martimfasantos's picture
Model save
9d61ede verified
raw
history blame
225 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"eval_steps": 100,
"global_step": 4164,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0007204610951008645,
"grad_norm": 2.336021900177002,
"learning_rate": 1.199040767386091e-10,
"logits/chosen": -1.3860063552856445,
"logits/rejected": -1.3949532508850098,
"logps/chosen": -34.621925354003906,
"logps/rejected": -37.30891418457031,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.007204610951008645,
"grad_norm": 2.7931599617004395,
"learning_rate": 1.199040767386091e-09,
"logits/chosen": -1.546767234802246,
"logits/rejected": -1.5282517671585083,
"logps/chosen": -42.52494812011719,
"logps/rejected": -44.546756744384766,
"loss": 0.6932,
"rewards/accuracies": 0.3680555522441864,
"rewards/chosen": -0.00010908626427408308,
"rewards/margins": -0.00013866486551705748,
"rewards/rejected": 2.95786012429744e-05,
"step": 10
},
{
"epoch": 0.01440922190201729,
"grad_norm": 2.9333579540252686,
"learning_rate": 2.398081534772182e-09,
"logits/chosen": -1.5552335977554321,
"logits/rejected": -1.5412750244140625,
"logps/chosen": -44.075599670410156,
"logps/rejected": -46.59809112548828,
"loss": 0.6931,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 9.86563172773458e-05,
"rewards/margins": 0.00012048264034092426,
"rewards/rejected": -2.182633033953607e-05,
"step": 20
},
{
"epoch": 0.021613832853025938,
"grad_norm": 3.4909462928771973,
"learning_rate": 3.597122302158273e-09,
"logits/chosen": -1.511649250984192,
"logits/rejected": -1.5045579671859741,
"logps/chosen": -47.84404754638672,
"logps/rejected": -50.805335998535156,
"loss": 0.6931,
"rewards/accuracies": 0.5,
"rewards/chosen": 4.7573317715432495e-05,
"rewards/margins": 0.00018799836107064039,
"rewards/rejected": -0.0001404250506311655,
"step": 30
},
{
"epoch": 0.02881844380403458,
"grad_norm": 2.575885772705078,
"learning_rate": 4.796163069544364e-09,
"logits/chosen": -1.5584311485290527,
"logits/rejected": -1.5544531345367432,
"logps/chosen": -43.06354522705078,
"logps/rejected": -45.552642822265625,
"loss": 0.6932,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.0001178598977276124,
"rewards/margins": -6.472436507465318e-05,
"rewards/rejected": -5.313555084285326e-05,
"step": 40
},
{
"epoch": 0.03602305475504323,
"grad_norm": 2.653576374053955,
"learning_rate": 5.995203836930456e-09,
"logits/chosen": -1.469327688217163,
"logits/rejected": -1.4684107303619385,
"logps/chosen": -42.99556350708008,
"logps/rejected": -44.81259536743164,
"loss": 0.6932,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.00011682316107908264,
"rewards/margins": -2.396003037574701e-05,
"rewards/rejected": 0.0001407831732649356,
"step": 50
},
{
"epoch": 0.043227665706051875,
"grad_norm": 3.936546564102173,
"learning_rate": 7.194244604316546e-09,
"logits/chosen": -1.567378044128418,
"logits/rejected": -1.5606577396392822,
"logps/chosen": -50.69051742553711,
"logps/rejected": -52.026954650878906,
"loss": 0.6931,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -3.9090933569241315e-05,
"rewards/margins": 1.4172494047670625e-05,
"rewards/rejected": -5.326343307388015e-05,
"step": 60
},
{
"epoch": 0.05043227665706052,
"grad_norm": 2.3031318187713623,
"learning_rate": 8.393285371702639e-09,
"logits/chosen": -1.5360815525054932,
"logits/rejected": -1.5282552242279053,
"logps/chosen": -50.07221221923828,
"logps/rejected": -52.78315353393555,
"loss": 0.6931,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -4.2442261474207044e-05,
"rewards/margins": 4.4568816520040855e-05,
"rewards/rejected": -8.701106708031148e-05,
"step": 70
},
{
"epoch": 0.05763688760806916,
"grad_norm": 3.459690570831299,
"learning_rate": 9.592326139088728e-09,
"logits/chosen": -1.5700651407241821,
"logits/rejected": -1.5622011423110962,
"logps/chosen": -51.09720230102539,
"logps/rejected": -52.700775146484375,
"loss": 0.693,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.00014371283759828657,
"rewards/margins": 0.0003250584995839745,
"rewards/rejected": -0.00018134564743377268,
"step": 80
},
{
"epoch": 0.06484149855907781,
"grad_norm": 2.787858009338379,
"learning_rate": 1.0791366906474819e-08,
"logits/chosen": -1.5036344528198242,
"logits/rejected": -1.5007749795913696,
"logps/chosen": -49.02019500732422,
"logps/rejected": -51.165855407714844,
"loss": 0.6932,
"rewards/accuracies": 0.4375,
"rewards/chosen": -3.3910677302628756e-05,
"rewards/margins": -8.141305443132296e-05,
"rewards/rejected": 4.750239895656705e-05,
"step": 90
},
{
"epoch": 0.07204610951008646,
"grad_norm": 2.978163957595825,
"learning_rate": 1.1990407673860912e-08,
"logits/chosen": -1.5843650102615356,
"logits/rejected": -1.5736749172210693,
"logps/chosen": -45.69003677368164,
"logps/rejected": -48.75359344482422,
"loss": 0.6932,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 5.446890099847224e-06,
"rewards/margins": -4.437283132574521e-05,
"rewards/rejected": 4.981973324902356e-05,
"step": 100
},
{
"epoch": 0.0792507204610951,
"grad_norm": 2.3196158409118652,
"learning_rate": 1.3189448441247003e-08,
"logits/chosen": -1.4547462463378906,
"logits/rejected": -1.4310463666915894,
"logps/chosen": -48.958641052246094,
"logps/rejected": -51.16803741455078,
"loss": 0.6933,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": -0.00026367080863565207,
"rewards/margins": -0.0002573660749476403,
"rewards/rejected": -6.304704584181309e-06,
"step": 110
},
{
"epoch": 0.08645533141210375,
"grad_norm": 2.3112876415252686,
"learning_rate": 1.4388489208633092e-08,
"logits/chosen": -1.4834020137786865,
"logits/rejected": -1.480193853378296,
"logps/chosen": -44.25951385498047,
"logps/rejected": -46.622459411621094,
"loss": 0.6931,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -6.177094473969191e-05,
"rewards/margins": 0.00016041506023611873,
"rewards/rejected": -0.00022218600497581065,
"step": 120
},
{
"epoch": 0.0936599423631124,
"grad_norm": 3.208045482635498,
"learning_rate": 1.5587529976019183e-08,
"logits/chosen": -1.5762228965759277,
"logits/rejected": -1.570908784866333,
"logps/chosen": -49.414695739746094,
"logps/rejected": -51.2851448059082,
"loss": 0.6933,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": -8.6869033111725e-05,
"rewards/margins": -0.0002423443365842104,
"rewards/rejected": 0.000155475310748443,
"step": 130
},
{
"epoch": 0.10086455331412104,
"grad_norm": 2.822239875793457,
"learning_rate": 1.6786570743405277e-08,
"logits/chosen": -1.4522497653961182,
"logits/rejected": -1.4427827596664429,
"logps/chosen": -45.83789825439453,
"logps/rejected": -50.24224090576172,
"loss": 0.6931,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 8.89547445694916e-05,
"rewards/margins": 4.058218473801389e-05,
"rewards/rejected": 4.837256346945651e-05,
"step": 140
},
{
"epoch": 0.10806916426512968,
"grad_norm": 3.587710380554199,
"learning_rate": 1.7985611510791365e-08,
"logits/chosen": -1.4713795185089111,
"logits/rejected": -1.4646179676055908,
"logps/chosen": -48.30494689941406,
"logps/rejected": -51.4416389465332,
"loss": 0.6931,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -4.7554804041283205e-05,
"rewards/margins": 4.275305400369689e-05,
"rewards/rejected": -9.030787623487413e-05,
"step": 150
},
{
"epoch": 0.11527377521613832,
"grad_norm": 2.4577479362487793,
"learning_rate": 1.9184652278177456e-08,
"logits/chosen": -1.5048809051513672,
"logits/rejected": -1.4864928722381592,
"logps/chosen": -41.26644515991211,
"logps/rejected": -44.58869171142578,
"loss": 0.6932,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.00011363021621946245,
"rewards/margins": -0.00019988985150121152,
"rewards/rejected": 8.625965710962191e-05,
"step": 160
},
{
"epoch": 0.12247838616714697,
"grad_norm": 3.1715707778930664,
"learning_rate": 2.038369304556355e-08,
"logits/chosen": -1.5167441368103027,
"logits/rejected": -1.4980051517486572,
"logps/chosen": -44.847110748291016,
"logps/rejected": -46.87194061279297,
"loss": 0.6931,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 6.15198805462569e-05,
"rewards/margins": 4.4902008085045964e-05,
"rewards/rejected": 1.661786882323213e-05,
"step": 170
},
{
"epoch": 0.12968299711815562,
"grad_norm": 2.6304078102111816,
"learning_rate": 2.1582733812949638e-08,
"logits/chosen": -1.5800201892852783,
"logits/rejected": -1.5682014226913452,
"logps/chosen": -45.10023880004883,
"logps/rejected": -46.83604049682617,
"loss": 0.6932,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": -8.165388862835243e-05,
"rewards/margins": -0.00017628518980927765,
"rewards/rejected": 9.46312939049676e-05,
"step": 180
},
{
"epoch": 0.13688760806916425,
"grad_norm": 2.9077770709991455,
"learning_rate": 2.278177458033573e-08,
"logits/chosen": -1.5893758535385132,
"logits/rejected": -1.5856701135635376,
"logps/chosen": -42.25619125366211,
"logps/rejected": -45.408897399902344,
"loss": 0.6931,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": 5.6146480346797034e-05,
"rewards/margins": 9.998455789173022e-05,
"rewards/rejected": -4.3838103010784835e-05,
"step": 190
},
{
"epoch": 0.1440922190201729,
"grad_norm": 3.549778699874878,
"learning_rate": 2.3980815347721823e-08,
"logits/chosen": -1.5360238552093506,
"logits/rejected": -1.5294934511184692,
"logps/chosen": -43.458560943603516,
"logps/rejected": -47.107208251953125,
"loss": 0.6931,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.00025255040964111686,
"rewards/margins": 0.00011713804269675165,
"rewards/rejected": 0.00013541239604819566,
"step": 200
},
{
"epoch": 0.15129682997118155,
"grad_norm": 3.0489232540130615,
"learning_rate": 2.517985611510791e-08,
"logits/chosen": -1.5652011632919312,
"logits/rejected": -1.55386221408844,
"logps/chosen": -43.03472137451172,
"logps/rejected": -43.407081604003906,
"loss": 0.6931,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": 2.3424910978064872e-05,
"rewards/margins": 8.2639220636338e-05,
"rewards/rejected": -5.9214326029177755e-05,
"step": 210
},
{
"epoch": 0.1585014409221902,
"grad_norm": 2.841609001159668,
"learning_rate": 2.6378896882494006e-08,
"logits/chosen": -1.480421781539917,
"logits/rejected": -1.4737306833267212,
"logps/chosen": -47.355350494384766,
"logps/rejected": -52.514892578125,
"loss": 0.6932,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": 1.830189285101369e-05,
"rewards/margins": -0.00019653179333545268,
"rewards/rejected": 0.00021483367891050875,
"step": 220
},
{
"epoch": 0.16570605187319884,
"grad_norm": 2.58459734916687,
"learning_rate": 2.7577937649880097e-08,
"logits/chosen": -1.531930685043335,
"logits/rejected": -1.530457615852356,
"logps/chosen": -44.46145248413086,
"logps/rejected": -48.232242584228516,
"loss": 0.6932,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 8.398554928135127e-05,
"rewards/margins": -8.254259591922164e-05,
"rewards/rejected": 0.00016652815975248814,
"step": 230
},
{
"epoch": 0.1729106628242075,
"grad_norm": 3.012089252471924,
"learning_rate": 2.8776978417266184e-08,
"logits/chosen": -1.579641342163086,
"logits/rejected": -1.5689789056777954,
"logps/chosen": -49.132789611816406,
"logps/rejected": -51.18694305419922,
"loss": 0.6932,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 9.1132904344704e-05,
"rewards/margins": -7.982960960362107e-05,
"rewards/rejected": 0.00017096252122428268,
"step": 240
},
{
"epoch": 0.18011527377521613,
"grad_norm": 4.0211029052734375,
"learning_rate": 2.997601918465228e-08,
"logits/chosen": -1.4552438259124756,
"logits/rejected": -1.4445832967758179,
"logps/chosen": -49.68731689453125,
"logps/rejected": -50.77144241333008,
"loss": 0.693,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.00020649081852752715,
"rewards/margins": 0.0003089832025580108,
"rewards/rejected": -0.00010249239858239889,
"step": 250
},
{
"epoch": 0.1873198847262248,
"grad_norm": 3.18790864944458,
"learning_rate": 3.1175059952038366e-08,
"logits/chosen": -1.4960715770721436,
"logits/rejected": -1.490350365638733,
"logps/chosen": -50.00872802734375,
"logps/rejected": -51.44108200073242,
"loss": 0.6932,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 4.101751983398572e-06,
"rewards/margins": -8.562284347135574e-05,
"rewards/rejected": 8.972459909273311e-05,
"step": 260
},
{
"epoch": 0.19452449567723343,
"grad_norm": 3.316842794418335,
"learning_rate": 3.237410071942446e-08,
"logits/chosen": -1.5841925144195557,
"logits/rejected": -1.5689890384674072,
"logps/chosen": -46.749019622802734,
"logps/rejected": -49.31301498413086,
"loss": 0.6931,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.00019280468404758722,
"rewards/margins": 8.136655378621072e-05,
"rewards/rejected": 0.00011143812298541889,
"step": 270
},
{
"epoch": 0.2017291066282421,
"grad_norm": 2.9199252128601074,
"learning_rate": 3.3573141486810555e-08,
"logits/chosen": -1.5495785474777222,
"logits/rejected": -1.5337769985198975,
"logps/chosen": -42.458438873291016,
"logps/rejected": -43.72154235839844,
"loss": 0.6931,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.00012140088074374944,
"rewards/margins": 0.00014567100151907653,
"rewards/rejected": -2.4270115318358876e-05,
"step": 280
},
{
"epoch": 0.20893371757925072,
"grad_norm": 2.7867319583892822,
"learning_rate": 3.477218225419664e-08,
"logits/chosen": -1.5798470973968506,
"logits/rejected": -1.5693657398223877,
"logps/chosen": -44.48418426513672,
"logps/rejected": -45.79877471923828,
"loss": 0.6929,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0002844139817170799,
"rewards/margins": 0.0004200335533823818,
"rewards/rejected": -0.0001356196153210476,
"step": 290
},
{
"epoch": 0.21613832853025935,
"grad_norm": 3.150078058242798,
"learning_rate": 3.597122302158273e-08,
"logits/chosen": -1.556154489517212,
"logits/rejected": -1.539389967918396,
"logps/chosen": -48.21941375732422,
"logps/rejected": -50.685508728027344,
"loss": 0.6933,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.0001608836610103026,
"rewards/margins": -0.0002904186840169132,
"rewards/rejected": 0.00012953505211044103,
"step": 300
},
{
"epoch": 0.22334293948126802,
"grad_norm": 3.015779972076416,
"learning_rate": 3.717026378896883e-08,
"logits/chosen": -1.4409435987472534,
"logits/rejected": -1.4314186573028564,
"logps/chosen": -48.024295806884766,
"logps/rejected": -50.0726432800293,
"loss": 0.693,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.00016624320414848626,
"rewards/margins": 0.000264329748461023,
"rewards/rejected": -9.808655886445194e-05,
"step": 310
},
{
"epoch": 0.23054755043227665,
"grad_norm": 2.547288417816162,
"learning_rate": 3.836930455635491e-08,
"logits/chosen": -1.553330659866333,
"logits/rejected": -1.5384843349456787,
"logps/chosen": -47.11131286621094,
"logps/rejected": -51.53942108154297,
"loss": 0.6933,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 3.6099205317441374e-05,
"rewards/margins": -0.00022887022350914776,
"rewards/rejected": 0.0002649694215506315,
"step": 320
},
{
"epoch": 0.2377521613832853,
"grad_norm": 2.309831380844116,
"learning_rate": 3.9568345323741003e-08,
"logits/chosen": -1.5248470306396484,
"logits/rejected": -1.5183141231536865,
"logps/chosen": -50.730499267578125,
"logps/rejected": -49.556068420410156,
"loss": 0.6931,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.00026169692864641547,
"rewards/margins": 0.0001796955766621977,
"rewards/rejected": 8.200139563996345e-05,
"step": 330
},
{
"epoch": 0.24495677233429394,
"grad_norm": 2.6688296794891357,
"learning_rate": 4.07673860911271e-08,
"logits/chosen": -1.5785892009735107,
"logits/rejected": -1.5679844617843628,
"logps/chosen": -51.07538604736328,
"logps/rejected": -52.128135681152344,
"loss": 0.6929,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.00021654777810908854,
"rewards/margins": 0.0004967943532392383,
"rewards/rejected": -0.00028024654602631927,
"step": 340
},
{
"epoch": 0.2521613832853026,
"grad_norm": 3.666226387023926,
"learning_rate": 4.1966426858513185e-08,
"logits/chosen": -1.5095466375350952,
"logits/rejected": -1.5047543048858643,
"logps/chosen": -45.6449089050293,
"logps/rejected": -48.59443664550781,
"loss": 0.693,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.00023484873236157,
"rewards/margins": 0.0001965599658433348,
"rewards/rejected": 3.82887119485531e-05,
"step": 350
},
{
"epoch": 0.25936599423631124,
"grad_norm": 3.478895664215088,
"learning_rate": 4.3165467625899276e-08,
"logits/chosen": -1.5015857219696045,
"logits/rejected": -1.4911534786224365,
"logps/chosen": -53.9161376953125,
"logps/rejected": -56.55745315551758,
"loss": 0.6931,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0001681277499301359,
"rewards/margins": 0.00018387913587503135,
"rewards/rejected": -1.575138230691664e-05,
"step": 360
},
{
"epoch": 0.2665706051873199,
"grad_norm": 3.750058650970459,
"learning_rate": 4.4364508393285374e-08,
"logits/chosen": -1.4821351766586304,
"logits/rejected": -1.480381727218628,
"logps/chosen": -48.33544158935547,
"logps/rejected": -53.026039123535156,
"loss": 0.6931,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 8.031638571992517e-05,
"rewards/margins": 0.00012994577991776168,
"rewards/rejected": -4.96293832839001e-05,
"step": 370
},
{
"epoch": 0.2737752161383285,
"grad_norm": 2.4797353744506836,
"learning_rate": 4.556354916067146e-08,
"logits/chosen": -1.5723177194595337,
"logits/rejected": -1.565124750137329,
"logps/chosen": -47.00983428955078,
"logps/rejected": -48.0451774597168,
"loss": 0.6932,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": -9.197367762681097e-05,
"rewards/margins": -0.00010503224621061236,
"rewards/rejected": 1.3058568583801389e-05,
"step": 380
},
{
"epoch": 0.28097982708933716,
"grad_norm": 2.919807195663452,
"learning_rate": 4.676258992805755e-08,
"logits/chosen": -1.544861078262329,
"logits/rejected": -1.5350227355957031,
"logps/chosen": -48.17707824707031,
"logps/rejected": -51.51404571533203,
"loss": 0.693,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.00019117488409392536,
"rewards/margins": 0.00030245125526562333,
"rewards/rejected": -0.0001112763857236132,
"step": 390
},
{
"epoch": 0.2881844380403458,
"grad_norm": 3.4181175231933594,
"learning_rate": 4.796163069544365e-08,
"logits/chosen": -1.5584533214569092,
"logits/rejected": -1.553442120552063,
"logps/chosen": -44.867557525634766,
"logps/rejected": -46.099464416503906,
"loss": 0.6931,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.0003581286873668432,
"rewards/margins": 5.371281804400496e-05,
"rewards/rejected": 0.0003044158511329442,
"step": 400
},
{
"epoch": 0.2953890489913545,
"grad_norm": 3.621582269668579,
"learning_rate": 4.916067146282973e-08,
"logits/chosen": -1.5090839862823486,
"logits/rejected": -1.5060720443725586,
"logps/chosen": -47.50264358520508,
"logps/rejected": -49.63032531738281,
"loss": 0.693,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.00016462437633890659,
"rewards/margins": 0.00023242369934450835,
"rewards/rejected": -6.779931572964415e-05,
"step": 410
},
{
"epoch": 0.3025936599423631,
"grad_norm": 2.4397823810577393,
"learning_rate": 4.999992091672379e-08,
"logits/chosen": -1.4693224430084229,
"logits/rejected": -1.4796779155731201,
"logps/chosen": -45.638816833496094,
"logps/rejected": -48.969810485839844,
"loss": 0.693,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.000272547040367499,
"rewards/margins": 0.0002472659107297659,
"rewards/rejected": 2.5281125999754295e-05,
"step": 420
},
{
"epoch": 0.30979827089337175,
"grad_norm": 2.2847747802734375,
"learning_rate": 4.999851500573209e-08,
"logits/chosen": -1.497286081314087,
"logits/rejected": -1.4975759983062744,
"logps/chosen": -46.069549560546875,
"logps/rejected": -46.195926666259766,
"loss": 0.6931,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.00028530642157420516,
"rewards/margins": 5.241289545665495e-05,
"rewards/rejected": 0.00023289353703148663,
"step": 430
},
{
"epoch": 0.3170028818443804,
"grad_norm": 2.459535598754883,
"learning_rate": 4.999535180235972e-08,
"logits/chosen": -1.4981845617294312,
"logits/rejected": -1.4901264905929565,
"logps/chosen": -46.00641632080078,
"logps/rejected": -49.42933654785156,
"loss": 0.6931,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.00040736678056418896,
"rewards/margins": 0.00017191791266668588,
"rewards/rejected": 0.00023544885334558785,
"step": 440
},
{
"epoch": 0.3242074927953891,
"grad_norm": 3.167510747909546,
"learning_rate": 4.9990431528966836e-08,
"logits/chosen": -1.5115673542022705,
"logits/rejected": -1.4903723001480103,
"logps/chosen": -53.19926071166992,
"logps/rejected": -51.2760124206543,
"loss": 0.6931,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.00023766886442899704,
"rewards/margins": 7.205204019555822e-05,
"rewards/rejected": 0.00016561683150939643,
"step": 450
},
{
"epoch": 0.3314121037463977,
"grad_norm": 3.773123264312744,
"learning_rate": 4.9983754531428326e-08,
"logits/chosen": -1.5162955522537231,
"logits/rejected": -1.4986896514892578,
"logps/chosen": -53.790428161621094,
"logps/rejected": -55.66992950439453,
"loss": 0.6929,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.0004971676971763372,
"rewards/margins": 0.0004848612588830292,
"rewards/rejected": 1.2306480130064301e-05,
"step": 460
},
{
"epoch": 0.33861671469740634,
"grad_norm": 3.7350404262542725,
"learning_rate": 4.997532127910954e-08,
"logits/chosen": -1.5784637928009033,
"logits/rejected": -1.5491013526916504,
"logps/chosen": -52.63544464111328,
"logps/rejected": -53.17797088623047,
"loss": 0.6931,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.00048227087245322764,
"rewards/margins": 0.00017049835878424346,
"rewards/rejected": 0.00031177254277281463,
"step": 470
},
{
"epoch": 0.345821325648415,
"grad_norm": 3.6733524799346924,
"learning_rate": 4.996513236483331e-08,
"logits/chosen": -1.6473217010498047,
"logits/rejected": -1.633281946182251,
"logps/chosen": -42.51477813720703,
"logps/rejected": -45.43607711791992,
"loss": 0.6929,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.00038208425394259393,
"rewards/margins": 0.000435996160376817,
"rewards/rejected": -5.39118773303926e-05,
"step": 480
},
{
"epoch": 0.3530259365994236,
"grad_norm": 4.256612777709961,
"learning_rate": 4.9953188504838225e-08,
"logits/chosen": -1.5243208408355713,
"logits/rejected": -1.5126068592071533,
"logps/chosen": -46.43769836425781,
"logps/rejected": -49.511756896972656,
"loss": 0.693,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.0005211577517911792,
"rewards/margins": 0.00022547971457242966,
"rewards/rejected": 0.0002956780372187495,
"step": 490
},
{
"epoch": 0.36023054755043227,
"grad_norm": 2.8351917266845703,
"learning_rate": 4.993949053872834e-08,
"logits/chosen": -1.5284124612808228,
"logits/rejected": -1.5051486492156982,
"logps/chosen": -42.646305084228516,
"logps/rejected": -45.89906311035156,
"loss": 0.6929,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0006723391707055271,
"rewards/margins": 0.0005468688905239105,
"rewards/rejected": 0.00012547028018161654,
"step": 500
},
{
"epoch": 0.36743515850144093,
"grad_norm": 2.8914971351623535,
"learning_rate": 4.9924039429414086e-08,
"logits/chosen": -1.6392762660980225,
"logits/rejected": -1.6216766834259033,
"logps/chosen": -46.00267791748047,
"logps/rejected": -47.987403869628906,
"loss": 0.6928,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0007155792554840446,
"rewards/margins": 0.0006390741327777505,
"rewards/rejected": 7.650510087842122e-05,
"step": 510
},
{
"epoch": 0.3746397694524496,
"grad_norm": 3.5407791137695312,
"learning_rate": 4.990683626304467e-08,
"logits/chosen": -1.534824013710022,
"logits/rejected": -1.5296251773834229,
"logps/chosen": -53.9051399230957,
"logps/rejected": -56.04487991333008,
"loss": 0.6929,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.000572909542825073,
"rewards/margins": 0.00040510791586712003,
"rewards/rejected": 0.0001678016851656139,
"step": 520
},
{
"epoch": 0.3818443804034582,
"grad_norm": 3.2243497371673584,
"learning_rate": 4.9887882248931646e-08,
"logits/chosen": -1.458703637123108,
"logits/rejected": -1.4379961490631104,
"logps/chosen": -46.455928802490234,
"logps/rejected": -47.554115295410156,
"loss": 0.6931,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.0005672247498296201,
"rewards/margins": 0.00012306116695981473,
"rewards/rejected": 0.00044416356831789017,
"step": 530
},
{
"epoch": 0.38904899135446686,
"grad_norm": 3.160867929458618,
"learning_rate": 4.986717871946393e-08,
"logits/chosen": -1.485264778137207,
"logits/rejected": -1.4642503261566162,
"logps/chosen": -45.888916015625,
"logps/rejected": -47.80854797363281,
"loss": 0.6929,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.0008015796774998307,
"rewards/margins": 0.0005912004271522164,
"rewards/rejected": 0.0002103792503476143,
"step": 540
},
{
"epoch": 0.3962536023054755,
"grad_norm": 3.078800916671753,
"learning_rate": 4.984472713001416e-08,
"logits/chosen": -1.4299088716506958,
"logits/rejected": -1.4215964078903198,
"logps/chosen": -48.37324905395508,
"logps/rejected": -48.35413360595703,
"loss": 0.693,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0007138814544305205,
"rewards/margins": 0.0003910651430487633,
"rewards/rejected": 0.0003228162822779268,
"step": 550
},
{
"epoch": 0.4034582132564842,
"grad_norm": 3.2446839809417725,
"learning_rate": 4.982052905883637e-08,
"logits/chosen": -1.5735007524490356,
"logits/rejected": -1.563230276107788,
"logps/chosen": -48.497169494628906,
"logps/rejected": -49.916038513183594,
"loss": 0.693,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.0005788643029518425,
"rewards/margins": 0.00021954579278826714,
"rewards/rejected": 0.0003593183937482536,
"step": 560
},
{
"epoch": 0.4106628242074928,
"grad_norm": 2.9157779216766357,
"learning_rate": 4.979458620695505e-08,
"logits/chosen": -1.5526814460754395,
"logits/rejected": -1.523316740989685,
"logps/chosen": -52.486785888671875,
"logps/rejected": -54.491912841796875,
"loss": 0.6928,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0009498519939370453,
"rewards/margins": 0.000704078353010118,
"rewards/rejected": 0.00024577361182309687,
"step": 570
},
{
"epoch": 0.41786743515850144,
"grad_norm": 3.174314022064209,
"learning_rate": 4.976690039804555e-08,
"logits/chosen": -1.5769068002700806,
"logits/rejected": -1.5633710622787476,
"logps/chosen": -42.617549896240234,
"logps/rejected": -44.070777893066406,
"loss": 0.6929,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.0007929207058623433,
"rewards/margins": 0.0005030486499890685,
"rewards/rejected": 0.0002898719976656139,
"step": 580
},
{
"epoch": 0.4250720461095101,
"grad_norm": 2.7369155883789062,
"learning_rate": 4.973747357830592e-08,
"logits/chosen": -1.5272729396820068,
"logits/rejected": -1.5258095264434814,
"logps/chosen": -47.52259063720703,
"logps/rejected": -53.1329345703125,
"loss": 0.6928,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.0009988851379603148,
"rewards/margins": 0.0007000741316005588,
"rewards/rejected": 0.00029881083173677325,
"step": 590
},
{
"epoch": 0.4322766570605187,
"grad_norm": 2.792400360107422,
"learning_rate": 4.970630781632009e-08,
"logits/chosen": -1.6299186944961548,
"logits/rejected": -1.6194143295288086,
"logps/chosen": -45.39707565307617,
"logps/rejected": -49.081199645996094,
"loss": 0.6927,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.001002423232421279,
"rewards/margins": 0.0009681530063971877,
"rewards/rejected": 3.427025876590051e-05,
"step": 600
},
{
"epoch": 0.43948126801152737,
"grad_norm": 3.9282000064849854,
"learning_rate": 4.967340530291242e-08,
"logits/chosen": -1.5342658758163452,
"logits/rejected": -1.5172860622406006,
"logps/chosen": -50.43634796142578,
"logps/rejected": -51.06001281738281,
"loss": 0.693,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.0008850853191688657,
"rewards/margins": 0.00033199749304912984,
"rewards/rejected": 0.0005530878552235663,
"step": 610
},
{
"epoch": 0.44668587896253603,
"grad_norm": 2.692594289779663,
"learning_rate": 4.9638768350993755e-08,
"logits/chosen": -1.5681965351104736,
"logits/rejected": -1.5538604259490967,
"logps/chosen": -42.41255569458008,
"logps/rejected": -44.433387756347656,
"loss": 0.6931,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.0007358186412602663,
"rewards/margins": 9.383581345900893e-05,
"rewards/rejected": 0.0006419828278012574,
"step": 620
},
{
"epoch": 0.4538904899135447,
"grad_norm": 2.3626010417938232,
"learning_rate": 4.9602399395398786e-08,
"logits/chosen": -1.571014642715454,
"logits/rejected": -1.5637868642807007,
"logps/chosen": -43.03638458251953,
"logps/rejected": -46.532352447509766,
"loss": 0.6928,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0012007402256131172,
"rewards/margins": 0.000781078590080142,
"rewards/rejected": 0.00041966172284446657,
"step": 630
},
{
"epoch": 0.4610951008645533,
"grad_norm": 2.919173002243042,
"learning_rate": 4.9564300992714914e-08,
"logits/chosen": -1.4291613101959229,
"logits/rejected": -1.4245188236236572,
"logps/chosen": -45.36858367919922,
"logps/rejected": -48.030757904052734,
"loss": 0.6926,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.001236187876202166,
"rewards/margins": 0.0011594726238399744,
"rewards/rejected": 7.671533967368305e-05,
"step": 640
},
{
"epoch": 0.46829971181556196,
"grad_norm": 3.464174747467041,
"learning_rate": 4.952447582110253e-08,
"logits/chosen": -1.6131465435028076,
"logits/rejected": -1.584380865097046,
"logps/chosen": -45.44336700439453,
"logps/rejected": -45.41891860961914,
"loss": 0.6927,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.0014582508010789752,
"rewards/margins": 0.0009963499614968896,
"rewards/rejected": 0.0004619006940629333,
"step": 650
},
{
"epoch": 0.4755043227665706,
"grad_norm": 3.4227089881896973,
"learning_rate": 4.948292668010676e-08,
"logits/chosen": -1.5425691604614258,
"logits/rejected": -1.5405323505401611,
"logps/chosen": -47.156578063964844,
"logps/rejected": -50.03830337524414,
"loss": 0.6926,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0010936754988506436,
"rewards/margins": 0.0011404131073504686,
"rewards/rejected": -4.6737539378227666e-05,
"step": 660
},
{
"epoch": 0.4827089337175792,
"grad_norm": 3.4312925338745117,
"learning_rate": 4.943965649046064e-08,
"logits/chosen": -1.5018900632858276,
"logits/rejected": -1.4744553565979004,
"logps/chosen": -49.83372116088867,
"logps/rejected": -51.13875198364258,
"loss": 0.693,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.0009614540031179786,
"rewards/margins": 0.0003561762277968228,
"rewards/rejected": 0.0006052777171134949,
"step": 670
},
{
"epoch": 0.4899135446685879,
"grad_norm": 4.607568740844727,
"learning_rate": 4.9394668293879835e-08,
"logits/chosen": -1.4445058107376099,
"logits/rejected": -1.431208848953247,
"logps/chosen": -49.71855926513672,
"logps/rejected": -49.596534729003906,
"loss": 0.6926,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0016611143946647644,
"rewards/margins": 0.001147769158706069,
"rewards/rejected": 0.0005133452359586954,
"step": 680
},
{
"epoch": 0.49711815561959655,
"grad_norm": 3.329970598220825,
"learning_rate": 4.93479652528488e-08,
"logits/chosen": -1.5312420129776,
"logits/rejected": -1.5205342769622803,
"logps/chosen": -47.82133102416992,
"logps/rejected": -50.619808197021484,
"loss": 0.6927,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.0012991353869438171,
"rewards/margins": 0.000877738930284977,
"rewards/rejected": 0.00042139639845117927,
"step": 690
},
{
"epoch": 0.5043227665706052,
"grad_norm": 2.7837576866149902,
"learning_rate": 4.929955065039848e-08,
"logits/chosen": -1.54481840133667,
"logits/rejected": -1.5314353704452515,
"logps/chosen": -46.46809005737305,
"logps/rejected": -49.25910949707031,
"loss": 0.6927,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.0013074753805994987,
"rewards/margins": 0.0009138353052549064,
"rewards/rejected": 0.00039364007534459233,
"step": 700
},
{
"epoch": 0.5115273775216138,
"grad_norm": 2.817760705947876,
"learning_rate": 4.92494278898755e-08,
"logits/chosen": -1.5252554416656494,
"logits/rejected": -1.509037733078003,
"logps/chosen": -41.35044860839844,
"logps/rejected": -43.40355682373047,
"loss": 0.6926,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0012825509766116738,
"rewards/margins": 0.000999027630314231,
"rewards/rejected": 0.00028352331719361246,
"step": 710
},
{
"epoch": 0.5187319884726225,
"grad_norm": 3.3338074684143066,
"learning_rate": 4.9197600494702955e-08,
"logits/chosen": -1.4963575601577759,
"logits/rejected": -1.4812816381454468,
"logps/chosen": -49.3184700012207,
"logps/rejected": -52.44392013549805,
"loss": 0.6928,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0011441984679549932,
"rewards/margins": 0.000659407174680382,
"rewards/rejected": 0.00048479135148227215,
"step": 720
},
{
"epoch": 0.5259365994236311,
"grad_norm": 2.8599021434783936,
"learning_rate": 4.9144072108132725e-08,
"logits/chosen": -1.5103414058685303,
"logits/rejected": -1.4909043312072754,
"logps/chosen": -48.92387771606445,
"logps/rejected": -51.066978454589844,
"loss": 0.6927,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.0015517466235905886,
"rewards/margins": 0.0009926140774041414,
"rewards/rejected": 0.0005591326043941081,
"step": 730
},
{
"epoch": 0.5331412103746398,
"grad_norm": 2.899857521057129,
"learning_rate": 4.908884649298937e-08,
"logits/chosen": -1.5037367343902588,
"logits/rejected": -1.5001060962677002,
"logps/chosen": -46.71796417236328,
"logps/rejected": -46.27198791503906,
"loss": 0.6929,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.0012326440773904324,
"rewards/margins": 0.0005613928660750389,
"rewards/rejected": 0.0006712513277307153,
"step": 740
},
{
"epoch": 0.5403458213256485,
"grad_norm": 2.8896918296813965,
"learning_rate": 4.903192753140557e-08,
"logits/chosen": -1.5269062519073486,
"logits/rejected": -1.510358214378357,
"logps/chosen": -48.91157150268555,
"logps/rejected": -50.0924072265625,
"loss": 0.6925,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.001482138060964644,
"rewards/margins": 0.001297360984608531,
"rewards/rejected": 0.0001847770472522825,
"step": 750
},
{
"epoch": 0.547550432276657,
"grad_norm": 3.3254764080047607,
"learning_rate": 4.897331922454931e-08,
"logits/chosen": -1.4530668258666992,
"logits/rejected": -1.451474666595459,
"logps/chosen": -45.5156135559082,
"logps/rejected": -48.615623474121094,
"loss": 0.6926,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.001580315874889493,
"rewards/margins": 0.001087161828763783,
"rewards/rejected": 0.0004931538132950664,
"step": 760
},
{
"epoch": 0.5547550432276657,
"grad_norm": 3.2673330307006836,
"learning_rate": 4.891302569234256e-08,
"logits/chosen": -1.47372305393219,
"logits/rejected": -1.4675675630569458,
"logps/chosen": -43.22402572631836,
"logps/rejected": -45.916168212890625,
"loss": 0.6923,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": 0.0018752291798591614,
"rewards/margins": 0.0017130204942077398,
"rewards/rejected": 0.00016220868565142155,
"step": 770
},
{
"epoch": 0.5619596541786743,
"grad_norm": 2.8122148513793945,
"learning_rate": 4.8851051173171656e-08,
"logits/chosen": -1.49901282787323,
"logits/rejected": -1.4895904064178467,
"logps/chosen": -48.40452575683594,
"logps/rejected": -50.17125701904297,
"loss": 0.6926,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0016933809965848923,
"rewards/margins": 0.0010203216224908829,
"rewards/rejected": 0.0006730594905093312,
"step": 780
},
{
"epoch": 0.569164265129683,
"grad_norm": 2.8876476287841797,
"learning_rate": 4.87874000235894e-08,
"logits/chosen": -1.5472230911254883,
"logits/rejected": -1.537274956703186,
"logps/chosen": -49.910091400146484,
"logps/rejected": -53.432830810546875,
"loss": 0.6926,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.001698293723165989,
"rewards/margins": 0.0011685159988701344,
"rewards/rejected": 0.00052977807354182,
"step": 790
},
{
"epoch": 0.5763688760806917,
"grad_norm": 3.3339130878448486,
"learning_rate": 4.872207671800876e-08,
"logits/chosen": -1.5249817371368408,
"logits/rejected": -1.5136332511901855,
"logps/chosen": -46.88755416870117,
"logps/rejected": -47.896827697753906,
"loss": 0.6925,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0017408020794391632,
"rewards/margins": 0.0012650018325075507,
"rewards/rejected": 0.00047580021782778203,
"step": 800
},
{
"epoch": 0.5835734870317003,
"grad_norm": 2.6908926963806152,
"learning_rate": 4.865508584838841e-08,
"logits/chosen": -1.5177758932113647,
"logits/rejected": -1.5210834741592407,
"logps/chosen": -44.743553161621094,
"logps/rejected": -47.87196731567383,
"loss": 0.6926,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.001408976735547185,
"rewards/margins": 0.0010177220683544874,
"rewards/rejected": 0.00039125472540035844,
"step": 810
},
{
"epoch": 0.590778097982709,
"grad_norm": 2.710764169692993,
"learning_rate": 4.858643212390985e-08,
"logits/chosen": -1.5524133443832397,
"logits/rejected": -1.5305286645889282,
"logps/chosen": -46.923805236816406,
"logps/rejected": -47.56487274169922,
"loss": 0.6923,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.0020391198340803385,
"rewards/margins": 0.001677272142842412,
"rewards/rejected": 0.00036184763303026557,
"step": 820
},
{
"epoch": 0.5979827089337176,
"grad_norm": 2.6352500915527344,
"learning_rate": 4.851612037064643e-08,
"logits/chosen": -1.5101532936096191,
"logits/rejected": -1.5033133029937744,
"logps/chosen": -41.79671859741211,
"logps/rejected": -44.642845153808594,
"loss": 0.6924,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0022669571917504072,
"rewards/margins": 0.0015860407147556543,
"rewards/rejected": 0.0006809166516177356,
"step": 830
},
{
"epoch": 0.6051873198847262,
"grad_norm": 2.259105682373047,
"learning_rate": 4.8444155531224065e-08,
"logits/chosen": -1.5194116830825806,
"logits/rejected": -1.5120502710342407,
"logps/chosen": -47.2054328918457,
"logps/rejected": -47.48673629760742,
"loss": 0.6923,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.0022509084083139896,
"rewards/margins": 0.0016614611959084868,
"rewards/rejected": 0.0005894473288208246,
"step": 840
},
{
"epoch": 0.6123919308357348,
"grad_norm": 3.8317296504974365,
"learning_rate": 4.8370542664473805e-08,
"logits/chosen": -1.5284236669540405,
"logits/rejected": -1.517364263534546,
"logps/chosen": -47.17852783203125,
"logps/rejected": -50.44862747192383,
"loss": 0.6922,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.002234196290373802,
"rewards/margins": 0.001899405149742961,
"rewards/rejected": 0.00033479099511168897,
"step": 850
},
{
"epoch": 0.6195965417867435,
"grad_norm": 2.7945921421051025,
"learning_rate": 4.829528694507624e-08,
"logits/chosen": -1.5346529483795166,
"logits/rejected": -1.5194079875946045,
"logps/chosen": -56.83978271484375,
"logps/rejected": -56.75861358642578,
"loss": 0.6922,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.0023796563036739826,
"rewards/margins": 0.0018080968875437975,
"rewards/rejected": 0.0005715594161301851,
"step": 860
},
{
"epoch": 0.6268011527377522,
"grad_norm": 3.107956647872925,
"learning_rate": 4.821839366319768e-08,
"logits/chosen": -1.5742651224136353,
"logits/rejected": -1.5632001161575317,
"logps/chosen": -47.60409927368164,
"logps/rejected": -50.62064743041992,
"loss": 0.6922,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.001870837644673884,
"rewards/margins": 0.0018480487633496523,
"rewards/rejected": 2.2788974092691205e-05,
"step": 870
},
{
"epoch": 0.6340057636887608,
"grad_norm": 3.052370309829712,
"learning_rate": 4.813986822411833e-08,
"logits/chosen": -1.5950231552124023,
"logits/rejected": -1.5870921611785889,
"logps/chosen": -46.46802520751953,
"logps/rejected": -47.61067581176758,
"loss": 0.6924,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0018212845316156745,
"rewards/margins": 0.0015648396220058203,
"rewards/rejected": 0.0002564448514021933,
"step": 880
},
{
"epoch": 0.6412103746397695,
"grad_norm": 2.974740505218506,
"learning_rate": 4.805971614785231e-08,
"logits/chosen": -1.5934207439422607,
"logits/rejected": -1.5831716060638428,
"logps/chosen": -44.20450973510742,
"logps/rejected": -45.84693145751953,
"loss": 0.6923,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.002194278407841921,
"rewards/margins": 0.0017778873443603516,
"rewards/rejected": 0.000416390917962417,
"step": 890
},
{
"epoch": 0.6484149855907781,
"grad_norm": 3.176281690597534,
"learning_rate": 4.797794306875963e-08,
"logits/chosen": -1.4426735639572144,
"logits/rejected": -1.4458198547363281,
"logps/chosen": -52.887428283691406,
"logps/rejected": -56.02144241333008,
"loss": 0.6924,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0018950074445456266,
"rewards/margins": 0.0015232457080855966,
"rewards/rejected": 0.0003717617364600301,
"step": 900
},
{
"epoch": 0.6556195965417867,
"grad_norm": 3.130652904510498,
"learning_rate": 4.7894554735150076e-08,
"logits/chosen": -1.4940052032470703,
"logits/rejected": -1.4864368438720703,
"logps/chosen": -50.41922378540039,
"logps/rejected": -51.92753982543945,
"loss": 0.6925,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.0017981244018301368,
"rewards/margins": 0.0012389495968818665,
"rewards/rejected": 0.0005591747467406094,
"step": 910
},
{
"epoch": 0.6628242074927954,
"grad_norm": 2.4986073970794678,
"learning_rate": 4.7809557008879185e-08,
"logits/chosen": -1.5263350009918213,
"logits/rejected": -1.5145517587661743,
"logps/chosen": -42.057559967041016,
"logps/rejected": -43.96914291381836,
"loss": 0.6921,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.0024824230931699276,
"rewards/margins": 0.0021964963525533676,
"rewards/rejected": 0.00028592668240889907,
"step": 920
},
{
"epoch": 0.670028818443804,
"grad_norm": 3.264617443084717,
"learning_rate": 4.772295586493613e-08,
"logits/chosen": -1.5924278497695923,
"logits/rejected": -1.5789930820465088,
"logps/chosen": -46.364784240722656,
"logps/rejected": -48.81365203857422,
"loss": 0.6922,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.0025308418553322554,
"rewards/margins": 0.001982019515708089,
"rewards/rejected": 0.0005488225724548101,
"step": 930
},
{
"epoch": 0.6772334293948127,
"grad_norm": 2.300140619277954,
"learning_rate": 4.763475739102374e-08,
"logits/chosen": -1.473141074180603,
"logits/rejected": -1.468505620956421,
"logps/chosen": -54.870941162109375,
"logps/rejected": -55.6484489440918,
"loss": 0.692,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.002576081547886133,
"rewards/margins": 0.0023079104721546173,
"rewards/rejected": 0.0002681712794583291,
"step": 940
},
{
"epoch": 0.6844380403458213,
"grad_norm": 2.9154794216156006,
"learning_rate": 4.754496778713054e-08,
"logits/chosen": -1.4293240308761597,
"logits/rejected": -1.4433194398880005,
"logps/chosen": -46.24222183227539,
"logps/rejected": -50.87912368774414,
"loss": 0.6924,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0029074286576360464,
"rewards/margins": 0.0015344502171501517,
"rewards/rejected": 0.0013729783240705729,
"step": 950
},
{
"epoch": 0.69164265129683,
"grad_norm": 2.978912830352783,
"learning_rate": 4.7453593365094926e-08,
"logits/chosen": -1.5649144649505615,
"logits/rejected": -1.5566378831863403,
"logps/chosen": -48.9169807434082,
"logps/rejected": -51.32488250732422,
"loss": 0.6923,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.0020612571388483047,
"rewards/margins": 0.0017103856662288308,
"rewards/rejected": 0.00035087167634628713,
"step": 960
},
{
"epoch": 0.6988472622478387,
"grad_norm": 3.66062068939209,
"learning_rate": 4.736064054816145e-08,
"logits/chosen": -1.579488754272461,
"logits/rejected": -1.5711945295333862,
"logps/chosen": -44.430999755859375,
"logps/rejected": -47.700523376464844,
"loss": 0.6919,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": 0.0031583986710757017,
"rewards/margins": 0.0025061373598873615,
"rewards/rejected": 0.0006522616604343057,
"step": 970
},
{
"epoch": 0.7060518731988472,
"grad_norm": 2.772946357727051,
"learning_rate": 4.726611587052933e-08,
"logits/chosen": -1.4304895401000977,
"logits/rejected": -1.4331814050674438,
"logps/chosen": -50.69477462768555,
"logps/rejected": -55.800628662109375,
"loss": 0.6924,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0024115366395562887,
"rewards/margins": 0.0015366144943982363,
"rewards/rejected": 0.0008749221451580524,
"step": 980
},
{
"epoch": 0.7132564841498559,
"grad_norm": 3.8805480003356934,
"learning_rate": 4.71700259768931e-08,
"logits/chosen": -1.5389564037322998,
"logits/rejected": -1.5328103303909302,
"logps/chosen": -50.487335205078125,
"logps/rejected": -51.9775390625,
"loss": 0.6923,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.002420094795525074,
"rewards/margins": 0.001630530459806323,
"rewards/rejected": 0.0007895643939264119,
"step": 990
},
{
"epoch": 0.7204610951008645,
"grad_norm": 2.787208080291748,
"learning_rate": 4.707237762197549e-08,
"logits/chosen": -1.5227080583572388,
"logits/rejected": -1.511089563369751,
"logps/chosen": -47.02164840698242,
"logps/rejected": -49.01847839355469,
"loss": 0.6922,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.0033734552562236786,
"rewards/margins": 0.0018900551367551088,
"rewards/rejected": 0.0014834003522992134,
"step": 1000
},
{
"epoch": 0.7276657060518732,
"grad_norm": 3.9273159503936768,
"learning_rate": 4.697317767005265e-08,
"logits/chosen": -1.530651330947876,
"logits/rejected": -1.519158959388733,
"logps/chosen": -43.01877212524414,
"logps/rejected": -44.823909759521484,
"loss": 0.692,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.0028608546126633883,
"rewards/margins": 0.002323360648006201,
"rewards/rejected": 0.0005374938482418656,
"step": 1010
},
{
"epoch": 0.7348703170028819,
"grad_norm": 2.8428092002868652,
"learning_rate": 4.6872433094471577e-08,
"logits/chosen": -1.5484195947647095,
"logits/rejected": -1.5358097553253174,
"logps/chosen": -46.577796936035156,
"logps/rejected": -48.426883697509766,
"loss": 0.6924,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.0024873907677829266,
"rewards/margins": 0.0014864472905173898,
"rewards/rejected": 0.0010009428951889277,
"step": 1020
},
{
"epoch": 0.7420749279538905,
"grad_norm": 2.6796295642852783,
"learning_rate": 4.677015097715994e-08,
"logits/chosen": -1.4803837537765503,
"logits/rejected": -1.4727258682250977,
"logps/chosen": -43.49967575073242,
"logps/rejected": -46.801334381103516,
"loss": 0.6917,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.003408711403608322,
"rewards/margins": 0.0028357082046568394,
"rewards/rejected": 0.0005730032571591437,
"step": 1030
},
{
"epoch": 0.7492795389048992,
"grad_norm": 2.420775890350342,
"learning_rate": 4.666633850812825e-08,
"logits/chosen": -1.523738145828247,
"logits/rejected": -1.507525086402893,
"logps/chosen": -46.27619552612305,
"logps/rejected": -48.205955505371094,
"loss": 0.6922,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.002326422370970249,
"rewards/margins": 0.0019570752047002316,
"rewards/rejected": 0.00036934722447767854,
"step": 1040
},
{
"epoch": 0.7564841498559077,
"grad_norm": 2.361326217651367,
"learning_rate": 4.656100298496439e-08,
"logits/chosen": -1.4339568614959717,
"logits/rejected": -1.4203391075134277,
"logps/chosen": -41.24075698852539,
"logps/rejected": -44.33374786376953,
"loss": 0.6916,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0036104682367295027,
"rewards/margins": 0.003143618581816554,
"rewards/rejected": 0.0004668496549129486,
"step": 1050
},
{
"epoch": 0.7636887608069164,
"grad_norm": 3.083638906478882,
"learning_rate": 4.6454151812320715e-08,
"logits/chosen": -1.5105210542678833,
"logits/rejected": -1.48483407497406,
"logps/chosen": -47.20132064819336,
"logps/rejected": -48.682518005371094,
"loss": 0.6916,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0033471949864178896,
"rewards/margins": 0.003038776572793722,
"rewards/rejected": 0.0003084186464548111,
"step": 1060
},
{
"epoch": 0.770893371757925,
"grad_norm": 3.846761465072632,
"learning_rate": 4.6345792501393434e-08,
"logits/chosen": -1.4992892742156982,
"logits/rejected": -1.4937262535095215,
"logps/chosen": -53.687469482421875,
"logps/rejected": -57.72792434692383,
"loss": 0.6916,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.003973551094532013,
"rewards/margins": 0.003104776842519641,
"rewards/rejected": 0.0008687739027664065,
"step": 1070
},
{
"epoch": 0.7780979827089337,
"grad_norm": 3.0516538619995117,
"learning_rate": 4.6235932669394676e-08,
"logits/chosen": -1.5073257684707642,
"logits/rejected": -1.4990612268447876,
"logps/chosen": -48.099449157714844,
"logps/rejected": -51.08936309814453,
"loss": 0.6916,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": 0.004019264131784439,
"rewards/margins": 0.003163648769259453,
"rewards/rejected": 0.0008556157117709517,
"step": 1080
},
{
"epoch": 0.7853025936599424,
"grad_norm": 3.5501997470855713,
"learning_rate": 4.612458003901698e-08,
"logits/chosen": -1.5299046039581299,
"logits/rejected": -1.5243635177612305,
"logps/chosen": -52.48954391479492,
"logps/rejected": -56.07282257080078,
"loss": 0.6914,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0027964431792497635,
"rewards/margins": 0.0034413684625178576,
"rewards/rejected": -0.0006449251086451113,
"step": 1090
},
{
"epoch": 0.792507204610951,
"grad_norm": 3.1133222579956055,
"learning_rate": 4.6011742437890476e-08,
"logits/chosen": -1.5386149883270264,
"logits/rejected": -1.515916347503662,
"logps/chosen": -47.325008392333984,
"logps/rejected": -48.78601837158203,
"loss": 0.692,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.002987902145832777,
"rewards/margins": 0.0024031363427639008,
"rewards/rejected": 0.0005847656866535544,
"step": 1100
},
{
"epoch": 0.7997118155619597,
"grad_norm": 2.2972168922424316,
"learning_rate": 4.589742779803259e-08,
"logits/chosen": -1.547487735748291,
"logits/rejected": -1.5349434614181519,
"logps/chosen": -46.49794387817383,
"logps/rejected": -48.665802001953125,
"loss": 0.6919,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.0025873552076518536,
"rewards/margins": 0.0025315138045698404,
"rewards/rejected": 5.5841472203610465e-05,
"step": 1110
},
{
"epoch": 0.8069164265129684,
"grad_norm": 2.634446620941162,
"learning_rate": 4.5781644155290486e-08,
"logits/chosen": -1.4874818325042725,
"logits/rejected": -1.4774929285049438,
"logps/chosen": -45.514652252197266,
"logps/rejected": -46.51914596557617,
"loss": 0.6914,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.004184176214039326,
"rewards/margins": 0.0034972827415913343,
"rewards/rejected": 0.000686893705278635,
"step": 1120
},
{
"epoch": 0.8141210374639769,
"grad_norm": 2.841136932373047,
"learning_rate": 4.566439964877613e-08,
"logits/chosen": -1.522335410118103,
"logits/rejected": -1.5161679983139038,
"logps/chosen": -43.42735290527344,
"logps/rejected": -45.222328186035156,
"loss": 0.6921,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.002366938628256321,
"rewards/margins": 0.0020665768533945084,
"rewards/rejected": 0.00030036186217330396,
"step": 1130
},
{
"epoch": 0.8213256484149856,
"grad_norm": 2.926987409591675,
"learning_rate": 4.554570252029421e-08,
"logits/chosen": -1.5708011388778687,
"logits/rejected": -1.5619392395019531,
"logps/chosen": -46.80385971069336,
"logps/rejected": -49.117591857910156,
"loss": 0.6911,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.004434407688677311,
"rewards/margins": 0.004192848689854145,
"rewards/rejected": 0.00024155872233677655,
"step": 1140
},
{
"epoch": 0.8285302593659942,
"grad_norm": 2.6681439876556396,
"learning_rate": 4.542556111376274e-08,
"logits/chosen": -1.5655062198638916,
"logits/rejected": -1.5541120767593384,
"logps/chosen": -48.73239517211914,
"logps/rejected": -50.90290069580078,
"loss": 0.6917,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.003030436811968684,
"rewards/margins": 0.0029130682814866304,
"rewards/rejected": 0.00011736871965695173,
"step": 1150
},
{
"epoch": 0.8357348703170029,
"grad_norm": 3.0899651050567627,
"learning_rate": 4.5303983874626506e-08,
"logits/chosen": -1.5412328243255615,
"logits/rejected": -1.5294244289398193,
"logps/chosen": -50.64223098754883,
"logps/rejected": -51.3464469909668,
"loss": 0.6919,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.002868660492822528,
"rewards/margins": 0.002512868959456682,
"rewards/rejected": 0.00035579182440415025,
"step": 1160
},
{
"epoch": 0.8429394812680115,
"grad_norm": 3.553450345993042,
"learning_rate": 4.518097934926339e-08,
"logits/chosen": -1.4599460363388062,
"logits/rejected": -1.4345753192901611,
"logps/chosen": -46.791717529296875,
"logps/rejected": -46.883766174316406,
"loss": 0.6916,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.003114109393209219,
"rewards/margins": 0.0031202633399516344,
"rewards/rejected": -6.154028142191237e-06,
"step": 1170
},
{
"epoch": 0.8501440922190202,
"grad_norm": 3.8637311458587646,
"learning_rate": 4.505655618438363e-08,
"logits/chosen": -1.4245116710662842,
"logits/rejected": -1.4106541872024536,
"logps/chosen": -48.68230438232422,
"logps/rejected": -49.66895294189453,
"loss": 0.6918,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.002959758508950472,
"rewards/margins": 0.0026471882592886686,
"rewards/rejected": 0.0003125699586234987,
"step": 1180
},
{
"epoch": 0.8573487031700289,
"grad_norm": 2.923635244369507,
"learning_rate": 4.4930723126421945e-08,
"logits/chosen": -1.5897529125213623,
"logits/rejected": -1.5666835308074951,
"logps/chosen": -49.047271728515625,
"logps/rejected": -50.365779876708984,
"loss": 0.6915,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.003261597827076912,
"rewards/margins": 0.0032364025246351957,
"rewards/rejected": 2.519541885703802e-05,
"step": 1190
},
{
"epoch": 0.8645533141210374,
"grad_norm": 3.259798049926758,
"learning_rate": 4.48034890209227e-08,
"logits/chosen": -1.4653351306915283,
"logits/rejected": -1.4455980062484741,
"logps/chosen": -51.658111572265625,
"logps/rejected": -53.60918426513672,
"loss": 0.6915,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.002871064469218254,
"rewards/margins": 0.0033232111018151045,
"rewards/rejected": -0.0004521465743891895,
"step": 1200
},
{
"epoch": 0.8717579250720461,
"grad_norm": 2.646080493927002,
"learning_rate": 4.4674862811918155e-08,
"logits/chosen": -1.4465787410736084,
"logits/rejected": -1.4438260793685913,
"logps/chosen": -43.387939453125,
"logps/rejected": -46.49829864501953,
"loss": 0.6917,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0034131291322410107,
"rewards/margins": 0.00300223333761096,
"rewards/rejected": 0.0004108964931219816,
"step": 1210
},
{
"epoch": 0.8789625360230547,
"grad_norm": 3.445535659790039,
"learning_rate": 4.454485354129966e-08,
"logits/chosen": -1.4952054023742676,
"logits/rejected": -1.489473819732666,
"logps/chosen": -46.59187698364258,
"logps/rejected": -50.117591857910156,
"loss": 0.6912,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0038944887928664684,
"rewards/margins": 0.0038705854676663876,
"rewards/rejected": 2.390358167758677e-05,
"step": 1220
},
{
"epoch": 0.8861671469740634,
"grad_norm": 2.979490041732788,
"learning_rate": 4.4413470348182124e-08,
"logits/chosen": -1.4463694095611572,
"logits/rejected": -1.4229583740234375,
"logps/chosen": -48.736202239990234,
"logps/rejected": -50.797584533691406,
"loss": 0.6912,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0038722821045666933,
"rewards/margins": 0.0038511925376951694,
"rewards/rejected": 2.1089776055305265e-05,
"step": 1230
},
{
"epoch": 0.8933717579250721,
"grad_norm": 3.6062870025634766,
"learning_rate": 4.42807224682615e-08,
"logits/chosen": -1.503015160560608,
"logits/rejected": -1.490422248840332,
"logps/chosen": -42.87591552734375,
"logps/rejected": -45.96302032470703,
"loss": 0.6909,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0038226053584367037,
"rewards/margins": 0.004490352235734463,
"rewards/rejected": -0.0006677471101284027,
"step": 1240
},
{
"epoch": 0.9005763688760807,
"grad_norm": 2.5306854248046875,
"learning_rate": 4.4146619233165604e-08,
"logits/chosen": -1.551155686378479,
"logits/rejected": -1.5466840267181396,
"logps/chosen": -50.65498733520508,
"logps/rejected": -54.01371383666992,
"loss": 0.6917,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.002975714858621359,
"rewards/margins": 0.0028749865014106035,
"rewards/rejected": 0.00010072816803585738,
"step": 1250
},
{
"epoch": 0.9077809798270894,
"grad_norm": 3.129319906234741,
"learning_rate": 4.4011170069798126e-08,
"logits/chosen": -1.5055732727050781,
"logits/rejected": -1.5217446088790894,
"logps/chosen": -46.49773025512695,
"logps/rejected": -53.42799758911133,
"loss": 0.6917,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0027830186299979687,
"rewards/margins": 0.00282860454171896,
"rewards/rejected": -4.558537693810649e-05,
"step": 1260
},
{
"epoch": 0.9149855907780979,
"grad_norm": 3.1904845237731934,
"learning_rate": 4.387438449967594e-08,
"logits/chosen": -1.454516053199768,
"logits/rejected": -1.4410918951034546,
"logps/chosen": -45.38161087036133,
"logps/rejected": -47.975074768066406,
"loss": 0.6907,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.004763299599289894,
"rewards/margins": 0.00504193315282464,
"rewards/rejected": -0.00027863297145813704,
"step": 1270
},
{
"epoch": 0.9221902017291066,
"grad_norm": 3.450941324234009,
"learning_rate": 4.373627213825983e-08,
"logits/chosen": -1.6091142892837524,
"logits/rejected": -1.5993342399597168,
"logps/chosen": -46.181339263916016,
"logps/rejected": -49.70648193359375,
"loss": 0.6904,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0049495515413582325,
"rewards/margins": 0.005530401133000851,
"rewards/rejected": -0.0005808495916426182,
"step": 1280
},
{
"epoch": 0.9293948126801153,
"grad_norm": 2.4806911945343018,
"learning_rate": 4.359684269427848e-08,
"logits/chosen": -1.5663807392120361,
"logits/rejected": -1.563518762588501,
"logps/chosen": -45.587608337402344,
"logps/rejected": -49.11951446533203,
"loss": 0.6911,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.004525287542492151,
"rewards/margins": 0.004190221894532442,
"rewards/rejected": 0.0003350655024405569,
"step": 1290
},
{
"epoch": 0.9365994236311239,
"grad_norm": 3.0177292823791504,
"learning_rate": 4.34561059690461e-08,
"logits/chosen": -1.609378457069397,
"logits/rejected": -1.6081126928329468,
"logps/chosen": -47.325008392333984,
"logps/rejected": -48.884395599365234,
"loss": 0.6925,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.0018209271365776658,
"rewards/margins": 0.0013891037087887526,
"rewards/rejected": 0.0004318233986850828,
"step": 1300
},
{
"epoch": 0.9438040345821326,
"grad_norm": 2.676215171813965,
"learning_rate": 4.3314071855773314e-08,
"logits/chosen": -1.5698845386505127,
"logits/rejected": -1.5718752145767212,
"logps/chosen": -41.882728576660156,
"logps/rejected": -45.07906723022461,
"loss": 0.6913,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.0035904725082218647,
"rewards/margins": 0.0036595568526536226,
"rewards/rejected": -6.908467912580818e-05,
"step": 1310
},
{
"epoch": 0.9510086455331412,
"grad_norm": 3.0929179191589355,
"learning_rate": 4.3170750338871806e-08,
"logits/chosen": -1.5062233209609985,
"logits/rejected": -1.491188406944275,
"logps/chosen": -46.517478942871094,
"logps/rejected": -49.685264587402344,
"loss": 0.6909,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.004791845567524433,
"rewards/margins": 0.004565626382827759,
"rewards/rejected": 0.00022621969401370734,
"step": 1320
},
{
"epoch": 0.9582132564841499,
"grad_norm": 3.0005862712860107,
"learning_rate": 4.3026151493252414e-08,
"logits/chosen": -1.5496270656585693,
"logits/rejected": -1.5283520221710205,
"logps/chosen": -51.4772834777832,
"logps/rejected": -52.89916229248047,
"loss": 0.6908,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.004191336687654257,
"rewards/margins": 0.0047586942091584206,
"rewards/rejected": -0.0005673574050888419,
"step": 1330
},
{
"epoch": 0.9654178674351584,
"grad_norm": 3.5196878910064697,
"learning_rate": 4.2880285483616895e-08,
"logits/chosen": -1.5336337089538574,
"logits/rejected": -1.530709981918335,
"logps/chosen": -45.714988708496094,
"logps/rejected": -49.015140533447266,
"loss": 0.6912,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.003844538237899542,
"rewards/margins": 0.0039407783187925816,
"rewards/rejected": -9.623957157600671e-05,
"step": 1340
},
{
"epoch": 0.9726224783861671,
"grad_norm": 2.7437596321105957,
"learning_rate": 4.273316256374342e-08,
"logits/chosen": -1.4035046100616455,
"logits/rejected": -1.398699164390564,
"logps/chosen": -52.25505447387695,
"logps/rejected": -53.14892578125,
"loss": 0.6912,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.004184540826827288,
"rewards/margins": 0.0038869075942784548,
"rewards/rejected": 0.0002976330870296806,
"step": 1350
},
{
"epoch": 0.9798270893371758,
"grad_norm": 3.3687148094177246,
"learning_rate": 4.258479307576576e-08,
"logits/chosen": -1.5007444620132446,
"logits/rejected": -1.4955756664276123,
"logps/chosen": -43.784767150878906,
"logps/rejected": -45.702003479003906,
"loss": 0.6903,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.005375358276069164,
"rewards/margins": 0.005684197880327702,
"rewards/rejected": -0.00030883969157002866,
"step": 1360
},
{
"epoch": 0.9870317002881844,
"grad_norm": 2.7493855953216553,
"learning_rate": 4.243518744944626e-08,
"logits/chosen": -1.5043120384216309,
"logits/rejected": -1.5005708932876587,
"logps/chosen": -43.27696228027344,
"logps/rejected": -47.130455017089844,
"loss": 0.6909,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.003993698861449957,
"rewards/margins": 0.004580583423376083,
"rewards/rejected": -0.0005868846783414483,
"step": 1370
},
{
"epoch": 0.9942363112391931,
"grad_norm": 3.398181915283203,
"learning_rate": 4.22843562014427e-08,
"logits/chosen": -1.449882984161377,
"logits/rejected": -1.4400885105133057,
"logps/chosen": -46.8693962097168,
"logps/rejected": -49.02799987792969,
"loss": 0.6918,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.004130179528146982,
"rewards/margins": 0.0028191660530865192,
"rewards/rejected": 0.001311013475060463,
"step": 1380
},
{
"epoch": 1.0014409221902016,
"grad_norm": 3.157179832458496,
"learning_rate": 4.2132309934569e-08,
"logits/chosen": -1.5672855377197266,
"logits/rejected": -1.562207579612732,
"logps/chosen": -43.753822326660156,
"logps/rejected": -46.12738800048828,
"loss": 0.6913,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0052392082288861275,
"rewards/margins": 0.0038300647865980864,
"rewards/rejected": 0.0014091433258727193,
"step": 1390
},
{
"epoch": 1.0086455331412103,
"grad_norm": 2.479654312133789,
"learning_rate": 4.197905933704989e-08,
"logits/chosen": -1.4313812255859375,
"logits/rejected": -1.4214115142822266,
"logps/chosen": -47.267311096191406,
"logps/rejected": -49.97428512573242,
"loss": 0.6896,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": 0.005089783109724522,
"rewards/margins": 0.007234425283968449,
"rewards/rejected": -0.002144642174243927,
"step": 1400
},
{
"epoch": 1.015850144092219,
"grad_norm": 2.7182703018188477,
"learning_rate": 4.1824615181769577e-08,
"logits/chosen": -1.486061453819275,
"logits/rejected": -1.492903232574463,
"logps/chosen": -43.82571792602539,
"logps/rejected": -47.75173568725586,
"loss": 0.6902,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.004842250142246485,
"rewards/margins": 0.005944275297224522,
"rewards/rejected": -0.0011020256206393242,
"step": 1410
},
{
"epoch": 1.0230547550432276,
"grad_norm": 3.1151888370513916,
"learning_rate": 4.1668988325514434e-08,
"logits/chosen": -1.5238953828811646,
"logits/rejected": -1.5135290622711182,
"logps/chosen": -49.40985107421875,
"logps/rejected": -52.1383171081543,
"loss": 0.6898,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.005019115749746561,
"rewards/margins": 0.006819140166044235,
"rewards/rejected": -0.0018000241834670305,
"step": 1420
},
{
"epoch": 1.0302593659942363,
"grad_norm": 3.011781930923462,
"learning_rate": 4.1512189708209844e-08,
"logits/chosen": -1.573925256729126,
"logits/rejected": -1.5644539594650269,
"logps/chosen": -38.25946807861328,
"logps/rejected": -39.431949615478516,
"loss": 0.6908,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.004985497798770666,
"rewards/margins": 0.0047707511112093925,
"rewards/rejected": 0.0002147464983863756,
"step": 1430
},
{
"epoch": 1.037463976945245,
"grad_norm": 3.615877866744995,
"learning_rate": 4.1354230352151143e-08,
"logits/chosen": -1.504152536392212,
"logits/rejected": -1.4912039041519165,
"logps/chosen": -56.4444694519043,
"logps/rejected": -56.67643356323242,
"loss": 0.6906,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.004153282381594181,
"rewards/margins": 0.005227471701800823,
"rewards/rejected": -0.0010741890873759985,
"step": 1440
},
{
"epoch": 1.0446685878962536,
"grad_norm": 2.5767173767089844,
"learning_rate": 4.119512136122882e-08,
"logits/chosen": -1.6086022853851318,
"logits/rejected": -1.6209228038787842,
"logps/chosen": -42.2785530090332,
"logps/rejected": -48.44477081298828,
"loss": 0.6897,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.004324558191001415,
"rewards/margins": 0.006945625878870487,
"rewards/rejected": -0.002621066989377141,
"step": 1450
},
{
"epoch": 1.0518731988472623,
"grad_norm": 3.475733757019043,
"learning_rate": 4.103487392014795e-08,
"logits/chosen": -1.475531816482544,
"logits/rejected": -1.4568547010421753,
"logps/chosen": -46.373329162597656,
"logps/rejected": -51.0313835144043,
"loss": 0.6888,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.005704123992472887,
"rewards/margins": 0.008804896846413612,
"rewards/rejected": -0.0031007737852633,
"step": 1460
},
{
"epoch": 1.059077809798271,
"grad_norm": 2.9147305488586426,
"learning_rate": 4.087349929364192e-08,
"logits/chosen": -1.5652912855148315,
"logits/rejected": -1.544355869293213,
"logps/chosen": -42.559722900390625,
"logps/rejected": -45.870784759521484,
"loss": 0.6897,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.004617183469235897,
"rewards/margins": 0.0069281430914998055,
"rewards/rejected": -0.002310959156602621,
"step": 1470
},
{
"epoch": 1.0662824207492796,
"grad_norm": 2.4540820121765137,
"learning_rate": 4.0711008825680645e-08,
"logits/chosen": -1.5043457746505737,
"logits/rejected": -1.4851245880126953,
"logps/chosen": -47.31243896484375,
"logps/rejected": -50.1568489074707,
"loss": 0.6905,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.004572049714624882,
"rewards/margins": 0.005410096608102322,
"rewards/rejected": -0.0008380465442314744,
"step": 1480
},
{
"epoch": 1.0734870317002883,
"grad_norm": 3.5471794605255127,
"learning_rate": 4.054741393867306e-08,
"logits/chosen": -1.4756460189819336,
"logits/rejected": -1.4643748998641968,
"logps/chosen": -54.061676025390625,
"logps/rejected": -55.685035705566406,
"loss": 0.6903,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.004487854428589344,
"rewards/margins": 0.005812914576381445,
"rewards/rejected": -0.0013250606134533882,
"step": 1490
},
{
"epoch": 1.080691642651297,
"grad_norm": 2.9531631469726562,
"learning_rate": 4.038272613266419e-08,
"logits/chosen": -1.5454641580581665,
"logits/rejected": -1.5215752124786377,
"logps/chosen": -44.894840240478516,
"logps/rejected": -47.37709045410156,
"loss": 0.6903,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": 0.004894919227808714,
"rewards/margins": 0.005796266719698906,
"rewards/rejected": -0.000901346851605922,
"step": 1500
},
{
"epoch": 1.0878962536023056,
"grad_norm": 3.38706111907959,
"learning_rate": 4.0216956984526784e-08,
"logits/chosen": -1.5542356967926025,
"logits/rejected": -1.550040364265442,
"logps/chosen": -42.899436950683594,
"logps/rejected": -45.57080078125,
"loss": 0.6897,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.005091445986181498,
"rewards/margins": 0.0070404186844825745,
"rewards/rejected": -0.0019489733967930079,
"step": 1510
},
{
"epoch": 1.0951008645533142,
"grad_norm": 3.101982831954956,
"learning_rate": 4.0050118147147446e-08,
"logits/chosen": -1.5179659128189087,
"logits/rejected": -1.509966254234314,
"logps/chosen": -53.40665817260742,
"logps/rejected": -52.11634063720703,
"loss": 0.6921,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.0031676869839429855,
"rewards/margins": 0.0020833946764469147,
"rewards/rejected": 0.0010842925403267145,
"step": 1520
},
{
"epoch": 1.1023054755043227,
"grad_norm": 3.091132640838623,
"learning_rate": 3.988222134860755e-08,
"logits/chosen": -1.563659906387329,
"logits/rejected": -1.5506489276885986,
"logps/chosen": -47.28032684326172,
"logps/rejected": -51.66279983520508,
"loss": 0.6899,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.005120830610394478,
"rewards/margins": 0.0066344826482236385,
"rewards/rejected": -0.0015136522706598043,
"step": 1530
},
{
"epoch": 1.1095100864553313,
"grad_norm": 3.1105594635009766,
"learning_rate": 3.9713278391358724e-08,
"logits/chosen": -1.5746492147445679,
"logits/rejected": -1.5628042221069336,
"logps/chosen": -45.93760681152344,
"logps/rejected": -49.18937301635742,
"loss": 0.6903,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.005176537670195103,
"rewards/margins": 0.005757451057434082,
"rewards/rejected": -0.0005809130962006748,
"step": 1540
},
{
"epoch": 1.11671469740634,
"grad_norm": 2.4670963287353516,
"learning_rate": 3.954330115139328e-08,
"logits/chosen": -1.5431309938430786,
"logits/rejected": -1.5327889919281006,
"logps/chosen": -46.76984405517578,
"logps/rejected": -48.848941802978516,
"loss": 0.6899,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.004580738488584757,
"rewards/margins": 0.006636142730712891,
"rewards/rejected": -0.0020554042421281338,
"step": 1550
},
{
"epoch": 1.1239193083573487,
"grad_norm": 4.051297187805176,
"learning_rate": 3.937230157740931e-08,
"logits/chosen": -1.5918155908584595,
"logits/rejected": -1.5730479955673218,
"logps/chosen": -47.90183639526367,
"logps/rejected": -51.474891662597656,
"loss": 0.6892,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.006330148316919804,
"rewards/margins": 0.007905492559075356,
"rewards/rejected": -0.0015753433108329773,
"step": 1560
},
{
"epoch": 1.1311239193083573,
"grad_norm": 2.364239454269409,
"learning_rate": 3.920029168997077e-08,
"logits/chosen": -1.5559653043746948,
"logits/rejected": -1.5427926778793335,
"logps/chosen": -48.723899841308594,
"logps/rejected": -51.545448303222656,
"loss": 0.6903,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.004865005612373352,
"rewards/margins": 0.005895170383155346,
"rewards/rejected": -0.001030164072290063,
"step": 1570
},
{
"epoch": 1.138328530259366,
"grad_norm": 3.9678614139556885,
"learning_rate": 3.9027283580662476e-08,
"logits/chosen": -1.5198521614074707,
"logits/rejected": -1.5075480937957764,
"logps/chosen": -49.6241340637207,
"logps/rejected": -52.78132247924805,
"loss": 0.6888,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.005017532967031002,
"rewards/margins": 0.008875850588083267,
"rewards/rejected": -0.003858317853882909,
"step": 1580
},
{
"epoch": 1.1455331412103746,
"grad_norm": 3.9845945835113525,
"learning_rate": 3.885328941124014e-08,
"logits/chosen": -1.5015017986297607,
"logits/rejected": -1.488577127456665,
"logps/chosen": -45.922607421875,
"logps/rejected": -50.63169860839844,
"loss": 0.6896,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.005352140404284,
"rewards/margins": 0.007262455765157938,
"rewards/rejected": -0.0019103146623820066,
"step": 1590
},
{
"epoch": 1.1527377521613833,
"grad_norm": 3.0348169803619385,
"learning_rate": 3.867832141277539e-08,
"logits/chosen": -1.5485866069793701,
"logits/rejected": -1.5293524265289307,
"logps/chosen": -49.10395812988281,
"logps/rejected": -51.272361755371094,
"loss": 0.69,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.004336564801633358,
"rewards/margins": 0.006302339024841785,
"rewards/rejected": -0.0019657742232084274,
"step": 1600
},
{
"epoch": 1.159942363112392,
"grad_norm": 3.4307823181152344,
"learning_rate": 3.850239188479606e-08,
"logits/chosen": -1.4621508121490479,
"logits/rejected": -1.4584615230560303,
"logps/chosen": -46.74929428100586,
"logps/rejected": -49.1221809387207,
"loss": 0.69,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.004755138885229826,
"rewards/margins": 0.006490757223218679,
"rewards/rejected": -0.0017356185708194971,
"step": 1610
},
{
"epoch": 1.1671469740634006,
"grad_norm": 3.5623254776000977,
"learning_rate": 3.832551319442151e-08,
"logits/chosen": -1.5857195854187012,
"logits/rejected": -1.5824106931686401,
"logps/chosen": -49.597537994384766,
"logps/rejected": -53.8158073425293,
"loss": 0.6899,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.00521459337323904,
"rewards/margins": 0.006661839783191681,
"rewards/rejected": -0.0014472461771219969,
"step": 1620
},
{
"epoch": 1.1743515850144093,
"grad_norm": 4.247158527374268,
"learning_rate": 3.81476977754933e-08,
"logits/chosen": -1.3996719121932983,
"logits/rejected": -1.3850539922714233,
"logps/chosen": -51.30238723754883,
"logps/rejected": -50.659263610839844,
"loss": 0.6901,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.004295485559850931,
"rewards/margins": 0.006129544228315353,
"rewards/rejected": -0.0018340591341257095,
"step": 1630
},
{
"epoch": 1.181556195965418,
"grad_norm": 2.404221773147583,
"learning_rate": 3.796895812770114e-08,
"logits/chosen": -1.5022691488265991,
"logits/rejected": -1.4923193454742432,
"logps/chosen": -45.77785110473633,
"logps/rejected": -47.2498664855957,
"loss": 0.6894,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.005898059345781803,
"rewards/margins": 0.007563650608062744,
"rewards/rejected": -0.00166559056378901,
"step": 1640
},
{
"epoch": 1.1887608069164266,
"grad_norm": 3.0766351222991943,
"learning_rate": 3.7789306815704216e-08,
"logits/chosen": -1.5271815061569214,
"logits/rejected": -1.5167516469955444,
"logps/chosen": -40.93082046508789,
"logps/rejected": -42.0340461730957,
"loss": 0.6908,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.0035404022783041,
"rewards/margins": 0.004842911381274462,
"rewards/rejected": -0.0013025096850469708,
"step": 1650
},
{
"epoch": 1.195965417867435,
"grad_norm": 2.643650531768799,
"learning_rate": 3.760875646824795e-08,
"logits/chosen": -1.395875334739685,
"logits/rejected": -1.3972784280776978,
"logps/chosen": -46.130271911621094,
"logps/rejected": -48.43110656738281,
"loss": 0.6896,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.003639611881226301,
"rewards/margins": 0.007139836438000202,
"rewards/rejected": -0.00350022385828197,
"step": 1660
},
{
"epoch": 1.2031700288184437,
"grad_norm": 3.6421399116516113,
"learning_rate": 3.742731977727623e-08,
"logits/chosen": -1.5392658710479736,
"logits/rejected": -1.5320463180541992,
"logps/chosen": -45.195159912109375,
"logps/rejected": -49.122718811035156,
"loss": 0.6894,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.006233454681932926,
"rewards/margins": 0.007511253468692303,
"rewards/rejected": -0.0012777980882674456,
"step": 1670
},
{
"epoch": 1.2103746397694524,
"grad_norm": 3.6778624057769775,
"learning_rate": 3.7245009497039244e-08,
"logits/chosen": -1.4356926679611206,
"logits/rejected": -1.4204113483428955,
"logps/chosen": -45.431312561035156,
"logps/rejected": -49.4974479675293,
"loss": 0.6892,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0038647849578410387,
"rewards/margins": 0.007974617183208466,
"rewards/rejected": -0.0041098324581980705,
"step": 1680
},
{
"epoch": 1.217579250720461,
"grad_norm": 2.6512179374694824,
"learning_rate": 3.7061838443196886e-08,
"logits/chosen": -1.5117601156234741,
"logits/rejected": -1.5023605823516846,
"logps/chosen": -50.03330612182617,
"logps/rejected": -52.160560607910156,
"loss": 0.6885,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.006138836033642292,
"rewards/margins": 0.009467920288443565,
"rewards/rejected": -0.003329083789139986,
"step": 1690
},
{
"epoch": 1.2247838616714697,
"grad_norm": 3.0043179988861084,
"learning_rate": 3.68778194919179e-08,
"logits/chosen": -1.4723100662231445,
"logits/rejected": -1.4666262865066528,
"logps/chosen": -50.07857894897461,
"logps/rejected": -53.3220100402832,
"loss": 0.6878,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.008099446073174477,
"rewards/margins": 0.010907831601798534,
"rewards/rejected": -0.0028083862271159887,
"step": 1700
},
{
"epoch": 1.2319884726224783,
"grad_norm": 3.6292684078216553,
"learning_rate": 3.66929655789747e-08,
"logits/chosen": -1.571274995803833,
"logits/rejected": -1.5532617568969727,
"logps/chosen": -41.95909118652344,
"logps/rejected": -46.51947784423828,
"loss": 0.6889,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": 0.005592281464487314,
"rewards/margins": 0.008521712385118008,
"rewards/rejected": -0.002929429989308119,
"step": 1710
},
{
"epoch": 1.239193083573487,
"grad_norm": 2.337404727935791,
"learning_rate": 3.6507289698834064e-08,
"logits/chosen": -1.471616506576538,
"logits/rejected": -1.4550542831420898,
"logps/chosen": -43.53514862060547,
"logps/rejected": -46.06405258178711,
"loss": 0.689,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.005335145629942417,
"rewards/margins": 0.008476309478282928,
"rewards/rejected": -0.0031411624513566494,
"step": 1720
},
{
"epoch": 1.2463976945244957,
"grad_norm": 4.029110431671143,
"learning_rate": 3.6320804903743684e-08,
"logits/chosen": -1.5162017345428467,
"logits/rejected": -1.5111761093139648,
"logps/chosen": -45.39423370361328,
"logps/rejected": -49.17229080200195,
"loss": 0.6887,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.0036242641508579254,
"rewards/margins": 0.009098870679736137,
"rewards/rejected": -0.005474607460200787,
"step": 1730
},
{
"epoch": 1.2536023054755043,
"grad_norm": 2.6486778259277344,
"learning_rate": 3.61335243028146e-08,
"logits/chosen": -1.495854377746582,
"logits/rejected": -1.4898358583450317,
"logps/chosen": -48.8822135925293,
"logps/rejected": -51.64506912231445,
"loss": 0.689,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.004486840683966875,
"rewards/margins": 0.00841403380036354,
"rewards/rejected": -0.003927193582057953,
"step": 1740
},
{
"epoch": 1.260806916426513,
"grad_norm": 3.2701268196105957,
"learning_rate": 3.5945461061099736e-08,
"logits/chosen": -1.4389398097991943,
"logits/rejected": -1.4077644348144531,
"logps/chosen": -50.76543426513672,
"logps/rejected": -49.68498992919922,
"loss": 0.6871,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.006598909851163626,
"rewards/margins": 0.012235610745847225,
"rewards/rejected": -0.0056367008946835995,
"step": 1750
},
{
"epoch": 1.2680115273775217,
"grad_norm": 2.968621253967285,
"learning_rate": 3.5756628398668446e-08,
"logits/chosen": -1.5589392185211182,
"logits/rejected": -1.5580534934997559,
"logps/chosen": -51.238853454589844,
"logps/rejected": -53.6939697265625,
"loss": 0.6885,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.003833180759102106,
"rewards/margins": 0.009422466158866882,
"rewards/rejected": -0.005589285399764776,
"step": 1760
},
{
"epoch": 1.2752161383285303,
"grad_norm": 2.652916193008423,
"learning_rate": 3.556703958967716e-08,
"logits/chosen": -1.557422399520874,
"logits/rejected": -1.543720006942749,
"logps/chosen": -44.34943771362305,
"logps/rejected": -47.97236251831055,
"loss": 0.6897,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.0036465474404394627,
"rewards/margins": 0.007039936725050211,
"rewards/rejected": -0.003393388818949461,
"step": 1770
},
{
"epoch": 1.282420749279539,
"grad_norm": 4.006654739379883,
"learning_rate": 3.5376707961436297e-08,
"logits/chosen": -1.5315957069396973,
"logits/rejected": -1.5152517557144165,
"logps/chosen": -53.33189010620117,
"logps/rejected": -53.46404266357422,
"loss": 0.6905,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.004476548638194799,
"rewards/margins": 0.005307585000991821,
"rewards/rejected": -0.0008310364792123437,
"step": 1780
},
{
"epoch": 1.2896253602305476,
"grad_norm": 2.423352003097534,
"learning_rate": 3.51856468934734e-08,
"logits/chosen": -1.4921244382858276,
"logits/rejected": -1.4953742027282715,
"logps/chosen": -46.34716033935547,
"logps/rejected": -48.64600372314453,
"loss": 0.6912,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.004633346572518349,
"rewards/margins": 0.00402105925604701,
"rewards/rejected": 0.0006122869672253728,
"step": 1790
},
{
"epoch": 1.2968299711815563,
"grad_norm": 3.358710289001465,
"learning_rate": 3.499386981659262e-08,
"logits/chosen": -1.5787866115570068,
"logits/rejected": -1.570294737815857,
"logps/chosen": -45.49131774902344,
"logps/rejected": -51.6963996887207,
"loss": 0.6891,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.006135477218776941,
"rewards/margins": 0.008345494046807289,
"rewards/rejected": -0.002210016595199704,
"step": 1800
},
{
"epoch": 1.304034582132565,
"grad_norm": 2.5438549518585205,
"learning_rate": 3.480139021193057e-08,
"logits/chosen": -1.4628260135650635,
"logits/rejected": -1.4626216888427734,
"logps/chosen": -46.477264404296875,
"logps/rejected": -49.938323974609375,
"loss": 0.6895,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.004061608109623194,
"rewards/margins": 0.007341681979596615,
"rewards/rejected": -0.003280073869973421,
"step": 1810
},
{
"epoch": 1.3112391930835736,
"grad_norm": 4.097665786743164,
"learning_rate": 3.4608221610008666e-08,
"logits/chosen": -1.554359793663025,
"logits/rejected": -1.5441957712173462,
"logps/chosen": -40.67582321166992,
"logps/rejected": -45.34405517578125,
"loss": 0.6876,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.005710783414542675,
"rewards/margins": 0.011185348965227604,
"rewards/rejected": -0.005474564619362354,
"step": 1820
},
{
"epoch": 1.318443804034582,
"grad_norm": 2.3061835765838623,
"learning_rate": 3.4414377589782e-08,
"logits/chosen": -1.489579439163208,
"logits/rejected": -1.4888752698898315,
"logps/chosen": -44.27809143066406,
"logps/rejected": -46.678611755371094,
"loss": 0.689,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.003133392659947276,
"rewards/margins": 0.008380794897675514,
"rewards/rejected": -0.0052474020048975945,
"step": 1830
},
{
"epoch": 1.3256484149855907,
"grad_norm": 2.2423462867736816,
"learning_rate": 3.4219871777684745e-08,
"logits/chosen": -1.5047807693481445,
"logits/rejected": -1.4803636074066162,
"logps/chosen": -48.244483947753906,
"logps/rejected": -49.67208480834961,
"loss": 0.6888,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.004430696833878756,
"rewards/margins": 0.008876333944499493,
"rewards/rejected": -0.004445637576282024,
"step": 1840
},
{
"epoch": 1.3328530259365994,
"grad_norm": 3.133010149002075,
"learning_rate": 3.4024717846672364e-08,
"logits/chosen": -1.55430006980896,
"logits/rejected": -1.5407884120941162,
"logps/chosen": -43.84967803955078,
"logps/rejected": -47.088844299316406,
"loss": 0.6885,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.003224353538826108,
"rewards/margins": 0.009443853981792927,
"rewards/rejected": -0.0062195006757974625,
"step": 1850
},
{
"epoch": 1.340057636887608,
"grad_norm": 3.2332677841186523,
"learning_rate": 3.382892951526036e-08,
"logits/chosen": -1.5088350772857666,
"logits/rejected": -1.4985536336898804,
"logps/chosen": -48.5511589050293,
"logps/rejected": -53.52519607543945,
"loss": 0.688,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.004791603423655033,
"rewards/margins": 0.01051848940551281,
"rewards/rejected": -0.005726885981857777,
"step": 1860
},
{
"epoch": 1.3472622478386167,
"grad_norm": 3.0702571868896484,
"learning_rate": 3.3632520546559974e-08,
"logits/chosen": -1.477571964263916,
"logits/rejected": -1.4503300189971924,
"logps/chosen": -42.1833610534668,
"logps/rejected": -46.2810173034668,
"loss": 0.6883,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.005192961078137159,
"rewards/margins": 0.0098468828946352,
"rewards/rejected": -0.004653920419514179,
"step": 1870
},
{
"epoch": 1.3544668587896254,
"grad_norm": 3.413985013961792,
"learning_rate": 3.34355047473107e-08,
"logits/chosen": -1.509441614151001,
"logits/rejected": -1.4929606914520264,
"logps/chosen": -49.14087677001953,
"logps/rejected": -50.32634353637695,
"loss": 0.6892,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.003523029386997223,
"rewards/margins": 0.00808628834784031,
"rewards/rejected": -0.004563258029520512,
"step": 1880
},
{
"epoch": 1.361671469740634,
"grad_norm": 3.261613607406616,
"learning_rate": 3.323789596690971e-08,
"logits/chosen": -1.4438692331314087,
"logits/rejected": -1.4406040906906128,
"logps/chosen": -46.01984405517578,
"logps/rejected": -50.35694122314453,
"loss": 0.6885,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.004411002155393362,
"rewards/margins": 0.009367231279611588,
"rewards/rejected": -0.0049562300555408,
"step": 1890
},
{
"epoch": 1.3688760806916427,
"grad_norm": 2.0379562377929688,
"learning_rate": 3.303970809643828e-08,
"logits/chosen": -1.5255157947540283,
"logits/rejected": -1.5278053283691406,
"logps/chosen": -45.327415466308594,
"logps/rejected": -49.04973220825195,
"loss": 0.689,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.005483666434884071,
"rewards/margins": 0.008489950560033321,
"rewards/rejected": -0.0030062845908105373,
"step": 1900
},
{
"epoch": 1.3760806916426513,
"grad_norm": 2.9973742961883545,
"learning_rate": 3.2840955067685356e-08,
"logits/chosen": -1.563256859779358,
"logits/rejected": -1.5630435943603516,
"logps/chosen": -46.008609771728516,
"logps/rejected": -50.46318054199219,
"loss": 0.6875,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": 0.0052226148545742035,
"rewards/margins": 0.011490847915410995,
"rewards/rejected": -0.006268233992159367,
"step": 1910
},
{
"epoch": 1.38328530259366,
"grad_norm": 2.7613437175750732,
"learning_rate": 3.264165085216817e-08,
"logits/chosen": -1.580041766166687,
"logits/rejected": -1.5727280378341675,
"logps/chosen": -38.53578186035156,
"logps/rejected": -43.83596420288086,
"loss": 0.6887,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0044464850798249245,
"rewards/margins": 0.009077567607164383,
"rewards/rejected": -0.004631082061678171,
"step": 1920
},
{
"epoch": 1.3904899135446687,
"grad_norm": 4.180810451507568,
"learning_rate": 3.244180946015008e-08,
"logits/chosen": -1.4442589282989502,
"logits/rejected": -1.435718297958374,
"logps/chosen": -52.15314483642578,
"logps/rejected": -53.91064453125,
"loss": 0.6896,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.004610007628798485,
"rewards/margins": 0.007277486380189657,
"rewards/rejected": -0.0026674787513911724,
"step": 1930
},
{
"epoch": 1.397694524495677,
"grad_norm": 2.520510196685791,
"learning_rate": 3.224144493965578e-08,
"logits/chosen": -1.5796912908554077,
"logits/rejected": -1.5778270959854126,
"logps/chosen": -43.635459899902344,
"logps/rejected": -45.74983215332031,
"loss": 0.6891,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.003754237201064825,
"rewards/margins": 0.008241266012191772,
"rewards/rejected": -0.00448702834546566,
"step": 1940
},
{
"epoch": 1.4048991354466858,
"grad_norm": 2.8245351314544678,
"learning_rate": 3.204057137548371e-08,
"logits/chosen": -1.5312103033065796,
"logits/rejected": -1.5232504606246948,
"logps/chosen": -43.69184112548828,
"logps/rejected": -47.32173156738281,
"loss": 0.6876,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": 0.004152680281549692,
"rewards/margins": 0.011306528933346272,
"rewards/rejected": -0.007153847720474005,
"step": 1950
},
{
"epoch": 1.4121037463976944,
"grad_norm": 3.71762752532959,
"learning_rate": 3.183920288821597e-08,
"logits/chosen": -1.4898974895477295,
"logits/rejected": -1.4814989566802979,
"logps/chosen": -45.30360412597656,
"logps/rejected": -49.977867126464844,
"loss": 0.6872,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.004725744016468525,
"rewards/margins": 0.012062130495905876,
"rewards/rejected": -0.0073363869450986385,
"step": 1960
},
{
"epoch": 1.419308357348703,
"grad_norm": 3.813885450363159,
"learning_rate": 3.1637353633225735e-08,
"logits/chosen": -1.5400512218475342,
"logits/rejected": -1.5293104648590088,
"logps/chosen": -41.25631332397461,
"logps/rejected": -45.72794723510742,
"loss": 0.6869,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.003288590582087636,
"rewards/margins": 0.012727511115372181,
"rewards/rejected": -0.009438920766115189,
"step": 1970
},
{
"epoch": 1.4265129682997117,
"grad_norm": 3.278331995010376,
"learning_rate": 3.143503779968213e-08,
"logits/chosen": -1.50669264793396,
"logits/rejected": -1.506869912147522,
"logps/chosen": -45.4344367980957,
"logps/rejected": -49.771766662597656,
"loss": 0.689,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.001246375497430563,
"rewards/margins": 0.00855537410825491,
"rewards/rejected": -0.0073089986108243465,
"step": 1980
},
{
"epoch": 1.4337175792507204,
"grad_norm": 3.326122760772705,
"learning_rate": 3.1232269609552875e-08,
"logits/chosen": -1.5182191133499146,
"logits/rejected": -1.5079318284988403,
"logps/chosen": -43.68280792236328,
"logps/rejected": -46.17193603515625,
"loss": 0.6887,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.004092085175216198,
"rewards/margins": 0.009019319899380207,
"rewards/rejected": -0.004927235189825296,
"step": 1990
},
{
"epoch": 1.440922190201729,
"grad_norm": 2.173727035522461,
"learning_rate": 3.102906331660444e-08,
"logits/chosen": -1.5566580295562744,
"logits/rejected": -1.542860507965088,
"logps/chosen": -41.94898223876953,
"logps/rejected": -48.24005126953125,
"loss": 0.6868,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": 0.005175874102860689,
"rewards/margins": 0.012963724322617054,
"rewards/rejected": -0.007787850685417652,
"step": 2000
},
{
"epoch": 1.4481268011527377,
"grad_norm": 3.172349214553833,
"learning_rate": 3.082543320540015e-08,
"logits/chosen": -1.4700515270233154,
"logits/rejected": -1.4550468921661377,
"logps/chosen": -43.87895965576172,
"logps/rejected": -47.5235710144043,
"loss": 0.6883,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0023733568377792835,
"rewards/margins": 0.009878008626401424,
"rewards/rejected": -0.007504651788622141,
"step": 2010
},
{
"epoch": 1.4553314121037464,
"grad_norm": 4.150721549987793,
"learning_rate": 3.062139359029599e-08,
"logits/chosen": -1.557370901107788,
"logits/rejected": -1.553252100944519,
"logps/chosen": -46.466400146484375,
"logps/rejected": -48.92326736450195,
"loss": 0.6881,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.003435027552768588,
"rewards/margins": 0.010258909314870834,
"rewards/rejected": -0.006823881529271603,
"step": 2020
},
{
"epoch": 1.462536023054755,
"grad_norm": 3.4097900390625,
"learning_rate": 3.041695881443437e-08,
"logits/chosen": -1.5758157968521118,
"logits/rejected": -1.5673719644546509,
"logps/chosen": -46.34037399291992,
"logps/rejected": -50.32112121582031,
"loss": 0.69,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.0035826407838612795,
"rewards/margins": 0.0064225392416119576,
"rewards/rejected": -0.0028398979920893908,
"step": 2030
},
{
"epoch": 1.4697406340057637,
"grad_norm": 4.047444820404053,
"learning_rate": 3.0212143248735886e-08,
"logits/chosen": -1.531328797340393,
"logits/rejected": -1.5289762020111084,
"logps/chosen": -49.777374267578125,
"logps/rejected": -54.45976638793945,
"loss": 0.6876,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0041795228607952595,
"rewards/margins": 0.01155170239508152,
"rewards/rejected": -0.007372179068624973,
"step": 2040
},
{
"epoch": 1.4769452449567724,
"grad_norm": 3.1559255123138428,
"learning_rate": 3.0006961290889077e-08,
"logits/chosen": -1.5212230682373047,
"logits/rejected": -1.4961992502212524,
"logps/chosen": -50.666038513183594,
"logps/rejected": -53.179840087890625,
"loss": 0.6875,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.005002585239708424,
"rewards/margins": 0.011495384387671947,
"rewards/rejected": -0.006492799613624811,
"step": 2050
},
{
"epoch": 1.484149855907781,
"grad_norm": 2.672394037246704,
"learning_rate": 2.980142736433833e-08,
"logits/chosen": -1.5468004941940308,
"logits/rejected": -1.522716760635376,
"logps/chosen": -44.2849235534668,
"logps/rejected": -44.554542541503906,
"loss": 0.6883,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.0016432057600468397,
"rewards/margins": 0.009843757376074791,
"rewards/rejected": -0.008200552314519882,
"step": 2060
},
{
"epoch": 1.4913544668587897,
"grad_norm": 4.110817909240723,
"learning_rate": 2.9595555917269997e-08,
"logits/chosen": -1.556030035018921,
"logits/rejected": -1.5290006399154663,
"logps/chosen": -51.58045196533203,
"logps/rejected": -53.32768630981445,
"loss": 0.6883,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.0018507679924368858,
"rewards/margins": 0.00991340633481741,
"rewards/rejected": -0.008062639273703098,
"step": 2070
},
{
"epoch": 1.4985590778097984,
"grad_norm": 3.1407599449157715,
"learning_rate": 2.9389361421596725e-08,
"logits/chosen": -1.4347405433654785,
"logits/rejected": -1.4316017627716064,
"logps/chosen": -49.19403839111328,
"logps/rejected": -53.545204162597656,
"loss": 0.6871,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0042905667796730995,
"rewards/margins": 0.012316063046455383,
"rewards/rejected": -0.00802549533545971,
"step": 2080
},
{
"epoch": 1.505763688760807,
"grad_norm": 2.570767879486084,
"learning_rate": 2.9182858371940126e-08,
"logits/chosen": -1.5318742990493774,
"logits/rejected": -1.5175909996032715,
"logps/chosen": -42.606483459472656,
"logps/rejected": -46.17628860473633,
"loss": 0.6866,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.0029234064277261496,
"rewards/margins": 0.013527562841773033,
"rewards/rejected": -0.01060415618121624,
"step": 2090
},
{
"epoch": 1.5129682997118157,
"grad_norm": 3.642561674118042,
"learning_rate": 2.8976061284611908e-08,
"logits/chosen": -1.4698221683502197,
"logits/rejected": -1.4809458255767822,
"logps/chosen": -41.60303497314453,
"logps/rejected": -45.31582260131836,
"loss": 0.6876,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0049819787964224815,
"rewards/margins": 0.011425621807575226,
"rewards/rejected": -0.006443643011152744,
"step": 2100
},
{
"epoch": 1.5201729106628243,
"grad_norm": 3.2771670818328857,
"learning_rate": 2.8768984696593384e-08,
"logits/chosen": -1.4798085689544678,
"logits/rejected": -1.4632951021194458,
"logps/chosen": -44.77037811279297,
"logps/rejected": -47.931278228759766,
"loss": 0.6868,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.005007242318242788,
"rewards/margins": 0.013035112991929054,
"rewards/rejected": -0.008027870208024979,
"step": 2110
},
{
"epoch": 1.527377521613833,
"grad_norm": 3.2423009872436523,
"learning_rate": 2.8561643164513637e-08,
"logits/chosen": -1.334378957748413,
"logits/rejected": -1.318719744682312,
"logps/chosen": -51.95469284057617,
"logps/rejected": -54.27800369262695,
"loss": 0.6888,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.003995453007519245,
"rewards/margins": 0.008951379917562008,
"rewards/rejected": -0.00495592737570405,
"step": 2120
},
{
"epoch": 1.5345821325648417,
"grad_norm": 3.3206255435943604,
"learning_rate": 2.8354051263626227e-08,
"logits/chosen": -1.46604323387146,
"logits/rejected": -1.466449499130249,
"logps/chosen": -50.30503463745117,
"logps/rejected": -52.849761962890625,
"loss": 0.6885,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0029683273751288652,
"rewards/margins": 0.00942517165094614,
"rewards/rejected": -0.006456844508647919,
"step": 2130
},
{
"epoch": 1.54178674351585,
"grad_norm": 4.676600456237793,
"learning_rate": 2.8146223586784573e-08,
"logits/chosen": -1.4574966430664062,
"logits/rejected": -1.4443973302841187,
"logps/chosen": -52.07725143432617,
"logps/rejected": -55.014495849609375,
"loss": 0.687,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.003784618806093931,
"rewards/margins": 0.012685844674706459,
"rewards/rejected": -0.00890122540295124,
"step": 2140
},
{
"epoch": 1.5489913544668588,
"grad_norm": 3.3971612453460693,
"learning_rate": 2.7938174743416205e-08,
"logits/chosen": -1.3624401092529297,
"logits/rejected": -1.3555347919464111,
"logps/chosen": -51.51061248779297,
"logps/rejected": -55.449851989746094,
"loss": 0.6877,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.0027002261485904455,
"rewards/margins": 0.0111739132553339,
"rewards/rejected": -0.008473685942590237,
"step": 2150
},
{
"epoch": 1.5561959654178674,
"grad_norm": 3.0555315017700195,
"learning_rate": 2.7729919358495728e-08,
"logits/chosen": -1.5037503242492676,
"logits/rejected": -1.4941872358322144,
"logps/chosen": -52.352867126464844,
"logps/rejected": -53.544029235839844,
"loss": 0.687,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.0034458369482308626,
"rewards/margins": 0.012528976425528526,
"rewards/rejected": -0.009083138778805733,
"step": 2160
},
{
"epoch": 1.563400576368876,
"grad_norm": 3.8352720737457275,
"learning_rate": 2.7521472071516772e-08,
"logits/chosen": -1.4726879596710205,
"logits/rejected": -1.4662566184997559,
"logps/chosen": -43.68360900878906,
"logps/rejected": -47.45799255371094,
"loss": 0.6883,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.005658530630171299,
"rewards/margins": 0.0100140031427145,
"rewards/rejected": -0.004355472978204489,
"step": 2170
},
{
"epoch": 1.5706051873198847,
"grad_norm": 3.8918683528900146,
"learning_rate": 2.731284753546289e-08,
"logits/chosen": -1.480924367904663,
"logits/rejected": -1.4743003845214844,
"logps/chosen": -53.0228157043457,
"logps/rejected": -56.795387268066406,
"loss": 0.6894,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.0002322282234672457,
"rewards/margins": 0.007790303323417902,
"rewards/rejected": -0.008022531867027283,
"step": 2180
},
{
"epoch": 1.5778097982708934,
"grad_norm": 4.045823097229004,
"learning_rate": 2.710406041577751e-08,
"logits/chosen": -1.5512325763702393,
"logits/rejected": -1.5479477643966675,
"logps/chosen": -47.99006271362305,
"logps/rejected": -53.79365921020508,
"loss": 0.6881,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.00434616394340992,
"rewards/margins": 0.010301290079951286,
"rewards/rejected": -0.005955126136541367,
"step": 2190
},
{
"epoch": 1.585014409221902,
"grad_norm": 3.306387424468994,
"learning_rate": 2.6895125389333017e-08,
"logits/chosen": -1.5372812747955322,
"logits/rejected": -1.5224703550338745,
"logps/chosen": -48.444435119628906,
"logps/rejected": -52.62398147583008,
"loss": 0.6851,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.007179601583629847,
"rewards/margins": 0.016431204974651337,
"rewards/rejected": -0.009251603856682777,
"step": 2200
},
{
"epoch": 1.5922190201729105,
"grad_norm": 3.1701338291168213,
"learning_rate": 2.6686057143399028e-08,
"logits/chosen": -1.5057289600372314,
"logits/rejected": -1.4977985620498657,
"logps/chosen": -48.5191764831543,
"logps/rejected": -50.016353607177734,
"loss": 0.6882,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.004845013376325369,
"rewards/margins": 0.010277556255459785,
"rewards/rejected": -0.005432543810456991,
"step": 2210
},
{
"epoch": 1.5994236311239192,
"grad_norm": 3.6945786476135254,
"learning_rate": 2.647687037460996e-08,
"logits/chosen": -1.4846514463424683,
"logits/rejected": -1.4771755933761597,
"logps/chosen": -52.859474182128906,
"logps/rejected": -58.45153045654297,
"loss": 0.6862,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.006427395157516003,
"rewards/margins": 0.01427508145570755,
"rewards/rejected": -0.007847686298191547,
"step": 2220
},
{
"epoch": 1.6066282420749278,
"grad_norm": 3.2228472232818604,
"learning_rate": 2.626757978793187e-08,
"logits/chosen": -1.50605046749115,
"logits/rejected": -1.4987943172454834,
"logps/chosen": -48.89452362060547,
"logps/rejected": -52.482139587402344,
"loss": 0.6889,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.0004428673128131777,
"rewards/margins": 0.008724686689674854,
"rewards/rejected": -0.009167554788291454,
"step": 2230
},
{
"epoch": 1.6138328530259365,
"grad_norm": 2.9732460975646973,
"learning_rate": 2.6058200095628797e-08,
"logits/chosen": -1.5061275959014893,
"logits/rejected": -1.5063308477401733,
"logps/chosen": -40.85835266113281,
"logps/rejected": -46.76417541503906,
"loss": 0.6849,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.00457608001306653,
"rewards/margins": 0.016967391595244408,
"rewards/rejected": -0.012391313910484314,
"step": 2240
},
{
"epoch": 1.6210374639769451,
"grad_norm": 3.2725095748901367,
"learning_rate": 2.584874601622854e-08,
"logits/chosen": -1.5664986371994019,
"logits/rejected": -1.5493196249008179,
"logps/chosen": -49.408843994140625,
"logps/rejected": -53.2461051940918,
"loss": 0.689,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.002165078418329358,
"rewards/margins": 0.00868967454880476,
"rewards/rejected": -0.00652459729462862,
"step": 2250
},
{
"epoch": 1.6282420749279538,
"grad_norm": 3.0531673431396484,
"learning_rate": 2.5639232273487993e-08,
"logits/chosen": -1.4604244232177734,
"logits/rejected": -1.4406094551086426,
"logps/chosen": -44.316959381103516,
"logps/rejected": -47.68523025512695,
"loss": 0.6875,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.004358319565653801,
"rewards/margins": 0.01151906605809927,
"rewards/rejected": -0.007160746958106756,
"step": 2260
},
{
"epoch": 1.6354466858789625,
"grad_norm": 3.623836040496826,
"learning_rate": 2.5429673595358142e-08,
"logits/chosen": -1.5242135524749756,
"logits/rejected": -1.5093494653701782,
"logps/chosen": -45.795074462890625,
"logps/rejected": -48.57317352294922,
"loss": 0.6877,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.001714804908260703,
"rewards/margins": 0.0111106988042593,
"rewards/rejected": -0.009395892731845379,
"step": 2270
},
{
"epoch": 1.6426512968299711,
"grad_norm": 3.297855854034424,
"learning_rate": 2.5220084712948764e-08,
"logits/chosen": -1.4578922986984253,
"logits/rejected": -1.445966362953186,
"logps/chosen": -52.103614807128906,
"logps/rejected": -55.21147918701172,
"loss": 0.6898,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0028083932120352983,
"rewards/margins": 0.006879979278892279,
"rewards/rejected": -0.004071586299687624,
"step": 2280
},
{
"epoch": 1.6498559077809798,
"grad_norm": 3.762510061264038,
"learning_rate": 2.5010480359492838e-08,
"logits/chosen": -1.463905692100525,
"logits/rejected": -1.4516466856002808,
"logps/chosen": -49.459896087646484,
"logps/rejected": -49.4564094543457,
"loss": 0.686,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.0037110510747879744,
"rewards/margins": 0.014671266078948975,
"rewards/rejected": -0.010960215702652931,
"step": 2290
},
{
"epoch": 1.6570605187319885,
"grad_norm": 2.9843027591705322,
"learning_rate": 2.480087526931091e-08,
"logits/chosen": -1.5036156177520752,
"logits/rejected": -1.4841887950897217,
"logps/chosen": -43.37828826904297,
"logps/rejected": -45.4171142578125,
"loss": 0.6862,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.003712053643539548,
"rewards/margins": 0.01433002669364214,
"rewards/rejected": -0.010617973282933235,
"step": 2300
},
{
"epoch": 1.6642651296829971,
"grad_norm": 3.441652297973633,
"learning_rate": 2.4591284176775326e-08,
"logits/chosen": -1.4469178915023804,
"logits/rejected": -1.4350392818450928,
"logps/chosen": -55.11457061767578,
"logps/rejected": -56.50977325439453,
"loss": 0.6891,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.003367173718288541,
"rewards/margins": 0.008350704796612263,
"rewards/rejected": -0.004983530845493078,
"step": 2310
},
{
"epoch": 1.6714697406340058,
"grad_norm": 2.8864128589630127,
"learning_rate": 2.4381721815274443e-08,
"logits/chosen": -1.5199902057647705,
"logits/rejected": -1.5132654905319214,
"logps/chosen": -43.241661071777344,
"logps/rejected": -46.4239616394043,
"loss": 0.687,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0014256946742534637,
"rewards/margins": 0.012737279757857323,
"rewards/rejected": -0.011311585083603859,
"step": 2320
},
{
"epoch": 1.6786743515850144,
"grad_norm": 3.1540913581848145,
"learning_rate": 2.4172202916176936e-08,
"logits/chosen": -1.5634369850158691,
"logits/rejected": -1.5546420812606812,
"logps/chosen": -43.037837982177734,
"logps/rejected": -47.856117248535156,
"loss": 0.686,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.0006109640235081315,
"rewards/margins": 0.014766094274818897,
"rewards/rejected": -0.0141551299020648,
"step": 2330
},
{
"epoch": 1.685878962536023,
"grad_norm": 3.647859573364258,
"learning_rate": 2.3962742207796268e-08,
"logits/chosen": -1.4479600191116333,
"logits/rejected": -1.4379609823226929,
"logps/chosen": -41.66529083251953,
"logps/rejected": -45.63011932373047,
"loss": 0.685,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.004694860894232988,
"rewards/margins": 0.016697410494089127,
"rewards/rejected": -0.012002546340227127,
"step": 2340
},
{
"epoch": 1.6930835734870318,
"grad_norm": 3.6262919902801514,
"learning_rate": 2.3753354414355334e-08,
"logits/chosen": -1.4229185581207275,
"logits/rejected": -1.400246024131775,
"logps/chosen": -53.5576171875,
"logps/rejected": -55.23441696166992,
"loss": 0.6872,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0002725389786064625,
"rewards/margins": 0.012355529703199863,
"rewards/rejected": -0.012082991190254688,
"step": 2350
},
{
"epoch": 1.7002881844380404,
"grad_norm": 3.3609025478363037,
"learning_rate": 2.3544054254951408e-08,
"logits/chosen": -1.4647839069366455,
"logits/rejected": -1.4447661638259888,
"logps/chosen": -42.85536193847656,
"logps/rejected": -48.42116165161133,
"loss": 0.6842,
"rewards/accuracies": 0.6812499761581421,
"rewards/chosen": 0.00428318977355957,
"rewards/margins": 0.018426140770316124,
"rewards/rejected": -0.014142952859401703,
"step": 2360
},
{
"epoch": 1.707492795389049,
"grad_norm": 3.3678860664367676,
"learning_rate": 2.3334856442521435e-08,
"logits/chosen": -1.5585176944732666,
"logits/rejected": -1.5401690006256104,
"logps/chosen": -51.256141662597656,
"logps/rejected": -51.39955520629883,
"loss": 0.688,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0024927393533289433,
"rewards/margins": 0.010601336136460304,
"rewards/rejected": -0.008108596317470074,
"step": 2370
},
{
"epoch": 1.7146974063400577,
"grad_norm": 3.430454730987549,
"learning_rate": 2.3125775682807826e-08,
"logits/chosen": -1.5538517236709595,
"logits/rejected": -1.552392601966858,
"logps/chosen": -49.91531753540039,
"logps/rejected": -53.7569465637207,
"loss": 0.686,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.0030004787258803844,
"rewards/margins": 0.01481213141232729,
"rewards/rejected": -0.011811653152108192,
"step": 2380
},
{
"epoch": 1.7219020172910664,
"grad_norm": 2.7333407402038574,
"learning_rate": 2.291682667332464e-08,
"logits/chosen": -1.612623929977417,
"logits/rejected": -1.5989640951156616,
"logps/chosen": -46.485713958740234,
"logps/rejected": -49.609825134277344,
"loss": 0.6888,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.000499680289067328,
"rewards/margins": 0.008998895063996315,
"rewards/rejected": -0.008499214425683022,
"step": 2390
},
{
"epoch": 1.729106628242075,
"grad_norm": 2.9348888397216797,
"learning_rate": 2.2708024102324454e-08,
"logits/chosen": -1.5304839611053467,
"logits/rejected": -1.5251142978668213,
"logps/chosen": -46.70942306518555,
"logps/rejected": -51.74241256713867,
"loss": 0.6851,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.00332896551117301,
"rewards/margins": 0.016565581783652306,
"rewards/rejected": -0.013236616738140583,
"step": 2400
},
{
"epoch": 1.7363112391930837,
"grad_norm": 3.6855247020721436,
"learning_rate": 2.2499382647765797e-08,
"logits/chosen": -1.4964635372161865,
"logits/rejected": -1.4977600574493408,
"logps/chosen": -48.41642379760742,
"logps/rejected": -52.0125617980957,
"loss": 0.6874,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 3.586681486922316e-05,
"rewards/margins": 0.011704354546964169,
"rewards/rejected": -0.011668487451970577,
"step": 2410
},
{
"epoch": 1.7435158501440924,
"grad_norm": 2.8829994201660156,
"learning_rate": 2.2290916976281427e-08,
"logits/chosen": -1.4776675701141357,
"logits/rejected": -1.4631723165512085,
"logps/chosen": -43.69971466064453,
"logps/rejected": -46.09686279296875,
"loss": 0.6861,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.00013883263454772532,
"rewards/margins": 0.014544370584189892,
"rewards/rejected": -0.014683201909065247,
"step": 2420
},
{
"epoch": 1.7507204610951008,
"grad_norm": 3.6001625061035156,
"learning_rate": 2.2082641742147238e-08,
"logits/chosen": -1.4721145629882812,
"logits/rejected": -1.4632716178894043,
"logps/chosen": -45.69694519042969,
"logps/rejected": -51.60881805419922,
"loss": 0.6866,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.000785698473919183,
"rewards/margins": 0.013660689815878868,
"rewards/rejected": -0.012874990701675415,
"step": 2430
},
{
"epoch": 1.7579250720461095,
"grad_norm": 3.089428424835205,
"learning_rate": 2.1874571586252177e-08,
"logits/chosen": -1.5453965663909912,
"logits/rejected": -1.5336331129074097,
"logps/chosen": -45.585872650146484,
"logps/rejected": -48.34822082519531,
"loss": 0.6877,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0007279099081642926,
"rewards/margins": 0.01115675363689661,
"rewards/rejected": -0.010428843088448048,
"step": 2440
},
{
"epoch": 1.7651296829971181,
"grad_norm": 2.425853967666626,
"learning_rate": 2.1666721135069037e-08,
"logits/chosen": -1.5158154964447021,
"logits/rejected": -1.502139687538147,
"logps/chosen": -49.875404357910156,
"logps/rejected": -51.33538055419922,
"loss": 0.6875,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.001414857804775238,
"rewards/margins": 0.011725572869181633,
"rewards/rejected": -0.010310716927051544,
"step": 2450
},
{
"epoch": 1.7723342939481268,
"grad_norm": 2.6500022411346436,
"learning_rate": 2.145910499962628e-08,
"logits/chosen": -1.5762279033660889,
"logits/rejected": -1.5551462173461914,
"logps/chosen": -44.01633834838867,
"logps/rejected": -46.197105407714844,
"loss": 0.6846,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.0036026693414896727,
"rewards/margins": 0.017717977985739708,
"rewards/rejected": -0.014115308411419392,
"step": 2460
},
{
"epoch": 1.7795389048991355,
"grad_norm": 3.8977699279785156,
"learning_rate": 2.1251737774480915e-08,
"logits/chosen": -1.5487562417984009,
"logits/rejected": -1.5394015312194824,
"logps/chosen": -53.25288772583008,
"logps/rejected": -55.375213623046875,
"loss": 0.6872,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0014735452132299542,
"rewards/margins": 0.01222173310816288,
"rewards/rejected": -0.010748187080025673,
"step": 2470
},
{
"epoch": 1.7867435158501441,
"grad_norm": 2.550679922103882,
"learning_rate": 2.104463403669264e-08,
"logits/chosen": -1.47690749168396,
"logits/rejected": -1.4564216136932373,
"logps/chosen": -49.03275680541992,
"logps/rejected": -51.274871826171875,
"loss": 0.6863,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.0006415749667212367,
"rewards/margins": 0.01415687520056963,
"rewards/rejected": -0.013515301048755646,
"step": 2480
},
{
"epoch": 1.7939481268011528,
"grad_norm": 2.6638097763061523,
"learning_rate": 2.0837808344799028e-08,
"logits/chosen": -1.4521634578704834,
"logits/rejected": -1.4361908435821533,
"logps/chosen": -43.8492431640625,
"logps/rejected": -47.552284240722656,
"loss": 0.6837,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": 0.006942611187696457,
"rewards/margins": 0.019251275807619095,
"rewards/rejected": -0.012308661825954914,
"step": 2490
},
{
"epoch": 1.8011527377521612,
"grad_norm": 3.212559700012207,
"learning_rate": 2.063127523779219e-08,
"logits/chosen": -1.4297634363174438,
"logits/rejected": -1.4289541244506836,
"logps/chosen": -44.858970642089844,
"logps/rejected": -51.319175720214844,
"loss": 0.6835,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": 0.002955908887088299,
"rewards/margins": 0.019690891727805138,
"rewards/rejected": -0.016734981909394264,
"step": 2500
},
{
"epoch": 1.8083573487031699,
"grad_norm": 3.859661102294922,
"learning_rate": 2.0425049234096737e-08,
"logits/chosen": -1.485381841659546,
"logits/rejected": -1.4706742763519287,
"logps/chosen": -49.1226921081543,
"logps/rejected": -51.82140350341797,
"loss": 0.6861,
"rewards/accuracies": 0.53125,
"rewards/chosen": 9.506577771389857e-05,
"rewards/margins": 0.014713233336806297,
"rewards/rejected": -0.014618167653679848,
"step": 2510
},
{
"epoch": 1.8155619596541785,
"grad_norm": 2.6423633098602295,
"learning_rate": 2.0219144830549163e-08,
"logits/chosen": -1.4601681232452393,
"logits/rejected": -1.4504835605621338,
"logps/chosen": -49.000816345214844,
"logps/rejected": -52.69274139404297,
"loss": 0.6847,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0008226001518778503,
"rewards/margins": 0.017380500212311745,
"rewards/rejected": -0.016557898372411728,
"step": 2520
},
{
"epoch": 1.8227665706051872,
"grad_norm": 2.8962371349334717,
"learning_rate": 2.0013576501378823e-08,
"logits/chosen": -1.4357713460922241,
"logits/rejected": -1.4260128736495972,
"logps/chosen": -44.66715621948242,
"logps/rejected": -48.878177642822266,
"loss": 0.6816,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.00762571394443512,
"rewards/margins": 0.02372068539261818,
"rewards/rejected": -0.01609497331082821,
"step": 2530
},
{
"epoch": 1.8299711815561959,
"grad_norm": 3.5678348541259766,
"learning_rate": 1.9808358697190426e-08,
"logits/chosen": -1.4616868495941162,
"logits/rejected": -1.4613463878631592,
"logps/chosen": -40.0428466796875,
"logps/rejected": -45.382728576660156,
"loss": 0.6845,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.0007027705432847142,
"rewards/margins": 0.017785217612981796,
"rewards/rejected": -0.01848798617720604,
"step": 2540
},
{
"epoch": 1.8371757925072045,
"grad_norm": 3.052267551422119,
"learning_rate": 1.9603505843948214e-08,
"logits/chosen": -1.4895626306533813,
"logits/rejected": -1.4689642190933228,
"logps/chosen": -41.04018020629883,
"logps/rejected": -46.34733200073242,
"loss": 0.6863,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.0010787765495479107,
"rewards/margins": 0.014137683436274529,
"rewards/rejected": -0.01305890642106533,
"step": 2550
},
{
"epoch": 1.8443804034582132,
"grad_norm": 3.034510374069214,
"learning_rate": 1.9399032341961886e-08,
"logits/chosen": -1.4604467153549194,
"logits/rejected": -1.4406030178070068,
"logps/chosen": -44.063720703125,
"logps/rejected": -45.94016647338867,
"loss": 0.6869,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0032738572917878628,
"rewards/margins": 0.013115392997860909,
"rewards/rejected": -0.009841536171734333,
"step": 2560
},
{
"epoch": 1.8515850144092219,
"grad_norm": 3.6909120082855225,
"learning_rate": 1.9194952564874323e-08,
"logits/chosen": -1.4899951219558716,
"logits/rejected": -1.4772499799728394,
"logps/chosen": -49.38003158569336,
"logps/rejected": -52.764427185058594,
"loss": 0.6858,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.0014510613400489092,
"rewards/margins": 0.015302592888474464,
"rewards/rejected": -0.013851528987288475,
"step": 2570
},
{
"epoch": 1.8587896253602305,
"grad_norm": 2.995894432067871,
"learning_rate": 1.8991280858651157e-08,
"logits/chosen": -1.465537190437317,
"logits/rejected": -1.4445680379867554,
"logps/chosen": -48.079158782958984,
"logps/rejected": -49.74862289428711,
"loss": 0.6861,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.0006449748761951923,
"rewards/margins": 0.014429867267608643,
"rewards/rejected": -0.013784890063107014,
"step": 2580
},
{
"epoch": 1.8659942363112392,
"grad_norm": 3.794341802597046,
"learning_rate": 1.8788031540572327e-08,
"logits/chosen": -1.4323540925979614,
"logits/rejected": -1.4187084436416626,
"logps/chosen": -43.358890533447266,
"logps/rejected": -47.263160705566406,
"loss": 0.6846,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": 0.0011780399363487959,
"rewards/margins": 0.017769720405340195,
"rewards/rejected": -0.01659167930483818,
"step": 2590
},
{
"epoch": 1.8731988472622478,
"grad_norm": 3.5173821449279785,
"learning_rate": 1.858521889822565e-08,
"logits/chosen": -1.4809000492095947,
"logits/rejected": -1.4711607694625854,
"logps/chosen": -44.77101135253906,
"logps/rejected": -47.36913299560547,
"loss": 0.6873,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.00217301188968122,
"rewards/margins": 0.012151877395808697,
"rewards/rejected": -0.009978866204619408,
"step": 2600
},
{
"epoch": 1.8804034582132565,
"grad_norm": 3.0861990451812744,
"learning_rate": 1.8382857188502422e-08,
"logits/chosen": -1.479575514793396,
"logits/rejected": -1.4646644592285156,
"logps/chosen": -43.409400939941406,
"logps/rejected": -46.272499084472656,
"loss": 0.6855,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.0008953431388363242,
"rewards/margins": 0.015901099890470505,
"rewards/rejected": -0.015005757100880146,
"step": 2610
},
{
"epoch": 1.8876080691642652,
"grad_norm": 3.0519216060638428,
"learning_rate": 1.8180960636595234e-08,
"logits/chosen": -1.434323787689209,
"logits/rejected": -1.4235074520111084,
"logps/chosen": -45.47016525268555,
"logps/rejected": -48.84101104736328,
"loss": 0.6842,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0007277448894456029,
"rewards/margins": 0.018448855727910995,
"rewards/rejected": -0.017721110954880714,
"step": 2620
},
{
"epoch": 1.8948126801152738,
"grad_norm": 2.6402575969696045,
"learning_rate": 1.7979543434998015e-08,
"logits/chosen": -1.5166761875152588,
"logits/rejected": -1.5122129917144775,
"logps/chosen": -54.10516357421875,
"logps/rejected": -55.752830505371094,
"loss": 0.6886,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.002384190447628498,
"rewards/margins": 0.009313153102993965,
"rewards/rejected": -0.011697344481945038,
"step": 2630
},
{
"epoch": 1.9020172910662825,
"grad_norm": 3.219054698944092,
"learning_rate": 1.7778619742508345e-08,
"logits/chosen": -1.4986519813537598,
"logits/rejected": -1.4785531759262085,
"logps/chosen": -48.8413200378418,
"logps/rejected": -50.63405990600586,
"loss": 0.6861,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.001953819999471307,
"rewards/margins": 0.014699439518153667,
"rewards/rejected": -0.016653258353471756,
"step": 2640
},
{
"epoch": 1.9092219020172911,
"grad_norm": 5.369049549102783,
"learning_rate": 1.757820368323213e-08,
"logits/chosen": -1.4477155208587646,
"logits/rejected": -1.431592345237732,
"logps/chosen": -55.639549255371094,
"logps/rejected": -60.893531799316406,
"loss": 0.6852,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.00035453608143143356,
"rewards/margins": 0.01644885540008545,
"rewards/rejected": -0.01680339314043522,
"step": 2650
},
{
"epoch": 1.9164265129682998,
"grad_norm": 2.708446741104126,
"learning_rate": 1.7378309345590803e-08,
"logits/chosen": -1.518417477607727,
"logits/rejected": -1.5215131044387817,
"logps/chosen": -48.15247344970703,
"logps/rejected": -51.795204162597656,
"loss": 0.6858,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": 0.0003538710006978363,
"rewards/margins": 0.01514039933681488,
"rewards/rejected": -0.01478652935475111,
"step": 2660
},
{
"epoch": 1.9236311239193085,
"grad_norm": 3.0533721446990967,
"learning_rate": 1.717895078133088e-08,
"logits/chosen": -1.5372436046600342,
"logits/rejected": -1.5273748636245728,
"logps/chosen": -45.74188995361328,
"logps/rejected": -50.93372344970703,
"loss": 0.6844,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": 0.001131549128331244,
"rewards/margins": 0.018048815429210663,
"rewards/rejected": -0.016917267814278603,
"step": 2670
},
{
"epoch": 1.9308357348703171,
"grad_norm": 2.965850830078125,
"learning_rate": 1.698014200453624e-08,
"logits/chosen": -1.5121673345565796,
"logits/rejected": -1.5155454874038696,
"logps/chosen": -48.55657196044922,
"logps/rejected": -53.24010467529297,
"loss": 0.6893,
"rewards/accuracies": 0.59375,
"rewards/chosen": -5.956319910183083e-06,
"rewards/margins": 0.00807130616158247,
"rewards/rejected": -0.008077261969447136,
"step": 2680
},
{
"epoch": 1.9380403458213258,
"grad_norm": 3.131131410598755,
"learning_rate": 1.6781896990642964e-08,
"logits/chosen": -1.4162065982818604,
"logits/rejected": -1.4067705869674683,
"logps/chosen": -53.68306350708008,
"logps/rejected": -55.64463424682617,
"loss": 0.6873,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.0006439717253670096,
"rewards/margins": 0.012254132889211178,
"rewards/rejected": -0.011610162444412708,
"step": 2690
},
{
"epoch": 1.9452449567723344,
"grad_norm": 3.7046432495117188,
"learning_rate": 1.658422967545693e-08,
"logits/chosen": -1.5447876453399658,
"logits/rejected": -1.5232436656951904,
"logps/chosen": -46.61565017700195,
"logps/rejected": -48.878089904785156,
"loss": 0.6864,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.002516025211662054,
"rewards/margins": 0.013911202549934387,
"rewards/rejected": -0.016427230089902878,
"step": 2700
},
{
"epoch": 1.952449567723343,
"grad_norm": 3.3615024089813232,
"learning_rate": 1.638715395417418e-08,
"logits/chosen": -1.515842318534851,
"logits/rejected": -1.4996858835220337,
"logps/chosen": -47.76423645019531,
"logps/rejected": -50.23698806762695,
"loss": 0.6872,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.0017537068342790008,
"rewards/margins": 0.012197253294289112,
"rewards/rejected": -0.013950960710644722,
"step": 2710
},
{
"epoch": 1.9596541786743515,
"grad_norm": 3.4271042346954346,
"learning_rate": 1.619068368040416e-08,
"logits/chosen": -1.5035583972930908,
"logits/rejected": -1.4936153888702393,
"logps/chosen": -42.36717987060547,
"logps/rejected": -48.00276565551758,
"loss": 0.6848,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0006282638642005622,
"rewards/margins": 0.017301367595791817,
"rewards/rejected": -0.016673101112246513,
"step": 2720
},
{
"epoch": 1.9668587896253602,
"grad_norm": 3.3180654048919678,
"learning_rate": 1.5994832665195853e-08,
"logits/chosen": -1.4341150522232056,
"logits/rejected": -1.4278676509857178,
"logps/chosen": -46.42858123779297,
"logps/rejected": -48.82233428955078,
"loss": 0.6873,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.00017052926705218852,
"rewards/margins": 0.01208487804979086,
"rewards/rejected": -0.011914348229765892,
"step": 2730
},
{
"epoch": 1.9740634005763689,
"grad_norm": 3.3026318550109863,
"learning_rate": 1.5799614676066906e-08,
"logits/chosen": -1.5610630512237549,
"logits/rejected": -1.5564082860946655,
"logps/chosen": -42.62709045410156,
"logps/rejected": -47.11577606201172,
"loss": 0.6852,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.001482530147768557,
"rewards/margins": 0.01645783707499504,
"rewards/rejected": -0.017940368503332138,
"step": 2740
},
{
"epoch": 1.9812680115273775,
"grad_norm": 2.802278518676758,
"learning_rate": 1.560504343603587e-08,
"logits/chosen": -1.4570392370224,
"logits/rejected": -1.4598525762557983,
"logps/chosen": -47.614715576171875,
"logps/rejected": -53.1636962890625,
"loss": 0.6862,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.0013699980918318033,
"rewards/margins": 0.014168500900268555,
"rewards/rejected": -0.012798503041267395,
"step": 2750
},
{
"epoch": 1.9884726224783862,
"grad_norm": 2.706557512283325,
"learning_rate": 1.541113262265748e-08,
"logits/chosen": -1.5580412149429321,
"logits/rejected": -1.5532734394073486,
"logps/chosen": -47.86179733276367,
"logps/rejected": -52.08073043823242,
"loss": 0.6855,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": 0.0002923453284893185,
"rewards/margins": 0.015893833711743355,
"rewards/rejected": -0.015601487830281258,
"step": 2760
},
{
"epoch": 1.9956772334293948,
"grad_norm": 2.82700777053833,
"learning_rate": 1.5217895867061227e-08,
"logits/chosen": -1.4797217845916748,
"logits/rejected": -1.46820068359375,
"logps/chosen": -49.087162017822266,
"logps/rejected": -51.762550354003906,
"loss": 0.6855,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.0006480285665020347,
"rewards/margins": 0.015934782102704048,
"rewards/rejected": -0.016582807525992393,
"step": 2770
},
{
"epoch": 2.0028818443804033,
"grad_norm": 3.225586414337158,
"learning_rate": 1.5025346752993098e-08,
"logits/chosen": -1.473181128501892,
"logits/rejected": -1.4776010513305664,
"logps/chosen": -47.254737854003906,
"logps/rejected": -51.428993225097656,
"loss": 0.6887,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.0032721899915486574,
"rewards/margins": 0.009254529140889645,
"rewards/rejected": -0.012526720762252808,
"step": 2780
},
{
"epoch": 2.010086455331412,
"grad_norm": 3.1906368732452393,
"learning_rate": 1.4833498815860756e-08,
"logits/chosen": -1.6033878326416016,
"logits/rejected": -1.594612717628479,
"logps/chosen": -44.750770568847656,
"logps/rejected": -49.424808502197266,
"loss": 0.6835,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.000528767064679414,
"rewards/margins": 0.019820012152194977,
"rewards/rejected": -0.019291242584586143,
"step": 2790
},
{
"epoch": 2.0172910662824206,
"grad_norm": 3.4794318675994873,
"learning_rate": 1.4642365541781993e-08,
"logits/chosen": -1.4186664819717407,
"logits/rejected": -1.4020098447799683,
"logps/chosen": -46.425323486328125,
"logps/rejected": -51.33763885498047,
"loss": 0.685,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -0.002099757781252265,
"rewards/margins": 0.01701737754046917,
"rewards/rejected": -0.019117135554552078,
"step": 2800
},
{
"epoch": 2.0244956772334293,
"grad_norm": 3.546409845352173,
"learning_rate": 1.4451960366636745e-08,
"logits/chosen": -1.5050753355026245,
"logits/rejected": -1.509311318397522,
"logps/chosen": -50.25437545776367,
"logps/rejected": -54.8748779296875,
"loss": 0.686,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -2.9070675736875273e-05,
"rewards/margins": 0.014723686501383781,
"rewards/rejected": -0.014752757735550404,
"step": 2810
},
{
"epoch": 2.031700288184438,
"grad_norm": 2.9926185607910156,
"learning_rate": 1.4262296675122592e-08,
"logits/chosen": -1.505336880683899,
"logits/rejected": -1.4901078939437866,
"logps/chosen": -43.937278747558594,
"logps/rejected": -48.619163513183594,
"loss": 0.6849,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0006590264965780079,
"rewards/margins": 0.016862262040376663,
"rewards/rejected": -0.017521290108561516,
"step": 2820
},
{
"epoch": 2.0389048991354466,
"grad_norm": 3.474058151245117,
"learning_rate": 1.407338779981389e-08,
"logits/chosen": -1.4738878011703491,
"logits/rejected": -1.4622704982757568,
"logps/chosen": -41.43745040893555,
"logps/rejected": -46.45849609375,
"loss": 0.683,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.001070915604941547,
"rewards/margins": 0.020852236077189445,
"rewards/rejected": -0.02192315086722374,
"step": 2830
},
{
"epoch": 2.0461095100864553,
"grad_norm": 3.1651461124420166,
"learning_rate": 1.3885247020224534e-08,
"logits/chosen": -1.4730961322784424,
"logits/rejected": -1.4625325202941895,
"logps/chosen": -40.98516082763672,
"logps/rejected": -44.32816696166992,
"loss": 0.6834,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0006273157196119428,
"rewards/margins": 0.020260414108633995,
"rewards/rejected": -0.0196330975741148,
"step": 2840
},
{
"epoch": 2.053314121037464,
"grad_norm": 2.788933277130127,
"learning_rate": 1.369788756187445e-08,
"logits/chosen": -1.5226420164108276,
"logits/rejected": -1.5108869075775146,
"logps/chosen": -46.892662048339844,
"logps/rejected": -48.15214920043945,
"loss": 0.6879,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.0032084626145660877,
"rewards/margins": 0.011149941943585873,
"rewards/rejected": -0.014358404092490673,
"step": 2850
},
{
"epoch": 2.0605187319884726,
"grad_norm": 3.175995111465454,
"learning_rate": 1.3511322595359925e-08,
"logits/chosen": -1.5318832397460938,
"logits/rejected": -1.5205862522125244,
"logps/chosen": -43.323387145996094,
"logps/rejected": -49.04631042480469,
"loss": 0.6839,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.0012531958054751158,
"rewards/margins": 0.019172416999936104,
"rewards/rejected": -0.02042561024427414,
"step": 2860
},
{
"epoch": 2.0677233429394812,
"grad_norm": 3.237109422683716,
"learning_rate": 1.3325565235427716e-08,
"logits/chosen": -1.5520436763763428,
"logits/rejected": -1.5436103343963623,
"logps/chosen": -45.29417037963867,
"logps/rejected": -49.32756423950195,
"loss": 0.6843,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.0010891546262428164,
"rewards/margins": 0.01812928169965744,
"rewards/rejected": -0.019218437373638153,
"step": 2870
},
{
"epoch": 2.07492795389049,
"grad_norm": 3.6626226902008057,
"learning_rate": 1.3140628540053218e-08,
"logits/chosen": -1.4575971364974976,
"logits/rejected": -1.4558467864990234,
"logps/chosen": -45.92155838012695,
"logps/rejected": -49.4353141784668,
"loss": 0.6855,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.0020993829239159822,
"rewards/margins": 0.01558808796107769,
"rewards/rejected": -0.0134887071326375,
"step": 2880
},
{
"epoch": 2.0821325648414986,
"grad_norm": 3.993279218673706,
"learning_rate": 1.2956525509522451e-08,
"logits/chosen": -1.434325933456421,
"logits/rejected": -1.43949556350708,
"logps/chosen": -47.83916473388672,
"logps/rejected": -51.43572998046875,
"loss": 0.6872,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": 0.001653014332987368,
"rewards/margins": 0.01248462125658989,
"rewards/rejected": -0.01083160750567913,
"step": 2890
},
{
"epoch": 2.089337175792507,
"grad_norm": 3.8572018146514893,
"learning_rate": 1.2773269085518267e-08,
"logits/chosen": -1.5160229206085205,
"logits/rejected": -1.5104598999023438,
"logps/chosen": -52.54634475708008,
"logps/rejected": -56.1825065612793,
"loss": 0.6865,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.0702252641058294e-06,
"rewards/margins": 0.013869213871657848,
"rewards/rejected": -0.013871285133063793,
"step": 2900
},
{
"epoch": 2.096541786743516,
"grad_norm": 2.667027235031128,
"learning_rate": 1.2590872150210574e-08,
"logits/chosen": -1.5915756225585938,
"logits/rejected": -1.575210452079773,
"logps/chosen": -45.68871307373047,
"logps/rejected": -47.84299087524414,
"loss": 0.6837,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.004803013987839222,
"rewards/margins": 0.019728917628526688,
"rewards/rejected": -0.024531930685043335,
"step": 2910
},
{
"epoch": 2.1037463976945245,
"grad_norm": 2.844864845275879,
"learning_rate": 1.2409347525350775e-08,
"logits/chosen": -1.4988892078399658,
"logits/rejected": -1.480571985244751,
"logps/chosen": -47.45149230957031,
"logps/rejected": -51.376197814941406,
"loss": 0.6831,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.0007760589360259473,
"rewards/margins": 0.020668352022767067,
"rewards/rejected": -0.019892293959856033,
"step": 2920
},
{
"epoch": 2.110951008645533,
"grad_norm": 3.4284279346466064,
"learning_rate": 1.2228707971370421e-08,
"logits/chosen": -1.4963008165359497,
"logits/rejected": -1.477461576461792,
"logps/chosen": -42.0821647644043,
"logps/rejected": -44.58640670776367,
"loss": 0.6839,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.0020004522521048784,
"rewards/margins": 0.01911274902522564,
"rewards/rejected": -0.01711229607462883,
"step": 2930
},
{
"epoch": 2.118155619596542,
"grad_norm": 4.411471843719482,
"learning_rate": 1.2048966186484282e-08,
"logits/chosen": -1.525394320487976,
"logits/rejected": -1.49559485912323,
"logps/chosen": -52.1498908996582,
"logps/rejected": -55.0552864074707,
"loss": 0.6861,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.0017288762610405684,
"rewards/margins": 0.014565639197826385,
"rewards/rejected": -0.016294512897729874,
"step": 2940
},
{
"epoch": 2.1253602305475505,
"grad_norm": 3.3227975368499756,
"learning_rate": 1.187013480579762e-08,
"logits/chosen": -1.489044189453125,
"logits/rejected": -1.4827308654785156,
"logps/chosen": -45.42039489746094,
"logps/rejected": -49.41254425048828,
"loss": 0.6841,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.0047860899940133095,
"rewards/margins": 0.019023966044187546,
"rewards/rejected": -0.02381005696952343,
"step": 2950
},
{
"epoch": 2.132564841498559,
"grad_norm": 4.395474433898926,
"learning_rate": 1.1692226400418073e-08,
"logits/chosen": -1.4131033420562744,
"logits/rejected": -1.4049489498138428,
"logps/chosen": -49.05443572998047,
"logps/rejected": -52.0718994140625,
"loss": 0.6855,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.004469198640435934,
"rewards/margins": 0.016008742153644562,
"rewards/rejected": -0.020477941259741783,
"step": 2960
},
{
"epoch": 2.139769452449568,
"grad_norm": 2.495917558670044,
"learning_rate": 1.1515253476571923e-08,
"logits/chosen": -1.4480403661727905,
"logits/rejected": -1.4422080516815186,
"logps/chosen": -44.43804931640625,
"logps/rejected": -51.04991149902344,
"loss": 0.684,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.003528184024617076,
"rewards/margins": 0.018853966146707535,
"rewards/rejected": -0.02238215133547783,
"step": 2970
},
{
"epoch": 2.1469740634005765,
"grad_norm": 3.3481624126434326,
"learning_rate": 1.133922847472496e-08,
"logits/chosen": -1.4908777475357056,
"logits/rejected": -1.4869946241378784,
"logps/chosen": -52.52874755859375,
"logps/rejected": -55.038414001464844,
"loss": 0.685,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.0005473472410812974,
"rewards/margins": 0.017001759260892868,
"rewards/rejected": -0.01754910498857498,
"step": 2980
},
{
"epoch": 2.154178674351585,
"grad_norm": 3.3193447589874268,
"learning_rate": 1.1164163768707952e-08,
"logits/chosen": -1.4650285243988037,
"logits/rejected": -1.4538644552230835,
"logps/chosen": -47.43016052246094,
"logps/rejected": -51.636924743652344,
"loss": 0.6824,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0007126646814867854,
"rewards/margins": 0.02234521321952343,
"rewards/rejected": -0.023057879880070686,
"step": 2990
},
{
"epoch": 2.161383285302594,
"grad_norm": 3.359654426574707,
"learning_rate": 1.0990071664846861e-08,
"logits/chosen": -1.4393339157104492,
"logits/rejected": -1.429203748703003,
"logps/chosen": -48.76907730102539,
"logps/rejected": -53.98026657104492,
"loss": 0.6819,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": 0.00010694740194594488,
"rewards/margins": 0.023278547450900078,
"rewards/rejected": -0.023171598091721535,
"step": 3000
},
{
"epoch": 2.1685878962536025,
"grad_norm": 3.054330348968506,
"learning_rate": 1.0816964401097739e-08,
"logits/chosen": -1.4826459884643555,
"logits/rejected": -1.472663164138794,
"logps/chosen": -43.025970458984375,
"logps/rejected": -45.86864471435547,
"loss": 0.6854,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.0005824376130476594,
"rewards/margins": 0.016256345435976982,
"rewards/rejected": -0.016838783398270607,
"step": 3010
},
{
"epoch": 2.175792507204611,
"grad_norm": 3.929755926132202,
"learning_rate": 1.0644854146186406e-08,
"logits/chosen": -1.5146058797836304,
"logits/rejected": -1.49686598777771,
"logps/chosen": -48.23189926147461,
"logps/rejected": -52.9786491394043,
"loss": 0.6827,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.0024410493206232786,
"rewards/margins": 0.021744880825281143,
"rewards/rejected": -0.02418592758476734,
"step": 3020
},
{
"epoch": 2.18299711815562,
"grad_norm": 3.215834856033325,
"learning_rate": 1.0473752998753114e-08,
"logits/chosen": -1.4946410655975342,
"logits/rejected": -1.4729467630386353,
"logps/chosen": -48.56220626831055,
"logps/rejected": -51.715721130371094,
"loss": 0.6823,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": 0.0010552393505349755,
"rewards/margins": 0.022361181676387787,
"rewards/rejected": -0.02130594104528427,
"step": 3030
},
{
"epoch": 2.1902017291066285,
"grad_norm": 2.973313093185425,
"learning_rate": 1.030367298650201e-08,
"logits/chosen": -1.4930821657180786,
"logits/rejected": -1.492546796798706,
"logps/chosen": -48.8010368347168,
"logps/rejected": -53.67090606689453,
"loss": 0.6873,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.004332934506237507,
"rewards/margins": 0.012475891038775444,
"rewards/rejected": -0.016808826476335526,
"step": 3040
},
{
"epoch": 2.1974063400576367,
"grad_norm": 3.869215726852417,
"learning_rate": 1.0134626065355675e-08,
"logits/chosen": -1.5941343307495117,
"logits/rejected": -1.5830258131027222,
"logps/chosen": -49.36521530151367,
"logps/rejected": -52.964378356933594,
"loss": 0.6825,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.0014757805038243532,
"rewards/margins": 0.02204863727092743,
"rewards/rejected": -0.020572859793901443,
"step": 3050
},
{
"epoch": 2.2046109510086453,
"grad_norm": 3.5000784397125244,
"learning_rate": 9.966624118614611e-09,
"logits/chosen": -1.491275668144226,
"logits/rejected": -1.471986174583435,
"logps/chosen": -52.319740295410156,
"logps/rejected": -55.521949768066406,
"loss": 0.6837,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": 0.002091693691909313,
"rewards/margins": 0.019565019756555557,
"rewards/rejected": -0.01747332513332367,
"step": 3060
},
{
"epoch": 2.211815561959654,
"grad_norm": 2.422412395477295,
"learning_rate": 9.799678956121976e-09,
"logits/chosen": -1.436295747756958,
"logits/rejected": -1.4197183847427368,
"logps/chosen": -45.91994094848633,
"logps/rejected": -48.53052520751953,
"loss": 0.6871,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.002334743272513151,
"rewards/margins": 0.012487743981182575,
"rewards/rejected": -0.014822488650679588,
"step": 3070
},
{
"epoch": 2.2190201729106627,
"grad_norm": 3.5557055473327637,
"learning_rate": 9.633802313433314e-09,
"logits/chosen": -1.4133331775665283,
"logits/rejected": -1.4094794988632202,
"logps/chosen": -48.400875091552734,
"logps/rejected": -50.9952278137207,
"loss": 0.685,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.0015882644802331924,
"rewards/margins": 0.016813434660434723,
"rewards/rejected": -0.018401699140667915,
"step": 3080
},
{
"epoch": 2.2262247838616713,
"grad_norm": 2.77536940574646,
"learning_rate": 9.469005850991705e-09,
"logits/chosen": -1.4840987920761108,
"logits/rejected": -1.4713189601898193,
"logps/chosen": -47.20307159423828,
"logps/rejected": -48.701622009277344,
"loss": 0.6842,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.0018709308933466673,
"rewards/margins": 0.01850222796201706,
"rewards/rejected": -0.020373158156871796,
"step": 3090
},
{
"epoch": 2.23342939481268,
"grad_norm": 3.2006888389587402,
"learning_rate": 9.305301153307949e-09,
"logits/chosen": -1.4952600002288818,
"logits/rejected": -1.499306321144104,
"logps/chosen": -40.018882751464844,
"logps/rejected": -44.07697296142578,
"loss": 0.6833,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.005154062993824482,
"rewards/margins": 0.020358018577098846,
"rewards/rejected": -0.025512080639600754,
"step": 3100
},
{
"epoch": 2.2406340057636887,
"grad_norm": 2.7660207748413086,
"learning_rate": 9.142699728146336e-09,
"logits/chosen": -1.4331061840057373,
"logits/rejected": -1.4245890378952026,
"logps/chosen": -46.127166748046875,
"logps/rejected": -51.11559295654297,
"loss": 0.6843,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.0028220864478498697,
"rewards/margins": 0.018410906195640564,
"rewards/rejected": -0.02123299241065979,
"step": 3110
},
{
"epoch": 2.2478386167146973,
"grad_norm": 2.9993467330932617,
"learning_rate": 8.981213005715627e-09,
"logits/chosen": -1.501518726348877,
"logits/rejected": -1.5010533332824707,
"logps/chosen": -44.233001708984375,
"logps/rejected": -49.084808349609375,
"loss": 0.6841,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0025021065957844257,
"rewards/margins": 0.01857968047261238,
"rewards/rejected": -0.021081790328025818,
"step": 3120
},
{
"epoch": 2.255043227665706,
"grad_norm": 3.6337778568267822,
"learning_rate": 8.820852337865611e-09,
"logits/chosen": -1.5516728162765503,
"logits/rejected": -1.536139965057373,
"logps/chosen": -45.085365295410156,
"logps/rejected": -48.6732063293457,
"loss": 0.6844,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -0.0014631979865953326,
"rewards/margins": 0.01818043366074562,
"rewards/rejected": -0.0196436308324337,
"step": 3130
},
{
"epoch": 2.2622478386167146,
"grad_norm": 2.8269283771514893,
"learning_rate": 8.661628997289044e-09,
"logits/chosen": -1.4337918758392334,
"logits/rejected": -1.420588731765747,
"logps/chosen": -45.37371826171875,
"logps/rejected": -49.878807067871094,
"loss": 0.6839,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.001038391375914216,
"rewards/margins": 0.019226137548685074,
"rewards/rejected": -0.020264528691768646,
"step": 3140
},
{
"epoch": 2.2694524495677233,
"grad_norm": 2.8211851119995117,
"learning_rate": 8.503554176729341e-09,
"logits/chosen": -1.4207173585891724,
"logits/rejected": -1.4155914783477783,
"logps/chosen": -45.44511032104492,
"logps/rejected": -49.06669998168945,
"loss": 0.6832,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.0009460471337661147,
"rewards/margins": 0.0205953698605299,
"rewards/rejected": -0.01964932307600975,
"step": 3150
},
{
"epoch": 2.276657060518732,
"grad_norm": 3.823362350463867,
"learning_rate": 8.346638988193636e-09,
"logits/chosen": -1.4726402759552002,
"logits/rejected": -1.4704288244247437,
"logps/chosen": -40.57404708862305,
"logps/rejected": -46.3819580078125,
"loss": 0.6835,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.0009540968458168209,
"rewards/margins": 0.02004236914217472,
"rewards/rejected": -0.020996464416384697,
"step": 3160
},
{
"epoch": 2.2838616714697406,
"grad_norm": 4.203202247619629,
"learning_rate": 8.19089446217176e-09,
"logits/chosen": -1.4289653301239014,
"logits/rejected": -1.4073545932769775,
"logps/chosen": -45.77039337158203,
"logps/rejected": -51.22684860229492,
"loss": 0.6797,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": 0.002206823555752635,
"rewards/margins": 0.02788766846060753,
"rewards/rejected": -0.025680843740701675,
"step": 3170
},
{
"epoch": 2.2910662824207493,
"grad_norm": 3.061116933822632,
"learning_rate": 8.036331546860777e-09,
"logits/chosen": -1.4556572437286377,
"logits/rejected": -1.453061819076538,
"logps/chosen": -45.42163848876953,
"logps/rejected": -48.2816276550293,
"loss": 0.6875,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.0038195946253836155,
"rewards/margins": 0.012059660628437996,
"rewards/rejected": -0.015879254788160324,
"step": 3180
},
{
"epoch": 2.298270893371758,
"grad_norm": 3.667120933532715,
"learning_rate": 7.882961107395416e-09,
"logits/chosen": -1.4970252513885498,
"logits/rejected": -1.4873993396759033,
"logps/chosen": -52.32844924926758,
"logps/rejected": -52.70637893676758,
"loss": 0.6869,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.006024113390594721,
"rewards/margins": 0.013099935837090015,
"rewards/rejected": -0.019124049693346024,
"step": 3190
},
{
"epoch": 2.3054755043227666,
"grad_norm": 4.6425042152404785,
"learning_rate": 7.73079392508428e-09,
"logits/chosen": -1.421281099319458,
"logits/rejected": -1.4252192974090576,
"logps/chosen": -49.780357360839844,
"logps/rejected": -56.728782653808594,
"loss": 0.6824,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.004081489983946085,
"rewards/margins": 0.02246311493217945,
"rewards/rejected": -0.026544606313109398,
"step": 3200
},
{
"epoch": 2.3126801152737753,
"grad_norm": 3.641294479370117,
"learning_rate": 7.579840696651938e-09,
"logits/chosen": -1.5132240056991577,
"logits/rejected": -1.5067884922027588,
"logps/chosen": -42.3286247253418,
"logps/rejected": -45.735172271728516,
"loss": 0.6841,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.004373098723590374,
"rewards/margins": 0.01880384422838688,
"rewards/rejected": -0.02317694202065468,
"step": 3210
},
{
"epoch": 2.319884726224784,
"grad_norm": 4.251941680908203,
"learning_rate": 7.43011203348704e-09,
"logits/chosen": -1.3565527200698853,
"logits/rejected": -1.3517991304397583,
"logps/chosen": -53.13508224487305,
"logps/rejected": -53.89665603637695,
"loss": 0.6857,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.005780863109976053,
"rewards/margins": 0.015468914993107319,
"rewards/rejected": -0.02124977670609951,
"step": 3220
},
{
"epoch": 2.3270893371757926,
"grad_norm": 3.2468185424804688,
"learning_rate": 7.281618460896344e-09,
"logits/chosen": -1.4836031198501587,
"logits/rejected": -1.4732134342193604,
"logps/chosen": -46.293495178222656,
"logps/rejected": -50.772377014160156,
"loss": 0.6844,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.0019498964538797736,
"rewards/margins": 0.018127668648958206,
"rewards/rejected": -0.02007756568491459,
"step": 3230
},
{
"epoch": 2.3342939481268012,
"grad_norm": 2.987872362136841,
"learning_rate": 7.134370417364849e-09,
"logits/chosen": -1.431056022644043,
"logits/rejected": -1.4228883981704712,
"logps/chosen": -45.278480529785156,
"logps/rejected": -48.09427261352539,
"loss": 0.6868,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.007163041736930609,
"rewards/margins": 0.013337318785488605,
"rewards/rejected": -0.020500360056757927,
"step": 3240
},
{
"epoch": 2.34149855907781,
"grad_norm": 4.016878604888916,
"learning_rate": 6.988378253821981e-09,
"logits/chosen": -1.4581267833709717,
"logits/rejected": -1.4503519535064697,
"logps/chosen": -51.406394958496094,
"logps/rejected": -54.95398712158203,
"loss": 0.6867,
"rewards/accuracies": 0.59375,
"rewards/chosen": 8.269222598755732e-05,
"rewards/margins": 0.013455493375658989,
"rewards/rejected": -0.013372799381613731,
"step": 3250
},
{
"epoch": 2.3487031700288186,
"grad_norm": 3.061265468597412,
"learning_rate": 6.8436522329140186e-09,
"logits/chosen": -1.4396600723266602,
"logits/rejected": -1.445936918258667,
"logps/chosen": -46.99631881713867,
"logps/rejected": -50.84511184692383,
"loss": 0.6853,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.0030071113724261522,
"rewards/margins": 0.016658034175634384,
"rewards/rejected": -0.019665146246552467,
"step": 3260
},
{
"epoch": 2.3559077809798272,
"grad_norm": 3.53556752204895,
"learning_rate": 6.700202528282603e-09,
"logits/chosen": -1.4168142080307007,
"logits/rejected": -1.3971381187438965,
"logps/chosen": -48.588462829589844,
"logps/rejected": -51.615745544433594,
"loss": 0.6833,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.00601479597389698,
"rewards/margins": 0.020779959857463837,
"rewards/rejected": -0.026794757694005966,
"step": 3270
},
{
"epoch": 2.363112391930836,
"grad_norm": 3.7148492336273193,
"learning_rate": 6.558039223849668e-09,
"logits/chosen": -1.5100008249282837,
"logits/rejected": -1.4907362461090088,
"logps/chosen": -46.292266845703125,
"logps/rejected": -52.69337844848633,
"loss": 0.6816,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0020173946395516396,
"rewards/margins": 0.024240782484412193,
"rewards/rejected": -0.026258179917931557,
"step": 3280
},
{
"epoch": 2.3703170028818445,
"grad_norm": 2.8915648460388184,
"learning_rate": 6.417172313108471e-09,
"logits/chosen": -1.4239482879638672,
"logits/rejected": -1.4123411178588867,
"logps/chosen": -44.13205337524414,
"logps/rejected": -47.372459411621094,
"loss": 0.6852,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.008008310571312904,
"rewards/margins": 0.016741162165999413,
"rewards/rejected": -0.024749474599957466,
"step": 3290
},
{
"epoch": 2.377521613832853,
"grad_norm": 2.96709942817688,
"learning_rate": 6.277611698421179e-09,
"logits/chosen": -1.5538231134414673,
"logits/rejected": -1.5337212085723877,
"logps/chosen": -39.08244323730469,
"logps/rejected": -45.03202438354492,
"loss": 0.6812,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.003737329039722681,
"rewards/margins": 0.02486787550151348,
"rewards/rejected": -0.028605204075574875,
"step": 3300
},
{
"epoch": 2.3847262247838614,
"grad_norm": 4.765800476074219,
"learning_rate": 6.139367190322714e-09,
"logits/chosen": -1.4922102689743042,
"logits/rejected": -1.4921929836273193,
"logps/chosen": -52.74082565307617,
"logps/rejected": -58.16496658325195,
"loss": 0.6855,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0037182599771767855,
"rewards/margins": 0.01609138958156109,
"rewards/rejected": -0.019809648394584656,
"step": 3310
},
{
"epoch": 2.39193083573487,
"grad_norm": 2.5080461502075195,
"learning_rate": 6.002448506831171e-09,
"logits/chosen": -1.4787399768829346,
"logits/rejected": -1.474265217781067,
"logps/chosen": -44.079654693603516,
"logps/rejected": -49.27398681640625,
"loss": 0.6843,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.003509215312078595,
"rewards/margins": 0.0185092780739069,
"rewards/rejected": -0.022018492221832275,
"step": 3320
},
{
"epoch": 2.3991354466858787,
"grad_norm": 3.059837818145752,
"learning_rate": 5.866865272764607e-09,
"logits/chosen": -1.4946706295013428,
"logits/rejected": -1.4880635738372803,
"logps/chosen": -46.43208694458008,
"logps/rejected": -50.517845153808594,
"loss": 0.6854,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.0064870258793234825,
"rewards/margins": 0.016077209264039993,
"rewards/rejected": -0.0225642379373312,
"step": 3330
},
{
"epoch": 2.4063400576368874,
"grad_norm": 4.633550643920898,
"learning_rate": 5.7326270190645595e-09,
"logits/chosen": -1.327772855758667,
"logits/rejected": -1.3225994110107422,
"logps/chosen": -49.93898391723633,
"logps/rejected": -52.00776290893555,
"loss": 0.6846,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.0053224144503474236,
"rewards/margins": 0.017836768180131912,
"rewards/rejected": -0.02315918542444706,
"step": 3340
},
{
"epoch": 2.413544668587896,
"grad_norm": 3.6820759773254395,
"learning_rate": 5.599743182125938e-09,
"logits/chosen": -1.537332534790039,
"logits/rejected": -1.536694049835205,
"logps/chosen": -48.791236877441406,
"logps/rejected": -54.156349182128906,
"loss": 0.6848,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.0013721368741244078,
"rewards/margins": 0.01742737926542759,
"rewards/rejected": -0.018799515441060066,
"step": 3350
},
{
"epoch": 2.4207492795389047,
"grad_norm": 3.4776570796966553,
"learning_rate": 5.46822310313379e-09,
"logits/chosen": -1.5602588653564453,
"logits/rejected": -1.5649378299713135,
"logps/chosen": -49.44075012207031,
"logps/rejected": -52.79970169067383,
"loss": 0.6874,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.0048035006038844585,
"rewards/margins": 0.012090938165783882,
"rewards/rejected": -0.01689443737268448,
"step": 3360
},
{
"epoch": 2.4279538904899134,
"grad_norm": 3.6628270149230957,
"learning_rate": 5.33807602740658e-09,
"logits/chosen": -1.5563075542449951,
"logits/rejected": -1.5405874252319336,
"logps/chosen": -41.89576721191406,
"logps/rejected": -47.50288009643555,
"loss": 0.6799,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.0014247519429773092,
"rewards/margins": 0.027519360184669495,
"rewards/rejected": -0.0289441104978323,
"step": 3370
},
{
"epoch": 2.435158501440922,
"grad_norm": 3.8053536415100098,
"learning_rate": 5.209311103746334e-09,
"logits/chosen": -1.475118637084961,
"logits/rejected": -1.4708943367004395,
"logps/chosen": -47.132999420166016,
"logps/rejected": -52.47126007080078,
"loss": 0.6836,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.003361554117873311,
"rewards/margins": 0.019802602007985115,
"rewards/rejected": -0.023164156824350357,
"step": 3380
},
{
"epoch": 2.4423631123919307,
"grad_norm": 4.187764644622803,
"learning_rate": 5.081937383795484e-09,
"logits/chosen": -1.4638832807540894,
"logits/rejected": -1.45353102684021,
"logps/chosen": -44.22564697265625,
"logps/rejected": -48.906394958496094,
"loss": 0.6821,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.0011400593211874366,
"rewards/margins": 0.022827504202723503,
"rewards/rejected": -0.02396756038069725,
"step": 3390
},
{
"epoch": 2.4495677233429394,
"grad_norm": 3.6951913833618164,
"learning_rate": 4.955963821400599e-09,
"logits/chosen": -1.524279236793518,
"logits/rejected": -1.506208062171936,
"logps/chosen": -46.91497802734375,
"logps/rejected": -49.75968933105469,
"loss": 0.6827,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -0.002611330011859536,
"rewards/margins": 0.0218039583414793,
"rewards/rejected": -0.024415289983153343,
"step": 3400
},
{
"epoch": 2.456772334293948,
"grad_norm": 2.797675371170044,
"learning_rate": 4.831399271982928e-09,
"logits/chosen": -1.3963180780410767,
"logits/rejected": -1.380124568939209,
"logps/chosen": -49.73040008544922,
"logps/rejected": -52.8338737487793,
"loss": 0.6831,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.0018096374114975333,
"rewards/margins": 0.02105477824807167,
"rewards/rejected": -0.022864414379000664,
"step": 3410
},
{
"epoch": 2.4639769452449567,
"grad_norm": 3.9951834678649902,
"learning_rate": 4.708252491915951e-09,
"logits/chosen": -1.4993171691894531,
"logits/rejected": -1.4892971515655518,
"logps/chosen": -47.080963134765625,
"logps/rejected": -51.67246627807617,
"loss": 0.683,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.003982001915574074,
"rewards/margins": 0.021508801728487015,
"rewards/rejected": -0.02549080178141594,
"step": 3420
},
{
"epoch": 2.4711815561959654,
"grad_norm": 2.9166736602783203,
"learning_rate": 4.58653213790981e-09,
"logits/chosen": -1.4970018863677979,
"logits/rejected": -1.4785791635513306,
"logps/chosen": -47.441192626953125,
"logps/rejected": -52.0790901184082,
"loss": 0.6837,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.002383728977292776,
"rewards/margins": 0.019747789949178696,
"rewards/rejected": -0.02213152125477791,
"step": 3430
},
{
"epoch": 2.478386167146974,
"grad_norm": 3.378357172012329,
"learning_rate": 4.466246766402773e-09,
"logits/chosen": -1.4705661535263062,
"logits/rejected": -1.4513781070709229,
"logps/chosen": -48.759010314941406,
"logps/rejected": -52.48942947387695,
"loss": 0.6819,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.0021346856374293566,
"rewards/margins": 0.023418936878442764,
"rewards/rejected": -0.025553623214364052,
"step": 3440
},
{
"epoch": 2.4855907780979827,
"grad_norm": 3.730109453201294,
"learning_rate": 4.347404832959775e-09,
"logits/chosen": -1.5252504348754883,
"logits/rejected": -1.5139144659042358,
"logps/chosen": -44.665199279785156,
"logps/rejected": -48.89989471435547,
"loss": 0.6839,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.004804068244993687,
"rewards/margins": 0.019236544147133827,
"rewards/rejected": -0.02404061332345009,
"step": 3450
},
{
"epoch": 2.4927953890489913,
"grad_norm": 3.4961752891540527,
"learning_rate": 4.230014691678016e-09,
"logits/chosen": -1.4771640300750732,
"logits/rejected": -1.4784762859344482,
"logps/chosen": -49.46748352050781,
"logps/rejected": -51.16899871826172,
"loss": 0.6866,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.006382669322192669,
"rewards/margins": 0.013793786987662315,
"rewards/rejected": -0.020176459103822708,
"step": 3460
},
{
"epoch": 2.5,
"grad_norm": 3.152644395828247,
"learning_rate": 4.114084594599707e-09,
"logits/chosen": -1.4625837802886963,
"logits/rejected": -1.4395456314086914,
"logps/chosen": -45.60460662841797,
"logps/rejected": -51.585113525390625,
"loss": 0.682,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": -0.0029743132181465626,
"rewards/margins": 0.02330465242266655,
"rewards/rejected": -0.026278968900442123,
"step": 3470
},
{
"epoch": 2.5072046109510087,
"grad_norm": 3.1570560932159424,
"learning_rate": 3.9996226911319546e-09,
"logits/chosen": -1.4798153638839722,
"logits/rejected": -1.457363486289978,
"logps/chosen": -45.63786315917969,
"logps/rejected": -48.74847412109375,
"loss": 0.6838,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.0032687117345631123,
"rewards/margins": 0.01941683515906334,
"rewards/rejected": -0.022685546427965164,
"step": 3480
},
{
"epoch": 2.5144092219020173,
"grad_norm": 3.3458943367004395,
"learning_rate": 3.886637027473949e-09,
"logits/chosen": -1.512085199356079,
"logits/rejected": -1.5080697536468506,
"logps/chosen": -47.53780746459961,
"logps/rejected": -51.640533447265625,
"loss": 0.6836,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.0049394266679883,
"rewards/margins": 0.01969726011157036,
"rewards/rejected": -0.024636687710881233,
"step": 3490
},
{
"epoch": 2.521613832853026,
"grad_norm": 3.2042598724365234,
"learning_rate": 3.775135546051295e-09,
"logits/chosen": -1.4048144817352295,
"logits/rejected": -1.4050971269607544,
"logps/chosen": -46.04724884033203,
"logps/rejected": -50.54692077636719,
"loss": 0.6819,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.004532016348093748,
"rewards/margins": 0.023406367748975754,
"rewards/rejected": -0.02793838456273079,
"step": 3500
},
{
"epoch": 2.5288184438040346,
"grad_norm": 3.420597553253174,
"learning_rate": 3.665126084957723e-09,
"logits/chosen": -1.469005823135376,
"logits/rejected": -1.4607031345367432,
"logps/chosen": -50.867950439453125,
"logps/rejected": -51.225799560546875,
"loss": 0.6844,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.005590965040028095,
"rewards/margins": 0.01840902492403984,
"rewards/rejected": -0.023999987170100212,
"step": 3510
},
{
"epoch": 2.5360230547550433,
"grad_norm": 3.1464619636535645,
"learning_rate": 3.556616377404101e-09,
"logits/chosen": -1.5020486116409302,
"logits/rejected": -1.4901654720306396,
"logps/chosen": -51.85783004760742,
"logps/rejected": -55.8673210144043,
"loss": 0.6817,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -0.0066805132664740086,
"rewards/margins": 0.02360719069838524,
"rewards/rejected": -0.03028770722448826,
"step": 3520
},
{
"epoch": 2.543227665706052,
"grad_norm": 3.3954811096191406,
"learning_rate": 3.4496140511748125e-09,
"logits/chosen": -1.485480546951294,
"logits/rejected": -1.4669150114059448,
"logps/chosen": -48.15165710449219,
"logps/rejected": -51.049583435058594,
"loss": 0.6833,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.007361284457147121,
"rewards/margins": 0.02040119096636772,
"rewards/rejected": -0.027762476354837418,
"step": 3530
},
{
"epoch": 2.5504322766570606,
"grad_norm": 3.955996513366699,
"learning_rate": 3.3441266280915427e-09,
"logits/chosen": -1.4491957426071167,
"logits/rejected": -1.446547269821167,
"logps/chosen": -53.73142623901367,
"logps/rejected": -57.25432205200195,
"loss": 0.6852,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.0023315693251788616,
"rewards/margins": 0.01659976877272129,
"rewards/rejected": -0.01893133856356144,
"step": 3540
},
{
"epoch": 2.5576368876080693,
"grad_norm": 3.521127462387085,
"learning_rate": 3.2401615234845693e-09,
"logits/chosen": -1.492701530456543,
"logits/rejected": -1.475007176399231,
"logps/chosen": -54.01853561401367,
"logps/rejected": -57.351951599121094,
"loss": 0.6819,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.0070812939666211605,
"rewards/margins": 0.023319412022829056,
"rewards/rejected": -0.030400704592466354,
"step": 3550
},
{
"epoch": 2.564841498559078,
"grad_norm": 3.1193737983703613,
"learning_rate": 3.1377260456714375e-09,
"logits/chosen": -1.3230210542678833,
"logits/rejected": -1.3108537197113037,
"logps/chosen": -49.07139587402344,
"logps/rejected": -54.1447868347168,
"loss": 0.6835,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.007112964056432247,
"rewards/margins": 0.020381804555654526,
"rewards/rejected": -0.027494769543409348,
"step": 3560
},
{
"epoch": 2.5720461095100866,
"grad_norm": 3.66926908493042,
"learning_rate": 3.0368273954432698e-09,
"logits/chosen": -1.5296647548675537,
"logits/rejected": -1.5019731521606445,
"logps/chosen": -51.096221923828125,
"logps/rejected": -53.320556640625,
"loss": 0.6845,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.006328393705189228,
"rewards/margins": 0.018091212958097458,
"rewards/rejected": -0.02441960759460926,
"step": 3570
},
{
"epoch": 2.5792507204610953,
"grad_norm": 3.0272419452667236,
"learning_rate": 2.937472665558541e-09,
"logits/chosen": -1.5538597106933594,
"logits/rejected": -1.5464608669281006,
"logps/chosen": -45.452919006347656,
"logps/rejected": -47.75053024291992,
"loss": 0.6817,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.008346345275640488,
"rewards/margins": 0.02381196618080139,
"rewards/rejected": -0.03215831145644188,
"step": 3580
},
{
"epoch": 2.586455331412104,
"grad_norm": 4.039034843444824,
"learning_rate": 2.8396688402445053e-09,
"logits/chosen": -1.574406385421753,
"logits/rejected": -1.557455062866211,
"logps/chosen": -45.44925308227539,
"logps/rejected": -51.695777893066406,
"loss": 0.6816,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.010429826565086842,
"rewards/margins": 0.024167021736502647,
"rewards/rejected": -0.03459685295820236,
"step": 3590
},
{
"epoch": 2.5936599423631126,
"grad_norm": 4.070069313049316,
"learning_rate": 2.7434227947062324e-09,
"logits/chosen": -1.526296854019165,
"logits/rejected": -1.5150690078735352,
"logps/chosen": -53.84295654296875,
"logps/rejected": -57.382408142089844,
"loss": 0.6856,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.005254354793578386,
"rewards/margins": 0.015645707026124,
"rewards/rejected": -0.02090005949139595,
"step": 3600
},
{
"epoch": 2.6008645533141213,
"grad_norm": 3.0419559478759766,
"learning_rate": 2.6487412946432976e-09,
"logits/chosen": -1.4455441236495972,
"logits/rejected": -1.4326177835464478,
"logps/chosen": -49.550575256347656,
"logps/rejected": -52.355995178222656,
"loss": 0.6822,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.013003657571971416,
"rewards/margins": 0.02292976900935173,
"rewards/rejected": -0.03593342751264572,
"step": 3610
},
{
"epoch": 2.60806916426513,
"grad_norm": 3.4363460540771484,
"learning_rate": 2.5556309957742024e-09,
"logits/chosen": -1.4442135095596313,
"logits/rejected": -1.4347007274627686,
"logps/chosen": -44.96186447143555,
"logps/rejected": -52.17518997192383,
"loss": 0.6803,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": 0.0013826603535562754,
"rewards/margins": 0.02670016512274742,
"rewards/rejected": -0.02531750500202179,
"step": 3620
},
{
"epoch": 2.6152737752161386,
"grad_norm": 3.359717607498169,
"learning_rate": 2.4640984433684758e-09,
"logits/chosen": -1.5575648546218872,
"logits/rejected": -1.543250560760498,
"logps/chosen": -50.973533630371094,
"logps/rejected": -53.1517448425293,
"loss": 0.6838,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.004719638731330633,
"rewards/margins": 0.019542286172509193,
"rewards/rejected": -0.024261925369501114,
"step": 3630
},
{
"epoch": 2.6224783861671472,
"grad_norm": 3.6610288619995117,
"learning_rate": 2.3741500717865987e-09,
"logits/chosen": -1.4447298049926758,
"logits/rejected": -1.456498384475708,
"logps/chosen": -47.38273620605469,
"logps/rejected": -52.238502502441406,
"loss": 0.684,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.0018823295831680298,
"rewards/margins": 0.01906423084437847,
"rewards/rejected": -0.0209465604275465,
"step": 3640
},
{
"epoch": 2.629682997118156,
"grad_norm": 3.1195054054260254,
"learning_rate": 2.285792204027678e-09,
"logits/chosen": -1.4207738637924194,
"logits/rejected": -1.4102163314819336,
"logps/chosen": -47.51162338256836,
"logps/rejected": -54.73249053955078,
"loss": 0.6823,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.004845681134611368,
"rewards/margins": 0.02257709763944149,
"rewards/rejected": -0.02742278017103672,
"step": 3650
},
{
"epoch": 2.636887608069164,
"grad_norm": 3.726067066192627,
"learning_rate": 2.199031051284972e-09,
"logits/chosen": -1.4994523525238037,
"logits/rejected": -1.4987690448760986,
"logps/chosen": -48.3768196105957,
"logps/rejected": -52.434532165527344,
"loss": 0.6837,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.004970946349203587,
"rewards/margins": 0.020046500489115715,
"rewards/rejected": -0.025017445906996727,
"step": 3660
},
{
"epoch": 2.6440922190201728,
"grad_norm": 3.8218469619750977,
"learning_rate": 2.113872712509254e-09,
"logits/chosen": -1.4067411422729492,
"logits/rejected": -1.3974635601043701,
"logps/chosen": -56.22896194458008,
"logps/rejected": -59.406829833984375,
"loss": 0.6831,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.009745483286678791,
"rewards/margins": 0.020852208137512207,
"rewards/rejected": -0.030597690492868423,
"step": 3670
},
{
"epoch": 2.6512968299711814,
"grad_norm": 3.4928946495056152,
"learning_rate": 2.0303231739801143e-09,
"logits/chosen": -1.410017490386963,
"logits/rejected": -1.3968251943588257,
"logps/chosen": -50.764427185058594,
"logps/rejected": -55.013816833496094,
"loss": 0.6845,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.009240074083209038,
"rewards/margins": 0.01817367412149906,
"rewards/rejected": -0.02741374634206295,
"step": 3680
},
{
"epoch": 2.65850144092219,
"grad_norm": 3.9155476093292236,
"learning_rate": 1.948388308885102e-09,
"logits/chosen": -1.573972225189209,
"logits/rejected": -1.558861255645752,
"logps/chosen": -50.125404357910156,
"logps/rejected": -53.09186553955078,
"loss": 0.6849,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.003737266408279538,
"rewards/margins": 0.01738080568611622,
"rewards/rejected": -0.02111807093024254,
"step": 3690
},
{
"epoch": 2.6657060518731988,
"grad_norm": 3.190427780151367,
"learning_rate": 1.86807387690692e-09,
"logits/chosen": -1.5523760318756104,
"logits/rejected": -1.5453031063079834,
"logps/chosen": -50.241458892822266,
"logps/rejected": -57.71672821044922,
"loss": 0.6787,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.0018135461723431945,
"rewards/margins": 0.03013269044458866,
"rewards/rejected": -0.03194623440504074,
"step": 3700
},
{
"epoch": 2.6729106628242074,
"grad_norm": 3.522392511367798,
"learning_rate": 1.789385523818493e-09,
"logits/chosen": -1.4759515523910522,
"logits/rejected": -1.4781670570373535,
"logps/chosen": -45.24992370605469,
"logps/rejected": -51.2096061706543,
"loss": 0.6822,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.0035810978151857853,
"rewards/margins": 0.022562062367796898,
"rewards/rejected": -0.026143159717321396,
"step": 3710
},
{
"epoch": 2.680115273775216,
"grad_norm": 3.612794876098633,
"learning_rate": 1.712328781086131e-09,
"logits/chosen": -1.5478874444961548,
"logits/rejected": -1.531994104385376,
"logps/chosen": -51.03803634643555,
"logps/rejected": -53.24555206298828,
"loss": 0.6864,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.007410322315990925,
"rewards/margins": 0.014138393104076385,
"rewards/rejected": -0.021548714488744736,
"step": 3720
},
{
"epoch": 2.6873198847262247,
"grad_norm": 3.423640727996826,
"learning_rate": 1.6369090654806543e-09,
"logits/chosen": -1.5726535320281982,
"logits/rejected": -1.5603439807891846,
"logps/chosen": -46.89988708496094,
"logps/rejected": -51.72241973876953,
"loss": 0.6846,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.008618971332907677,
"rewards/margins": 0.017804089933633804,
"rewards/rejected": -0.026423057541251183,
"step": 3730
},
{
"epoch": 2.6945244956772334,
"grad_norm": 3.223950147628784,
"learning_rate": 1.5631316786966498e-09,
"logits/chosen": -1.4826900959014893,
"logits/rejected": -1.4667797088623047,
"logps/chosen": -45.12305450439453,
"logps/rejected": -48.53042984008789,
"loss": 0.6849,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.006777583155781031,
"rewards/margins": 0.01729729399085045,
"rewards/rejected": -0.024074876680970192,
"step": 3740
},
{
"epoch": 2.701729106628242,
"grad_norm": 4.128338813781738,
"learning_rate": 1.491001806979772e-09,
"logits/chosen": -1.5129797458648682,
"logits/rejected": -1.498471736907959,
"logps/chosen": -50.177757263183594,
"logps/rejected": -54.36769485473633,
"loss": 0.6838,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.0021677894983440638,
"rewards/margins": 0.019533688202500343,
"rewards/rejected": -0.021701481193304062,
"step": 3750
},
{
"epoch": 2.7089337175792507,
"grad_norm": 3.762078285217285,
"learning_rate": 1.4205245207621508e-09,
"logits/chosen": -1.4353492259979248,
"logits/rejected": -1.4193631410598755,
"logps/chosen": -52.8912239074707,
"logps/rejected": -55.724334716796875,
"loss": 0.682,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.0020478595979511738,
"rewards/margins": 0.023334335535764694,
"rewards/rejected": -0.02538219466805458,
"step": 3760
},
{
"epoch": 2.7161383285302594,
"grad_norm": 3.871379852294922,
"learning_rate": 1.3517047743059978e-09,
"logits/chosen": -1.5186526775360107,
"logits/rejected": -1.5207148790359497,
"logps/chosen": -49.53696060180664,
"logps/rejected": -55.4676513671875,
"loss": 0.6838,
"rewards/accuracies": 0.6812499761581421,
"rewards/chosen": -0.0054653664119541645,
"rewards/margins": 0.01927504874765873,
"rewards/rejected": -0.024740414693951607,
"step": 3770
},
{
"epoch": 2.723342939481268,
"grad_norm": 3.295902729034424,
"learning_rate": 1.2845474053553156e-09,
"logits/chosen": -1.5167324542999268,
"logits/rejected": -1.5084768533706665,
"logps/chosen": -43.50970458984375,
"logps/rejected": -47.10810089111328,
"loss": 0.685,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.007235602475702763,
"rewards/margins": 0.01710323989391327,
"rewards/rejected": -0.024338845163583755,
"step": 3780
},
{
"epoch": 2.7305475504322767,
"grad_norm": 2.8286335468292236,
"learning_rate": 1.2190571347958422e-09,
"logits/chosen": -1.5425716638565063,
"logits/rejected": -1.5455642938613892,
"logps/chosen": -43.28416061401367,
"logps/rejected": -50.13677978515625,
"loss": 0.6843,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.0002638758742250502,
"rewards/margins": 0.018405336886644363,
"rewards/rejected": -0.018669212237000465,
"step": 3790
},
{
"epoch": 2.7377521613832854,
"grad_norm": 2.9479782581329346,
"learning_rate": 1.1552385663231634e-09,
"logits/chosen": -1.4791892766952515,
"logits/rejected": -1.457729458808899,
"logps/chosen": -48.17145538330078,
"logps/rejected": -50.11743927001953,
"loss": 0.6851,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.005142406094819307,
"rewards/margins": 0.016895312815904617,
"rewards/rejected": -0.022037718445062637,
"step": 3800
},
{
"epoch": 2.744956772334294,
"grad_norm": 3.1124942302703857,
"learning_rate": 1.0930961861191302e-09,
"logits/chosen": -1.4413386583328247,
"logits/rejected": -1.4406673908233643,
"logps/chosen": -46.42212677001953,
"logps/rejected": -49.962947845458984,
"loss": 0.6863,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.007161378860473633,
"rewards/margins": 0.014661896042525768,
"rewards/rejected": -0.021823275834321976,
"step": 3810
},
{
"epoch": 2.7521613832853027,
"grad_norm": 3.0346012115478516,
"learning_rate": 1.0326343625364608e-09,
"logits/chosen": -1.4345784187316895,
"logits/rejected": -1.4189555644989014,
"logps/chosen": -47.113319396972656,
"logps/rejected": -52.565155029296875,
"loss": 0.6807,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.005450129974633455,
"rewards/margins": 0.02591022290289402,
"rewards/rejected": -0.03136035054922104,
"step": 3820
},
{
"epoch": 2.7593659942363113,
"grad_norm": 2.6539506912231445,
"learning_rate": 9.738573457917066e-10,
"logits/chosen": -1.5480695962905884,
"logits/rejected": -1.5419654846191406,
"logps/chosen": -41.202308654785156,
"logps/rejected": -47.30192184448242,
"loss": 0.6825,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.005542878992855549,
"rewards/margins": 0.022122148424386978,
"rewards/rejected": -0.0276650283485651,
"step": 3830
},
{
"epoch": 2.76657060518732,
"grad_norm": 3.010223865509033,
"learning_rate": 9.16769267666434e-10,
"logits/chosen": -1.466994047164917,
"logits/rejected": -1.4610474109649658,
"logps/chosen": -46.274566650390625,
"logps/rejected": -48.29325866699219,
"loss": 0.6887,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.006465951446443796,
"rewards/margins": 0.009437872096896172,
"rewards/rejected": -0.01590382307767868,
"step": 3840
},
{
"epoch": 2.7737752161383287,
"grad_norm": 3.2893083095550537,
"learning_rate": 8.613741412168113e-10,
"logits/chosen": -1.4859510660171509,
"logits/rejected": -1.4807151556015015,
"logps/chosen": -54.31746292114258,
"logps/rejected": -58.5187873840332,
"loss": 0.6831,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.003899764269590378,
"rewards/margins": 0.02083163894712925,
"rewards/rejected": -0.024731403216719627,
"step": 3850
},
{
"epoch": 2.7809798270893373,
"grad_norm": 3.3477933406829834,
"learning_rate": 8.076758604914802e-10,
"logits/chosen": -1.4456332921981812,
"logits/rejected": -1.4328854084014893,
"logps/chosen": -43.15898895263672,
"logps/rejected": -46.71881866455078,
"loss": 0.6843,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.0022692340426146984,
"rewards/margins": 0.018418530002236366,
"rewards/rejected": -0.020687762647867203,
"step": 3860
},
{
"epoch": 2.7881844380403455,
"grad_norm": 4.6873016357421875,
"learning_rate": 7.55678200257856e-10,
"logits/chosen": -1.442856788635254,
"logits/rejected": -1.4301444292068481,
"logps/chosen": -50.084381103515625,
"logps/rejected": -55.53865432739258,
"loss": 0.6827,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.006987369619309902,
"rewards/margins": 0.02188916876912117,
"rewards/rejected": -0.028876539319753647,
"step": 3870
},
{
"epoch": 2.795389048991354,
"grad_norm": 3.2509396076202393,
"learning_rate": 7.053848157367315e-10,
"logits/chosen": -1.4659183025360107,
"logits/rejected": -1.4521931409835815,
"logps/chosen": -48.176673889160156,
"logps/rejected": -53.21508026123047,
"loss": 0.6827,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.001605423167347908,
"rewards/margins": 0.021672086790204048,
"rewards/rejected": -0.023277509957551956,
"step": 3880
},
{
"epoch": 2.802593659942363,
"grad_norm": 2.586207389831543,
"learning_rate": 6.567992423453794e-10,
"logits/chosen": -1.4940398931503296,
"logits/rejected": -1.4875811338424683,
"logps/chosen": -43.39032745361328,
"logps/rejected": -46.69108581542969,
"loss": 0.6838,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.004483362659811974,
"rewards/margins": 0.019497577100992203,
"rewards/rejected": -0.023980939760804176,
"step": 3890
},
{
"epoch": 2.8097982708933715,
"grad_norm": 3.191453695297241,
"learning_rate": 6.099248954489794e-10,
"logits/chosen": -1.4086828231811523,
"logits/rejected": -1.4071089029312134,
"logps/chosen": -47.94374465942383,
"logps/rejected": -53.171348571777344,
"loss": 0.6833,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.007214091718196869,
"rewards/margins": 0.020340237766504288,
"rewards/rejected": -0.027554329484701157,
"step": 3900
},
{
"epoch": 2.81700288184438,
"grad_norm": 3.6769521236419678,
"learning_rate": 5.647650701205653e-10,
"logits/chosen": -1.5014244318008423,
"logits/rejected": -1.4815281629562378,
"logps/chosen": -54.42211151123047,
"logps/rejected": -58.34537887573242,
"loss": 0.6805,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.00018626952078193426,
"rewards/margins": 0.026389459148049355,
"rewards/rejected": -0.026575729250907898,
"step": 3910
},
{
"epoch": 2.824207492795389,
"grad_norm": 3.143864631652832,
"learning_rate": 5.213229409093856e-10,
"logits/chosen": -1.5347833633422852,
"logits/rejected": -1.5241453647613525,
"logps/chosen": -52.7486572265625,
"logps/rejected": -57.841835021972656,
"loss": 0.6815,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.004956572316586971,
"rewards/margins": 0.02456669509410858,
"rewards/rejected": -0.029523268342018127,
"step": 3920
},
{
"epoch": 2.8314121037463975,
"grad_norm": 4.357008457183838,
"learning_rate": 4.796015616177401e-10,
"logits/chosen": -1.4576940536499023,
"logits/rejected": -1.4457906484603882,
"logps/chosen": -51.851417541503906,
"logps/rejected": -55.67144012451172,
"loss": 0.6854,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.007598734460771084,
"rewards/margins": 0.016234243288636208,
"rewards/rejected": -0.023832976818084717,
"step": 3930
},
{
"epoch": 2.838616714697406,
"grad_norm": 3.335747480392456,
"learning_rate": 4.3960386508631595e-10,
"logits/chosen": -1.3839704990386963,
"logits/rejected": -1.383461356163025,
"logps/chosen": -42.68264389038086,
"logps/rejected": -46.677146911621094,
"loss": 0.6852,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.007006730884313583,
"rewards/margins": 0.016936389729380608,
"rewards/rejected": -0.02394312247633934,
"step": 3940
},
{
"epoch": 2.845821325648415,
"grad_norm": 4.871637344360352,
"learning_rate": 4.013326629880243e-10,
"logits/chosen": -1.4295424222946167,
"logits/rejected": -1.4122328758239746,
"logps/chosen": -50.189979553222656,
"logps/rejected": -54.05989456176758,
"loss": 0.6825,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.007883711718022823,
"rewards/margins": 0.022242117673158646,
"rewards/rejected": -0.030125826597213745,
"step": 3950
},
{
"epoch": 2.8530259365994235,
"grad_norm": 3.459960460662842,
"learning_rate": 3.64790645630339e-10,
"logits/chosen": -1.3913426399230957,
"logits/rejected": -1.3863009214401245,
"logps/chosen": -53.34258270263672,
"logps/rejected": -55.69043731689453,
"loss": 0.6872,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.0019184326520189643,
"rewards/margins": 0.012400278821587563,
"rewards/rejected": -0.014318712055683136,
"step": 3960
},
{
"epoch": 2.860230547550432,
"grad_norm": 4.9613542556762695,
"learning_rate": 3.2998038176619e-10,
"logits/chosen": -1.4524660110473633,
"logits/rejected": -1.4362642765045166,
"logps/chosen": -51.490867614746094,
"logps/rejected": -54.9944953918457,
"loss": 0.6849,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.007123004645109177,
"rewards/margins": 0.01731189154088497,
"rewards/rejected": -0.024434898048639297,
"step": 3970
},
{
"epoch": 2.867435158501441,
"grad_norm": 3.5082051753997803,
"learning_rate": 2.969043184133907e-10,
"logits/chosen": -1.5575920343399048,
"logits/rejected": -1.5562455654144287,
"logps/chosen": -45.0047721862793,
"logps/rejected": -53.420326232910156,
"loss": 0.6815,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.000279056781437248,
"rewards/margins": 0.02413121983408928,
"rewards/rejected": -0.024410273879766464,
"step": 3980
},
{
"epoch": 2.8746397694524495,
"grad_norm": 3.8809545040130615,
"learning_rate": 2.6556478068261447e-10,
"logits/chosen": -1.4493497610092163,
"logits/rejected": -1.4356361627578735,
"logps/chosen": -44.4965934753418,
"logps/rejected": -48.013553619384766,
"loss": 0.6796,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": 0.0006090818787924945,
"rewards/margins": 0.028188396245241165,
"rewards/rejected": -0.02757931686937809,
"step": 3990
},
{
"epoch": 2.881844380403458,
"grad_norm": 3.4924869537353516,
"learning_rate": 2.3596397161395607e-10,
"logits/chosen": -1.558559775352478,
"logits/rejected": -1.5367920398712158,
"logps/chosen": -49.601959228515625,
"logps/rejected": -54.7277946472168,
"loss": 0.6809,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": 0.0002869130694307387,
"rewards/margins": 0.025376971811056137,
"rewards/rejected": -0.025090059265494347,
"step": 4000
},
{
"epoch": 2.889048991354467,
"grad_norm": 4.832521915435791,
"learning_rate": 2.0810397202206399e-10,
"logits/chosen": -1.4147446155548096,
"logits/rejected": -1.4100733995437622,
"logps/chosen": -49.89708709716797,
"logps/rejected": -53.32837677001953,
"loss": 0.6846,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": 0.0006277470965869725,
"rewards/margins": 0.01780160702764988,
"rewards/rejected": -0.01717386022210121,
"step": 4010
},
{
"epoch": 2.8962536023054755,
"grad_norm": 3.2794227600097656,
"learning_rate": 1.819867403498737e-10,
"logits/chosen": -1.5649325847625732,
"logits/rejected": -1.5557774305343628,
"logps/chosen": -47.87010955810547,
"logps/rejected": -51.79607391357422,
"loss": 0.6835,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.00826673861593008,
"rewards/margins": 0.020225917920470238,
"rewards/rejected": -0.028492655605077744,
"step": 4020
},
{
"epoch": 2.903458213256484,
"grad_norm": 3.472938299179077,
"learning_rate": 1.5761411253092382e-10,
"logits/chosen": -1.430612325668335,
"logits/rejected": -1.4091012477874756,
"logps/chosen": -46.053260803222656,
"logps/rejected": -48.203346252441406,
"loss": 0.6843,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.006684750318527222,
"rewards/margins": 0.01832910254597664,
"rewards/rejected": -0.02501385286450386,
"step": 4030
},
{
"epoch": 2.910662824207493,
"grad_norm": 3.624494791030884,
"learning_rate": 1.3498780186031455e-10,
"logits/chosen": -1.4939110279083252,
"logits/rejected": -1.4849644899368286,
"logps/chosen": -53.622901916503906,
"logps/rejected": -57.3227653503418,
"loss": 0.6845,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.006433118134737015,
"rewards/margins": 0.018010510131716728,
"rewards/rejected": -0.024443628266453743,
"step": 4040
},
{
"epoch": 2.9178674351585014,
"grad_norm": 3.310802936553955,
"learning_rate": 1.1410939887425141e-10,
"logits/chosen": -1.49782133102417,
"logits/rejected": -1.4898258447647095,
"logps/chosen": -47.12517166137695,
"logps/rejected": -49.63444137573242,
"loss": 0.6862,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.010405481792986393,
"rewards/margins": 0.014516057446599007,
"rewards/rejected": -0.024921538308262825,
"step": 4050
},
{
"epoch": 2.92507204610951,
"grad_norm": 2.949910879135132,
"learning_rate": 9.498037123825686e-11,
"logits/chosen": -1.5098581314086914,
"logits/rejected": -1.498915433883667,
"logps/chosen": -45.20121383666992,
"logps/rejected": -49.48499298095703,
"loss": 0.6833,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.004668011330068111,
"rewards/margins": 0.0204728152602911,
"rewards/rejected": -0.025140831246972084,
"step": 4060
},
{
"epoch": 2.9322766570605188,
"grad_norm": 3.276839256286621,
"learning_rate": 7.760206364398614e-11,
"logits/chosen": -1.5862383842468262,
"logits/rejected": -1.5651108026504517,
"logps/chosen": -49.96502685546875,
"logps/rejected": -53.223594665527344,
"loss": 0.6837,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.009547281078994274,
"rewards/margins": 0.019979404285550117,
"rewards/rejected": -0.029526684433221817,
"step": 4070
},
{
"epoch": 2.9394812680115274,
"grad_norm": 3.859468698501587,
"learning_rate": 6.19756977147029e-11,
"logits/chosen": -1.4423165321350098,
"logits/rejected": -1.4334721565246582,
"logps/chosen": -47.239994049072266,
"logps/rejected": -54.47491455078125,
"loss": 0.6829,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.009559379890561104,
"rewards/margins": 0.021419430151581764,
"rewards/rejected": -0.03097880817949772,
"step": 4080
},
{
"epoch": 2.946685878962536,
"grad_norm": 2.8501765727996826,
"learning_rate": 4.810237191940625e-11,
"logits/chosen": -1.4424220323562622,
"logits/rejected": -1.4332430362701416,
"logps/chosen": -46.96299362182617,
"logps/rejected": -49.96550369262695,
"loss": 0.6856,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.008839382790029049,
"rewards/margins": 0.01596887595951557,
"rewards/rejected": -0.024808257818222046,
"step": 4090
},
{
"epoch": 2.9538904899135447,
"grad_norm": 3.3934149742126465,
"learning_rate": 3.5983061495617476e-11,
"logits/chosen": -1.5262759923934937,
"logits/rejected": -1.5253779888153076,
"logps/chosen": -51.821449279785156,
"logps/rejected": -57.49546432495117,
"loss": 0.684,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.005071837455034256,
"rewards/margins": 0.019206812605261803,
"rewards/rejected": -0.02427865006029606,
"step": 4100
},
{
"epoch": 2.9610951008645534,
"grad_norm": 3.1180672645568848,
"learning_rate": 2.5618618380812694e-11,
"logits/chosen": -1.520560622215271,
"logits/rejected": -1.5057575702667236,
"logps/chosen": -42.085235595703125,
"logps/rejected": -47.49565887451172,
"loss": 0.6811,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -0.0040741474367678165,
"rewards/margins": 0.02492835372686386,
"rewards/rejected": -0.029002506285905838,
"step": 4110
},
{
"epoch": 2.968299711815562,
"grad_norm": 3.4017624855041504,
"learning_rate": 1.700977115254576e-11,
"logits/chosen": -1.4638285636901855,
"logits/rejected": -1.454246163368225,
"logps/chosen": -46.245887756347656,
"logps/rejected": -51.503753662109375,
"loss": 0.6831,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.007428421638906002,
"rewards/margins": 0.02083088457584381,
"rewards/rejected": -0.028259307146072388,
"step": 4120
},
{
"epoch": 2.9755043227665707,
"grad_norm": 2.984016180038452,
"learning_rate": 1.0157124977230868e-11,
"logits/chosen": -1.4343056678771973,
"logits/rejected": -1.4250215291976929,
"logps/chosen": -43.65608596801758,
"logps/rejected": -47.83146667480469,
"loss": 0.6845,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.0026060601230710745,
"rewards/margins": 0.01812109351158142,
"rewards/rejected": -0.02072715200483799,
"step": 4130
},
{
"epoch": 2.9827089337175794,
"grad_norm": 3.555567741394043,
"learning_rate": 5.061161567596061e-12,
"logits/chosen": -1.4683890342712402,
"logits/rejected": -1.4556920528411865,
"logps/chosen": -47.792022705078125,
"logps/rejected": -50.41429901123047,
"loss": 0.6844,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0011950184125453234,
"rewards/margins": 0.018146729096770287,
"rewards/rejected": -0.01934174820780754,
"step": 4140
},
{
"epoch": 2.989913544668588,
"grad_norm": 3.3906075954437256,
"learning_rate": 1.7222391488297406e-12,
"logits/chosen": -1.5164161920547485,
"logits/rejected": -1.5046594142913818,
"logps/chosen": -53.657630920410156,
"logps/rejected": -58.71654510498047,
"loss": 0.6793,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.003981665708124638,
"rewards/margins": 0.029033973813056946,
"rewards/rejected": -0.03301564231514931,
"step": 4150
},
{
"epoch": 2.9971181556195967,
"grad_norm": 3.9542465209960938,
"learning_rate": 1.4059243338693238e-13,
"logits/chosen": -1.4423478841781616,
"logits/rejected": -1.4315330982208252,
"logps/chosen": -48.669158935546875,
"logps/rejected": -53.47356033325195,
"loss": 0.6825,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.0017929436871781945,
"rewards/margins": 0.02205517329275608,
"rewards/rejected": -0.023848116397857666,
"step": 4160
},
{
"epoch": 3.0,
"step": 4164,
"total_flos": 0.0,
"train_loss": 0.6880504754617968,
"train_runtime": 7518.2292,
"train_samples_per_second": 8.859,
"train_steps_per_second": 0.554
}
],
"logging_steps": 10,
"max_steps": 4164,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"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": 8,
"trial_name": null,
"trial_params": null
}