233.1659393310547,
      "logps/ref_chosen": -83.37364959716797,
      "logps/ref_rejected": -85.28791809082031,
      "logps/rejected": -321.4836730957031,
      "loss": 1.0818,
      "margin_dpo/margin_mean": 86.4034423828125,
      "margin_dpo/margin_std": 75.89598083496094,
      "step": 500
    },
    {
      "epoch": 0.7573696145124716,
      "grad_norm": 13.901487350463867,
      "learning_rate": 8.528784436016878e-08,
      "logits/chosen": 0.9171326756477356,
      "logits/rejected": 0.9103375673294067,
      "logps/chosen": -201.10220336914062,
      "logps/ref_chosen": -71.76341247558594,
      "logps/ref_rejected": -77.99368286132812,
      "logps/rejected": -288.089599609375,
      "loss": 0.9773,
      "margin_dpo/margin_mean": 80.75711059570312,
      "margin_dpo/margin_std": 61.29353713989258,
      "step": 501
    },
    {
      "epoch": 0.7588813303099018,
      "grad_norm": 16.59943389892578,
      "learning_rate": 8.4295479559726e-08,
      "logits/chosen": 0.9081599712371826,
      "logits/rejected": 0.8390029072761536,
      "logps/chosen": -177.0962677001953,
      "logps/ref_chosen": -60.16824722290039,
      "logps/ref_rejected": -88.61361694335938,
      "logps/rejected": -263.456787109375,
      "loss": 1.0216,
      "margin_dpo/margin_mean": 57.91515350341797,
      "margin_dpo/margin_std": 97.9842300415039,
      "step": 502
    },
    {
      "epoch": 0.7603930461073318,
      "grad_norm": 15.813238143920898,
      "learning_rate": 8.330774987092712e-08,
      "logits/chosen": 0.9412652254104614,
      "logits/rejected": 0.9647856950759888,
      "logps/chosen": -200.7902374267578,
      "logps/ref_chosen": -64.67495727539062,
      "logps/ref_rejected": -68.66828918457031,
      "logps/rejected": -268.06390380859375,
      "loss": 1.0442,
      "margin_dpo/margin_mean": 63.280311584472656,
      "margin_dpo/margin_std": 97.19824981689453,
      "step": 503
    },
    {
      "epoch": 0.7619047619047619,
      "grad_norm": 15.342427253723145,
      "learning_rate": 8.232468292269479e-08,
      "logits/chosen": 0.9019661545753479,
      "logits/rejected": 0.8831943273544312,
      "logps/chosen": -195.31814575195312,
      "logps/ref_chosen": -59.16814041137695,
      "logps/ref_rejected": -80.79418182373047,
      "logps/rejected": -313.3367919921875,
      "loss": 0.88,
      "margin_dpo/margin_mean": 96.392578125,
      "margin_dpo/margin_std": 102.2872314453125,
      "step": 504
    },
    {
      "epoch": 0.763416477702192,
      "grad_norm": 18.31743621826172,
      "learning_rate": 8.134630621352483e-08,
      "logits/chosen": 0.8820549249649048,
      "logits/rejected": 0.8513251543045044,
      "logps/chosen": -191.5747528076172,
      "logps/ref_chosen": -70.2003173828125,
      "logps/ref_rejected": -96.7729721069336,
      "logps/rejected": -323.912841796875,
      "loss": 1.1311,
      "margin_dpo/margin_mean": 105.76544189453125,
      "margin_dpo/margin_std": 129.75132751464844,
      "step": 505
    },
    {
      "epoch": 0.764928193499622,
      "grad_norm": 19.81454849243164,
      "learning_rate": 8.037264711071698e-08,
      "logits/chosen": 0.9372712969779968,
      "logits/rejected": 0.9639301300048828,
      "logps/chosen": -213.41952514648438,
      "logps/ref_chosen": -77.78474426269531,
      "logps/ref_rejected": -76.35060119628906,
      "logps/rejected": -262.79083251953125,
      "loss": 1.1566,
      "margin_dpo/margin_mean": 50.80542755126953,
      "margin_dpo/margin_std": 70.04209899902344,
      "step": 506
    },
    {
      "epoch": 0.7664399092970522,
      "grad_norm": 16.390592575073242,
      "learning_rate": 7.940373284960933e-08,
      "logits/chosen": 0.8365110158920288,
      "logits/rejected": 0.869307279586792,
      "logps/chosen": -263.01031494140625,
      "logps/ref_chosen": -102.45763397216797,
      "logps/ref_rejected": -107.32395935058594,
      "logps/rejected": -325.25531005859375,
      "loss": 1.0336,
      "margin_dpo/margin_mean": 57.37864685058594,
      "margin_dpo/margin_std": 107.75471496582031,
      "step": 507
    },
    {
      "epoch": 0.7679516250944822,
      "grad_norm": 18.062192916870117,
      "learning_rate": 7.843959053281663e-08,
      "logits/chosen": 0.9670290946960449,
      "logits/rejected": 0.8693418502807617,
      "logps/chosen": -160.03451538085938,
      "logps/ref_chosen": -54.93028259277344,
      "logps/ref_rejected": -94.2146987915039,
      "logps/rejected": -347.9501953125,
      "loss": 0.9714,
      "margin_dpo/margin_mean": 148.63128662109375,
      "margin_dpo/margin_std": 169.13424682617188,
      "step": 508
    },
    {
      "epoch": 0.7694633408919124,
      "grad_norm": 21.243953704833984,
      "learning_rate": 7.748024712947204e-08,
      "logits/chosen": 0.7485306262969971,
      "logits/rejected": 0.8197051286697388,
      "logps/chosen": -231.2813262939453,
      "logps/ref_chosen": -68.31385803222656,
      "logps/ref_rejected": -49.03086853027344,
      "logps/rejected": -233.74880981445312,
      "loss": 1.0812,
      "margin_dpo/margin_mean": 21.75048065185547,
      "margin_dpo/margin_std": 92.87283325195312,
      "step": 509
    },
    {
      "epoch": 0.7709750566893424,
      "grad_norm": 18.86398696899414,
      "learning_rate": 7.652572947447272e-08,
      "logits/chosen": 0.8953909873962402,
      "logits/rejected": 0.8276166319847107,
      "logps/chosen": -233.6016845703125,
      "logps/ref_chosen": -73.07958984375,
      "logps/ref_rejected": -96.33815002441406,
      "logps/rejected": -307.2120666503906,
      "loss": 1.0692,
      "margin_dpo/margin_mean": 50.35181427001953,
      "margin_dpo/margin_std": 74.205322265625,
      "step": 510
    },
    {
      "epoch": 0.7724867724867724,
      "grad_norm": 21.314716339111328,
      "learning_rate": 7.557606426772961e-08,
      "logits/chosen": 1.0289305448532104,
      "logits/rejected": 0.9097828269004822,
      "logps/chosen": -169.629638671875,
      "logps/ref_chosen": -42.50914764404297,
      "logps/ref_rejected": -75.06625366210938,
      "logps/rejected": -301.7984619140625,
      "loss": 0.9783,
      "margin_dpo/margin_mean": 99.61170959472656,
      "margin_dpo/margin_std": 98.32453155517578,
      "step": 511
    },
    {
      "epoch": 0.7739984882842026,
      "grad_norm": 31.400192260742188,
      "learning_rate": 7.463127807341966e-08,
      "logits/chosen": 0.9768285155296326,
      "logits/rejected": 1.0176584720611572,
      "logps/chosen": -181.43753051757812,
      "logps/ref_chosen": -56.2095947265625,
      "logps/ref_rejected": -53.511173248291016,
      "logps/rejected": -220.98828125,
      "loss": 1.1303,
      "margin_dpo/margin_mean": 42.249168395996094,
      "margin_dpo/margin_std": 113.38427734375,
      "step": 512
    },
    {
      "epoch": 0.7755102040816326,
      "grad_norm": 18.380573272705078,
      "learning_rate": 7.369139731924401e-08,
      "logits/chosen": 1.0797477960586548,
      "logits/rejected": 1.0573055744171143,
      "logps/chosen": -140.5891571044922,
      "logps/ref_chosen": -38.419105529785156,
      "logps/ref_rejected": -52.706207275390625,
      "logps/rejected": -238.41525268554688,
      "loss": 0.9919,
      "margin_dpo/margin_mean": 83.53899383544922,
      "margin_dpo/margin_std": 82.92984008789062,
      "step": 513
    },
    {
      "epoch": 0.7770219198790628,
      "grad_norm": 18.939552307128906,
      "learning_rate": 7.275644829568747e-08,
      "logits/chosen": 0.9963523149490356,
      "logits/rejected": 0.9776211977005005,
      "logps/chosen": -203.09329223632812,
      "logps/ref_chosen": -62.999786376953125,
      "logps/ref_rejected": -66.38030242919922,
      "logps/rejected": -280.66339111328125,
      "loss": 1.0044,
      "margin_dpo/margin_mean": 74.18960571289062,
      "margin_dpo/margin_std": 45.82038116455078,
      "step": 514
    },
    {
      "epoch": 0.7785336356764928,
      "grad_norm": 18.924774169921875,
      "learning_rate": 7.182645715528435e-08,
      "logits/chosen": 1.1489193439483643,
      "logits/rejected": 1.0843408107757568,
      "logps/chosen": -159.77142333984375,
      "logps/ref_chosen": -38.9024658203125,
      "logps/ref_rejected": -66.47667694091797,
      "logps/rejected": -273.12286376953125,
      "loss": 1.0422,
      "margin_dpo/margin_mean": 85.77719116210938,
      "margin_dpo/margin_std": 76.35650634765625,
      "step": 515
    },
    {
      "epoch": 0.780045351473923,
      "grad_norm": 17.709348678588867,
      "learning_rate": 7.090144991188568e-08,
      "logits/chosen": 0.980697751045227,
      "logits/rejected": 0.9438163042068481,
      "logps/chosen": -185.6435546875,
      "logps/ref_chosen": -56.46565246582031,
      "logps/ref_rejected": -69.74240112304688,
      "logps/rejected": -303.88299560546875,
      "loss": 1.0754,
      "margin_dpo/margin_mean": 104.96269989013672,
      "margin_dpo/margin_std": 115.3629150390625,
      "step": 516
    },
    {
      "epoch": 0.781557067271353,
      "grad_norm": 19.174524307250977,
      "learning_rate": 6.998145243993284e-08,
      "logits/chosen": 1.0660595893859863,
      "logits/rejected": 1.0764927864074707,
      "logps/chosen": -227.71185302734375,
      "logps/ref_chosen": -54.366859436035156,
      "logps/ref_rejected": -43.79294204711914,
      "logps/rejected": -237.5216064453125,
      "loss": 1.1571,
      "margin_dpo/margin_mean": 20.383655548095703,
      "margin_dpo/margin_std": 77.95108032226562,
      "step": 517
    },
    {
      "epoch": 0.783068783068783,
      "grad_norm": 17.056671142578125,
      "learning_rate": 6.906649047373245e-08,
      "logits/chosen": 0.927460789680481,
      "logits/rejected": 0.9620521068572998,
      "logps/chosen": -204.52810668945312,
      "logps/ref_chosen": -62.31662368774414,
      "logps/ref_rejected": -56.51953887939453,
      "logps/rejected": -246.00726318359375,
      "loss": 1.0561,
      "margin_dpo/margin_mean": 47.2762451171875,
      "margin_dpo/margin_std": 91.59005737304688,
      "step": 518
    },
    {
      "epoch": 0.7845804988662132,
      "grad_norm": 21.632381439208984,
      "learning_rate": 6.815658960673781e-08,
      "logits/chosen": 0.9049865007400513,
      "logits/rejected": 0.8612052202224731,
      "logps/chosen": -216.9791717529297,
      "logps/ref_chosen": -56.87085723876953,
      "logps/ref_rejected": -81.93241882324219,
      "logps/rejected": -281.4555969238281,
      "loss": 1.3521,
      "margin_dpo/margin_mean": 39.41484069824219,
      "margin_dpo/margin_std": 104.36680603027344,
      "step": 519
    },
    {
      "epoch": 0.7860922146636432,
      "grad_norm": 17.4406795501709,
      "learning_rate": 6.725177529083209e-08,
      "logits/chosen": 1.0133739709854126,
      "logits/rejected": 0.9440522789955139,
      "logps/chosen": -191.7762908935547,
      "logps/ref_chosen": -57.21955108642578,
      "logps/ref_rejected": -66.96128845214844,
      "logps/rejected": -295.70098876953125,
      "loss": 1.0819,
      "margin_dpo/margin_mean": 94.18293762207031,
      "margin_dpo/margin_std": 142.52069091796875,
      "step": 520
    },
    {
      "epoch": 0.7876039304610734,
      "grad_norm": 15.527779579162598,
      "learning_rate": 6.63520728356167e-08,
      "logits/chosen": 0.6215152740478516,
      "logits/rejected": 0.6250849962234497,
      "logps/chosen": -193.13436889648438,
      "logps/ref_chosen": -82.06413269042969,
      "logps/ref_rejected": -93.37128448486328,
      "logps/rejected": -282.43743896484375,
      "loss": 0.9091,
      "margin_dpo/margin_mean": 77.99594116210938,
      "margin_dpo/margin_std": 60.643978118896484,
      "step": 521
    },
    {
      "epoch": 0.7891156462585034,
      "grad_norm": 17.787490844726562,
      "learning_rate": 6.545750740770336e-08,
      "logits/chosen": 0.9250505566596985,
      "logits/rejected": 0.8676795959472656,
      "logps/chosen": -171.79269409179688,
      "logps/ref_chosen": -36.52648162841797,
      "logps/ref_rejected": -69.4611587524414,
      "logps/rejected": -259.8004455566406,
      "loss": 1.2173,
      "margin_dpo/margin_mean": 55.07307434082031,
      "margin_dpo/margin_std": 102.25482940673828,
      "step": 522
    },
    {
      "epoch": 0.7906273620559335,
      "grad_norm": 19.3404483795166,
      "learning_rate": 6.456810403001012e-08,
      "logits/chosen": 1.0570783615112305,
      "logits/rejected": 0.8110660910606384,
      "logps/chosen": -182.0962371826172,
      "logps/ref_chosen": -43.33647918701172,
      "logps/ref_rejected": -82.51911926269531,
      "logps/rejected": -322.3846130371094,
      "loss": 1.0791,
      "margin_dpo/margin_mean": 101.10575103759766,
      "margin_dpo/margin_std": 114.18901062011719,
      "step": 523
    },
    {
      "epoch": 0.7921390778533636,
      "grad_norm": 18.590288162231445,
      "learning_rate": 6.368388758106134e-08,
      "logits/chosen": 0.7667090892791748,
      "logits/rejected": 0.7514083385467529,
      "logps/chosen": -188.31198120117188,
      "logps/ref_chosen": -71.81331634521484,
      "logps/ref_rejected": -80.51419067382812,
      "logps/rejected": -263.0672302246094,
      "loss": 1.0755,
      "margin_dpo/margin_mean": 66.05437469482422,
      "margin_dpo/margin_std": 116.67527770996094,
      "step": 524
    },
    {
      "epoch": 0.7936507936507936,
      "grad_norm": 19.926000595092773,
      "learning_rate": 6.280488279429185e-08,
      "logits/chosen": 0.6996693015098572,
      "logits/rejected": 0.661113977432251,
      "logps/chosen": -237.54904174804688,
      "logps/ref_chosen": -86.49774169921875,
      "logps/ref_rejected": -98.95057678222656,
      "logps/rejected": -275.597412109375,
      "loss": 1.1514,
      "margin_dpo/margin_mean": 25.595535278320312,
      "margin_dpo/margin_std": 67.40962219238281,
      "step": 525
    },
    {
      "epoch": 0.7951625094482238,
      "grad_norm": 18.31951332092285,
      "learning_rate": 6.193111425735515e-08,
      "logits/chosen": 0.8428486585617065,
      "logits/rejected": 0.7279390096664429,
      "logps/chosen": -215.59396362304688,
      "logps/ref_chosen": -71.5922622680664,
      "logps/ref_rejected": -89.43500518798828,
      "logps/rejected": -250.84075927734375,
      "loss": 1.1416,
      "margin_dpo/margin_mean": 17.404056549072266,
      "margin_dpo/margin_std": 81.23014831542969,
      "step": 526
    },
    {
      "epoch": 0.7966742252456538,
      "grad_norm": 21.556499481201172,
      "learning_rate": 6.106260641143546e-08,
      "logits/chosen": 1.0022237300872803,
      "logits/rejected": 0.8509221076965332,
      "logps/chosen": -152.88330078125,
      "logps/ref_chosen": -39.413856506347656,
      "logps/ref_rejected": -96.52778625488281,
      "logps/rejected": -277.23040771484375,
      "loss": 1.2379,
      "margin_dpo/margin_mean": 67.23316192626953,
      "margin_dpo/margin_std": 81.01995086669922,
      "step": 527
    },
    {
      "epoch": 0.7981859410430839,
      "grad_norm": 22.758813858032227,
      "learning_rate": 6.019938355056422e-08,
      "logits/chosen": 0.6834633350372314,
      "logits/rejected": 0.8445563316345215,
      "logps/chosen": -205.08164978027344,
      "logps/ref_chosen": -77.37605285644531,
      "logps/ref_rejected": -46.57087707519531,
      "logps/rejected": -181.67904663085938,
      "loss": 1.249,
      "margin_dpo/margin_mean": 7.4025726318359375,
      "margin_dpo/margin_std": 108.43580627441406,
      "step": 528
    },
    {
      "epoch": 0.799697656840514,
      "grad_norm": 14.703264236450195,
      "learning_rate": 5.934146982094049e-08,
      "logits/chosen": 0.8407646417617798,
      "logits/rejected": 0.8265160322189331,
      "logps/chosen": -189.27963256835938,
      "logps/ref_chosen": -58.75701141357422,
      "logps/ref_rejected": -71.7961196899414,
      "logps/rejected": -265.8705749511719,
      "loss": 0.8373,
      "margin_dpo/margin_mean": 63.551815032958984,
      "margin_dpo/margin_std": 80.8453598022461,
      "step": 529
    },
    {
      "epoch": 0.8012093726379441,
      "grad_norm": 16.264793395996094,
      "learning_rate": 5.848888922025552e-08,
      "logits/chosen": 1.0589945316314697,
      "logits/rejected": 0.912827730178833,
      "logps/chosen": -172.25611877441406,
      "logps/ref_chosen": -45.41581726074219,
      "logps/ref_rejected": -80.25865173339844,
      "logps/rejected": -290.4316711425781,
      "loss": 1.1008,
      "margin_dpo/margin_mean": 83.33270263671875,
      "margin_dpo/margin_std": 107.66976165771484,
      "step": 530
    },
    {
      "epoch": 0.8027210884353742,
      "grad_norm": 18.56505584716797,
      "learning_rate": 5.7641665597021435e-08,
      "logits/chosen": 1.0082396268844604,
      "logits/rejected": 0.9188197255134583,
      "logps/chosen": -170.84205627441406,
      "logps/ref_chosen": -49.827571868896484,
      "logps/ref_rejected": -87.68862915039062,
      "logps/rejected": -295.68585205078125,
      "loss": 1.1211,
      "margin_dpo/margin_mean": 86.98274993896484,
      "margin_dpo/margin_std": 96.60000610351562,
      "step": 531
    },
    {
      "epoch": 0.8042328042328042,
      "grad_norm": 18.886518478393555,
      "learning_rate": 5.679982264990424e-08,
      "logits/chosen": 0.8886632919311523,
      "logits/rejected": 0.7267423868179321,
      "logps/chosen": -214.15121459960938,
      "logps/ref_chosen": -59.289215087890625,
      "logps/ref_rejected": -88.06941223144531,
      "logps/rejected": -338.8141174316406,
      "loss": 1.069,
      "margin_dpo/margin_mean": 95.8827133178711,
      "margin_dpo/margin_std": 99.42604064941406,
      "step": 532
    },
    {
      "epoch": 0.8057445200302343,
      "grad_norm": 16.084077835083008,
      "learning_rate": 5.596338392706076e-08,
      "logits/chosen": 1.04569411277771,
      "logits/rejected": 1.0000990629196167,
      "logps/chosen": -164.958740234375,
      "logps/ref_chosen": -57.924049377441406,
      "logps/ref_rejected": -70.69439697265625,
      "logps/rejected": -268.8208312988281,
      "loss": 0.9852,
      "margin_dpo/margin_mean": 91.09176635742188,
      "margin_dpo/margin_std": 97.24915313720703,
      "step": 533
    },
    {
      "epoch": 0.8072562358276644,
      "grad_norm": 18.018505096435547,
      "learning_rate": 5.513237282548033e-08,
      "logits/chosen": 0.8638824820518494,
      "logits/rejected": 0.6540743708610535,
      "logps/chosen": -193.56890869140625,
      "logps/ref_chosen": -59.63468551635742,
      "logps/ref_rejected": -109.55838012695312,
      "logps/rejected": -294.1387634277344,
      "loss": 1.0613,
      "margin_dpo/margin_mean": 50.64615249633789,
      "margin_dpo/margin_std": 116.63580322265625,
      "step": 534
    },
    {
      "epoch": 0.8087679516250945,
      "grad_norm": 19.90399932861328,
      "learning_rate": 5.430681259032957e-08,
      "logits/chosen": 0.7901930212974548,
      "logits/rejected": 0.7147120237350464,
      "logps/chosen": -205.218994140625,
      "logps/ref_chosen": -45.231201171875,
      "logps/ref_rejected": -70.03333282470703,
      "logps/rejected": -260.12701416015625,
      "loss": 1.1115,
      "margin_dpo/margin_mean": 30.105876922607422,
      "margin_dpo/margin_std": 91.64392852783203,
      "step": 535
    },
    {
      "epoch": 0.8102796674225246,
      "grad_norm": 12.067679405212402,
      "learning_rate": 5.3486726314303175e-08,
      "logits/chosen": 0.9899567365646362,
      "logits/rejected": 0.9399088621139526,
      "logps/chosen": -187.06640625,
      "logps/ref_chosen": -61.66278839111328,
      "logps/ref_rejected": -78.1522445678711,
      "logps/rejected": -277.64288330078125,
      "loss": 0.8678,
      "margin_dpo/margin_mean": 74.0870361328125,
      "margin_dpo/margin_std": 70.92481994628906,
      "step": 536
    },
    {
      "epoch": 0.8117913832199547,
      "grad_norm": 18.532299041748047,
      "learning_rate": 5.267213693697695e-08,
      "logits/chosen": 0.9581376314163208,
      "logits/rejected": 0.8182666301727295,
      "logps/chosen": -162.33416748046875,
      "logps/ref_chosen": -39.777610778808594,
      "logps/ref_rejected": -122.8561782836914,
      "logps/rejected": -347.7552490234375,
      "loss": 1.0719,
      "margin_dpo/margin_mean": 102.342529296875,
      "margin_dpo/margin_std": 55.462013244628906,
      "step": 537
    },
    {
      "epoch": 0.8133030990173847,
      "grad_norm": 15.579018592834473,
      "learning_rate": 5.1863067244167144e-08,
      "logits/chosen": 1.0169093608856201,
      "logits/rejected": 1.08309006690979,
      "logps/chosen": -190.21852111816406,
      "logps/ref_chosen": -53.97972106933594,
      "logps/ref_rejected": -40.60531997680664,
      "logps/rejected": -223.557373046875,
      "loss": 0.9732,
      "margin_dpo/margin_mean": 46.71326446533203,
      "margin_dpo/margin_std": 72.07386779785156,
      "step": 538
    },
    {
      "epoch": 0.8148148148148148,
      "grad_norm": 18.707054138183594,
      "learning_rate": 5.105953986729195e-08,
      "logits/chosen": 0.7336928844451904,
      "logits/rejected": 0.6928164958953857,
      "logps/chosen": -191.83001708984375,
      "logps/ref_chosen": -75.5906753540039,
      "logps/ref_rejected": -90.88420104980469,
      "logps/rejected": -264.90667724609375,
      "loss": 1.0761,
      "margin_dpo/margin_mean": 57.78313064575195,
      "margin_dpo/margin_std": 57.98158264160156,
      "step": 539
    },
    {
      "epoch": 0.8163265306122449,
      "grad_norm": 17.138023376464844,
      "learning_rate": 5.026157728273966e-08,
      "logits/chosen": 1.0368293523788452,
      "logits/rejected": 0.9761728048324585,
      "logps/chosen": -202.86624145507812,
      "logps/ref_chosen": -53.29787063598633,
      "logps/ref_rejected": -80.79837799072266,
      "logps/rejected": -307.9012451171875,
      "loss": 0.9363,
      "margin_dpo/margin_mean": 77.53450012207031,
      "margin_dpo/margin_std": 124.25914764404297,
      "step": 540
    },
    {
      "epoch": 0.817838246409675,
      "grad_norm": 18.846561431884766,
      "learning_rate": 4.9469201811239035e-08,
      "logits/chosen": 0.8673323392868042,
      "logits/rejected": 1.0076346397399902,
      "logps/chosen": -233.70118713378906,
      "logps/ref_chosen": -89.37522888183594,
      "logps/ref_rejected": -50.772430419921875,
      "logps/rejected": -251.73175048828125,
      "loss": 0.9834,
      "margin_dpo/margin_mean": 56.633365631103516,
      "margin_dpo/margin_std": 85.56578063964844,
      "step": 541
    },
    {
      "epoch": 0.8193499622071051,
      "grad_norm": 13.740031242370605,
      "learning_rate": 4.868243561723534e-08,
      "logits/chosen": 0.8200675845146179,
      "logits/rejected": 0.7452304363250732,
      "logps/chosen": -162.4728240966797,
      "logps/ref_chosen": -53.12909698486328,
      "logps/ref_rejected": -85.17762756347656,
      "logps/rejected": -272.97418212890625,
      "loss": 0.9604,
      "margin_dpo/margin_mean": 78.45283508300781,
      "margin_dpo/margin_std": 69.77384948730469,
      "step": 542
    },
    {
      "epoch": 0.8208616780045351,
      "grad_norm": 15.152771949768066,
      "learning_rate": 4.790130070827028e-08,
      "logits/chosen": 0.8890964984893799,
      "logits/rejected": 0.8061411380767822,
      "logps/chosen": -171.46209716796875,
      "logps/ref_chosen": -49.52074432373047,
      "logps/ref_rejected": -89.69110870361328,
      "logps/rejected": -296.0673828125,
      "loss": 0.9858,
      "margin_dpo/margin_mean": 84.43494415283203,
      "margin_dpo/margin_std": 86.14839172363281,
      "step": 543
    },
    {
      "epoch": 0.8223733938019653,
      "grad_norm": 18.41205596923828,
      "learning_rate": 4.7125818934366454e-08,
      "logits/chosen": 0.8170167207717896,
      "logits/rejected": 0.8432954549789429,
      "logps/chosen": -270.7913818359375,
      "logps/ref_chosen": -79.90715026855469,
      "logps/ref_rejected": -65.11402130126953,
      "logps/rejected": -296.67626953125,
      "loss": 0.9642,
      "margin_dpo/margin_mean": 40.67804718017578,
      "margin_dpo/margin_std": 80.78584289550781,
      "step": 544
    },
    {
      "epoch": 0.8238851095993953,
      "grad_norm": 20.38388442993164,
      "learning_rate": 4.635601198741607e-08,
      "logits/chosen": 0.9498938322067261,
      "logits/rejected": 0.8978836536407471,
      "logps/chosen": -191.94818115234375,
      "logps/ref_chosen": -49.8741455078125,
      "logps/ref_rejected": -74.81036376953125,
      "logps/rejected": -259.6213073730469,
      "loss": 1.1264,
      "margin_dpo/margin_mean": 42.736907958984375,
      "margin_dpo/margin_std": 79.07319641113281,
      "step": 545
    },
    {
      "epoch": 0.8253968253968254,
      "grad_norm": 19.26015853881836,
      "learning_rate": 4.559190140057428e-08,
      "logits/chosen": 0.9439926743507385,
      "logits/rejected": 0.9385854601860046,
      "logps/chosen": -254.52891540527344,
      "logps/ref_chosen": -83.20497131347656,
      "logps/ref_rejected": -79.42691040039062,
      "logps/rejected": -289.6729736328125,
      "loss": 1.0739,
      "margin_dpo/margin_mean": 38.9221305847168,
      "margin_dpo/margin_std": 98.08988952636719,
      "step": 546
    },
    {
      "epoch": 0.8269085411942555,
      "grad_norm": 16.925718307495117,
      "learning_rate": 4.483350854765672e-08,
      "logits/chosen": 0.9709137678146362,
      "logits/rejected": 0.7829012870788574,
      "logps/chosen": -161.71536254882812,
      "logps/ref_chosen": -30.964988708496094,
      "logps/ref_rejected": -70.52467346191406,
      "logps/rejected": -272.62640380859375,
      "loss": 0.9703,
      "margin_dpo/margin_mean": 71.35137939453125,
      "margin_dpo/margin_std": 87.27641296386719,
      "step": 547
    },
    {
      "epoch": 0.8284202569916855,
      "grad_norm": 20.741260528564453,
      "learning_rate": 4.4080854642541826e-08,
      "logits/chosen": 0.7766395211219788,
      "logits/rejected": 0.7720953822135925,
      "logps/chosen": -190.9144287109375,
      "logps/ref_chosen": -63.495338439941406,
      "logps/ref_rejected": -73.84056091308594,
      "logps/rejected": -251.48435974121094,
      "loss": 1.1606,
      "margin_dpo/margin_mean": 50.224700927734375,
      "margin_dpo/margin_std": 94.93733215332031,
      "step": 548
    },
    {
      "epoch": 0.8299319727891157,
      "grad_norm": 27.655611038208008,
      "learning_rate": 4.333396073857723e-08,
      "logits/chosen": 0.9606146812438965,
      "logits/rejected": 0.9492313861846924,
      "logps/chosen": -232.9254608154297,
      "logps/ref_chosen": -73.39987182617188,
      "logps/ref_rejected": -76.37744140625,
      "logps/rejected": -271.0856628417969,
      "loss": 1.18,
      "margin_dpo/margin_mean": 35.182621002197266,
      "margin_dpo/margin_std": 107.7386703491211,
      "step": 549
    },
    {
      "epoch": 0.8314436885865457,
      "grad_norm": 17.47970199584961,
      "learning_rate": 4.259284772799099e-08,
      "logits/chosen": 1.0232315063476562,
      "logits/rejected": 1.0318495035171509,
      "logps/chosen": -138.86331176757812,
      "logps/ref_chosen": -41.076881408691406,
      "logps/ref_rejected": -47.863609313964844,
      "logps/rejected": -196.962158203125,
      "loss": 1.1886,
      "margin_dpo/margin_mean": 51.31211471557617,
      "margin_dpo/margin_std": 87.56314086914062,
      "step": 550
    },
    {
      "epoch": 0.8329554043839759,
      "grad_norm": 16.67303466796875,
      "learning_rate": 4.1857536341307176e-08,
      "logits/chosen": 1.0085614919662476,
      "logits/rejected": 0.9332855939865112,
      "logps/chosen": -187.7180633544922,
      "logps/ref_chosen": -43.32201385498047,
      "logps/ref_rejected": -80.82965850830078,
      "logps/rejected": -318.88507080078125,
      "loss": 1.0361,
      "margin_dpo/margin_mean": 93.65936279296875,
      "margin_dpo/margin_std": 136.92462158203125,
      "step": 551
    },
    {
      "epoch": 0.8344671201814059,
      "grad_norm": 20.140682220458984,
      "learning_rate": 4.112804714676593e-08,
      "logits/chosen": 1.030794382095337,
      "logits/rejected": 0.9187796115875244,
      "logps/chosen": -200.61541748046875,
      "logps/ref_chosen": -51.589683532714844,
      "logps/ref_rejected": -92.36204528808594,
      "logps/rejected": -306.091064453125,
      "loss": 1.0185,
      "margin_dpo/margin_mean": 64.70327758789062,
      "margin_dpo/margin_std": 70.84175109863281,
      "step": 552
    },
    {
      "epoch": 0.8359788359788359,
      "grad_norm": 18.41575050354004,
      "learning_rate": 4.0404400549748144e-08,
      "logits/chosen": 1.0242167711257935,
      "logits/rejected": 0.8356890082359314,
      "logps/chosen": -183.38101196289062,
      "logps/ref_chosen": -46.85206604003906,
      "logps/ref_rejected": -111.05335235595703,
      "logps/rejected": -361.5396728515625,
      "loss": 1.1174,
      "margin_dpo/margin_mean": 113.9573745727539,
      "margin_dpo/margin_std": 144.8773193359375,
      "step": 553
    },
    {
      "epoch": 0.8374905517762661,
      "grad_norm": 18.322969436645508,
      "learning_rate": 3.968661679220467e-08,
      "logits/chosen": 0.8415111303329468,
      "logits/rejected": 0.7810766696929932,
      "logps/chosen": -176.80612182617188,
      "logps/ref_chosen": -56.84330368041992,
      "logps/ref_rejected": -71.64338684082031,
      "logps/rejected": -253.65359497070312,
      "loss": 1.0157,
      "margin_dpo/margin_mean": 62.0473747253418,
      "margin_dpo/margin_std": 105.61851501464844,
      "step": 554
    },
    {
      "epoch": 0.8390022675736961,
      "grad_norm": 18.993671417236328,
      "learning_rate": 3.89747159520904e-08,
      "logits/chosen": 0.8262725472450256,
      "logits/rejected": 0.7698884010314941,
      "logps/chosen": -232.79461669921875,
      "logps/ref_chosen": -88.44906616210938,
      "logps/ref_rejected": -92.66055297851562,
      "logps/rejected": -271.56744384765625,
      "loss": 1.1031,
      "margin_dpo/margin_mean": 34.56132125854492,
      "margin_dpo/margin_std": 118.9439926147461,
      "step": 555
    },
    {
      "epoch": 0.8405139833711263,
      "grad_norm": 17.27111053466797,
      "learning_rate": 3.826871794280192e-08,
      "logits/chosen": 0.9412505030632019,
      "logits/rejected": 0.9116497039794922,
      "logps/chosen": -217.48727416992188,
      "logps/ref_chosen": -56.714725494384766,
      "logps/ref_rejected": -64.49860382080078,
      "logps/rejected": -274.18817138671875,
      "loss": 1.1464,
      "margin_dpo/margin_mean": 48.91699981689453,
      "margin_dpo/margin_std": 71.00090026855469,
      "step": 556
    },
    {
      "epoch": 0.8420256991685563,
      "grad_norm": 16.771560668945312,
      "learning_rate": 3.756864251262143e-08,
      "logits/chosen": 0.9594881534576416,
      "logits/rejected": 0.9018399715423584,
      "logps/chosen": -157.4939422607422,
      "logps/ref_chosen": -35.662818908691406,
      "logps/ref_rejected": -49.78130340576172,
      "logps/rejected": -229.19021606445312,
      "loss": 0.9731,
      "margin_dpo/margin_mean": 57.57780456542969,
      "margin_dpo/margin_std": 78.69767761230469,
      "step": 557
    },
    {
      "epoch": 0.8435374149659864,
      "grad_norm": 15.087602615356445,
      "learning_rate": 3.687450924416341e-08,
      "logits/chosen": 1.0110355615615845,
      "logits/rejected": 0.9612963199615479,
      "logps/chosen": -196.357666015625,
      "logps/ref_chosen": -67.20350646972656,
      "logps/ref_rejected": -90.46063232421875,
      "logps/rejected": -308.90771484375,
      "loss": 0.9023,
      "margin_dpo/margin_mean": 89.29290771484375,
      "margin_dpo/margin_std": 79.98139953613281,
      "step": 558
    },
    {
      "epoch": 0.8450491307634165,
      "grad_norm": 15.841297149658203,
      "learning_rate": 3.6186337553827743e-08,
      "logits/chosen": 0.8561170101165771,
      "logits/rejected": 0.7312831878662109,
      "logps/chosen": -225.37515258789062,
      "logps/ref_chosen": -82.38276672363281,
      "logps/ref_rejected": -148.32070922851562,
      "logps/rejected": -394.68408203125,
      "loss": 1.0114,
      "margin_dpo/margin_mean": 103.37101745605469,
      "margin_dpo/margin_std": 142.59449768066406,
      "step": 559
    },
    {
      "epoch": 0.8465608465608465,
      "grad_norm": 18.28428077697754,
      "learning_rate": 3.550414669125573e-08,
      "logits/chosen": 0.8422687649726868,
      "logits/rejected": 0.8341982364654541,
      "logps/chosen": -237.72918701171875,
      "logps/ref_chosen": -87.37367248535156,
      "logps/ref_rejected": -98.05863952636719,
      "logps/rejected": -306.2855529785156,
      "loss": 1.0279,
      "margin_dpo/margin_mean": 57.87139892578125,
      "margin_dpo/margin_std": 98.06793212890625,
      "step": 560
    },
    {
      "epoch": 0.8480725623582767,
      "grad_norm": 15.278297424316406,
      "learning_rate": 3.482795573879241e-08,
      "logits/chosen": 0.9140326976776123,
      "logits/rejected": 0.8790804743766785,
      "logps/chosen": -175.6079559326172,
      "logps/ref_chosen": -46.72880935668945,
      "logps/ref_rejected": -72.11878204345703,
      "logps/rejected": -249.57186889648438,
      "loss": 1.0484,
      "margin_dpo/margin_mean": 48.573936462402344,
      "margin_dpo/margin_std": 56.30522155761719,
      "step": 561
    },
    {
      "epoch": 0.8495842781557067,
      "grad_norm": 17.868244171142578,
      "learning_rate": 3.415778361095226e-08,
      "logits/chosen": 0.9695894718170166,
      "logits/rejected": 0.8319031000137329,
      "logps/chosen": -219.78314208984375,
      "logps/ref_chosen": -73.16322326660156,
      "logps/ref_rejected": -130.28079223632812,
      "logps/rejected": -364.35076904296875,
      "loss": 0.9357,
      "margin_dpo/margin_mean": 87.45001220703125,
      "margin_dpo/margin_std": 103.97313690185547,
      "step": 562
    },
    {
      "epoch": 0.8510959939531368,
      "grad_norm": 20.500097274780273,
      "learning_rate": 3.349364905389032e-08,
      "logits/chosen": 1.082218885421753,
      "logits/rejected": 1.0179955959320068,
      "logps/chosen": -173.17086791992188,
      "logps/ref_chosen": -52.58354187011719,
      "logps/ref_rejected": -81.23396301269531,
      "logps/rejected": -261.9834899902344,
      "loss": 1.0687,
      "margin_dpo/margin_mean": 60.16220474243164,
      "margin_dpo/margin_std": 83.78437805175781,
      "step": 563
    },
    {
      "epoch": 0.8526077097505669,
      "grad_norm": 20.62779426574707,
      "learning_rate": 3.283557064487785e-08,
      "logits/chosen": 0.9116697311401367,
      "logits/rejected": 0.9356800317764282,
      "logps/chosen": -168.27835083007812,
      "logps/ref_chosen": -54.56080627441406,
      "logps/ref_rejected": -52.1234245300293,
      "logps/rejected": -250.03018188476562,
      "loss": 0.9691,
      "margin_dpo/margin_mean": 84.18922424316406,
      "margin_dpo/margin_std": 93.12564849853516,
      "step": 564
    },
    {
      "epoch": 0.854119425547997,
      "grad_norm": 17.084941864013672,
      "learning_rate": 3.218356679178252e-08,
      "logits/chosen": 0.9586303234100342,
      "logits/rejected": 0.7509380578994751,
      "logps/chosen": -274.4070129394531,
      "logps/ref_chosen": -56.54259490966797,
      "logps/ref_rejected": -125.5247573852539,
      "logps/rejected": -408.77130126953125,
      "loss": 0.9844,
      "margin_dpo/margin_mean": 65.38215637207031,
      "margin_dpo/margin_std": 83.48878479003906,
      "step": 565
    },
    {
      "epoch": 0.8556311413454271,
      "grad_norm": 17.82448959350586,
      "learning_rate": 3.1537655732553764e-08,
      "logits/chosen": 0.9717875719070435,
      "logits/rejected": 0.9099733233451843,
      "logps/chosen": -195.46827697753906,
      "logps/ref_chosen": -66.43058776855469,
      "logps/ref_rejected": -88.7685775756836,
      "logps/rejected": -317.3566589355469,
      "loss": 1.0874,
      "margin_dpo/margin_mean": 99.5504150390625,
      "margin_dpo/margin_std": 77.15394592285156,
      "step": 566
    },
    {
      "epoch": 0.8571428571428571,
      "grad_norm": 15.454371452331543,
      "learning_rate": 3.089785553471233e-08,
      "logits/chosen": 0.8322715759277344,
      "logits/rejected": 0.8646911978721619,
      "logps/chosen": -187.09246826171875,
      "logps/ref_chosen": -61.010929107666016,
      "logps/ref_rejected": -59.36076736450195,
      "logps/rejected": -264.74932861328125,
      "loss": 0.9967,
      "margin_dpo/margin_mean": 79.30699920654297,
      "margin_dpo/margin_std": 74.8924789428711,
      "step": 567
    },
    {
      "epoch": 0.8586545729402872,
      "grad_norm": 16.217979431152344,
      "learning_rate": 3.026418409484513e-08,
      "logits/chosen": 1.0041697025299072,
      "logits/rejected": 0.9253759384155273,
      "logps/chosen": -160.50274658203125,
      "logps/ref_chosen": -39.93986892700195,
      "logps/ref_rejected": -80.44910430908203,
      "logps/rejected": -245.42349243164062,
      "loss": 0.9188,
      "margin_dpo/margin_mean": 44.4115104675293,
      "margin_dpo/margin_std": 74.29348754882812,
      "step": 568
    },
    {
      "epoch": 0.8601662887377173,
      "grad_norm": 20.45829963684082,
      "learning_rate": 2.963665913810451e-08,
      "logits/chosen": 0.9038550853729248,
      "logits/rejected": 0.9029619693756104,
      "logps/chosen": -228.15231323242188,
      "logps/ref_chosen": -81.28688049316406,
      "logps/ref_rejected": -83.6900405883789,
      "logps/rejected": -316.5882873535156,
      "loss": 1.1757,
      "margin_dpo/margin_mean": 86.03282165527344,
      "margin_dpo/margin_std": 147.73619079589844,
      "step": 569
    },
    {
      "epoch": 0.8616780045351474,
      "grad_norm": 17.267189025878906,
      "learning_rate": 2.9015298217712453e-08,
      "logits/chosen": 0.9757269620895386,
      "logits/rejected": 0.9647901058197021,
      "logps/chosen": -202.49285888671875,
      "logps/ref_chosen": -53.46790313720703,
      "logps/ref_rejected": -51.320770263671875,
      "logps/rejected": -260.1011962890625,
      "loss": 0.9055,
      "margin_dpo/margin_mean": 59.75544738769531,
      "margin_dpo/margin_std": 78.49115753173828,
      "step": 570
    },
    {
      "epoch": 0.8631897203325775,
      "grad_norm": 19.252838134765625,
      "learning_rate": 2.840011871446962e-08,
      "logits/chosen": 0.9464564323425293,
      "logits/rejected": 0.8572288751602173,
      "logps/chosen": -164.54974365234375,
      "logps/ref_chosen": -45.11099624633789,
      "logps/ref_rejected": -71.55215454101562,
      "logps/rejected": -285.15167236328125,
      "loss": 1.1555,
      "margin_dpo/margin_mean": 94.1607666015625,
      "margin_dpo/margin_std": 116.37611389160156,
      "step": 571
    },
    {
      "epoch": 0.8647014361300076,
      "grad_norm": 22.696313858032227,
      "learning_rate": 2.7791137836269158e-08,
      "logits/chosen": 0.9143853187561035,
      "logits/rejected": 0.863966703414917,
      "logps/chosen": -178.92416381835938,
      "logps/ref_chosen": -50.03799819946289,
      "logps/ref_rejected": -69.7855224609375,
      "logps/rejected": -282.4919738769531,
      "loss": 1.0403,
      "margin_dpo/margin_mean": 83.82028198242188,
      "margin_dpo/margin_std": 97.30913543701172,
      "step": 572
    },
    {
      "epoch": 0.8662131519274376,
      "grad_norm": 17.95223617553711,
      "learning_rate": 2.718837261761528e-08,
      "logits/chosen": 0.9436647891998291,
      "logits/rejected": 0.8891823291778564,
      "logps/chosen": -223.79197692871094,
      "logps/ref_chosen": -65.06491088867188,
      "logps/ref_rejected": -74.16371154785156,
      "logps/rejected": -237.40768432617188,
      "loss": 1.1207,
      "margin_dpo/margin_mean": 4.516890525817871,
      "margin_dpo/margin_std": 93.70068359375,
      "step": 573
    },
    {
      "epoch": 0.8677248677248677,
      "grad_norm": 13.765616416931152,
      "learning_rate": 2.659183991914696e-08,
      "logits/chosen": 0.9557688236236572,
      "logits/rejected": 0.9775044918060303,
      "logps/chosen": -214.97642517089844,
      "logps/ref_chosen": -50.88386917114258,
      "logps/ref_rejected": -51.98750686645508,
      "logps/rejected": -228.66259765625,
      "loss": 0.8728,
      "margin_dpo/margin_mean": 12.582534790039062,
      "margin_dpo/margin_std": 53.39247131347656,
      "step": 574
    },
    {
      "epoch": 0.8692365835222978,
      "grad_norm": 17.010929107666016,
      "learning_rate": 2.600155642716606e-08,
      "logits/chosen": 0.9337095022201538,
      "logits/rejected": 0.786598801612854,
      "logps/chosen": -208.71426391601562,
      "logps/ref_chosen": -61.41114044189453,
      "logps/ref_rejected": -112.07469177246094,
      "logps/rejected": -286.92864990234375,
      "loss": 1.1869,
      "margin_dpo/margin_mean": 27.55083465576172,
      "margin_dpo/margin_std": 119.59794616699219,
      "step": 575
    },
    {
      "epoch": 0.8707482993197279,
      "grad_norm": 15.698728561401367,
      "learning_rate": 2.5417538653170754e-08,
      "logits/chosen": 1.0223554372787476,
      "logits/rejected": 0.9235273599624634,
      "logps/chosen": -183.23175048828125,
      "logps/ref_chosen": -54.792724609375,
      "logps/ref_rejected": -84.78044128417969,
      "logps/rejected": -297.049072265625,
      "loss": 0.8786,
      "margin_dpo/margin_mean": 83.82960510253906,
      "margin_dpo/margin_std": 106.40802001953125,
      "step": 576
    },
    {
      "epoch": 0.872260015117158,
      "grad_norm": 17.254444122314453,
      "learning_rate": 2.4839802933393607e-08,
      "logits/chosen": 1.0162341594696045,
      "logits/rejected": 0.9787265062332153,
      "logps/chosen": -163.14552307128906,
      "logps/ref_chosen": -44.14771270751953,
      "logps/ref_rejected": -65.19963836669922,
      "logps/rejected": -230.81610107421875,
      "loss": 1.113,
      "margin_dpo/margin_mean": 46.61864471435547,
      "margin_dpo/margin_std": 70.7182846069336,
      "step": 577
    },
    {
      "epoch": 0.873771730914588,
      "grad_norm": 17.691556930541992,
      "learning_rate": 2.4268365428344733e-08,
      "logits/chosen": 1.0092999935150146,
      "logits/rejected": 0.9245076179504395,
      "logps/chosen": -200.7638702392578,
      "logps/ref_chosen": -47.827537536621094,
      "logps/ref_rejected": -84.31831359863281,
      "logps/rejected": -295.13720703125,
      "loss": 1.1273,
      "margin_dpo/margin_mean": 57.8825798034668,
      "margin_dpo/margin_std": 108.26670837402344,
      "step": 578
    },
    {
      "epoch": 0.8752834467120182,
      "grad_norm": 17.17439842224121,
      "learning_rate": 2.3703242122359357e-08,
      "logits/chosen": 0.9395872950553894,
      "logits/rejected": 0.883934497833252,
      "logps/chosen": -172.1771240234375,
      "logps/ref_chosen": -39.918418884277344,
      "logps/ref_rejected": -67.36483001708984,
      "logps/rejected": -293.5935974121094,
      "loss": 0.9021,
      "margin_dpo/margin_mean": 93.97006225585938,
      "margin_dpo/margin_std": 82.55453491210938,
      "step": 579
    },
    {
      "epoch": 0.8767951625094482,
      "grad_norm": 16.5377197265625,
      "learning_rate": 2.3144448823151392e-08,
      "logits/chosen": 0.8914264440536499,
      "logits/rejected": 0.8536027669906616,
      "logps/chosen": -203.43321228027344,
      "logps/ref_chosen": -57.42926025390625,
      "logps/ref_rejected": -71.1881332397461,
      "logps/rejected": -251.05111694335938,
      "loss": 1.0703,
      "margin_dpo/margin_mean": 33.859039306640625,
      "margin_dpo/margin_std": 79.80448150634766,
      "step": 580
    },
    {
      "epoch": 0.8783068783068783,
      "grad_norm": 16.735671997070312,
      "learning_rate": 2.259200116137039e-08,
      "logits/chosen": 0.9245070219039917,
      "logits/rejected": 0.8710756301879883,
      "logps/chosen": -207.47396850585938,
      "logps/ref_chosen": -64.48869323730469,
      "logps/ref_rejected": -80.07521057128906,
      "logps/rejected": -284.458251953125,
      "loss": 1.0719,
      "margin_dpo/margin_mean": 61.39775085449219,
      "margin_dpo/margin_std": 92.19239807128906,
      "step": 581
    },
    {
      "epoch": 0.8798185941043084,
      "grad_norm": 21.038320541381836,
      "learning_rate": 2.204591459016525e-08,
      "logits/chosen": 0.9050915241241455,
      "logits/rejected": 0.7998449802398682,
      "logps/chosen": -180.35040283203125,
      "logps/ref_chosen": -52.15564727783203,
      "logps/ref_rejected": -81.67626190185547,
      "logps/rejected": -324.54791259765625,
      "loss": 1.106,
      "margin_dpo/margin_mean": 114.67691802978516,
      "margin_dpo/margin_std": 95.79690551757812,
      "step": 582
    },
    {
      "epoch": 0.8813303099017384,
      "grad_norm": 21.99834442138672,
      "learning_rate": 2.1506204384751064e-08,
      "logits/chosen": 1.0146265029907227,
      "logits/rejected": 0.9865950345993042,
      "logps/chosen": -198.27951049804688,
      "logps/ref_chosen": -53.600677490234375,
      "logps/ref_rejected": -66.55783081054688,
      "logps/rejected": -231.4248046875,
      "loss": 1.05,
      "margin_dpo/margin_mean": 20.188133239746094,
      "margin_dpo/margin_std": 88.58358001708984,
      "step": 583
    },
    {
      "epoch": 0.8828420256991686,
      "grad_norm": 18.144161224365234,
      "learning_rate": 2.09728856419826e-08,
      "logits/chosen": 0.9950936436653137,
      "logits/rejected": 0.8881626129150391,
      "logps/chosen": -145.82467651367188,
      "logps/ref_chosen": -39.92349624633789,
      "logps/ref_rejected": -85.95851135253906,
      "logps/rejected": -260.9383544921875,
      "loss": 1.0972,
      "margin_dpo/margin_mean": 69.07866668701172,
      "margin_dpo/margin_std": 94.79948425292969,
      "step": 584
    },
    {
      "epoch": 0.8843537414965986,
      "grad_norm": 15.75501537322998,
      "learning_rate": 2.044597327993153e-08,
      "logits/chosen": 0.9527912139892578,
      "logits/rejected": 0.7973681092262268,
      "logps/chosen": -196.22564697265625,
      "logps/ref_chosen": -54.76570129394531,
      "logps/ref_rejected": -115.46517944335938,
      "logps/rejected": -307.131591796875,
      "loss": 1.1429,
      "margin_dpo/margin_mean": 50.206451416015625,
      "margin_dpo/margin_std": 92.417724609375,
      "step": 585
    },
    {
      "epoch": 0.8858654572940288,
      "grad_norm": 18.22370147705078,
      "learning_rate": 1.9925482037469187e-08,
      "logits/chosen": 0.9899888038635254,
      "logits/rejected": 1.0135765075683594,
      "logps/chosen": -180.9400634765625,
      "logps/ref_chosen": -62.72941207885742,
      "logps/ref_rejected": -46.449256896972656,
      "logps/rejected": -233.3268585205078,
      "loss": 0.9577,
      "margin_dpo/margin_mean": 68.66695404052734,
      "margin_dpo/margin_std": 74.32232666015625,
      "step": 586
    },
    {
      "epoch": 0.8873771730914588,
      "grad_norm": 21.38698387145996,
      "learning_rate": 1.9411426473854687e-08,
      "logits/chosen": 1.0005683898925781,
      "logits/rejected": 0.954791784286499,
      "logps/chosen": -218.3308868408203,
      "logps/ref_chosen": -58.70615768432617,
      "logps/ref_rejected": -70.26844024658203,
      "logps/rejected": -261.1861267089844,
      "loss": 1.0981,
      "margin_dpo/margin_mean": 31.292953491210938,
      "margin_dpo/margin_std": 83.6708984375,
      "step": 587
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 19.388957977294922,
      "learning_rate": 1.890382096832699e-08,
      "logits/chosen": 0.8030673265457153,
      "logits/rejected": 0.7511934041976929,
      "logps/chosen": -196.69418334960938,
      "logps/ref_chosen": -66.02383422851562,
      "logps/ref_rejected": -92.533203125,
      "logps/rejected": -316.143798828125,
      "loss": 1.0657,
      "margin_dpo/margin_mean": 92.94023895263672,
      "margin_dpo/margin_std": 82.5245361328125,
      "step": 588
    },
    {
      "epoch": 0.890400604686319,
      "grad_norm": 16.032787322998047,
      "learning_rate": 1.840267971970344e-08,
      "logits/chosen": 0.8890354633331299,
      "logits/rejected": 0.8605284690856934,
      "logps/chosen": -199.17193603515625,
      "logps/ref_chosen": -52.7588005065918,
      "logps/ref_rejected": -68.78264617919922,
      "logps/rejected": -266.2153015136719,
      "loss": 0.9575,
      "margin_dpo/margin_mean": 51.01952362060547,
      "margin_dpo/margin_std": 97.24200439453125,
      "step": 589
    },
    {
      "epoch": 0.891912320483749,
      "grad_norm": 16.486291885375977,
      "learning_rate": 1.7908016745981856e-08,
      "logits/chosen": 0.9132862091064453,
      "logits/rejected": 0.8946816921234131,
      "logps/chosen": -198.87548828125,
      "logps/ref_chosen": -53.84131622314453,
      "logps/ref_rejected": -64.9970703125,
      "logps/rejected": -240.9401092529297,
      "loss": 0.9877,
      "margin_dpo/margin_mean": 30.908885955810547,
      "margin_dpo/margin_std": 70.15216827392578,
      "step": 590
    },
    {
      "epoch": 0.8934240362811792,
      "grad_norm": 18.833675384521484,
      "learning_rate": 1.7419845883949098e-08,
      "logits/chosen": 0.9794512987136841,
      "logits/rejected": 0.9539260268211365,
      "logps/chosen": -170.65090942382812,
      "logps/ref_chosen": -50.85618591308594,
      "logps/ref_rejected": -61.08381271362305,
      "logps/rejected": -261.77154541015625,
      "loss": 0.9912,
      "margin_dpo/margin_mean": 80.89300537109375,
      "margin_dpo/margin_std": 80.72918701171875,
      "step": 591
    },
    {
      "epoch": 0.8949357520786092,
      "grad_norm": 18.000661849975586,
      "learning_rate": 1.6938180788793556e-08,
      "logits/chosen": 1.0009410381317139,
      "logits/rejected": 0.9112914800643921,
      "logps/chosen": -176.25234985351562,
      "logps/ref_chosen": -51.354652404785156,
      "logps/ref_rejected": -67.10594177246094,
      "logps/rejected": -271.6142578125,
      "loss": 0.985,
      "margin_dpo/margin_mean": 79.610595703125,
      "margin_dpo/margin_std": 81.41339111328125,
      "step": 592
    },
    {
      "epoch": 0.8964474678760394,
      "grad_norm": 18.60577392578125,
      "learning_rate": 1.6463034933723336e-08,
      "logits/chosen": 0.8909306526184082,
      "logits/rejected": 0.8740547299385071,
      "logps/chosen": -187.12136840820312,
      "logps/ref_chosen": -61.935089111328125,
      "logps/ref_rejected": -82.4512939453125,
      "logps/rejected": -286.46282958984375,
      "loss": 1.0216,
      "margin_dpo/margin_mean": 78.82524871826172,
      "margin_dpo/margin_std": 103.09696960449219,
      "step": 593
    },
    {
      "epoch": 0.8979591836734694,
      "grad_norm": 18.425268173217773,
      "learning_rate": 1.5994421609589385e-08,
      "logits/chosen": 0.8772367835044861,
      "logits/rejected": 0.8558773994445801,
      "logps/chosen": -211.46441650390625,
      "logps/ref_chosen": -63.053680419921875,
      "logps/ref_rejected": -68.9009017944336,
      "logps/rejected": -248.54051208496094,
      "loss": 1.0588,
      "margin_dpo/margin_mean": 31.228851318359375,
      "margin_dpo/margin_std": 52.22967529296875,
      "step": 594
    },
    {
      "epoch": 0.8994708994708994,
      "grad_norm": 14.107276916503906,
      "learning_rate": 1.553235392451377e-08,
      "logits/chosen": 1.0201103687286377,
      "logits/rejected": 0.9065714478492737,
      "logps/chosen": -164.76438903808594,
      "logps/ref_chosen": -36.28746032714844,
      "logps/ref_rejected": -76.62598419189453,
      "logps/rejected": -292.2261962890625,
      "loss": 0.9302,
      "margin_dpo/margin_mean": 87.12327575683594,
      "margin_dpo/margin_std": 96.01534271240234,
      "step": 595
    },
    {
      "epoch": 0.9009826152683296,
      "grad_norm": 16.283180236816406,
      "learning_rate": 1.507684480352292e-08,
      "logits/chosen": 0.8756155371665955,
      "logits/rejected": 0.8407232761383057,
      "logps/chosen": -196.04852294921875,
      "logps/ref_chosen": -58.18864440917969,
      "logps/ref_rejected": -69.7575454711914,
      "logps/rejected": -244.40557861328125,
      "loss": 1.2206,
      "margin_dpo/margin_mean": 36.78814697265625,
      "margin_dpo/margin_std": 108.71817016601562,
      "step": 596
    },
    {
      "epoch": 0.9024943310657596,
      "grad_norm": 14.265636444091797,
      "learning_rate": 1.4627906988186111e-08,
      "logits/chosen": 0.9384099245071411,
      "logits/rejected": 0.9882324934005737,
      "logps/chosen": -195.47030639648438,
      "logps/ref_chosen": -62.06855773925781,
      "logps/ref_rejected": -62.94855880737305,
      "logps/rejected": -221.61302185058594,
      "loss": 1.0398,
      "margin_dpo/margin_mean": 25.262706756591797,
      "margin_dpo/margin_std": 77.90994262695312,
      "step": 597
    },
    {
      "epoch": 0.9040060468631897,
      "grad_norm": 22.982269287109375,
      "learning_rate": 1.4185553036259095e-08,
      "logits/chosen": 0.8697057962417603,
      "logits/rejected": 0.8449388742446899,
      "logps/chosen": -206.47000122070312,
      "logps/ref_chosen": -48.21617889404297,
      "logps/ref_rejected": -68.27192687988281,
      "logps/rejected": -257.6476745605469,
      "loss": 1.2139,
      "margin_dpo/margin_mean": 31.121919631958008,
      "margin_dpo/margin_std": 93.97370147705078,
      "step": 598
    },
    {
      "epoch": 0.9055177626606198,
      "grad_norm": 17.03192138671875,
      "learning_rate": 1.3749795321332885e-08,
      "logits/chosen": 0.8180956244468689,
      "logits/rejected": 0.8315231800079346,
      "logps/chosen": -224.57041931152344,
      "logps/ref_chosen": -61.865814208984375,
      "logps/ref_rejected": -62.11842346191406,
      "logps/rejected": -258.51904296875,
      "loss": 1.146,
      "margin_dpo/margin_mean": 33.696022033691406,
      "margin_dpo/margin_std": 111.64860534667969,
      "step": 599
    },
    {
      "epoch": 0.9070294784580499,
      "grad_norm": 17.681379318237305,
      "learning_rate": 1.3320646032487393e-08,
      "logits/chosen": 0.9060215353965759,
      "logits/rejected": 0.9525216817855835,
      "logps/chosen": -173.75914001464844,
      "logps/ref_chosen": -55.92002868652344,
      "logps/ref_rejected": -53.79228210449219,
      "logps/rejected": -239.77134704589844,
      "loss": 1.1391,
      "margin_dpo/margin_mean": 68.13994598388672,
      "margin_dpo/margin_std": 82.62960815429688,
      "step": 600
    },
    {
      "epoch": 0.9070294784580499,
      "eval_logits/chosen": 0.8960238695144653,
      "eval_logits/rejected": 0.863480806350708,
      "eval_logps/chosen": -211.5150146484375,
      "eval_logps/ref_chosen": -75.30646514892578,
      "eval_logps/ref_rejected": -77.75511932373047,
      "eval_logps/rejected": -274.14215087890625,
      "eval_loss": 0.5347773432731628,
      "eval_margin_dpo/margin_mean": 60.178462982177734,
      "eval_margin_dpo/margin_std": 94.62100982666016,
      "eval_runtime": 37.4693,
      "eval_samples_per_second": 61.464,
      "eval_steps_per_second": 1.922,
      "step": 600
    },
    {
      "epoch": 0.90854119425548,
      "grad_norm": 14.842012405395508,
      "learning_rate": 1.2898117173950868e-08,
      "logits/chosen": 0.8835675716400146,
      "logits/rejected": 0.8951883912086487,
      "logps/chosen": -201.0157470703125,
      "logps/ref_chosen": -70.18791198730469,
      "logps/ref_rejected": -71.40547180175781,
      "logps/rejected": -291.8153381347656,
      "loss": 0.9772,
      "margin_dpo/margin_mean": 89.58202362060547,
      "margin_dpo/margin_std": 125.6627197265625,
      "step": 601
    },
    {
      "epoch": 0.91005291005291,
      "grad_norm": 18.32841682434082,
      "learning_rate": 1.2482220564763667e-08,
      "logits/chosen": 0.9220967292785645,
      "logits/rejected": 0.8645837306976318,
      "logps/chosen": -187.869873046875,
      "logps/ref_chosen": -70.41340637207031,
      "logps/ref_rejected": -97.12376403808594,
      "logps/rejected": -304.54473876953125,
      "loss": 0.989,
      "margin_dpo/margin_mean": 89.96450805664062,
      "margin_dpo/margin_std": 100.75538635253906,
      "step": 602
    },
    {
      "epoch": 0.9115646258503401,
      "grad_norm": 19.300460815429688,
      "learning_rate": 1.2072967838448051e-08,
      "logits/chosen": 0.9347689747810364,
      "logits/rejected": 0.8060243129730225,
      "logps/chosen": -211.57089233398438,
      "logps/ref_chosen": -62.89923858642578,
      "logps/ref_rejected": -108.35671997070312,
      "logps/rejected": -351.4485778808594,
      "loss": 0.9984,
      "margin_dpo/margin_mean": 94.42021179199219,
      "margin_dpo/margin_std": 152.901123046875,
      "step": 603
    },
    {
      "epoch": 0.9130763416477702,
      "grad_norm": 15.397801399230957,
      "learning_rate": 1.1670370442682459e-08,
      "logits/chosen": 0.9415032267570496,
      "logits/rejected": 0.9144819974899292,
      "logps/chosen": -188.33102416992188,
      "logps/ref_chosen": -62.4847412109375,
      "logps/ref_rejected": -62.371429443359375,
      "logps/rejected": -235.138427734375,
      "loss": 1.1185,
      "margin_dpo/margin_mean": 46.92070770263672,
      "margin_dpo/margin_std": 84.41587829589844,
      "step": 604
    },
    {
      "epoch": 0.9145880574452003,
      "grad_norm": 18.700698852539062,
      "learning_rate": 1.1274439638981532e-08,
      "logits/chosen": 1.0574252605438232,
      "logits/rejected": 0.938467264175415,
      "logps/chosen": -192.36363220214844,
      "logps/ref_chosen": -48.73389434814453,
      "logps/ref_rejected": -83.2470703125,
      "logps/rejected": -292.3554992675781,
      "loss": 1.0445,
      "margin_dpo/margin_mean": 65.47867584228516,
      "margin_dpo/margin_std": 119.60535430908203,
      "step": 605
    },
    {
      "epoch": 0.9160997732426304,
      "grad_norm": 17.30083465576172,
      "learning_rate": 1.0885186502381016e-08,
      "logits/chosen": 0.9336838722229004,
      "logits/rejected": 0.846339225769043,
      "logps/chosen": -190.03634643554688,
      "logps/ref_chosen": -53.08481979370117,
      "logps/ref_rejected": -80.11920166015625,
      "logps/rejected": -295.1481018066406,
      "loss": 0.989,
      "margin_dpo/margin_mean": 78.07736206054688,
      "margin_dpo/margin_std": 93.74771118164062,
      "step": 606
    },
    {
      "epoch": 0.9176114890400605,
      "grad_norm": 16.944414138793945,
      "learning_rate": 1.0502621921127774e-08,
      "logits/chosen": 0.7050179243087769,
      "logits/rejected": 0.8792567253112793,
      "logps/chosen": -264.5372009277344,
      "logps/ref_chosen": -97.28004455566406,
      "logps/ref_rejected": -86.22888946533203,
      "logps/rejected": -283.00103759765625,
      "loss": 1.0318,
      "margin_dpo/margin_mean": 29.514984130859375,
      "margin_dpo/margin_std": 90.30196380615234,
      "step": 607
    },
    {
      "epoch": 0.9191232048374905,
      "grad_norm": 18.15805435180664,
      "learning_rate": 1.0126756596375685e-08,
      "logits/chosen": 0.8514116406440735,
      "logits/rejected": 0.8852680921554565,
      "logps/chosen": -181.86837768554688,
      "logps/ref_chosen": -66.10746765136719,
      "logps/ref_rejected": -59.404441833496094,
      "logps/rejected": -243.4190673828125,
      "loss": 1.0596,
      "margin_dpo/margin_mean": 68.25370788574219,
      "margin_dpo/margin_std": 68.52870178222656,
      "step": 608
    },
    {
      "epoch": 0.9206349206349206,
      "grad_norm": 14.3372163772583,
      "learning_rate": 9.757601041885694e-09,
      "logits/chosen": 1.0790916681289673,
      "logits/rejected": 0.9650440216064453,
      "logps/chosen": -173.4683380126953,
      "logps/ref_chosen": -49.93000030517578,
      "logps/ref_rejected": -98.36846923828125,
      "logps/rejected": -298.0489501953125,
      "loss": 0.9481,
      "margin_dpo/margin_mean": 76.14215087890625,
      "margin_dpo/margin_std": 91.03385925292969,
      "step": 609
    },
    {
      "epoch": 0.9221466364323507,
      "grad_norm": 31.100370407104492,
      "learning_rate": 9.395165583732379e-09,
      "logits/chosen": 0.9039211869239807,
      "logits/rejected": 0.9379135370254517,
      "logps/chosen": -220.49267578125,
      "logps/ref_chosen": -71.89997863769531,
      "logps/ref_rejected": -89.20841979980469,
      "logps/rejected": -285.78515625,
      "loss": 1.1377,
      "margin_dpo/margin_mean": 47.98402404785156,
      "margin_dpo/margin_std": 88.36325073242188,
      "step": 610
    },
    {
      "epoch": 0.9236583522297808,
      "grad_norm": 18.051315307617188,
      "learning_rate": 9.03946036001449e-09,
      "logits/chosen": 0.9056074619293213,
      "logits/rejected": 0.9064819812774658,
      "logps/chosen": -214.71875,
      "logps/ref_chosen": -67.77259826660156,
      "logps/ref_rejected": -66.90919494628906,
      "logps/rejected": -263.32421875,
      "loss": 1.1144,
      "margin_dpo/margin_mean": 49.4688606262207,
      "margin_dpo/margin_std": 66.27508544921875,
      "step": 611
    },
    {
      "epoch": 0.9251700680272109,
      "grad_norm": 15.253009796142578,
      "learning_rate": 8.690495320571839e-09,
      "logits/chosen": 0.6356790065765381,
      "logits/rejected": 0.6141253113746643,
      "logps/chosen": -254.52713012695312,
      "logps/ref_chosen": -98.77779388427734,
      "logps/ref_rejected": -129.40658569335938,
      "logps/rejected": -327.4943542480469,
      "loss": 0.9714,
      "margin_dpo/margin_mean": 42.33841323852539,
      "margin_dpo/margin_std": 53.188514709472656,
      "step": 612
    },
    {
      "epoch": 0.926681783824641,
      "grad_norm": 22.37784767150879,
      "learning_rate": 8.348280226706722e-09,
      "logits/chosen": 0.9218869209289551,
      "logits/rejected": 0.9058539867401123,
      "logps/chosen": -176.6582794189453,
      "logps/ref_chosen": -39.422340393066406,
      "logps/ref_rejected": -48.479248046875,
      "logps/rejected": -244.13925170898438,
      "loss": 0.9836,
      "margin_dpo/margin_mean": 58.42406463623047,
      "margin_dpo/margin_std": 89.93438720703125,
      "step": 613
    },
    {
      "epoch": 0.9281934996220711,
      "grad_norm": 24.711509704589844,
      "learning_rate": 8.012824650910937e-09,
      "logits/chosen": 1.0547163486480713,
      "logits/rejected": 0.9173270463943481,
      "logps/chosen": -167.89730834960938,
      "logps/ref_chosen": -52.9261474609375,
      "logps/ref_rejected": -68.24092102050781,
      "logps/rejected": -296.45574951171875,
      "loss": 1.0583,
      "margin_dpo/margin_mean": 113.24365234375,
      "margin_dpo/margin_std": 79.29691314697266,
      "step": 614
    },
    {
      "epoch": 0.9297052154195011,
      "grad_norm": 25.62435531616211,
      "learning_rate": 7.684137976598088e-09,
      "logits/chosen": 0.9779610633850098,
      "logits/rejected": 0.9523606896400452,
      "logps/chosen": -217.04598999023438,
      "logps/ref_chosen": -70.00674438476562,
      "logps/ref_rejected": -101.93522644042969,
      "logps/rejected": -298.4114074707031,
      "loss": 1.0525,
      "margin_dpo/margin_mean": 49.43693542480469,
      "margin_dpo/margin_std": 92.1063003540039,
      "step": 615
    },
    {
      "epoch": 0.9312169312169312,
      "grad_norm": 16.771865844726562,
      "learning_rate": 7.36222939784098e-09,
      "logits/chosen": 0.9289509057998657,
      "logits/rejected": 0.8705403804779053,
      "logps/chosen": -209.6519012451172,
      "logps/ref_chosen": -66.25517272949219,
      "logps/ref_rejected": -92.31936645507812,
      "logps/rejected": -286.00390625,
      "loss": 1.0395,
      "margin_dpo/margin_mean": 50.28779220581055,
      "margin_dpo/margin_std": 81.98731994628906,
      "step": 616
    },
    {
      "epoch": 0.9327286470143613,
      "grad_norm": 19.335580825805664,
      "learning_rate": 7.047107919114586e-09,
      "logits/chosen": 0.9605733752250671,
      "logits/rejected": 0.9251487255096436,
      "logps/chosen": -212.73922729492188,
      "logps/ref_chosen": -70.3993911743164,
      "logps/ref_rejected": -90.81809997558594,
      "logps/rejected": -292.81744384765625,
      "loss": 1.0436,
      "margin_dpo/margin_mean": 59.65951919555664,
      "margin_dpo/margin_std": 60.16530227661133,
      "step": 617
    },
    {
      "epoch": 0.9342403628117913,
      "grad_norm": 19.181501388549805,
      "learning_rate": 6.738782355044048e-09,
      "logits/chosen": 0.9526699781417847,
      "logits/rejected": 0.8337723016738892,
      "logps/chosen": -169.63429260253906,
      "logps/ref_chosen": -62.809661865234375,
      "logps/ref_rejected": -109.43612670898438,
      "logps/rejected": -310.1464538574219,
      "loss": 0.9829,
      "margin_dpo/margin_mean": 93.88569641113281,
      "margin_dpo/margin_std": 91.3293228149414,
      "step": 618
    },
    {
      "epoch": 0.9357520786092215,
      "grad_norm": 16.765586853027344,
      "learning_rate": 6.437261330158206e-09,
      "logits/chosen": 0.9776492118835449,
      "logits/rejected": 0.9457724094390869,
      "logps/chosen": -158.1160888671875,
      "logps/ref_chosen": -47.85075378417969,
      "logps/ref_rejected": -72.59744262695312,
      "logps/rejected": -258.400634765625,
      "loss": 0.9783,
      "margin_dpo/margin_mean": 75.53782653808594,
      "margin_dpo/margin_std": 94.00806427001953,
      "step": 619
    },
    {
      "epoch": 0.9372637944066515,
      "grad_norm": 19.020383834838867,
      "learning_rate": 6.142553278648238e-09,
      "logits/chosen": 0.8768226504325867,
      "logits/rejected": 0.8743495345115662,
      "logps/chosen": -224.13729858398438,
      "logps/ref_chosen": -79.60867309570312,
      "logps/ref_rejected": -86.96884155273438,
      "logps/rejected": -288.10784912109375,
      "loss": 1.1095,
      "margin_dpo/margin_mean": 56.610382080078125,
      "margin_dpo/margin_std": 97.45252990722656,
      "step": 620
    },
    {
      "epoch": 0.9387755102040817,
      "grad_norm": 17.923816680908203,
      "learning_rate": 5.854666444131934e-09,
      "logits/chosen": 0.7942125797271729,
      "logits/rejected": 0.8802160620689392,
      "logps/chosen": -215.23654174804688,
      "logps/ref_chosen": -68.8500747680664,
      "logps/ref_rejected": -53.94578552246094,
      "logps/rejected": -231.7372589111328,
      "loss": 1.0581,
      "margin_dpo/margin_mean": 31.405014038085938,
      "margin_dpo/margin_std": 76.86921691894531,
      "step": 621
    },
    {
      "epoch": 0.9402872260015117,
      "grad_norm": 20.861454010009766,
      "learning_rate": 5.573608879422875e-09,
      "logits/chosen": 0.8228497505187988,
      "logits/rejected": 0.8221659660339355,
      "logps/chosen": -211.20352172851562,
      "logps/ref_chosen": -49.52813720703125,
      "logps/ref_rejected": -56.0748291015625,
      "logps/rejected": -251.8733673095703,
      "loss": 1.0707,
      "margin_dpo/margin_mean": 34.12314987182617,
      "margin_dpo/margin_std": 85.17213439941406,
      "step": 622
    },
    {
      "epoch": 0.9417989417989417,
      "grad_norm": 18.039180755615234,
      "learning_rate": 5.299388446305342e-09,
      "logits/chosen": 0.9442222118377686,
      "logits/rejected": 0.8388766050338745,
      "logps/chosen": -213.2373809814453,
      "logps/ref_chosen": -51.98558807373047,
      "logps/ref_rejected": -89.86474609375,
      "logps/rejected": -364.82171630859375,
      "loss": 0.9715,
      "margin_dpo/margin_mean": 113.70519256591797,
      "margin_dpo/margin_std": 154.77015686035156,
      "step": 623
    },
    {
      "epoch": 0.9433106575963719,
      "grad_norm": 16.79059410095215,
      "learning_rate": 5.03201281531429e-09,
      "logits/chosen": 0.9908918142318726,
      "logits/rejected": 0.9922984838485718,
      "logps/chosen": -189.19586181640625,
      "logps/ref_chosen": -61.39537811279297,
      "logps/ref_rejected": -64.496826171875,
      "logps/rejected": -265.69671630859375,
      "loss": 0.9914,
      "margin_dpo/margin_mean": 73.39938354492188,
      "margin_dpo/margin_std": 75.53822326660156,
      "step": 624
    },
    {
      "epoch": 0.9448223733938019,
      "grad_norm": 16.281400680541992,
      "learning_rate": 4.7714894655209174e-09,
      "logits/chosen": 1.096935510635376,
      "logits/rejected": 0.9919509887695312,
      "logps/chosen": -173.40103149414062,
      "logps/ref_chosen": -49.35209274291992,
      "logps/ref_rejected": -75.47923278808594,
      "logps/rejected": -266.97930908203125,
      "loss": 1.1236,
      "margin_dpo/margin_mean": 67.45115661621094,
      "margin_dpo/margin_std": 78.00559997558594,
      "step": 625
    },
    {
      "epoch": 0.9463340891912321,
      "grad_norm": 16.243318557739258,
      "learning_rate": 4.517825684323323e-09,
      "logits/chosen": 1.0832741260528564,
      "logits/rejected": 1.023711085319519,
      "logps/chosen": -122.06949615478516,
      "logps/ref_chosen": -32.55897521972656,
      "logps/ref_rejected": -71.60407257080078,
      "logps/rejected": -270.57061767578125,
      "loss": 1.0152,
      "margin_dpo/margin_mean": 109.45604705810547,
      "margin_dpo/margin_std": 145.14871215820312,
      "step": 626
    },
    {
      "epoch": 0.9478458049886621,
      "grad_norm": 18.35875701904297,
      "learning_rate": 4.271028567242818e-09,
      "logits/chosen": 0.8411812782287598,
      "logits/rejected": 0.708235502243042,
      "logps/chosen": -198.689453125,
      "logps/ref_chosen": -65.68330383300781,
      "logps/ref_rejected": -124.0399169921875,
      "logps/rejected": -374.5428161621094,
      "loss": 0.9747,
      "margin_dpo/margin_mean": 117.49674224853516,
      "margin_dpo/margin_std": 99.51753234863281,
      "step": 627
    },
    {
      "epoch": 0.9493575207860923,
      "grad_norm": 17.412473678588867,
      "learning_rate": 4.0311050177251895e-09,
      "logits/chosen": 0.857243001461029,
      "logits/rejected": 0.9196332693099976,
      "logps/chosen": -188.25320434570312,
      "logps/ref_chosen": -70.50054168701172,
      "logps/ref_rejected": -69.05288696289062,
      "logps/rejected": -286.83795166015625,
      "loss": 1.0335,
      "margin_dpo/margin_mean": 100.03239440917969,
      "margin_dpo/margin_std": 102.7291259765625,
      "step": 628
    },
    {
      "epoch": 0.9508692365835223,
      "grad_norm": 17.979310989379883,
      "learning_rate": 3.798061746947995e-09,
      "logits/chosen": 0.9377484321594238,
      "logits/rejected": 1.008423089981079,
      "logps/chosen": -184.6234893798828,
      "logps/ref_chosen": -68.93040466308594,
      "logps/ref_rejected": -49.09862518310547,
      "logps/rejected": -198.7704620361328,
      "loss": 1.0083,
      "margin_dpo/margin_mean": 33.978755950927734,
      "margin_dpo/margin_std": 70.85810089111328,
      "step": 629
    },
    {
      "epoch": 0.9523809523809523,
      "grad_norm": 17.69316864013672,
      "learning_rate": 3.5719052736323806e-09,
      "logits/chosen": 1.0553034543991089,
      "logits/rejected": 0.9892318844795227,
      "logps/chosen": -134.7643585205078,
      "logps/ref_chosen": -36.853294372558594,
      "logps/ref_rejected": -71.96033477783203,
      "logps/rejected": -252.25209045410156,
      "loss": 0.9744,
      "margin_dpo/margin_mean": 82.38069152832031,
      "margin_dpo/margin_std": 89.0606460571289,
      "step": 630
    },
    {
      "epoch": 0.9538926681783825,
      "grad_norm": 18.01328468322754,
      "learning_rate": 3.352641923861144e-09,
      "logits/chosen": 0.9821938872337341,
      "logits/rejected": 0.938322901725769,
      "logps/chosen": -195.1949462890625,
      "logps/ref_chosen": -64.66175079345703,
      "logps/ref_rejected": -89.40802001953125,
      "logps/rejected": -301.36553955078125,
      "loss": 0.9122,
      "margin_dpo/margin_mean": 81.42431640625,
      "margin_dpo/margin_std": 62.148040771484375,
      "step": 631
    },
    {
      "epoch": 0.9554043839758125,
      "grad_norm": 18.51304054260254,
      "learning_rate": 3.140277830901428e-09,
      "logits/chosen": 0.9517176151275635,
      "logits/rejected": 0.9564006328582764,
      "logps/chosen": -203.13833618164062,
      "logps/ref_chosen": -74.05264282226562,
      "logps/ref_rejected": -66.81067657470703,
      "logps/rejected": -266.5632629394531,
      "loss": 0.93,
      "margin_dpo/margin_mean": 70.66688537597656,
      "margin_dpo/margin_std": 84.40769958496094,
      "step": 632
    },
    {
      "epoch": 0.9569160997732427,
      "grad_norm": 17.859493255615234,
      "learning_rate": 2.9348189350335007e-09,
      "logits/chosen": 0.9254288077354431,
      "logits/rejected": 0.87281334400177,
      "logps/chosen": -146.23092651367188,
      "logps/ref_chosen": -45.47814178466797,
      "logps/ref_rejected": -66.22723388671875,
      "logps/rejected": -234.7810516357422,
      "loss": 1.0317,
      "margin_dpo/margin_mean": 67.80101776123047,
      "margin_dpo/margin_std": 67.22663879394531,
      "step": 633
    },
    {
      "epoch": 0.9584278155706727,
      "grad_norm": 23.980148315429688,
      "learning_rate": 2.736270983384276e-09,
      "logits/chosen": 0.8979266285896301,
      "logits/rejected": 0.8862963914871216,
      "logps/chosen": -212.658203125,
      "logps/ref_chosen": -57.611724853515625,
      "logps/ref_rejected": -54.337623596191406,
      "logps/rejected": -215.5278778076172,
      "loss": 1.3386,
      "margin_dpo/margin_mean": 6.1437668800354,
      "margin_dpo/margin_std": 45.94617462158203,
      "step": 634
    },
    {
      "epoch": 0.9599395313681028,
      "grad_norm": 18.18239974975586,
      "learning_rate": 2.5446395297668287e-09,
      "logits/chosen": 0.945531964302063,
      "logits/rejected": 0.9129035472869873,
      "logps/chosen": -211.33154296875,
      "logps/ref_chosen": -49.714500427246094,
      "logps/ref_rejected": -67.80001831054688,
      "logps/rejected": -269.0682067871094,
      "loss": 1.1286,
      "margin_dpo/margin_mean": 39.65115737915039,
      "margin_dpo/margin_std": 108.38533020019531,
      "step": 635
    },
    {
      "epoch": 0.9614512471655329,
      "grad_norm": 15.897844314575195,
      "learning_rate": 2.359929934524829e-09,
      "logits/chosen": 0.9161902666091919,
      "logits/rejected": 0.8111059069633484,
      "logps/chosen": -202.0909423828125,
      "logps/ref_chosen": -50.605045318603516,
      "logps/ref_rejected": -77.53902435302734,
      "logps/rejected": -293.85711669921875,
      "loss": 0.9728,
      "margin_dpo/margin_mean": 64.83220672607422,
      "margin_dpo/margin_std": 55.20343017578125,
      "step": 636
    },
    {
      "epoch": 0.9629629629629629,
      "grad_norm": 18.38071060180664,
      "learning_rate": 2.1821473643827137e-09,
      "logits/chosen": 0.9871467351913452,
      "logits/rejected": 0.8986793160438538,
      "logps/chosen": -245.16378784179688,
      "logps/ref_chosen": -74.90892028808594,
      "logps/ref_rejected": -97.07221984863281,
      "logps/rejected": -350.8941650390625,
      "loss": 1.0848,
      "margin_dpo/margin_mean": 83.56706237792969,
      "margin_dpo/margin_std": 141.6215057373047,
      "step": 637
    },
    {
      "epoch": 0.9644746787603931,
      "grad_norm": 16.282268524169922,
      "learning_rate": 2.0112967923011646e-09,
      "logits/chosen": 0.8492100238800049,
      "logits/rejected": 0.823014497756958,
      "logps/chosen": -249.89724731445312,
      "logps/ref_chosen": -67.56559753417969,
      "logps/ref_rejected": -83.56005859375,
      "logps/rejected": -284.5699157714844,
      "loss": 0.998,
      "margin_dpo/margin_mean": 18.678207397460938,
      "margin_dpo/margin_std": 95.15095520019531,
      "step": 638
    },
    {
      "epoch": 0.9659863945578231,
      "grad_norm": 18.59486198425293,
      "learning_rate": 1.847382997337943e-09,
      "logits/chosen": 0.9617091417312622,
      "logits/rejected": 0.8375701308250427,
      "logps/chosen": -166.54019165039062,
      "logps/ref_chosen": -40.15863037109375,
      "logps/ref_rejected": -65.15992736816406,
      "logps/rejected": -268.45574951171875,
      "loss": 0.9745,
      "margin_dpo/margin_mean": 76.91429138183594,
      "margin_dpo/margin_std": 78.58573913574219,
      "step": 639
    },
    {
      "epoch": 0.9674981103552532,
      "grad_norm": 19.1875,
      "learning_rate": 1.690410564514244e-09,
      "logits/chosen": 0.8923076391220093,
      "logits/rejected": 0.7876547574996948,
      "logps/chosen": -186.8981475830078,
      "logps/ref_chosen": -56.694557189941406,
      "logps/ref_rejected": -100.86506652832031,
      "logps/rejected": -306.697265625,
      "loss": 1.0631,
      "margin_dpo/margin_mean": 75.62860870361328,
      "margin_dpo/margin_std": 84.25181579589844,
      "step": 640
    },
    {
      "epoch": 0.9690098261526833,
      "grad_norm": 16.5269775390625,
      "learning_rate": 1.5403838846864692e-09,
      "logits/chosen": 0.9277850389480591,
      "logits/rejected": 1.0083123445510864,
      "logps/chosen": -200.206298828125,
      "logps/ref_chosen": -67.92603302001953,
      "logps/ref_rejected": -50.31890106201172,
      "logps/rejected": -228.96484375,
      "loss": 1.0154,
      "margin_dpo/margin_mean": 46.365638732910156,
      "margin_dpo/margin_std": 66.46011352539062,
      "step": 641
    },
    {
      "epoch": 0.9705215419501134,
      "grad_norm": 26.048898696899414,
      "learning_rate": 1.3973071544233218e-09,
      "logits/chosen": 0.9023500084877014,
      "logits/rejected": 0.9192708134651184,
      "logps/chosen": -210.30873107910156,
      "logps/ref_chosen": -66.0007553100586,
      "logps/ref_rejected": -71.08174133300781,
      "logps/rejected": -260.4029235839844,
      "loss": 1.1862,
      "margin_dpo/margin_mean": 45.013214111328125,
      "margin_dpo/margin_std": 130.67271423339844,
      "step": 642
    },
    {
      "epoch": 0.9720332577475435,
      "grad_norm": 18.16506004333496,
      "learning_rate": 1.261184375888541e-09,
      "logits/chosen": 0.8603549599647522,
      "logits/rejected": 0.8541243076324463,
      "logps/chosen": -236.55422973632812,
      "logps/ref_chosen": -85.6345443725586,
      "logps/ref_rejected": -86.74612426757812,
      "logps/rejected": -362.2666320800781,
      "loss": 1.0099,
      "margin_dpo/margin_mean": 124.60084533691406,
      "margin_dpo/margin_std": 134.28097534179688,
      "step": 643
    },
    {
      "epoch": 0.9735449735449735,
      "grad_norm": 15.612887382507324,
      "learning_rate": 1.1320193567288527e-09,
      "logits/chosen": 1.0640028715133667,
      "logits/rejected": 1.0360305309295654,
      "logps/chosen": -185.77064514160156,
      "logps/ref_chosen": -43.66929626464844,
      "logps/ref_rejected": -61.423343658447266,
      "logps/rejected": -251.08602905273438,
      "loss": 1.2055,
      "margin_dpo/margin_mean": 47.56130599975586,
      "margin_dpo/margin_std": 103.23611450195312,
      "step": 644
    },
    {
      "epoch": 0.9750566893424036,
      "grad_norm": 18.053247451782227,
      "learning_rate": 1.0098157099674987e-09,
      "logits/chosen": 0.9287192821502686,
      "logits/rejected": 0.9003345966339111,
      "logps/chosen": -219.2329864501953,
      "logps/ref_chosen": -58.112247467041016,
      "logps/ref_rejected": -62.135780334472656,
      "logps/rejected": -301.72528076171875,
      "loss": 1.0062,
      "margin_dpo/margin_mean": 78.46878051757812,
      "margin_dpo/margin_std": 99.51702880859375,
      "step": 645
    },
    {
      "epoch": 0.9765684051398337,
      "grad_norm": 19.156396865844727,
      "learning_rate": 8.945768539031783e-10,
      "logits/chosen": 0.8444064855575562,
      "logits/rejected": 0.7318651676177979,
      "logps/chosen": -187.85931396484375,
      "logps/ref_chosen": -48.80539321899414,
      "logps/ref_rejected": -79.34556579589844,
      "logps/rejected": -289.73748779296875,
      "loss": 1.1103,
      "margin_dpo/margin_mean": 71.3379898071289,
      "margin_dpo/margin_std": 87.06074523925781,
      "step": 646
    },
    {
      "epoch": 0.9780801209372638,
      "grad_norm": 22.503517150878906,
      "learning_rate": 7.863060120144316e-10,
      "logits/chosen": 0.9451459050178528,
      "logits/rejected": 0.7789862155914307,
      "logps/chosen": -220.54901123046875,
      "logps/ref_chosen": -63.22569274902344,
      "logps/ref_rejected": -149.20523071289062,
      "logps/rejected": -385.1798095703125,
      "loss": 0.8997,
      "margin_dpo/margin_mean": 78.65129089355469,
      "margin_dpo/margin_std": 98.49140930175781,
      "step": 647
    },
    {
      "epoch": 0.9795918367346939,
      "grad_norm": 15.99667739868164,
      "learning_rate": 6.850062128694045e-10,
      "logits/chosen": 0.8275370597839355,
      "logits/rejected": 0.7766451835632324,
      "logps/chosen": -226.41717529296875,
      "logps/ref_chosen": -67.55680084228516,
      "logps/ref_rejected": -78.43806457519531,
      "logps/rejected": -259.7667236328125,
      "loss": 1.0994,
      "margin_dpo/margin_mean": 22.468292236328125,
      "margin_dpo/margin_std": 122.872314453125,
      "step": 648
    },
    {
      "epoch": 0.981103552532124,
      "grad_norm": 26.316551208496094,
      "learning_rate": 5.906802900412788e-10,
      "logits/chosen": 0.9751644134521484,
      "logits/rejected": 0.926426887512207,
      "logps/chosen": -177.2069091796875,
      "logps/ref_chosen": -42.90663146972656,
      "logps/ref_rejected": -64.04318237304688,
      "logps/rejected": -287.71142578125,
      "loss": 1.1161,
      "margin_dpo/margin_mean": 89.36796569824219,
      "margin_dpo/margin_std": 82.30953979492188,
      "step": 649
    },
    {
      "epoch": 0.982615268329554,
      "grad_norm": 17.684829711914062,
      "learning_rate": 5.033308820289184e-10,
      "logits/chosen": 1.0313094854354858,
      "logits/rejected": 0.9337575435638428,
      "logps/chosen": -158.346923828125,
      "logps/ref_chosen": -44.43085479736328,
      "logps/ref_rejected": -88.01769256591797,
      "logps/rejected": -314.0736083984375,
      "loss": 1.0373,
      "margin_dpo/margin_mean": 112.13986206054688,
      "margin_dpo/margin_std": 143.96426391601562,
      "step": 650
    },
    {
      "epoch": 0.9841269841269841,
      "grad_norm": 16.09817886352539,
      "learning_rate": 4.2296043218295606e-10,
      "logits/chosen": 0.8436592221260071,
      "logits/rejected": 0.8099097013473511,
      "logps/chosen": -188.1514892578125,
      "logps/ref_chosen": -54.74213790893555,
      "logps/ref_rejected": -69.22908020019531,
      "logps/rejected": -218.41043090820312,
      "loss": 1.0272,
      "margin_dpo/margin_mean": 15.772006034851074,
      "margin_dpo/margin_std": 73.69939422607422,
      "step": 651
    },
    {
      "epoch": 0.9856386999244142,
      "grad_norm": 21.886402130126953,
      "learning_rate": 3.4957118863768176e-10,
      "logits/chosen": 0.9487487077713013,
      "logits/rejected": 0.9406764507293701,
      "logps/chosen": -158.75241088867188,
      "logps/ref_chosen": -39.43302917480469,
      "logps/ref_rejected": -45.17872619628906,
      "logps/rejected": -234.00820922851562,
      "loss": 1.1864,
      "margin_dpo/margin_mean": 69.51010131835938,
      "margin_dpo/margin_std": 92.36702728271484,
      "step": 652
    },
    {
      "epoch": 0.9871504157218443,
      "grad_norm": 16.707674026489258,
      "learning_rate": 2.831652042480093e-10,
      "logits/chosen": 0.7637461423873901,
      "logits/rejected": 0.8205589056015015,
      "logps/chosen": -223.41229248046875,
      "logps/ref_chosen": -68.25508117675781,
      "logps/ref_rejected": -78.01954650878906,
      "logps/rejected": -293.35223388671875,
      "loss": 0.9948,
      "margin_dpo/margin_mean": 60.175498962402344,
      "margin_dpo/margin_std": 88.66094970703125,
      "step": 653
    },
    {
      "epoch": 0.9886621315192744,
      "grad_norm": 18.35707664489746,
      "learning_rate": 2.2374433653205016e-10,
      "logits/chosen": 0.9188174605369568,
      "logits/rejected": 0.8119950890541077,
      "logps/chosen": -183.5347442626953,
      "logps/ref_chosen": -44.32666778564453,
      "logps/ref_rejected": -75.5877685546875,
      "logps/rejected": -248.86541748046875,
      "loss": 1.062,
      "margin_dpo/margin_mean": 34.06956100463867,
      "margin_dpo/margin_std": 89.29640197753906,
      "step": 654
    },
    {
      "epoch": 0.9901738473167044,
      "grad_norm": 15.2190580368042,
      "learning_rate": 1.7131024761923852e-10,
      "logits/chosen": 0.7755975723266602,
      "logits/rejected": 0.8196406364440918,
      "logps/chosen": -192.36354064941406,
      "logps/ref_chosen": -72.61821746826172,
      "logps/ref_rejected": -61.733245849609375,
      "logps/rejected": -256.1627197265625,
      "loss": 0.8594,
      "margin_dpo/margin_mean": 74.68414306640625,
      "margin_dpo/margin_std": 70.4334487915039,
      "step": 655
    },
    {
      "epoch": 0.9916855631141346,
      "grad_norm": 15.999876022338867,
      "learning_rate": 1.2586440420372934e-10,
      "logits/chosen": 0.7340442538261414,
      "logits/rejected": 0.7868098020553589,
      "logps/chosen": -193.5328369140625,
      "logps/ref_chosen": -88.81657409667969,
      "logps/ref_rejected": -67.2658462524414,
      "logps/rejected": -228.84466552734375,
      "loss": 0.9997,
      "margin_dpo/margin_mean": 56.862571716308594,
      "margin_dpo/margin_std": 70.70635986328125,
      "step": 656
    },
    {
      "epoch": 0.9931972789115646,
      "grad_norm": 20.399385452270508,
      "learning_rate": 8.740807750345913e-11,
      "logits/chosen": 1.0027674436569214,
      "logits/rejected": 0.8794127702713013,
      "logps/chosen": -157.93316650390625,
      "logps/ref_chosen": -34.980865478515625,
      "logps/ref_rejected": -85.61087036132812,
      "logps/rejected": -322.360595703125,
      "loss": 0.9962,
      "margin_dpo/margin_mean": 113.79745483398438,
      "margin_dpo/margin_std": 116.89727783203125,
      "step": 657
    },
    {
      "epoch": 0.9947089947089947,
      "grad_norm": 18.558809280395508,
      "learning_rate": 5.594234322453539e-11,
      "logits/chosen": 1.0621709823608398,
      "logits/rejected": 1.0080801248550415,
      "logps/chosen": -221.82943725585938,
      "logps/ref_chosen": -67.89546203613281,
      "logps/ref_rejected": -78.8687515258789,
      "logps/rejected": -280.9857177734375,
      "loss": 1.1266,
      "margin_dpo/margin_mean": 48.18297576904297,
      "margin_dpo/margin_std": 105.40860748291016,
      "step": 658
    },
    {
      "epoch": 0.9962207105064248,
      "grad_norm": 17.150558471679688,
      "learning_rate": 3.146808153123293e-11,
      "logits/chosen": 1.0832101106643677,
      "logits/rejected": 0.9403669834136963,
      "logps/chosen": -138.60635375976562,
      "logps/ref_chosen": -34.2476806640625,
      "logps/ref_rejected": -74.97065734863281,
      "logps/rejected": -273.552001953125,
      "loss": 1.091,
      "margin_dpo/margin_mean": 94.22268676757812,
      "margin_dpo/margin_std": 84.5680923461914,
      "step": 659
    },
    {
      "epoch": 0.9977324263038548,
      "grad_norm": 18.795955657958984,
      "learning_rate": 1.3985977021235829e-11,
      "logits/chosen": 0.9274425506591797,
      "logits/rejected": 0.9459260106086731,
      "logps/chosen": -189.62442016601562,
      "logps/ref_chosen": -57.3446044921875,
      "logps/ref_rejected": -62.27751922607422,
      "logps/rejected": -275.3188171386719,
      "loss": 0.8589,
      "margin_dpo/margin_mean": 80.76148223876953,
      "margin_dpo/margin_std": 117.38458251953125,
      "step": 660
    },
    {
      "epoch": 0.999244142101285,
      "grad_norm": 18.007423400878906,
      "learning_rate": 3.4965187065971735e-12,
      "logits/chosen": 0.883224368095398,
      "logits/rejected": 0.9031695127487183,
      "logps/chosen": -226.634521484375,
      "logps/ref_chosen": -63.84727478027344,
      "logps/ref_rejected": -49.75703430175781,
      "logps/rejected": -239.585205078125,
      "loss": 1.1426,
      "margin_dpo/margin_mean": 27.040931701660156,
      "margin_dpo/margin_std": 113.18049621582031,
      "step": 661
    },
    {
      "epoch": 0.999244142101285,
      "step": 661,
      "total_flos": 0.0,
      "train_loss": 1.1443162902220294,
      "train_runtime": 1702.0491,
      "train_samples_per_second": 24.874,
      "train_steps_per_second": 0.388
    }
  ],
  "logging_steps": 1,
  "max_steps": 661,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}