{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.44982428738773916,
  "eval_steps": 200,
  "global_step": 1800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.004998047637641547,
      "grad_norm": 15.0,
      "learning_rate": 4.7499999999999995e-08,
      "log_odds_chosen": 0.053503312170505524,
      "log_odds_ratio": -0.6757957935333252,
      "logits/chosen": 0.9201898574829102,
      "logits/rejected": 0.9253188967704773,
      "logps/chosen": -0.2984420955181122,
      "logps/rejected": -0.31250840425491333,
      "loss": 0.3322,
      "nll_loss": 0.2984420955181122,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": -0.014922104775905609,
      "rewards/margins": 0.0007033147849142551,
      "rewards/rejected": -0.015625419095158577,
      "step": 20
    },
    {
      "epoch": 0.009996095275283093,
      "grad_norm": 16.625,
      "learning_rate": 9.749999999999999e-08,
      "log_odds_chosen": 0.05107953026890755,
      "log_odds_ratio": -0.6783660054206848,
      "logits/chosen": 0.921713650226593,
      "logits/rejected": 0.9308159947395325,
      "logps/chosen": -0.3038911521434784,
      "logps/rejected": -0.31706634163856506,
      "loss": 0.3378,
      "nll_loss": 0.3038911521434784,
      "rewards/accuracies": 0.6343749761581421,
      "rewards/chosen": -0.01519455574452877,
      "rewards/margins": 0.000658758741337806,
      "rewards/rejected": -0.015853317454457283,
      "step": 40
    },
    {
      "epoch": 0.014994142912924639,
      "grad_norm": 17.125,
      "learning_rate": 1.475e-07,
      "log_odds_chosen": 0.07946861535310745,
      "log_odds_ratio": -0.6646280884742737,
      "logits/chosen": 0.9023857116699219,
      "logits/rejected": 0.9164944887161255,
      "logps/chosen": -0.29856765270233154,
      "logps/rejected": -0.3206278383731842,
      "loss": 0.3318,
      "nll_loss": 0.29856765270233154,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.014928383752703667,
      "rewards/margins": 0.0011030068853870034,
      "rewards/rejected": -0.01603139005601406,
      "step": 60
    },
    {
      "epoch": 0.019992190550566186,
      "grad_norm": 16.375,
      "learning_rate": 1.975e-07,
      "log_odds_chosen": 0.06295903027057648,
      "log_odds_ratio": -0.6717745661735535,
      "logits/chosen": 0.9514683485031128,
      "logits/rejected": 0.9563652873039246,
      "logps/chosen": -0.3006761074066162,
      "logps/rejected": -0.317982941865921,
      "loss": 0.3343,
      "nll_loss": 0.3006761074066162,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.01503380574285984,
      "rewards/margins": 0.0008653400582261384,
      "rewards/rejected": -0.01589914597570896,
      "step": 80
    },
    {
      "epoch": 0.024990238188207732,
      "grad_norm": 16.875,
      "learning_rate": 2.475e-07,
      "log_odds_chosen": 0.08118291199207306,
      "log_odds_ratio": -0.6621332764625549,
      "logits/chosen": 0.9076636433601379,
      "logits/rejected": 0.9208146929740906,
      "logps/chosen": -0.2926270067691803,
      "logps/rejected": -0.3143068552017212,
      "loss": 0.3257,
      "nll_loss": 0.2926270067691803,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.01463135052472353,
      "rewards/margins": 0.0010839940514415503,
      "rewards/rejected": -0.015715342015028,
      "step": 100
    },
    {
      "epoch": 0.029988285825849278,
      "grad_norm": 17.0,
      "learning_rate": 2.9749999999999996e-07,
      "log_odds_chosen": 0.06199132278561592,
      "log_odds_ratio": -0.6743236780166626,
      "logits/chosen": 0.9231328964233398,
      "logits/rejected": 0.9274512529373169,
      "logps/chosen": -0.29089081287384033,
      "logps/rejected": -0.30626899003982544,
      "loss": 0.3246,
      "nll_loss": 0.29089081287384033,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -0.014544541016221046,
      "rewards/margins": 0.0007689098711125553,
      "rewards/rejected": -0.015313449315726757,
      "step": 120
    },
    {
      "epoch": 0.034986333463490824,
      "grad_norm": 14.5,
      "learning_rate": 3.4749999999999996e-07,
      "log_odds_chosen": 0.04646921902894974,
      "log_odds_ratio": -0.6814665794372559,
      "logits/chosen": 0.9149503707885742,
      "logits/rejected": 0.917160153388977,
      "logps/chosen": -0.27925345301628113,
      "logps/rejected": -0.2919124662876129,
      "loss": 0.3133,
      "nll_loss": 0.27925345301628113,
      "rewards/accuracies": 0.640625,
      "rewards/chosen": -0.013962673023343086,
      "rewards/margins": 0.0006329504540190101,
      "rewards/rejected": -0.01459562312811613,
      "step": 140
    },
    {
      "epoch": 0.03998438110113237,
      "grad_norm": 13.625,
      "learning_rate": 3.975e-07,
      "log_odds_chosen": 0.10817976295948029,
      "log_odds_ratio": -0.6498610377311707,
      "logits/chosen": 0.9148440361022949,
      "logits/rejected": 0.927189826965332,
      "logps/chosen": -0.2689565122127533,
      "logps/rejected": -0.29589301347732544,
      "loss": 0.3014,
      "nll_loss": 0.2689565122127533,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.01344782579690218,
      "rewards/margins": 0.0013468241086229682,
      "rewards/rejected": -0.014794649556279182,
      "step": 160
    },
    {
      "epoch": 0.044982428738773915,
      "grad_norm": 15.5625,
      "learning_rate": 4.475e-07,
      "log_odds_chosen": 0.11356785148382187,
      "log_odds_ratio": -0.6494964361190796,
      "logits/chosen": 0.9352966547012329,
      "logits/rejected": 0.9427382349967957,
      "logps/chosen": -0.25873515009880066,
      "logps/rejected": -0.28691548109054565,
      "loss": 0.2912,
      "nll_loss": 0.25873515009880066,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.012936757877469063,
      "rewards/margins": 0.0014090158510953188,
      "rewards/rejected": -0.014345774427056313,
      "step": 180
    },
    {
      "epoch": 0.049980476376415464,
      "grad_norm": 16.25,
      "learning_rate": 4.975e-07,
      "log_odds_chosen": 0.06411220878362656,
      "log_odds_ratio": -0.6711533069610596,
      "logits/chosen": 0.9243879318237305,
      "logits/rejected": 0.9328471422195435,
      "logps/chosen": -0.2516902983188629,
      "logps/rejected": -0.2661595046520233,
      "loss": 0.2852,
      "nll_loss": 0.2516902983188629,
      "rewards/accuracies": 0.653124988079071,
      "rewards/chosen": -0.012584514915943146,
      "rewards/margins": 0.0007234609802253544,
      "rewards/rejected": -0.013307975605130196,
      "step": 200
    },
    {
      "epoch": 0.054978524014057006,
      "grad_norm": 13.8125,
      "learning_rate": 4.778184825674965e-07,
      "log_odds_chosen": 0.09629851579666138,
      "log_odds_ratio": -0.658815860748291,
      "logits/chosen": 0.9301077723503113,
      "logits/rejected": 0.9415726661682129,
      "logps/chosen": -0.236604243516922,
      "logps/rejected": -0.2575286328792572,
      "loss": 0.2695,
      "nll_loss": 0.236604243516922,
      "rewards/accuracies": 0.690625011920929,
      "rewards/chosen": -0.011830211617052555,
      "rewards/margins": 0.0010462201898917556,
      "rewards/rejected": -0.012876431457698345,
      "step": 220
    },
    {
      "epoch": 0.059976571651698556,
      "grad_norm": 14.375,
      "learning_rate": 4.573893537463482e-07,
      "log_odds_chosen": 0.14045165479183197,
      "log_odds_ratio": -0.6405442357063293,
      "logits/chosen": 0.9148332476615906,
      "logits/rejected": 0.927426815032959,
      "logps/chosen": -0.21869663894176483,
      "logps/rejected": -0.2480110377073288,
      "loss": 0.2507,
      "nll_loss": 0.21869663894176483,
      "rewards/accuracies": 0.7093750238418579,
      "rewards/chosen": -0.010934833437204361,
      "rewards/margins": 0.0014657191932201385,
      "rewards/rejected": -0.012400551699101925,
      "step": 240
    },
    {
      "epoch": 0.0649746192893401,
      "grad_norm": 14.0,
      "learning_rate": 4.393747751637468e-07,
      "log_odds_chosen": 0.12323787063360214,
      "log_odds_ratio": -0.6511827707290649,
      "logits/chosen": 0.9187728762626648,
      "logits/rejected": 0.9229263067245483,
      "logps/chosen": -0.21028056740760803,
      "logps/rejected": -0.23566849529743195,
      "loss": 0.2428,
      "nll_loss": 0.21028056740760803,
      "rewards/accuracies": 0.653124988079071,
      "rewards/chosen": -0.010514028370380402,
      "rewards/margins": 0.001269396860152483,
      "rewards/rejected": -0.011783424764871597,
      "step": 260
    },
    {
      "epoch": 0.06997266692698165,
      "grad_norm": 14.4375,
      "learning_rate": 4.233337566673016e-07,
      "log_odds_chosen": 0.13571910560131073,
      "log_odds_ratio": -0.6461990475654602,
      "logits/chosen": 0.9238992929458618,
      "logits/rejected": 0.9354574084281921,
      "logps/chosen": -0.19902026653289795,
      "logps/rejected": -0.22289612889289856,
      "loss": 0.2313,
      "nll_loss": 0.19902026653289795,
      "rewards/accuracies": 0.6968749761581421,
      "rewards/chosen": -0.009951015003025532,
      "rewards/margins": 0.001193791744299233,
      "rewards/rejected": -0.011144806630909443,
      "step": 280
    },
    {
      "epoch": 0.07497071456462319,
      "grad_norm": 16.5,
      "learning_rate": 4.089304100547653e-07,
      "log_odds_chosen": 0.1527879536151886,
      "log_odds_ratio": -0.6412774324417114,
      "logits/chosen": 0.9129140973091125,
      "logits/rejected": 0.9227436780929565,
      "logps/chosen": -0.18697939813137054,
      "logps/rejected": -0.21293890476226807,
      "loss": 0.219,
      "nll_loss": 0.18697938323020935,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.009348970837891102,
      "rewards/margins": 0.00129797519184649,
      "rewards/rejected": -0.010646944865584373,
      "step": 300
    },
    {
      "epoch": 0.07996876220226475,
      "grad_norm": 14.125,
      "learning_rate": 3.959037912324479e-07,
      "log_odds_chosen": 0.15243181586265564,
      "log_odds_ratio": -0.649452269077301,
      "logits/chosen": 0.9372090101242065,
      "logits/rejected": 0.9436160922050476,
      "logps/chosen": -0.17855599522590637,
      "logps/rejected": -0.20220260322093964,
      "loss": 0.211,
      "nll_loss": 0.17855599522590637,
      "rewards/accuracies": 0.6656249761581421,
      "rewards/chosen": -0.008927800692617893,
      "rewards/margins": 0.0011823308886960149,
      "rewards/rejected": -0.010110130533576012,
      "step": 320
    },
    {
      "epoch": 0.08496680983990629,
      "grad_norm": 12.375,
      "learning_rate": 3.8404768632128423e-07,
      "log_odds_chosen": 0.14714181423187256,
      "log_odds_ratio": -0.6486814618110657,
      "logits/chosen": 0.9469276666641235,
      "logits/rejected": 0.9466784596443176,
      "logps/chosen": -0.17795135080814362,
      "logps/rejected": -0.19992610812187195,
      "loss": 0.2104,
      "nll_loss": 0.17795135080814362,
      "rewards/accuracies": 0.690625011920929,
      "rewards/chosen": -0.008897567167878151,
      "rewards/margins": 0.0010987392161041498,
      "rewards/rejected": -0.009996306151151657,
      "step": 340
    },
    {
      "epoch": 0.08996485747754783,
      "grad_norm": 12.125,
      "learning_rate": 3.7319668543103794e-07,
      "log_odds_chosen": 0.21195058524608612,
      "log_odds_ratio": -0.6215499639511108,
      "logits/chosen": 0.9279447793960571,
      "logits/rejected": 0.937017560005188,
      "logps/chosen": -0.15589554607868195,
      "logps/rejected": -0.18526491522789001,
      "loss": 0.187,
      "nll_loss": 0.15589554607868195,
      "rewards/accuracies": 0.7093750238418579,
      "rewards/chosen": -0.007794777397066355,
      "rewards/margins": 0.0014684679917991161,
      "rewards/rejected": -0.009263245388865471,
      "step": 360
    },
    {
      "epoch": 0.09496290511518939,
      "grad_norm": 15.6875,
      "learning_rate": 3.6321635614607396e-07,
      "log_odds_chosen": 0.2677266001701355,
      "log_odds_ratio": -0.6049823760986328,
      "logits/chosen": 0.8884922862052917,
      "logits/rejected": 0.9015036821365356,
      "logps/chosen": -0.15307621657848358,
      "logps/rejected": -0.1920001059770584,
      "loss": 0.1833,
      "nll_loss": 0.1530762016773224,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": -0.007653811015188694,
      "rewards/margins": 0.001946194446645677,
      "rewards/rejected": -0.009600005112588406,
      "step": 380
    },
    {
      "epoch": 0.09996095275283093,
      "grad_norm": 13.375,
      "learning_rate": 3.539961627023943e-07,
      "log_odds_chosen": 0.24993428587913513,
      "log_odds_ratio": -0.6095752716064453,
      "logits/chosen": 0.9298154711723328,
      "logits/rejected": 0.9397618174552917,
      "logps/chosen": -0.15187695622444153,
      "logps/rejected": -0.18737712502479553,
      "loss": 0.1824,
      "nll_loss": 0.15187695622444153,
      "rewards/accuracies": 0.746874988079071,
      "rewards/chosen": -0.007593848742544651,
      "rewards/margins": 0.0017750074621289968,
      "rewards/rejected": -0.009368856437504292,
      "step": 400
    },
    {
      "epoch": 0.10495900039047247,
      "grad_norm": 11.0,
      "learning_rate": 3.4544426792673337e-07,
      "log_odds_chosen": 0.21539661288261414,
      "log_odds_ratio": -0.6292492747306824,
      "logits/chosen": 0.9220725297927856,
      "logits/rejected": 0.9303799867630005,
      "logps/chosen": -0.14802944660186768,
      "logps/rejected": -0.17441526055335999,
      "loss": 0.1795,
      "nll_loss": 0.14802944660186768,
      "rewards/accuracies": 0.6781250238418579,
      "rewards/chosen": -0.007401472423225641,
      "rewards/margins": 0.0013192903716117144,
      "rewards/rejected": -0.008720763027668,
      "step": 420
    },
    {
      "epoch": 0.10995704802811401,
      "grad_norm": 12.375,
      "learning_rate": 3.374836535314135e-07,
      "log_odds_chosen": 0.3108893036842346,
      "log_odds_ratio": -0.5947648286819458,
      "logits/chosen": 0.9221057891845703,
      "logits/rejected": 0.9320783615112305,
      "logps/chosen": -0.13556332886219025,
      "logps/rejected": -0.1746472865343094,
      "loss": 0.1653,
      "nll_loss": 0.13556331396102905,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": -0.006778166629374027,
      "rewards/margins": 0.0019541983492672443,
      "rewards/rejected": -0.00873236358165741,
      "step": 440
    },
    {
      "epoch": 0.11495509566575557,
      "grad_norm": 9.75,
      "learning_rate": 3.3004918099222473e-07,
      "log_odds_chosen": 0.27864372730255127,
      "log_odds_ratio": -0.6085864901542664,
      "logits/chosen": 0.9199404716491699,
      "logits/rejected": 0.9265459179878235,
      "logps/chosen": -0.1416058987379074,
      "logps/rejected": -0.1737016886472702,
      "loss": 0.172,
      "nll_loss": 0.1416058987379074,
      "rewards/accuracies": 0.721875011920929,
      "rewards/chosen": -0.007080295123159885,
      "rewards/margins": 0.001604790217243135,
      "rewards/rejected": -0.008685084991157055,
      "step": 460
    },
    {
      "epoch": 0.11995314330339711,
      "grad_norm": 11.25,
      "learning_rate": 3.2308533487561923e-07,
      "log_odds_chosen": 0.26685529947280884,
      "log_odds_ratio": -0.6108108758926392,
      "logits/chosen": 0.9367237091064453,
      "logits/rejected": 0.941074013710022,
      "logps/chosen": -0.13906843960285187,
      "logps/rejected": -0.17184624075889587,
      "loss": 0.1696,
      "nll_loss": 0.13906843960285187,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.006953421980142593,
      "rewards/margins": 0.0016388905933126807,
      "rewards/rejected": -0.008592312224209309,
      "step": 480
    },
    {
      "epoch": 0.12495119094103865,
      "grad_norm": 10.875,
      "learning_rate": 3.1654446891645916e-07,
      "log_odds_chosen": 0.2782067060470581,
      "log_odds_ratio": -0.6120156049728394,
      "logits/chosen": 0.927230954170227,
      "logits/rejected": 0.93439120054245,
      "logps/chosen": -0.12512263655662537,
      "logps/rejected": -0.15220871567726135,
      "loss": 0.1557,
      "nll_loss": 0.12512263655662537,
      "rewards/accuracies": 0.721875011920929,
      "rewards/chosen": -0.006256132386624813,
      "rewards/margins": 0.0013543039094656706,
      "rewards/rejected": -0.007610436528921127,
      "step": 500
    },
    {
      "epoch": 0.1299492385786802,
      "grad_norm": 12.25,
      "learning_rate": 3.103854276221004e-07,
      "log_odds_chosen": 0.32244807481765747,
      "log_odds_ratio": -0.6038824915885925,
      "logits/chosen": 0.8974004983901978,
      "logits/rejected": 0.9039150476455688,
      "logps/chosen": -0.13300421833992004,
      "logps/rejected": -0.16509173810482025,
      "loss": 0.1632,
      "nll_loss": 0.13300421833992004,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.006650210823863745,
      "rewards/margins": 0.0016043769428506494,
      "rewards/rejected": -0.008254587650299072,
      "step": 520
    },
    {
      "epoch": 0.13494728621632174,
      "grad_norm": 10.5,
      "learning_rate": 3.0457245193658633e-07,
      "log_odds_chosen": 0.3255225121974945,
      "log_odds_ratio": -0.5922858119010925,
      "logits/chosen": 0.9137985110282898,
      "logits/rejected": 0.9264832735061646,
      "logps/chosen": -0.13399162888526917,
      "logps/rejected": -0.16814462840557098,
      "loss": 0.1636,
      "nll_loss": 0.13399162888526917,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.0066995820961892605,
      "rewards/margins": 0.0017076492076739669,
      "rewards/rejected": -0.008407230488955975,
      "step": 540
    },
    {
      "epoch": 0.1399453338539633,
      "grad_norm": 7.90625,
      "learning_rate": 2.9907430227728956e-07,
      "log_odds_chosen": 0.40904003381729126,
      "log_odds_ratio": -0.5781009793281555,
      "logits/chosen": 0.9187723398208618,
      "logits/rejected": 0.9299019575119019,
      "logps/chosen": -0.1179613471031189,
      "logps/rejected": -0.15843844413757324,
      "loss": 0.1469,
      "nll_loss": 0.1179613471031189,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.0058980667963624,
      "rewards/margins": 0.0020238549914211035,
      "rewards/rejected": -0.007921922020614147,
      "step": 560
    },
    {
      "epoch": 0.14494338149160485,
      "grad_norm": 8.0625,
      "learning_rate": 2.938635496921166e-07,
      "log_odds_chosen": 0.3449932932853699,
      "log_odds_ratio": -0.6018896102905273,
      "logits/chosen": 0.9525572061538696,
      "logits/rejected": 0.9547427296638489,
      "logps/chosen": -0.13223889470100403,
      "logps/rejected": -0.15701927244663239,
      "loss": 0.1623,
      "nll_loss": 0.13223889470100403,
      "rewards/accuracies": 0.6968749761581421,
      "rewards/chosen": -0.006611944641917944,
      "rewards/margins": 0.0012390186311677098,
      "rewards/rejected": -0.00785096362233162,
      "step": 580
    },
    {
      "epoch": 0.14994142912924638,
      "grad_norm": 9.3125,
      "learning_rate": 2.8891599832849384e-07,
      "log_odds_chosen": 0.34423956274986267,
      "log_odds_ratio": -0.5991203784942627,
      "logits/chosen": 0.9297738075256348,
      "logits/rejected": 0.9321788549423218,
      "logps/chosen": -0.12783615291118622,
      "logps/rejected": -0.15769797563552856,
      "loss": 0.1578,
      "nll_loss": 0.12783615291118622,
      "rewards/accuracies": 0.703125,
      "rewards/chosen": -0.006391808390617371,
      "rewards/margins": 0.0014930909965187311,
      "rewards/rejected": -0.007884899154305458,
      "step": 600
    },
    {
      "epoch": 0.15493947676688793,
      "grad_norm": 10.1875,
      "learning_rate": 2.842102114016023e-07,
      "log_odds_chosen": 0.42549681663513184,
      "log_odds_ratio": -0.5824495553970337,
      "logits/chosen": 0.943702220916748,
      "logits/rejected": 0.9487889409065247,
      "logps/chosen": -0.11176331341266632,
      "logps/rejected": -0.15435673296451569,
      "loss": 0.1409,
      "nll_loss": 0.11176331341266632,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": -0.005588165950030088,
      "rewards/margins": 0.002129671163856983,
      "rewards/rejected": -0.007717837579548359,
      "step": 620
    },
    {
      "epoch": 0.1599375244045295,
      "grad_norm": 8.3125,
      "learning_rate": 2.7972711943222965e-07,
      "log_odds_chosen": 0.40351733565330505,
      "log_odds_ratio": -0.5847312211990356,
      "logits/chosen": 0.9194564819335938,
      "logits/rejected": 0.9308915138244629,
      "logps/chosen": -0.12050439417362213,
      "logps/rejected": -0.15004847943782806,
      "loss": 0.1497,
      "nll_loss": 0.12050439417362213,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.006025219336152077,
      "rewards/margins": 0.001477204728871584,
      "rewards/rejected": -0.007502424065023661,
      "step": 640
    },
    {
      "epoch": 0.16493557204217102,
      "grad_norm": 9.3125,
      "learning_rate": 2.754496943956198e-07,
      "log_odds_chosen": 0.45548972487449646,
      "log_odds_ratio": -0.5643774271011353,
      "logits/chosen": 0.9225139617919922,
      "logits/rejected": 0.9325849413871765,
      "logps/chosen": -0.12030527740716934,
      "logps/rejected": -0.1618068367242813,
      "loss": 0.1485,
      "nll_loss": 0.12030526250600815,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.006015262566506863,
      "rewards/margins": 0.0020750784315168858,
      "rewards/rejected": -0.008090341463685036,
      "step": 660
    },
    {
      "epoch": 0.16993361967981258,
      "grad_norm": 6.75,
      "learning_rate": 2.7136267706462855e-07,
      "log_odds_chosen": 0.43778514862060547,
      "log_odds_ratio": -0.5712169408798218,
      "logits/chosen": 0.914314866065979,
      "logits/rejected": 0.9230707287788391,
      "logps/chosen": -0.10785359144210815,
      "logps/rejected": -0.14945395290851593,
      "loss": 0.1364,
      "nll_loss": 0.10785359144210815,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.00539267947897315,
      "rewards/margins": 0.0020800174679607153,
      "rewards/rejected": -0.007472698576748371,
      "step": 680
    },
    {
      "epoch": 0.17493166731745413,
      "grad_norm": 7.71875,
      "learning_rate": 2.674523475799702e-07,
      "log_odds_chosen": 0.4804418087005615,
      "log_odds_ratio": -0.5581243634223938,
      "logits/chosen": 0.8933164477348328,
      "logits/rejected": 0.9070658683776855,
      "logps/chosen": -0.11702468246221542,
      "logps/rejected": -0.15869295597076416,
      "loss": 0.1449,
      "nll_loss": 0.11702468246221542,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.005851234775036573,
      "rewards/margins": 0.0020834137685596943,
      "rewards/rejected": -0.007934647612273693,
      "step": 700
    },
    {
      "epoch": 0.17992971495509566,
      "grad_norm": 8.9375,
      "learning_rate": 2.637063313749449e-07,
      "log_odds_chosen": 0.44908443093299866,
      "log_odds_ratio": -0.5753530263900757,
      "logits/chosen": 0.9126005172729492,
      "logits/rejected": 0.9192832708358765,
      "logps/chosen": -0.12123344093561172,
      "logps/rejected": -0.15340758860111237,
      "loss": 0.15,
      "nll_loss": 0.12123344093561172,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": -0.006061672233045101,
      "rewards/margins": 0.0016087075928226113,
      "rewards/rejected": -0.007670379243791103,
      "step": 720
    },
    {
      "epoch": 0.18492776259273722,
      "grad_norm": 7.59375,
      "learning_rate": 2.6011343419135436e-07,
      "log_odds_chosen": 0.4241892695426941,
      "log_odds_ratio": -0.5738194584846497,
      "logits/chosen": 0.9139753580093384,
      "logits/rejected": 0.9213765859603882,
      "logps/chosen": -0.11084351688623428,
      "logps/rejected": -0.1464253067970276,
      "loss": 0.1395,
      "nll_loss": 0.11084351688623428,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.005542175378650427,
      "rewards/margins": 0.0017790896818041801,
      "rewards/rejected": -0.007321265526115894,
      "step": 740
    },
    {
      "epoch": 0.18992581023037877,
      "grad_norm": 7.46875,
      "learning_rate": 2.566635011696727e-07,
      "log_odds_chosen": 0.43390053510665894,
      "log_odds_ratio": -0.5840678811073303,
      "logits/chosen": 0.9058341979980469,
      "logits/rejected": 0.9128694534301758,
      "logps/chosen": -0.12015216052532196,
      "logps/rejected": -0.1576036810874939,
      "loss": 0.1494,
      "nll_loss": 0.12015216052532196,
      "rewards/accuracies": 0.6781250238418579,
      "rewards/chosen": -0.006007608026266098,
      "rewards/margins": 0.0018725760746747255,
      "rewards/rejected": -0.00788018386811018,
      "step": 760
    },
    {
      "epoch": 0.1949238578680203,
      "grad_norm": 6.5625,
      "learning_rate": 2.5334729596906997e-07,
      "log_odds_chosen": 0.44079160690307617,
      "log_odds_ratio": -0.5891875624656677,
      "logits/chosen": 0.8957408666610718,
      "logits/rejected": 0.9005910754203796,
      "logps/chosen": -0.12434019893407822,
      "logps/rejected": -0.1635545790195465,
      "loss": 0.1538,
      "nll_loss": 0.12434019148349762,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.006217009387910366,
      "rewards/margins": 0.001960719469934702,
      "rewards/rejected": -0.008177729323506355,
      "step": 780
    },
    {
      "epoch": 0.19992190550566186,
      "grad_norm": 8.125,
      "learning_rate": 2.5015639663712994e-07,
      "log_odds_chosen": 0.4944338798522949,
      "log_odds_ratio": -0.5524186491966248,
      "logits/chosen": 0.9060338735580444,
      "logits/rejected": 0.9127451181411743,
      "logps/chosen": -0.10685400664806366,
      "logps/rejected": -0.14750149846076965,
      "loss": 0.1345,
      "nll_loss": 0.10685399919748306,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.005342700053006411,
      "rewards/margins": 0.002032375428825617,
      "rewards/rejected": -0.0073750754818320274,
      "step": 800
    },
    {
      "epoch": 0.20491995314330338,
      "grad_norm": 7.09375,
      "learning_rate": 2.4708310555370036e-07,
      "log_odds_chosen": 0.5313096642494202,
      "log_odds_ratio": -0.5536187291145325,
      "logits/chosen": 0.918668270111084,
      "logits/rejected": 0.9237827062606812,
      "logps/chosen": -0.10433455556631088,
      "logps/rejected": -0.1460377722978592,
      "loss": 0.132,
      "nll_loss": 0.10433454811573029,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.005216727964580059,
      "rewards/margins": 0.002085160929709673,
      "rewards/rejected": -0.007301889359951019,
      "step": 820
    },
    {
      "epoch": 0.20991800078094494,
      "grad_norm": 6.8125,
      "learning_rate": 2.44120371254637e-07,
      "log_odds_chosen": 0.47817087173461914,
      "log_odds_ratio": -0.5633410215377808,
      "logits/chosen": 0.8913200497627258,
      "logits/rejected": 0.8975034952163696,
      "logps/chosen": -0.1225389614701271,
      "logps/rejected": -0.15963582694530487,
      "loss": 0.1507,
      "nll_loss": 0.1225389614701271,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.0061269486322999,
      "rewards/margins": 0.0018548432271927595,
      "rewards/rejected": -0.007981792092323303,
      "step": 840
    },
    {
      "epoch": 0.2149160484185865,
      "grad_norm": 6.34375,
      "learning_rate": 2.412617203266904e-07,
      "log_odds_chosen": 0.5440431833267212,
      "log_odds_ratio": -0.5503210425376892,
      "logits/chosen": 0.9195548892021179,
      "logits/rejected": 0.9298384785652161,
      "logps/chosen": -0.11318721622228622,
      "logps/rejected": -0.16293054819107056,
      "loss": 0.1407,
      "nll_loss": 0.11318721622228622,
      "rewards/accuracies": 0.7281249761581421,
      "rewards/chosen": -0.005659360904246569,
      "rewards/margins": 0.0024871663190424442,
      "rewards/rejected": -0.008146527223289013,
      "step": 860
    },
    {
      "epoch": 0.21991409605622803,
      "grad_norm": 7.65625,
      "learning_rate": 2.385011978752745e-07,
      "log_odds_chosen": 0.6000367999076843,
      "log_odds_ratio": -0.5213104486465454,
      "logits/chosen": 0.9104911684989929,
      "logits/rejected": 0.916944146156311,
      "logps/chosen": -0.10162152349948883,
      "logps/rejected": -0.15571561455726624,
      "loss": 0.1277,
      "nll_loss": 0.10162152349948883,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.005081075709313154,
      "rewards/margins": 0.0027047048788517714,
      "rewards/rejected": -0.007785779889672995,
      "step": 880
    },
    {
      "epoch": 0.22491214369386958,
      "grad_norm": 6.53125,
      "learning_rate": 2.3583331531828906e-07,
      "log_odds_chosen": 0.446669340133667,
      "log_odds_ratio": -0.5984799861907959,
      "logits/chosen": 0.913371205329895,
      "logits/rejected": 0.9163562059402466,
      "logps/chosen": -0.11884693056344986,
      "logps/rejected": -0.14721472561359406,
      "loss": 0.1488,
      "nll_loss": 0.11884693056344986,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.005942346528172493,
      "rewards/margins": 0.0014183905441313982,
      "rewards/rejected": -0.00736073637381196,
      "step": 900
    },
    {
      "epoch": 0.22991019133151114,
      "grad_norm": 7.3125,
      "learning_rate": 2.3325300446383042e-07,
      "log_odds_chosen": 0.5155836939811707,
      "log_odds_ratio": -0.5510798692703247,
      "logits/chosen": 0.9267963171005249,
      "logits/rejected": 0.9351229667663574,
      "logps/chosen": -0.10163624584674835,
      "logps/rejected": -0.14410734176635742,
      "loss": 0.1292,
      "nll_loss": 0.10163624584674835,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.005081811919808388,
      "rewards/margins": 0.002123554702848196,
      "rewards/rejected": -0.007205367088317871,
      "step": 920
    },
    {
      "epoch": 0.23490823896915267,
      "grad_norm": 6.65625,
      "learning_rate": 2.307555769970229e-07,
      "log_odds_chosen": 0.5363120436668396,
      "log_odds_ratio": -0.5466144680976868,
      "logits/chosen": 0.9063386917114258,
      "logits/rejected": 0.9168686866760254,
      "logps/chosen": -0.11152966320514679,
      "logps/rejected": -0.1580004245042801,
      "loss": 0.1389,
      "nll_loss": 0.11152966320514679,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.0055764829739928246,
      "rewards/margins": 0.0023235382977873087,
      "rewards/rejected": -0.007900021970272064,
      "step": 940
    },
    {
      "epoch": 0.23990628660679422,
      "grad_norm": 5.96875,
      "learning_rate": 2.283366886387715e-07,
      "log_odds_chosen": 0.547343909740448,
      "log_odds_ratio": -0.546667754650116,
      "logits/chosen": 0.8902915120124817,
      "logits/rejected": 0.9023626446723938,
      "logps/chosen": -0.11764903366565704,
      "logps/rejected": -0.15501850843429565,
      "loss": 0.145,
      "nll_loss": 0.11764900386333466,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.005882450845092535,
      "rewards/margins": 0.0018684742972254753,
      "rewards/rejected": -0.007750925607979298,
      "step": 960
    },
    {
      "epoch": 0.24490433424443578,
      "grad_norm": 7.1875,
      "learning_rate": 2.2599230735278433e-07,
      "log_odds_chosen": 0.43896007537841797,
      "log_odds_ratio": -0.584160327911377,
      "logits/chosen": 0.9090871810913086,
      "logits/rejected": 0.9111261367797852,
      "logps/chosen": -0.11605717241764069,
      "logps/rejected": -0.14885535836219788,
      "loss": 0.1453,
      "nll_loss": 0.11605717241764069,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.005802858155220747,
      "rewards/margins": 0.001639908878132701,
      "rewards/rejected": -0.007442767266184092,
      "step": 980
    },
    {
      "epoch": 0.2499023818820773,
      "grad_norm": 6.5625,
      "learning_rate": 2.2371868507134142e-07,
      "log_odds_chosen": 0.5110324621200562,
      "log_odds_ratio": -0.5605560541152954,
      "logits/chosen": 0.9306246042251587,
      "logits/rejected": 0.9422101974487305,
      "logps/chosen": -0.11277545988559723,
      "logps/rejected": -0.1539277285337448,
      "loss": 0.1408,
      "nll_loss": 0.11277545988559723,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.005638773087412119,
      "rewards/margins": 0.0020576133392751217,
      "rewards/rejected": -0.007696386426687241,
      "step": 1000
    },
    {
      "epoch": 0.25490042951971886,
      "grad_norm": 7.375,
      "learning_rate": 2.2151233248863027e-07,
      "log_odds_chosen": 0.5508419275283813,
      "log_odds_ratio": -0.5428467988967896,
      "logits/chosen": 0.9079995155334473,
      "logits/rejected": 0.9198476076126099,
      "logps/chosen": -0.10438521951436996,
      "logps/rejected": -0.15011721849441528,
      "loss": 0.1315,
      "nll_loss": 0.10438521951436996,
      "rewards/accuracies": 0.746874988079071,
      "rewards/chosen": -0.005219261161983013,
      "rewards/margins": 0.002286599949002266,
      "rewards/rejected": -0.007505861576646566,
      "step": 1020
    },
    {
      "epoch": 0.2598984771573604,
      "grad_norm": 7.1875,
      "learning_rate": 2.1936999653592838e-07,
      "log_odds_chosen": 0.505623459815979,
      "log_odds_ratio": -0.5574303865432739,
      "logits/chosen": 0.9124652147293091,
      "logits/rejected": 0.9180907011032104,
      "logps/chosen": -0.10750292241573334,
      "logps/rejected": -0.1413356065750122,
      "loss": 0.1354,
      "nll_loss": 0.10750291496515274,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": -0.005375145934522152,
      "rewards/margins": 0.0016916341846808791,
      "rewards/rejected": -0.0070667811669409275,
      "step": 1040
    },
    {
      "epoch": 0.264896524795002,
      "grad_norm": 7.09375,
      "learning_rate": 2.1728864020781585e-07,
      "log_odds_chosen": 0.5389599800109863,
      "log_odds_ratio": -0.5554670691490173,
      "logits/chosen": 0.9160731434822083,
      "logits/rejected": 0.9225283861160278,
      "logps/chosen": -0.10790044069290161,
      "logps/rejected": -0.1453263759613037,
      "loss": 0.1357,
      "nll_loss": 0.10790042579174042,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.005395021755248308,
      "rewards/margins": 0.0018712974851951003,
      "rewards/rejected": -0.007266318891197443,
      "step": 1060
    },
    {
      "epoch": 0.2698945724326435,
      "grad_norm": 6.125,
      "learning_rate": 2.1526542445480877e-07,
      "log_odds_chosen": 0.5067550539970398,
      "log_odds_ratio": -0.5616964101791382,
      "logits/chosen": 0.897504448890686,
      "logits/rejected": 0.9034231305122375,
      "logps/chosen": -0.10974349081516266,
      "logps/rejected": -0.1446974128484726,
      "loss": 0.1378,
      "nll_loss": 0.10974347591400146,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.005487173795700073,
      "rewards/margins": 0.0017476966604590416,
      "rewards/rejected": -0.007234870456159115,
      "step": 1080
    },
    {
      "epoch": 0.27489262007028503,
      "grad_norm": 5.875,
      "learning_rate": 2.1329769189683425e-07,
      "log_odds_chosen": 0.5055086016654968,
      "log_odds_ratio": -0.5575133562088013,
      "logits/chosen": 0.9033730626106262,
      "logits/rejected": 0.911507785320282,
      "logps/chosen": -0.1160176545381546,
      "logps/rejected": -0.15694013237953186,
      "loss": 0.1439,
      "nll_loss": 0.1160176545381546,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.0058008828200399876,
      "rewards/margins": 0.0020461236126720905,
      "rewards/rejected": -0.007847006432712078,
      "step": 1100
    },
    {
      "epoch": 0.2798906677079266,
      "grad_norm": 5.90625,
      "learning_rate": 2.113829521450379e-07,
      "log_odds_chosen": 0.4901196360588074,
      "log_odds_ratio": -0.5681068301200867,
      "logits/chosen": 0.9164495468139648,
      "logits/rejected": 0.9228304028511047,
      "logps/chosen": -0.11275438964366913,
      "logps/rejected": -0.14565686881542206,
      "loss": 0.1412,
      "nll_loss": 0.11275438964366913,
      "rewards/accuracies": 0.721875011920929,
      "rewards/chosen": -0.005637719761580229,
      "rewards/margins": 0.0016451241681352258,
      "rewards/rejected": -0.0072828433476388454,
      "step": 1120
    },
    {
      "epoch": 0.28488871534556814,
      "grad_norm": 6.78125,
      "learning_rate": 2.0951886854752806e-07,
      "log_odds_chosen": 0.4967380464076996,
      "log_odds_ratio": -0.5616387128829956,
      "logits/chosen": 0.9169589877128601,
      "logits/rejected": 0.9249241948127747,
      "logps/chosen": -0.11185993999242783,
      "logps/rejected": -0.14960147440433502,
      "loss": 0.1399,
      "nll_loss": 0.11185993999242783,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.0055929976515471935,
      "rewards/margins": 0.0018870767671614885,
      "rewards/rejected": -0.007480073720216751,
      "step": 1140
    },
    {
      "epoch": 0.2898867629832097,
      "grad_norm": 6.53125,
      "learning_rate": 2.0770324619863194e-07,
      "log_odds_chosen": 0.5465720891952515,
      "log_odds_ratio": -0.5451698303222656,
      "logits/chosen": 0.874528706073761,
      "logits/rejected": 0.884798526763916,
      "logps/chosen": -0.10728394985198975,
      "logps/rejected": -0.15278363227844238,
      "loss": 0.1345,
      "nll_loss": 0.10728394985198975,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.005364197306334972,
      "rewards/margins": 0.002274984959512949,
      "rewards/rejected": -0.007639183197170496,
      "step": 1160
    },
    {
      "epoch": 0.29488481062085126,
      "grad_norm": 6.15625,
      "learning_rate": 2.059340210717385e-07,
      "log_odds_chosen": 0.6263851523399353,
      "log_odds_ratio": -0.5190600156784058,
      "logits/chosen": 0.9336968660354614,
      "logits/rejected": 0.9423727989196777,
      "logps/chosen": -0.10160555690526962,
      "logps/rejected": -0.15376660227775574,
      "loss": 0.1276,
      "nll_loss": 0.10160555690526962,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.005080278031527996,
      "rewards/margins": 0.002608052920550108,
      "rewards/rejected": -0.007688330952078104,
      "step": 1180
    },
    {
      "epoch": 0.29988285825849276,
      "grad_norm": 7.46875,
      "learning_rate": 2.042092501533827e-07,
      "log_odds_chosen": 0.5330203771591187,
      "log_odds_ratio": -0.552624523639679,
      "logits/chosen": 0.8918447494506836,
      "logits/rejected": 0.8984712362289429,
      "logps/chosen": -0.10406164079904556,
      "logps/rejected": -0.14680352807044983,
      "loss": 0.1317,
      "nll_loss": 0.10406164079904556,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.005203082226216793,
      "rewards/margins": 0.002137094270437956,
      "rewards/rejected": -0.007340176496654749,
      "step": 1200
    },
    {
      "epoch": 0.3048809058961343,
      "grad_norm": 6.96875,
      "learning_rate": 2.0252710247134554e-07,
      "log_odds_chosen": 0.5235636234283447,
      "log_odds_ratio": -0.5710805654525757,
      "logits/chosen": 0.9045993089675903,
      "logits/rejected": 0.907692551612854,
      "logps/chosen": -0.11793967336416245,
      "logps/rejected": -0.15877366065979004,
      "loss": 0.1465,
      "nll_loss": 0.11793966591358185,
      "rewards/accuracies": 0.690625011920929,
      "rewards/chosen": -0.0058969841338694096,
      "rewards/margins": 0.0020416988991200924,
      "rewards/rejected": -0.007938683032989502,
      "step": 1220
    },
    {
      "epoch": 0.30987895353377587,
      "grad_norm": 6.71875,
      "learning_rate": 2.0088585092257693e-07,
      "log_odds_chosen": 0.538957953453064,
      "log_odds_ratio": -0.5514823198318481,
      "logits/chosen": 0.8724090456962585,
      "logits/rejected": 0.8843700289726257,
      "logps/chosen": -0.11092664301395416,
      "logps/rejected": -0.15797437727451324,
      "loss": 0.1385,
      "nll_loss": 0.11092665046453476,
      "rewards/accuracies": 0.703125,
      "rewards/chosen": -0.005546332336962223,
      "rewards/margins": 0.0023523871786892414,
      "rewards/rejected": -0.007898719981312752,
      "step": 1240
    },
    {
      "epoch": 0.3148770011714174,
      "grad_norm": 7.53125,
      "learning_rate": 1.9928386481802028e-07,
      "log_odds_chosen": 0.48481544852256775,
      "log_odds_ratio": -0.5818430185317993,
      "logits/chosen": 0.9090177416801453,
      "logits/rejected": 0.9191758036613464,
      "logps/chosen": -0.11104512214660645,
      "logps/rejected": -0.15532509982585907,
      "loss": 0.1401,
      "nll_loss": 0.11104512214660645,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.005552256479859352,
      "rewards/margins": 0.0022139991633594036,
      "rewards/rejected": -0.007766254246234894,
      "step": 1260
    },
    {
      "epoch": 0.319875048809059,
      "grad_norm": 8.5625,
      "learning_rate": 1.9771960307117736e-07,
      "log_odds_chosen": 0.5779954195022583,
      "log_odds_ratio": -0.5283289551734924,
      "logits/chosen": 0.8994030952453613,
      "logits/rejected": 0.909386157989502,
      "logps/chosen": -0.10013916343450546,
      "logps/rejected": -0.1448444128036499,
      "loss": 0.1266,
      "nll_loss": 0.10013916343450546,
      "rewards/accuracies": 0.753125011920929,
      "rewards/chosen": -0.005006958730518818,
      "rewards/margins": 0.0022352614905685186,
      "rewards/rejected": -0.00724222045391798,
      "step": 1280
    },
    {
      "epoch": 0.3248730964467005,
      "grad_norm": 6.625,
      "learning_rate": 1.961916079657376e-07,
      "log_odds_chosen": 0.5866302251815796,
      "log_odds_ratio": -0.5394390225410461,
      "logits/chosen": 0.912439227104187,
      "logits/rejected": 0.9187355041503906,
      "logps/chosen": -0.10509935766458511,
      "logps/rejected": -0.14786145091056824,
      "loss": 0.1321,
      "nll_loss": 0.10509936511516571,
      "rewards/accuracies": 0.7281249761581421,
      "rewards/chosen": -0.005254968069493771,
      "rewards/margins": 0.0021381047554314137,
      "rewards/rejected": -0.0073930732905864716,
      "step": 1300
    },
    {
      "epoch": 0.32987114408434204,
      "grad_norm": 7.21875,
      "learning_rate": 1.9469849944497735e-07,
      "log_odds_chosen": 0.5995177030563354,
      "log_odds_ratio": -0.5288160443305969,
      "logits/chosen": 0.8972539901733398,
      "logits/rejected": 0.9027105569839478,
      "logps/chosen": -0.1045871153473854,
      "logps/rejected": -0.14993666112422943,
      "loss": 0.131,
      "nll_loss": 0.1045871153473854,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.005229355301707983,
      "rewards/margins": 0.0022674778010696173,
      "rewards/rejected": -0.007496832869946957,
      "step": 1320
    },
    {
      "epoch": 0.3348691917219836,
      "grad_norm": 7.75,
      "learning_rate": 1.9323896987207783e-07,
      "log_odds_chosen": 0.5442622900009155,
      "log_odds_ratio": -0.546025812625885,
      "logits/chosen": 0.905092716217041,
      "logits/rejected": 0.9169044494628906,
      "logps/chosen": -0.11127519607543945,
      "logps/rejected": -0.1558954417705536,
      "loss": 0.1386,
      "nll_loss": 0.11127519607543945,
      "rewards/accuracies": 0.7281249761581421,
      "rewards/chosen": -0.00556376064196229,
      "rewards/margins": 0.002231011865660548,
      "rewards/rejected": -0.007794772274792194,
      "step": 1340
    },
    {
      "epoch": 0.33986723935962515,
      "grad_norm": 6.8125,
      "learning_rate": 1.9181177921614533e-07,
      "log_odds_chosen": 0.6376086473464966,
      "log_odds_ratio": -0.5051892995834351,
      "logits/chosen": 0.90375155210495,
      "logits/rejected": 0.9119564890861511,
      "logps/chosen": -0.10214891284704208,
      "logps/rejected": -0.15685433149337769,
      "loss": 0.1274,
      "nll_loss": 0.10214891284704208,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.0051074461080133915,
      "rewards/margins": 0.0027352706529200077,
      "rewards/rejected": -0.0078427167609334,
      "step": 1360
    },
    {
      "epoch": 0.3448652869972667,
      "grad_norm": 6.40625,
      "learning_rate": 1.9041575062365312e-07,
      "log_odds_chosen": 0.475393682718277,
      "log_odds_ratio": -0.5804024934768677,
      "logits/chosen": 0.907962441444397,
      "logits/rejected": 0.9179188013076782,
      "logps/chosen": -0.11482979357242584,
      "logps/rejected": -0.15231165289878845,
      "loss": 0.1438,
      "nll_loss": 0.11482979357242584,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.005741490051150322,
      "rewards/margins": 0.001874092617072165,
      "rewards/rejected": -0.007615582551807165,
      "step": 1380
    },
    {
      "epoch": 0.34986333463490826,
      "grad_norm": 6.90625,
      "learning_rate": 1.8904976633935942e-07,
      "log_odds_chosen": 0.5324280858039856,
      "log_odds_ratio": -0.5489314794540405,
      "logits/chosen": 0.9013870358467102,
      "logits/rejected": 0.913454532623291,
      "logps/chosen": -0.11568818986415863,
      "logps/rejected": -0.15911448001861572,
      "loss": 0.1431,
      "nll_loss": 0.11568818986415863,
      "rewards/accuracies": 0.7281249761581421,
      "rewards/chosen": -0.005784410052001476,
      "rewards/margins": 0.0021713145542889833,
      "rewards/rejected": -0.007955724373459816,
      "step": 1400
    },
    {
      "epoch": 0.35486138227254976,
      "grad_norm": 6.09375,
      "learning_rate": 1.8771276394456945e-07,
      "log_odds_chosen": 0.49037519097328186,
      "log_odds_ratio": -0.5825397372245789,
      "logits/chosen": 0.9016868472099304,
      "logits/rejected": 0.9083212018013,
      "logps/chosen": -0.11602238565683365,
      "logps/rejected": -0.15114490687847137,
      "loss": 0.1451,
      "nll_loss": 0.11602238565683365,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.005801118910312653,
      "rewards/margins": 0.0017561258282512426,
      "rewards/rejected": -0.007557244505733252,
      "step": 1420
    },
    {
      "epoch": 0.3598594299101913,
      "grad_norm": 6.46875,
      "learning_rate": 1.8640373288397021e-07,
      "log_odds_chosen": 0.4668620526790619,
      "log_odds_ratio": -0.5718046426773071,
      "logits/chosen": 0.9127907752990723,
      "logits/rejected": 0.9162998199462891,
      "logps/chosen": -0.11841784417629242,
      "logps/rejected": -0.15003439784049988,
      "loss": 0.147,
      "nll_loss": 0.11841784417629242,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.005920892581343651,
      "rewards/margins": 0.0015808267053216696,
      "rewards/rejected": -0.007501719985157251,
      "step": 1440
    },
    {
      "epoch": 0.3648574775478329,
      "grad_norm": 6.96875,
      "learning_rate": 1.851217112552365e-07,
      "log_odds_chosen": 0.4538698196411133,
      "log_odds_ratio": -0.5834226012229919,
      "logits/chosen": 0.8787198066711426,
      "logits/rejected": 0.8867900967597961,
      "logps/chosen": -0.12552694976329803,
      "logps/rejected": -0.16267669200897217,
      "loss": 0.1547,
      "nll_loss": 0.12552696466445923,
      "rewards/accuracies": 0.659375011920929,
      "rewards/chosen": -0.006276348140090704,
      "rewards/margins": 0.0018574863206595182,
      "rewards/rejected": -0.008133834227919579,
      "step": 1460
    },
    {
      "epoch": 0.36985552518547443,
      "grad_norm": 5.46875,
      "learning_rate": 1.8386578283823054e-07,
      "log_odds_chosen": 0.5172036290168762,
      "log_odds_ratio": -0.556606650352478,
      "logits/chosen": 0.8837558627128601,
      "logits/rejected": 0.893505871295929,
      "logps/chosen": -0.11744554340839386,
      "logps/rejected": -0.1555454432964325,
      "loss": 0.1453,
      "nll_loss": 0.11744552850723267,
      "rewards/accuracies": 0.721875011920929,
      "rewards/chosen": -0.0058722770772874355,
      "rewards/margins": 0.0019049948314204812,
      "rewards/rejected": -0.007777272257953882,
      "step": 1480
    },
    {
      "epoch": 0.374853572823116,
      "grad_norm": 7.125,
      "learning_rate": 1.8263507434294625e-07,
      "log_odds_chosen": 0.6737366914749146,
      "log_odds_ratio": -0.4980563223361969,
      "logits/chosen": 0.9022199511528015,
      "logits/rejected": 0.911712646484375,
      "logps/chosen": -0.09678754955530167,
      "logps/rejected": -0.15280142426490784,
      "loss": 0.1217,
      "nll_loss": 0.09678753465414047,
      "rewards/accuracies": 0.7906249761581421,
      "rewards/chosen": -0.004839377012103796,
      "rewards/margins": 0.002800694201141596,
      "rewards/rejected": -0.007640071213245392,
      "step": 1500
    },
    {
      "epoch": 0.37985162046075754,
      "grad_norm": 8.0625,
      "learning_rate": 1.8142875285741498e-07,
      "log_odds_chosen": 0.6103123426437378,
      "log_odds_ratio": -0.522514820098877,
      "logits/chosen": 0.9107071161270142,
      "logits/rejected": 0.9207186698913574,
      "logps/chosen": -0.09944355487823486,
      "logps/rejected": -0.1504182517528534,
      "loss": 0.1256,
      "nll_loss": 0.09944355487823486,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.004972177557647228,
      "rewards/margins": 0.0025487353559583426,
      "rewards/rejected": -0.0075209131464362144,
      "step": 1520
    },
    {
      "epoch": 0.38484966809839904,
      "grad_norm": 7.3125,
      "learning_rate": 1.8024602347862503e-07,
      "log_odds_chosen": 0.49091896414756775,
      "log_odds_ratio": -0.57284015417099,
      "logits/chosen": 0.9056397676467896,
      "logits/rejected": 0.911255955696106,
      "logps/chosen": -0.11540244519710541,
      "logps/rejected": -0.1518068015575409,
      "loss": 0.144,
      "nll_loss": 0.1154024600982666,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.0057701226323843,
      "rewards/margins": 0.0018202174687758088,
      "rewards/rejected": -0.007590339984744787,
      "step": 1540
    },
    {
      "epoch": 0.3898477157360406,
      "grad_norm": 7.0625,
      "learning_rate": 1.7908612711114454e-07,
      "log_odds_chosen": 0.5035929679870605,
      "log_odds_ratio": -0.5619618892669678,
      "logits/chosen": 0.8992635607719421,
      "logits/rejected": 0.9032192230224609,
      "logps/chosen": -0.11157232522964478,
      "logps/rejected": -0.1540084183216095,
      "loss": 0.1397,
      "nll_loss": 0.11157232522964478,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.005578616634011269,
      "rewards/margins": 0.0021218047477304935,
      "rewards/rejected": -0.007700420916080475,
      "step": 1560
    },
    {
      "epoch": 0.39484576337368216,
      "grad_norm": 7.65625,
      "learning_rate": 1.7794833841959615e-07,
      "log_odds_chosen": 0.6406098008155823,
      "log_odds_ratio": -0.5061673521995544,
      "logits/chosen": 0.9134248495101929,
      "logits/rejected": 0.9297193288803101,
      "logps/chosen": -0.09615035355091095,
      "logps/rejected": -0.153140589594841,
      "loss": 0.1215,
      "nll_loss": 0.09615035355091095,
      "rewards/accuracies": 0.778124988079071,
      "rewards/chosen": -0.004807517863810062,
      "rewards/margins": 0.002849512966349721,
      "rewards/rejected": -0.007657030131667852,
      "step": 1580
    },
    {
      "epoch": 0.3998438110113237,
      "grad_norm": 8.0,
      "learning_rate": 1.7683196392243516e-07,
      "log_odds_chosen": 0.5980408191680908,
      "log_odds_ratio": -0.5216068029403687,
      "logits/chosen": 0.9140481948852539,
      "logits/rejected": 0.9252773523330688,
      "logps/chosen": -0.10728876292705536,
      "logps/rejected": -0.15625333786010742,
      "loss": 0.1334,
      "nll_loss": 0.10728876292705536,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.0053644380532205105,
      "rewards/margins": 0.0024482286535203457,
      "rewards/rejected": -0.007812665775418282,
      "step": 1600
    },
    {
      "epoch": 0.40484185864896527,
      "grad_norm": 7.4375,
      "learning_rate": 1.7573634021564908e-07,
      "log_odds_chosen": 0.5844752788543701,
      "log_odds_ratio": -0.5344146490097046,
      "logits/chosen": 0.9050723314285278,
      "logits/rejected": 0.9147335886955261,
      "logps/chosen": -0.10331355035305023,
      "logps/rejected": -0.14794692397117615,
      "loss": 0.13,
      "nll_loss": 0.10331355035305023,
      "rewards/accuracies": 0.734375,
      "rewards/chosen": -0.005165677517652512,
      "rewards/margins": 0.0022316682152450085,
      "rewards/rejected": -0.007397345267236233,
      "step": 1620
    },
    {
      "epoch": 0.40983990628660677,
      "grad_norm": 6.25,
      "learning_rate": 1.7466083231604192e-07,
      "log_odds_chosen": 0.5599223971366882,
      "log_odds_ratio": -0.5451304912567139,
      "logits/chosen": 0.921318531036377,
      "logits/rejected": 0.9311621785163879,
      "logps/chosen": -0.11299822479486465,
      "logps/rejected": -0.15573526918888092,
      "loss": 0.1403,
      "nll_loss": 0.11299822479486465,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.005649910774081945,
      "rewards/margins": 0.0021368523593991995,
      "rewards/rejected": -0.0077867633663117886,
      "step": 1640
    },
    {
      "epoch": 0.4148379539242483,
      "grad_norm": 6.3125,
      "learning_rate": 1.7360483211470364e-07,
      "log_odds_chosen": 0.5110288858413696,
      "log_odds_ratio": -0.5735262632369995,
      "logits/chosen": 0.9024496078491211,
      "logits/rejected": 0.9109729528427124,
      "logps/chosen": -0.11019035428762436,
      "logps/rejected": -0.150790274143219,
      "loss": 0.1389,
      "nll_loss": 0.11019035428762436,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.005509518086910248,
      "rewards/margins": 0.0020299963653087616,
      "rewards/rejected": -0.007539515383541584,
      "step": 1660
    },
    {
      "epoch": 0.4198360015618899,
      "grad_norm": 7.59375,
      "learning_rate": 1.7256775693210743e-07,
      "log_odds_chosen": 0.516677975654602,
      "log_odds_ratio": -0.5585073232650757,
      "logits/chosen": 0.9147561192512512,
      "logits/rejected": 0.9188590049743652,
      "logps/chosen": -0.11596517264842987,
      "logps/rejected": -0.15386167168617249,
      "loss": 0.1439,
      "nll_loss": 0.11596517264842987,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.0057982588186860085,
      "rewards/margins": 0.0018948239739984274,
      "rewards/rejected": -0.0076930830255150795,
      "step": 1680
    },
    {
      "epoch": 0.42483404919953144,
      "grad_norm": 7.21875,
      "learning_rate": 1.7154904816703557e-07,
      "log_odds_chosen": 0.5469341278076172,
      "log_odds_ratio": -0.5440756678581238,
      "logits/chosen": 0.8829646110534668,
      "logits/rejected": 0.8898215293884277,
      "logps/chosen": -0.10766366869211197,
      "logps/rejected": -0.14634230732917786,
      "loss": 0.1349,
      "nll_loss": 0.10766365379095078,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.0053831832483410835,
      "rewards/margins": 0.001933931722305715,
      "rewards/rejected": -0.007317115552723408,
      "step": 1700
    },
    {
      "epoch": 0.429832096837173,
      "grad_norm": 6.84375,
      "learning_rate": 1.7054817003221693e-07,
      "log_odds_chosen": 0.5472453832626343,
      "log_odds_ratio": -0.5421464443206787,
      "logits/chosen": 0.9170600771903992,
      "logits/rejected": 0.9204696416854858,
      "logps/chosen": -0.10923080146312714,
      "logps/rejected": -0.1489257663488388,
      "loss": 0.1363,
      "nll_loss": 0.10923081636428833,
      "rewards/accuracies": 0.753125011920929,
      "rewards/chosen": -0.005461541004478931,
      "rewards/margins": 0.001984748523682356,
      "rewards/rejected": -0.0074462890625,
      "step": 1720
    },
    {
      "epoch": 0.43483014447481455,
      "grad_norm": 6.5,
      "learning_rate": 1.6956460837017516e-07,
      "log_odds_chosen": 0.48406925797462463,
      "log_odds_ratio": -0.5709875822067261,
      "logits/chosen": 0.8931280970573425,
      "logits/rejected": 0.8998786807060242,
      "logps/chosen": -0.1121620312333107,
      "logps/rejected": -0.15092679858207703,
      "loss": 0.1407,
      "nll_loss": 0.1121620163321495,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.005608101841062307,
      "rewards/margins": 0.0019382396712899208,
      "rewards/rejected": -0.007546341512352228,
      "step": 1740
    },
    {
      "epoch": 0.43982819211245605,
      "grad_norm": 7.09375,
      "learning_rate": 1.6859786954334187e-07,
      "log_odds_chosen": 0.49666181206703186,
      "log_odds_ratio": -0.5695202946662903,
      "logits/chosen": 0.9191263914108276,
      "logits/rejected": 0.9239128232002258,
      "logps/chosen": -0.11047865450382233,
      "logps/rejected": -0.14814090728759766,
      "loss": 0.139,
      "nll_loss": 0.11047865450382233,
      "rewards/accuracies": 0.721875011920929,
      "rewards/chosen": -0.005523933097720146,
      "rewards/margins": 0.001883113756775856,
      "rewards/rejected": -0.007407046854496002,
      "step": 1760
    },
    {
      "epoch": 0.4448262397500976,
      "grad_norm": 7.0625,
      "learning_rate": 1.6764747939299312e-07,
      "log_odds_chosen": 0.6367926001548767,
      "log_odds_ratio": -0.5216981768608093,
      "logits/chosen": 0.8996027112007141,
      "logits/rejected": 0.9154459238052368,
      "logps/chosen": -0.09838853031396866,
      "logps/rejected": -0.15085098147392273,
      "loss": 0.1245,
      "nll_loss": 0.09838853031396866,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.004919426515698433,
      "rewards/margins": 0.002623123349621892,
      "rewards/rejected": -0.007542549632489681,
      "step": 1780
    },
    {
      "epoch": 0.44982428738773916,
      "grad_norm": 7.0,
      "learning_rate": 1.667129822620214e-07,
      "log_odds_chosen": 0.5275170207023621,
      "log_odds_ratio": -0.5443486571311951,
      "logits/chosen": 0.9300762414932251,
      "logits/rejected": 0.9408272504806519,
      "logps/chosen": -0.1066521406173706,
      "logps/rejected": -0.14904627203941345,
      "loss": 0.1339,
      "nll_loss": 0.1066521406173706,
      "rewards/accuracies": 0.7406250238418579,
      "rewards/chosen": -0.005332607310265303,
      "rewards/margins": 0.0021197067108005285,
      "rewards/rejected": -0.0074523137882351875,
      "step": 1800
    }
  ],
  "logging_steps": 20,
  "max_steps": 2000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}