Open-RS-DR_GRPO-8B / trainer_state.json
kangdawei's picture
Model save
55409f7 verified
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.5714285714285714,
"eval_steps": 500,
"global_step": 500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"completion_length": 2334.9792137145996,
"epoch": 0.001142857142857143,
"grad_norm": 0.04168504849076271,
"kl": 0.0,
"learning_rate": 0.0,
"loss": 0.0343,
"reward": 0.5176793523132801,
"reward_std": 0.7803475465625525,
"rewards/cosine_scaled_reward": -0.011993663385510445,
"rewards/format_reward": 0.5416666772216558,
"step": 1
},
{
"completion_length": 2554.7708892822266,
"epoch": 0.002285714285714286,
"grad_norm": 0.0329468809068203,
"kl": 0.0,
"learning_rate": 2e-08,
"loss": 0.048,
"reward": 0.544984623324126,
"reward_std": 0.8168745078146458,
"rewards/cosine_scaled_reward": 0.05374230569577776,
"rewards/format_reward": 0.43750000186264515,
"step": 2
},
{
"completion_length": 2859.1875076293945,
"epoch": 0.0034285714285714284,
"grad_norm": 0.023289382457733154,
"kl": 5.08427619934082e-05,
"learning_rate": 4e-08,
"loss": 0.0051,
"reward": -0.02553212083876133,
"reward_std": 0.48618486896157265,
"rewards/cosine_scaled_reward": -0.1481827348470688,
"rewards/format_reward": 0.27083333395421505,
"step": 3
},
{
"completion_length": 1444.7500267028809,
"epoch": 0.004571428571428572,
"grad_norm": 0.03743300959467888,
"kl": 4.610419273376465e-05,
"learning_rate": 6e-08,
"loss": -0.0001,
"reward": 0.7746477648615837,
"reward_std": 0.7952391989529133,
"rewards/cosine_scaled_reward": -0.03975944733247161,
"rewards/format_reward": 0.8541666716337204,
"step": 4
},
{
"completion_length": 3115.9791870117188,
"epoch": 0.005714285714285714,
"grad_norm": 0.03854230046272278,
"kl": 5.6996941566467285e-05,
"learning_rate": 8e-08,
"loss": 0.0264,
"reward": -0.07823497732169926,
"reward_std": 0.7049717996269464,
"rewards/cosine_scaled_reward": -0.18495082901790738,
"rewards/format_reward": 0.29166666977107525,
"step": 5
},
{
"completion_length": 2771.9166946411133,
"epoch": 0.006857142857142857,
"grad_norm": 0.035347286611795425,
"kl": 5.025416612625122e-05,
"learning_rate": 1e-07,
"loss": 0.0071,
"reward": 0.07868877053260803,
"reward_std": 0.4701222777366638,
"rewards/cosine_scaled_reward": -0.17940562218427658,
"rewards/format_reward": 0.4375,
"step": 6
},
{
"completion_length": 2687.6458740234375,
"epoch": 0.008,
"grad_norm": 0.027172546833753586,
"kl": 3.434717655181885e-05,
"learning_rate": 1.2e-07,
"loss": 0.0422,
"reward": 0.45066535100340843,
"reward_std": 0.5497629679739475,
"rewards/cosine_scaled_reward": -0.055917332880198956,
"rewards/format_reward": 0.5625000018626451,
"step": 7
},
{
"completion_length": 2470.6458587646484,
"epoch": 0.009142857142857144,
"grad_norm": 0.039594072848558426,
"kl": 4.009902477264404e-05,
"learning_rate": 1.4e-07,
"loss": -0.0246,
"reward": 0.9416400007903576,
"reward_std": 0.8407168909907341,
"rewards/cosine_scaled_reward": 0.1583199892193079,
"rewards/format_reward": 0.6250000055879354,
"step": 8
},
{
"completion_length": 2876.791717529297,
"epoch": 0.010285714285714285,
"grad_norm": 0.043148480355739594,
"kl": 5.060434341430664e-05,
"learning_rate": 1.6e-07,
"loss": 0.0566,
"reward": 0.06330016255378723,
"reward_std": 0.8118741400539875,
"rewards/cosine_scaled_reward": -0.13501658383756876,
"rewards/format_reward": 0.33333333767950535,
"step": 9
},
{
"completion_length": 2594.583381652832,
"epoch": 0.011428571428571429,
"grad_norm": 0.038916755467653275,
"kl": 3.3721327781677246e-05,
"learning_rate": 1.8e-07,
"loss": 0.0571,
"reward": 0.6445693820714951,
"reward_std": 0.703135633841157,
"rewards/cosine_scaled_reward": 0.11395137198269367,
"rewards/format_reward": 0.4166666679084301,
"step": 10
},
{
"completion_length": 3401.9166870117188,
"epoch": 0.012571428571428572,
"grad_norm": 0.034626927226781845,
"kl": 4.112720489501953e-05,
"learning_rate": 2e-07,
"loss": 0.0237,
"reward": -0.26008480228483677,
"reward_std": 0.52357386238873,
"rewards/cosine_scaled_reward": -0.21337573695927858,
"rewards/format_reward": 0.1666666716337204,
"step": 11
},
{
"completion_length": 1942.1250228881836,
"epoch": 0.013714285714285714,
"grad_norm": 0.033392105251550674,
"kl": 4.201382398605347e-05,
"learning_rate": 2.1999999999999998e-07,
"loss": 0.0569,
"reward": 0.6705481559038162,
"reward_std": 0.7221638858318329,
"rewards/cosine_scaled_reward": -0.02930926624685526,
"rewards/format_reward": 0.7291666772216558,
"step": 12
},
{
"completion_length": 2874.3541946411133,
"epoch": 0.014857142857142857,
"grad_norm": 0.019545605406165123,
"kl": 3.758817911148071e-05,
"learning_rate": 2.4e-07,
"loss": 0.0097,
"reward": 0.141155153978616,
"reward_std": 0.42822966538369656,
"rewards/cosine_scaled_reward": -0.07525575812906027,
"rewards/format_reward": 0.2916666679084301,
"step": 13
},
{
"completion_length": 2432.3750534057617,
"epoch": 0.016,
"grad_norm": 0.017046086490154266,
"kl": 2.7783215045928955e-05,
"learning_rate": 2.6e-07,
"loss": 0.0137,
"reward": 0.04899838753044605,
"reward_std": 0.42438305931864306,
"rewards/cosine_scaled_reward": -0.20466747391037643,
"rewards/format_reward": 0.4583333358168602,
"step": 14
},
{
"completion_length": 2732.5625038146973,
"epoch": 0.017142857142857144,
"grad_norm": 0.03251831606030464,
"kl": 4.025548696517944e-05,
"learning_rate": 2.8e-07,
"loss": -0.0008,
"reward": 0.6431202292442322,
"reward_std": 0.4675060287117958,
"rewards/cosine_scaled_reward": 0.11322677787393332,
"rewards/format_reward": 0.4166666716337204,
"step": 15
},
{
"completion_length": 3581.7916870117188,
"epoch": 0.018285714285714287,
"grad_norm": 0.029273949563503265,
"kl": 4.838407039642334e-05,
"learning_rate": 3e-07,
"loss": 0.0011,
"reward": -0.35534902289509773,
"reward_std": 0.5198515467345715,
"rewards/cosine_scaled_reward": -0.1985078384168446,
"rewards/format_reward": 0.0416666679084301,
"step": 16
},
{
"completion_length": 2177.9792251586914,
"epoch": 0.019428571428571427,
"grad_norm": 0.03885098919272423,
"kl": 5.97536563873291e-05,
"learning_rate": 3.2e-07,
"loss": 0.0438,
"reward": 0.4878217186778784,
"reward_std": 0.6091510504484177,
"rewards/cosine_scaled_reward": -0.026922473683953285,
"rewards/format_reward": 0.5416666716337204,
"step": 17
},
{
"completion_length": 2967.854217529297,
"epoch": 0.02057142857142857,
"grad_norm": 0.04610211029648781,
"kl": 3.8489699363708496e-05,
"learning_rate": 3.4000000000000003e-07,
"loss": 0.0619,
"reward": 0.5143160903826356,
"reward_std": 0.820105541497469,
"rewards/cosine_scaled_reward": 0.02799136098474264,
"rewards/format_reward": 0.45833333767950535,
"step": 18
},
{
"completion_length": 2860.062530517578,
"epoch": 0.021714285714285714,
"grad_norm": 0.040472596883773804,
"kl": 3.56137752532959e-05,
"learning_rate": 3.6e-07,
"loss": 0.0495,
"reward": 0.7190339863300323,
"reward_std": 0.6223967634141445,
"rewards/cosine_scaled_reward": 0.15118366852402687,
"rewards/format_reward": 0.4166666716337204,
"step": 19
},
{
"completion_length": 1933.0625495910645,
"epoch": 0.022857142857142857,
"grad_norm": 0.028757991269230843,
"kl": 3.4768134355545044e-05,
"learning_rate": 3.7999999999999996e-07,
"loss": 0.017,
"reward": 0.5407668675179593,
"reward_std": 0.6035287424456328,
"rewards/cosine_scaled_reward": -0.08378324937075377,
"rewards/format_reward": 0.7083333376795053,
"step": 20
},
{
"completion_length": 2637.354190826416,
"epoch": 0.024,
"grad_norm": 0.021128255873918533,
"kl": 4.544854164123535e-05,
"learning_rate": 4e-07,
"loss": 0.0244,
"reward": 0.29663072153925896,
"reward_std": 0.34191627195104957,
"rewards/cosine_scaled_reward": -0.028767995536327362,
"rewards/format_reward": 0.3541666716337204,
"step": 21
},
{
"completion_length": 1485.1041946411133,
"epoch": 0.025142857142857144,
"grad_norm": 0.024596063420176506,
"kl": 3.890693187713623e-05,
"learning_rate": 4.1999999999999995e-07,
"loss": 0.019,
"reward": 0.7402454526163638,
"reward_std": 0.44790306128561497,
"rewards/cosine_scaled_reward": -0.04654395952820778,
"rewards/format_reward": 0.8333333395421505,
"step": 22
},
{
"completion_length": 2294.4166870117188,
"epoch": 0.026285714285714287,
"grad_norm": 0.0354381762444973,
"kl": 5.1334500312805176e-05,
"learning_rate": 4.3999999999999997e-07,
"loss": 0.0626,
"reward": 0.5391949899494648,
"reward_std": 0.8765826672315598,
"rewards/cosine_scaled_reward": 0.009180818684399128,
"rewards/format_reward": 0.5208333395421505,
"step": 23
},
{
"completion_length": 2430.979217529297,
"epoch": 0.027428571428571427,
"grad_norm": 0.03837364539504051,
"kl": 2.7820002287626266e-05,
"learning_rate": 4.6e-07,
"loss": 0.0828,
"reward": 0.6404989643488079,
"reward_std": 0.7054540691897273,
"rewards/cosine_scaled_reward": 0.018166150897741318,
"rewards/format_reward": 0.6041666734963655,
"step": 24
},
{
"completion_length": 2256.9583702087402,
"epoch": 0.02857142857142857,
"grad_norm": 0.026419438421726227,
"kl": 5.500763654708862e-05,
"learning_rate": 4.8e-07,
"loss": -0.012,
"reward": 0.27087176591157913,
"reward_std": 0.62013322673738,
"rewards/cosine_scaled_reward": -0.12498080357909203,
"rewards/format_reward": 0.520833333954215,
"step": 25
},
{
"completion_length": 2868.500030517578,
"epoch": 0.029714285714285714,
"grad_norm": 0.028371552005410194,
"kl": 4.79966402053833e-05,
"learning_rate": 5e-07,
"loss": 0.0107,
"reward": 0.24959533289074898,
"reward_std": 0.6043985933065414,
"rewards/cosine_scaled_reward": -0.10436900053173304,
"rewards/format_reward": 0.45833334140479565,
"step": 26
},
{
"completion_length": 2872.520851135254,
"epoch": 0.030857142857142857,
"grad_norm": 0.03675654157996178,
"kl": 5.967915058135986e-05,
"learning_rate": 5.2e-07,
"loss": 0.0283,
"reward": 0.0008413083851337433,
"reward_std": 0.5718730166554451,
"rewards/cosine_scaled_reward": -0.17666268534958363,
"rewards/format_reward": 0.3541666753590107,
"step": 27
},
{
"completion_length": 2665.0416870117188,
"epoch": 0.032,
"grad_norm": 0.039023082703351974,
"kl": 4.050694406032562e-05,
"learning_rate": 5.4e-07,
"loss": 0.0179,
"reward": 0.4467268669977784,
"reward_std": 0.684929771348834,
"rewards/cosine_scaled_reward": 0.00461340369656682,
"rewards/format_reward": 0.4375000074505806,
"step": 28
},
{
"completion_length": 2884.125030517578,
"epoch": 0.03314285714285714,
"grad_norm": 0.03789398819208145,
"kl": 4.595518112182617e-05,
"learning_rate": 5.6e-07,
"loss": 0.0371,
"reward": -0.05399216455407441,
"reward_std": 0.5582299008965492,
"rewards/cosine_scaled_reward": -0.1936627607792616,
"rewards/format_reward": 0.3333333469927311,
"step": 29
},
{
"completion_length": 2812.604217529297,
"epoch": 0.03428571428571429,
"grad_norm": 0.05193966627120972,
"kl": 4.1157007217407227e-05,
"learning_rate": 5.8e-07,
"loss": 0.0381,
"reward": 0.5208531729876995,
"reward_std": 1.0489587001502514,
"rewards/cosine_scaled_reward": 0.010426569730043411,
"rewards/format_reward": 0.5000000111758709,
"step": 30
},
{
"completion_length": 2949.8125228881836,
"epoch": 0.03542857142857143,
"grad_norm": 0.03635473921895027,
"kl": 4.1857361793518066e-05,
"learning_rate": 6e-07,
"loss": -0.0211,
"reward": 0.07118985429406166,
"reward_std": 0.7498712912201881,
"rewards/cosine_scaled_reward": -0.09982176125049591,
"rewards/format_reward": 0.2708333395421505,
"step": 31
},
{
"completion_length": 2791.0000610351562,
"epoch": 0.036571428571428574,
"grad_norm": 0.028467582538723946,
"kl": 4.823505878448486e-05,
"learning_rate": 6.2e-07,
"loss": 0.0302,
"reward": 0.47064110776409507,
"reward_std": 0.5171910058706999,
"rewards/cosine_scaled_reward": 0.016570553183555603,
"rewards/format_reward": 0.43750000558793545,
"step": 32
},
{
"completion_length": 3185.166702270508,
"epoch": 0.037714285714285714,
"grad_norm": 0.03910686448216438,
"kl": 4.856288433074951e-05,
"learning_rate": 6.4e-07,
"loss": 0.0154,
"reward": 0.18704182095825672,
"reward_std": 0.7629665322601795,
"rewards/cosine_scaled_reward": -0.10439574904739857,
"rewards/format_reward": 0.39583334140479565,
"step": 33
},
{
"completion_length": 2163.333351135254,
"epoch": 0.038857142857142854,
"grad_norm": 0.03700648993253708,
"kl": 4.020100459456444e-05,
"learning_rate": 6.6e-07,
"loss": 0.0315,
"reward": 1.0100164553150535,
"reward_std": 0.8100163154304028,
"rewards/cosine_scaled_reward": 0.21334155462682247,
"rewards/format_reward": 0.5833333432674408,
"step": 34
},
{
"completion_length": 3131.5208435058594,
"epoch": 0.04,
"grad_norm": 0.04378014802932739,
"kl": 6.04093074798584e-05,
"learning_rate": 6.800000000000001e-07,
"loss": 0.007,
"reward": -0.16884424164891243,
"reward_std": 0.6548870000988245,
"rewards/cosine_scaled_reward": -0.18858879059553146,
"rewards/format_reward": 0.2083333358168602,
"step": 35
},
{
"completion_length": 3189.0208435058594,
"epoch": 0.04114285714285714,
"grad_norm": 0.034937676042318344,
"kl": 5.710124969482422e-05,
"learning_rate": 7e-07,
"loss": 0.0264,
"reward": -0.3049023966304958,
"reward_std": 0.4520086422562599,
"rewards/cosine_scaled_reward": -0.25661787390708923,
"rewards/format_reward": 0.20833334140479565,
"step": 36
},
{
"completion_length": 3393.3541870117188,
"epoch": 0.04228571428571429,
"grad_norm": 0.031590525060892105,
"kl": 4.273653030395508e-05,
"learning_rate": 7.2e-07,
"loss": 0.0103,
"reward": -0.2297830693423748,
"reward_std": 0.4669038709253073,
"rewards/cosine_scaled_reward": -0.21905820816755295,
"rewards/format_reward": 0.2083333358168602,
"step": 37
},
{
"completion_length": 3292.7708587646484,
"epoch": 0.04342857142857143,
"grad_norm": 0.026573847979307175,
"kl": 4.4655054807662964e-05,
"learning_rate": 7.4e-07,
"loss": 0.0033,
"reward": -0.19998125731945038,
"reward_std": 0.48700529895722866,
"rewards/cosine_scaled_reward": -0.172907296102494,
"rewards/format_reward": 0.14583333395421505,
"step": 38
},
{
"completion_length": 2868.5417137145996,
"epoch": 0.044571428571428574,
"grad_norm": 0.03660503029823303,
"kl": 3.191828727722168e-05,
"learning_rate": 7.599999999999999e-07,
"loss": 0.0278,
"reward": 0.5126112666912377,
"reward_std": 0.5605221642181277,
"rewards/cosine_scaled_reward": 0.02713894983753562,
"rewards/format_reward": 0.4583333395421505,
"step": 39
},
{
"completion_length": 2619.3958892822266,
"epoch": 0.045714285714285714,
"grad_norm": 0.04862872138619423,
"kl": 3.953278064727783e-05,
"learning_rate": 7.799999999999999e-07,
"loss": 0.0356,
"reward": 0.4027645355090499,
"reward_std": 0.7209917511790991,
"rewards/cosine_scaled_reward": -0.05903442995622754,
"rewards/format_reward": 0.5208333469927311,
"step": 40
},
{
"completion_length": 3137.8750915527344,
"epoch": 0.046857142857142854,
"grad_norm": 0.04735239967703819,
"kl": 4.553794860839844e-05,
"learning_rate": 8e-07,
"loss": -0.0455,
"reward": 0.18608580250293016,
"reward_std": 0.7598953321576118,
"rewards/cosine_scaled_reward": -0.1152904350310564,
"rewards/format_reward": 0.4166666753590107,
"step": 41
},
{
"completion_length": 2854.625,
"epoch": 0.048,
"grad_norm": 0.018570901826024055,
"kl": 6.361305713653564e-05,
"learning_rate": 8.199999999999999e-07,
"loss": 0.0098,
"reward": -0.3359166495501995,
"reward_std": 0.2978406958281994,
"rewards/cosine_scaled_reward": -0.3137916624546051,
"rewards/format_reward": 0.2916666679084301,
"step": 42
},
{
"completion_length": 2866.2083740234375,
"epoch": 0.04914285714285714,
"grad_norm": 0.03492705151438713,
"kl": 4.682131111621857e-05,
"learning_rate": 8.399999999999999e-07,
"loss": 0.0498,
"reward": 0.017277991399168968,
"reward_std": 0.693747516721487,
"rewards/cosine_scaled_reward": -0.14761100709438324,
"rewards/format_reward": 0.3125000074505806,
"step": 43
},
{
"completion_length": 2410.562530517578,
"epoch": 0.05028571428571429,
"grad_norm": 0.020008008927106857,
"kl": 4.0531158447265625e-05,
"learning_rate": 8.599999999999999e-07,
"loss": 0.0014,
"reward": 0.5113720148801804,
"reward_std": 0.5214159078896046,
"rewards/cosine_scaled_reward": 0.04735266789793968,
"rewards/format_reward": 0.4166666679084301,
"step": 44
},
{
"completion_length": 3197.2291870117188,
"epoch": 0.05142857142857143,
"grad_norm": 0.03368664160370827,
"kl": 4.8667192459106445e-05,
"learning_rate": 8.799999999999999e-07,
"loss": 0.0349,
"reward": -0.1545441746711731,
"reward_std": 0.6294996133074164,
"rewards/cosine_scaled_reward": -0.20227208855794743,
"rewards/format_reward": 0.2500000037252903,
"step": 45
},
{
"completion_length": 3049.458366394043,
"epoch": 0.052571428571428575,
"grad_norm": 0.023678315803408623,
"kl": 4.738569259643555e-05,
"learning_rate": 9e-07,
"loss": 0.0258,
"reward": -0.08186815958470106,
"reward_std": 0.4881300590932369,
"rewards/cosine_scaled_reward": -0.17635074770078063,
"rewards/format_reward": 0.2708333395421505,
"step": 46
},
{
"completion_length": 2636.1458740234375,
"epoch": 0.053714285714285714,
"grad_norm": 0.04777181148529053,
"kl": 4.3764710426330566e-05,
"learning_rate": 9.2e-07,
"loss": 0.0374,
"reward": 0.3579790024086833,
"reward_std": 0.8919647391885519,
"rewards/cosine_scaled_reward": -0.08142716065049171,
"rewards/format_reward": 0.5208333432674408,
"step": 47
},
{
"completion_length": 2504.8125534057617,
"epoch": 0.054857142857142854,
"grad_norm": 0.040610894560813904,
"kl": 3.6403536796569824e-05,
"learning_rate": 9.399999999999999e-07,
"loss": 0.0161,
"reward": 0.3527611903846264,
"reward_std": 0.796766672283411,
"rewards/cosine_scaled_reward": -0.06320274842437357,
"rewards/format_reward": 0.47916666977107525,
"step": 48
},
{
"completion_length": 1906.791690826416,
"epoch": 0.056,
"grad_norm": 0.03556185960769653,
"kl": 2.3037195205688477e-05,
"learning_rate": 9.6e-07,
"loss": 0.0207,
"reward": 0.5905593386851251,
"reward_std": 0.6989730293862522,
"rewards/cosine_scaled_reward": -0.02763699647039175,
"rewards/format_reward": 0.645833345130086,
"step": 49
},
{
"completion_length": 3025.2500228881836,
"epoch": 0.05714285714285714,
"grad_norm": 0.03857988119125366,
"kl": 3.446638584136963e-05,
"learning_rate": 9.8e-07,
"loss": 0.0268,
"reward": 0.29927817918360233,
"reward_std": 0.555895734578371,
"rewards/cosine_scaled_reward": -0.017027591355144978,
"rewards/format_reward": 0.3333333395421505,
"step": 50
},
{
"completion_length": 2294.208335876465,
"epoch": 0.05828571428571429,
"grad_norm": 0.020732907578349113,
"kl": 4.1797757148742676e-05,
"learning_rate": 1e-06,
"loss": 0.0055,
"reward": 0.25655205361545086,
"reward_std": 0.4376220293343067,
"rewards/cosine_scaled_reward": -0.12172399461269379,
"rewards/format_reward": 0.5,
"step": 51
},
{
"completion_length": 2716.0834045410156,
"epoch": 0.05942857142857143,
"grad_norm": 0.0520336888730526,
"kl": 4.2897649109363556e-05,
"learning_rate": 9.999890338174275e-07,
"loss": 0.0296,
"reward": 0.6955370828509331,
"reward_std": 0.8111556768417358,
"rewards/cosine_scaled_reward": 0.07693520188331604,
"rewards/format_reward": 0.541666679084301,
"step": 52
},
{
"completion_length": 2616.000015258789,
"epoch": 0.060571428571428575,
"grad_norm": 0.04615948721766472,
"kl": 4.25875186920166e-05,
"learning_rate": 9.999561358041868e-07,
"loss": 0.0573,
"reward": 0.6107768043875694,
"reward_std": 1.0185654014348984,
"rewards/cosine_scaled_reward": 0.03455505130114034,
"rewards/format_reward": 0.5416666753590107,
"step": 53
},
{
"completion_length": 2352.0625381469727,
"epoch": 0.061714285714285715,
"grad_norm": 0.04578646644949913,
"kl": 2.933293581008911e-05,
"learning_rate": 9.999013075636804e-07,
"loss": 0.0947,
"reward": 0.9733524695038795,
"reward_std": 0.9531826749444008,
"rewards/cosine_scaled_reward": 0.1637595584616065,
"rewards/format_reward": 0.6458333432674408,
"step": 54
},
{
"completion_length": 2831.3958435058594,
"epoch": 0.06285714285714286,
"grad_norm": 0.04407193139195442,
"kl": 3.0428171157836914e-05,
"learning_rate": 9.998245517681593e-07,
"loss": 0.0082,
"reward": 0.6569849252700806,
"reward_std": 0.8474150542169809,
"rewards/cosine_scaled_reward": 0.08890912728384137,
"rewards/format_reward": 0.4791666753590107,
"step": 55
},
{
"completion_length": 3047.125015258789,
"epoch": 0.064,
"grad_norm": 0.026214739307761192,
"kl": 4.430115222930908e-05,
"learning_rate": 9.997258721585931e-07,
"loss": 0.0257,
"reward": -0.18962648510932922,
"reward_std": 0.5728549361228943,
"rewards/cosine_scaled_reward": -0.2302299104630947,
"rewards/format_reward": 0.2708333358168602,
"step": 56
},
{
"completion_length": 3043.875030517578,
"epoch": 0.06514285714285714,
"grad_norm": 0.03775238245725632,
"kl": 2.8438866138458252e-05,
"learning_rate": 9.996052735444862e-07,
"loss": 0.0011,
"reward": -0.02485821396112442,
"reward_std": 0.6774866338819265,
"rewards/cosine_scaled_reward": -0.1895124390721321,
"rewards/format_reward": 0.3541666679084301,
"step": 57
},
{
"completion_length": 2135.1042556762695,
"epoch": 0.06628571428571428,
"grad_norm": 0.04270089045166969,
"kl": 3.1717121601104736e-05,
"learning_rate": 9.994627618036452e-07,
"loss": 0.0935,
"reward": 0.6859063357114792,
"reward_std": 0.959196649491787,
"rewards/cosine_scaled_reward": 0.020036499947309494,
"rewards/format_reward": 0.6458333376795053,
"step": 58
},
{
"completion_length": 2852.1041717529297,
"epoch": 0.06742857142857143,
"grad_norm": 0.035348936915397644,
"kl": 2.6881694793701172e-05,
"learning_rate": 9.992983438818915e-07,
"loss": -0.0016,
"reward": -0.11129460483789444,
"reward_std": 0.6194134466350079,
"rewards/cosine_scaled_reward": -0.21189730428159237,
"rewards/format_reward": 0.31250000186264515,
"step": 59
},
{
"completion_length": 2817.291702270508,
"epoch": 0.06857142857142857,
"grad_norm": 0.044353168457746506,
"kl": 2.095848321914673e-05,
"learning_rate": 9.991120277927223e-07,
"loss": 0.0249,
"reward": 0.22440146282315254,
"reward_std": 0.9068924002349377,
"rewards/cosine_scaled_reward": -0.10654927371069789,
"rewards/format_reward": 0.43750000186264515,
"step": 60
},
{
"completion_length": 2894.875030517578,
"epoch": 0.06971428571428571,
"grad_norm": 0.0331726111471653,
"kl": 3.4576281905174255e-05,
"learning_rate": 9.989038226169207e-07,
"loss": 0.0075,
"reward": 0.6926130764186382,
"reward_std": 0.6958329901099205,
"rewards/cosine_scaled_reward": 0.07547319121658802,
"rewards/format_reward": 0.5416666716337204,
"step": 61
},
{
"completion_length": 2464.4791946411133,
"epoch": 0.07085714285714285,
"grad_norm": 0.03456748649477959,
"kl": 2.349168062210083e-05,
"learning_rate": 9.98673738502114e-07,
"loss": 0.0285,
"reward": 0.6481589786708355,
"reward_std": 0.683325620368123,
"rewards/cosine_scaled_reward": 0.06366281025111675,
"rewards/format_reward": 0.520833333954215,
"step": 62
},
{
"completion_length": 2122.625045776367,
"epoch": 0.072,
"grad_norm": 0.7239912748336792,
"kl": 2.844538539648056e-05,
"learning_rate": 9.98421786662277e-07,
"loss": 0.0578,
"reward": 0.7681009210646152,
"reward_std": 0.7113255299627781,
"rewards/cosine_scaled_reward": 0.019467118196189404,
"rewards/format_reward": 0.7291666828095913,
"step": 63
},
{
"completion_length": 2740.1875534057617,
"epoch": 0.07314285714285715,
"grad_norm": 0.037132617086172104,
"kl": 4.111975431442261e-05,
"learning_rate": 9.981479793771866e-07,
"loss": 0.0213,
"reward": 0.6051971167325974,
"reward_std": 0.7252629213035107,
"rewards/cosine_scaled_reward": 0.052598556503653526,
"rewards/format_reward": 0.5000000111758709,
"step": 64
},
{
"completion_length": 2566.7916946411133,
"epoch": 0.07428571428571429,
"grad_norm": 0.01891911029815674,
"kl": 2.2094696760177612e-05,
"learning_rate": 9.97852329991824e-07,
"loss": -0.0092,
"reward": -0.04870823957026005,
"reward_std": 0.48125316202640533,
"rewards/cosine_scaled_reward": -0.25352078676223755,
"rewards/format_reward": 0.4583333358168602,
"step": 65
},
{
"completion_length": 2060.4375038146973,
"epoch": 0.07542857142857143,
"grad_norm": 0.016883157193660736,
"kl": 2.9248185455799103e-05,
"learning_rate": 9.975348529157229e-07,
"loss": -0.0004,
"reward": 0.33483995124697685,
"reward_std": 0.5551962554454803,
"rewards/cosine_scaled_reward": -0.08258003555238247,
"rewards/format_reward": 0.5,
"step": 66
},
{
"completion_length": 3343.9791870117188,
"epoch": 0.07657142857142857,
"grad_norm": 0.019418727606534958,
"kl": 2.110004425048828e-05,
"learning_rate": 9.971955636222684e-07,
"loss": 0.0382,
"reward": -0.4314758640830405,
"reward_std": 0.37546689808368683,
"rewards/cosine_scaled_reward": -0.2678212709724903,
"rewards/format_reward": 0.10416666977107525,
"step": 67
},
{
"completion_length": 1701.6458358764648,
"epoch": 0.07771428571428571,
"grad_norm": 0.02836545556783676,
"kl": 2.9101967811584473e-05,
"learning_rate": 9.968344786479415e-07,
"loss": -0.0354,
"reward": 0.6903760433197021,
"reward_std": 0.8633946646004915,
"rewards/cosine_scaled_reward": -0.019395311828702688,
"rewards/format_reward": 0.7291666716337204,
"step": 68
},
{
"completion_length": 2116.8333740234375,
"epoch": 0.07885714285714286,
"grad_norm": 0.027675610035657883,
"kl": 2.561137080192566e-05,
"learning_rate": 9.964516155915151e-07,
"loss": 0.0049,
"reward": 0.3666526600718498,
"reward_std": 0.5472365878522396,
"rewards/cosine_scaled_reward": -0.12917369417846203,
"rewards/format_reward": 0.6250000037252903,
"step": 69
},
{
"completion_length": 2915.4583587646484,
"epoch": 0.08,
"grad_norm": 0.033280231058597565,
"kl": 2.2306106984615326e-05,
"learning_rate": 9.960469931131936e-07,
"loss": 0.0425,
"reward": 0.07917704619467258,
"reward_std": 0.48793724877759814,
"rewards/cosine_scaled_reward": -0.13749481900595129,
"rewards/format_reward": 0.3541666716337204,
"step": 70
},
{
"completion_length": 2776.1458587646484,
"epoch": 0.08114285714285714,
"grad_norm": 0.02456558682024479,
"kl": 1.961737871170044e-05,
"learning_rate": 9.956206309337066e-07,
"loss": 0.0067,
"reward": 0.22148176468908787,
"reward_std": 0.5928666815161705,
"rewards/cosine_scaled_reward": -0.04550914093852043,
"rewards/format_reward": 0.31250000186264515,
"step": 71
},
{
"completion_length": 2835.916702270508,
"epoch": 0.08228571428571428,
"grad_norm": 0.032275594770908356,
"kl": 4.030764102935791e-05,
"learning_rate": 9.951725498333448e-07,
"loss": 0.03,
"reward": 0.0798844201490283,
"reward_std": 0.5809437595307827,
"rewards/cosine_scaled_reward": -0.1788077955134213,
"rewards/format_reward": 0.43750000186264515,
"step": 72
},
{
"completion_length": 3413.3958740234375,
"epoch": 0.08342857142857144,
"grad_norm": 0.0283922478556633,
"kl": 2.5540590286254883e-05,
"learning_rate": 9.947027716509488e-07,
"loss": 0.0536,
"reward": -0.20378641900606453,
"reward_std": 0.5730555988848209,
"rewards/cosine_scaled_reward": -0.18522655218839645,
"rewards/format_reward": 0.16666666977107525,
"step": 73
},
{
"completion_length": 2542.875030517578,
"epoch": 0.08457142857142858,
"grad_norm": 0.05036113038659096,
"kl": 3.673252649605274e-05,
"learning_rate": 9.942113192828444e-07,
"loss": 0.1141,
"reward": 0.5641527697443962,
"reward_std": 0.7980917133390903,
"rewards/cosine_scaled_reward": 0.06332638207823038,
"rewards/format_reward": 0.4375000111758709,
"step": 74
},
{
"completion_length": 2902.291748046875,
"epoch": 0.08571428571428572,
"grad_norm": 0.030917614698410034,
"kl": 0.00010601989924907684,
"learning_rate": 9.93698216681727e-07,
"loss": -0.0053,
"reward": 0.5767261572182178,
"reward_std": 0.6028854753822088,
"rewards/cosine_scaled_reward": 0.0696130646392703,
"rewards/format_reward": 0.43750001303851604,
"step": 75
},
{
"completion_length": 2528.979217529297,
"epoch": 0.08685714285714285,
"grad_norm": 0.03214319422841072,
"kl": 2.886354923248291e-05,
"learning_rate": 9.931634888554935e-07,
"loss": 0.0212,
"reward": 0.16732670925557613,
"reward_std": 0.5912668146193027,
"rewards/cosine_scaled_reward": -0.17675332073122263,
"rewards/format_reward": 0.5208333395421505,
"step": 76
},
{
"completion_length": 2850.666679382324,
"epoch": 0.088,
"grad_norm": 0.0167290810495615,
"kl": 2.2485852241516113e-05,
"learning_rate": 9.926071618660237e-07,
"loss": -0.002,
"reward": -0.04175177216529846,
"reward_std": 0.4047280023805797,
"rewards/cosine_scaled_reward": -0.17712588026188314,
"rewards/format_reward": 0.3125,
"step": 77
},
{
"completion_length": 3262.0416717529297,
"epoch": 0.08914285714285715,
"grad_norm": 0.04398359730839729,
"kl": 1.8542632460594177e-05,
"learning_rate": 9.9202926282791e-07,
"loss": 0.0164,
"reward": 0.04341258108615875,
"reward_std": 0.8593644499778748,
"rewards/cosine_scaled_reward": -0.10329371684929356,
"rewards/format_reward": 0.25000000931322575,
"step": 78
},
{
"completion_length": 2167.2708435058594,
"epoch": 0.09028571428571429,
"grad_norm": 0.03241743519902229,
"kl": 3.7103891372680664e-06,
"learning_rate": 9.91429819907136e-07,
"loss": 0.0034,
"reward": 0.34780592005699873,
"reward_std": 0.5333101600408554,
"rewards/cosine_scaled_reward": -0.12818037904798985,
"rewards/format_reward": 0.6041666697710752,
"step": 79
},
{
"completion_length": 3245.750015258789,
"epoch": 0.09142857142857143,
"grad_norm": 0.025279633700847626,
"kl": 1.8217913748230785e-05,
"learning_rate": 9.908088623197048e-07,
"loss": 0.0211,
"reward": -0.14338213577866554,
"reward_std": 0.3987116068601608,
"rewards/cosine_scaled_reward": -0.19669106788933277,
"rewards/format_reward": 0.25000000558793545,
"step": 80
},
{
"completion_length": 3126.250030517578,
"epoch": 0.09257142857142857,
"grad_norm": 0.0291125550866127,
"kl": 3.684312105178833e-05,
"learning_rate": 9.901664203302124e-07,
"loss": 0.0013,
"reward": -0.2766275405883789,
"reward_std": 0.49616212770342827,
"rewards/cosine_scaled_reward": -0.25289710983633995,
"rewards/format_reward": 0.22916666977107525,
"step": 81
},
{
"completion_length": 2787.916732788086,
"epoch": 0.09371428571428571,
"grad_norm": 0.053675856441259384,
"kl": 2.8818845748901367e-05,
"learning_rate": 9.895025252503755e-07,
"loss": 0.0989,
"reward": 0.4680606871843338,
"reward_std": 0.9560514762997627,
"rewards/cosine_scaled_reward": 0.0152803435921669,
"rewards/format_reward": 0.4375000074505806,
"step": 82
},
{
"completion_length": 2561.979175567627,
"epoch": 0.09485714285714286,
"grad_norm": 0.025762459263205528,
"kl": 3.769155591726303e-05,
"learning_rate": 9.888172094375033e-07,
"loss": -0.0082,
"reward": 0.18416154105216265,
"reward_std": 0.45330405980348587,
"rewards/cosine_scaled_reward": -0.10583589226007462,
"rewards/format_reward": 0.39583333395421505,
"step": 83
},
{
"completion_length": 2878.229217529297,
"epoch": 0.096,
"grad_norm": 0.050619326531887054,
"kl": 2.41938978433609e-05,
"learning_rate": 9.881105062929221e-07,
"loss": 0.0362,
"reward": 0.36311598774045706,
"reward_std": 0.9684205129742622,
"rewards/cosine_scaled_reward": -0.03719199728220701,
"rewards/format_reward": 0.43750001303851604,
"step": 84
},
{
"completion_length": 3074.5416870117188,
"epoch": 0.09714285714285714,
"grad_norm": 0.03848778456449509,
"kl": 1.1827796697616577e-05,
"learning_rate": 9.873824502603459e-07,
"loss": 0.048,
"reward": 0.007899533025920391,
"reward_std": 0.7812178507447243,
"rewards/cosine_scaled_reward": -0.1523002306057606,
"rewards/format_reward": 0.3125000074505806,
"step": 85
},
{
"completion_length": 2757.7708587646484,
"epoch": 0.09828571428571428,
"grad_norm": 0.030428804457187653,
"kl": 3.374367952346802e-05,
"learning_rate": 9.866330768241983e-07,
"loss": -0.007,
"reward": 0.2510148920118809,
"reward_std": 0.6730287857353687,
"rewards/cosine_scaled_reward": -0.09324256191030145,
"rewards/format_reward": 0.4375000074505806,
"step": 86
},
{
"completion_length": 2643.520866394043,
"epoch": 0.09942857142857142,
"grad_norm": 0.03848496824502945,
"kl": 2.314150333404541e-05,
"learning_rate": 9.85862422507884e-07,
"loss": 0.0456,
"reward": 0.45505039021372795,
"reward_std": 0.7903980556875467,
"rewards/cosine_scaled_reward": -0.012058130465447903,
"rewards/format_reward": 0.47916668094694614,
"step": 87
},
{
"completion_length": 2060.8959197998047,
"epoch": 0.10057142857142858,
"grad_norm": 0.04052264988422394,
"kl": 2.466072328388691e-05,
"learning_rate": 9.850705248720068e-07,
"loss": 0.0715,
"reward": 0.655466640368104,
"reward_std": 0.89967380464077,
"rewards/cosine_scaled_reward": 0.025649975519627333,
"rewards/format_reward": 0.6041666734963655,
"step": 88
},
{
"completion_length": 2851.687515258789,
"epoch": 0.10171428571428572,
"grad_norm": 0.026882044970989227,
"kl": 1.2446194887161255e-05,
"learning_rate": 9.8425742251254e-07,
"loss": 0.0021,
"reward": 0.33032611198723316,
"reward_std": 0.6415284462273121,
"rewards/cosine_scaled_reward": -0.032753610983490944,
"rewards/format_reward": 0.39583333395421505,
"step": 89
},
{
"completion_length": 2466.0208892822266,
"epoch": 0.10285714285714286,
"grad_norm": 0.023797115311026573,
"kl": 5.2119605243206024e-05,
"learning_rate": 9.83423155058946e-07,
"loss": 0.031,
"reward": 0.10800119815394282,
"reward_std": 0.4383825846016407,
"rewards/cosine_scaled_reward": -0.18558274419046938,
"rewards/format_reward": 0.47916666977107525,
"step": 90
},
{
"completion_length": 3075.833335876465,
"epoch": 0.104,
"grad_norm": 0.05504177510738373,
"kl": 2.738088369369507e-06,
"learning_rate": 9.825677631722435e-07,
"loss": 0.0478,
"reward": 0.1143021360039711,
"reward_std": 0.8147715367376804,
"rewards/cosine_scaled_reward": -0.08868227154016495,
"rewards/format_reward": 0.29166666977107525,
"step": 91
},
{
"completion_length": 2582.791717529297,
"epoch": 0.10514285714285715,
"grad_norm": 0.08598244190216064,
"kl": 2.1602027118206024e-05,
"learning_rate": 9.816912885430258e-07,
"loss": -0.0326,
"reward": 0.117531917989254,
"reward_std": 0.5453543923795223,
"rewards/cosine_scaled_reward": -0.19123404938727617,
"rewards/format_reward": 0.5000000037252903,
"step": 92
},
{
"completion_length": 3582.6458435058594,
"epoch": 0.10628571428571429,
"grad_norm": 0.02585573121905327,
"kl": 2.4219392798841e-05,
"learning_rate": 9.807937738894303e-07,
"loss": 0.0008,
"reward": -0.4189921743236482,
"reward_std": 0.49103348329663277,
"rewards/cosine_scaled_reward": -0.23032942693680525,
"rewards/format_reward": 0.0416666679084301,
"step": 93
},
{
"completion_length": 2344.5416946411133,
"epoch": 0.10742857142857143,
"grad_norm": 0.01883937232196331,
"kl": 3.026425838470459e-05,
"learning_rate": 9.798752629550546e-07,
"loss": -0.0014,
"reward": 0.2969733662903309,
"reward_std": 0.4309884384274483,
"rewards/cosine_scaled_reward": -0.10151330940425396,
"rewards/format_reward": 0.5,
"step": 94
},
{
"completion_length": 3309.8958435058594,
"epoch": 0.10857142857142857,
"grad_norm": 0.039376407861709595,
"kl": 1.1745840311050415e-05,
"learning_rate": 9.78935800506826e-07,
"loss": 0.0141,
"reward": -0.019916832447052002,
"reward_std": 0.6842259094119072,
"rewards/cosine_scaled_reward": -0.1349584199488163,
"rewards/format_reward": 0.25,
"step": 95
},
{
"completion_length": 2395.979217529297,
"epoch": 0.10971428571428571,
"grad_norm": 0.03585467487573624,
"kl": 4.587695002555847e-06,
"learning_rate": 9.779754323328192e-07,
"loss": 0.0035,
"reward": 0.3708239998668432,
"reward_std": 0.6535003744065762,
"rewards/cosine_scaled_reward": -0.08542133821174502,
"rewards/format_reward": 0.5416666679084301,
"step": 96
},
{
"completion_length": 3134.062545776367,
"epoch": 0.11085714285714286,
"grad_norm": 0.05161568522453308,
"kl": 2.466142177581787e-05,
"learning_rate": 9.769942052400235e-07,
"loss": 0.079,
"reward": 0.09229021240025759,
"reward_std": 0.7892664521932602,
"rewards/cosine_scaled_reward": -0.08927156520076096,
"rewards/format_reward": 0.27083334140479565,
"step": 97
},
{
"completion_length": 2911.7708892822266,
"epoch": 0.112,
"grad_norm": 0.032276708632707596,
"kl": 7.407739758491516e-06,
"learning_rate": 9.759921670520634e-07,
"loss": 0.0629,
"reward": 0.311330777592957,
"reward_std": 0.5608916245400906,
"rewards/cosine_scaled_reward": -0.05266796611249447,
"rewards/format_reward": 0.41666667349636555,
"step": 98
},
{
"completion_length": 2801.041702270508,
"epoch": 0.11314285714285714,
"grad_norm": 0.029017232358455658,
"kl": 9.931332897394896e-06,
"learning_rate": 9.749693666068663e-07,
"loss": 0.0092,
"reward": 0.1490808641538024,
"reward_std": 0.5960719883441925,
"rewards/cosine_scaled_reward": -0.08170956326648593,
"rewards/format_reward": 0.31250000186264515,
"step": 99
},
{
"completion_length": 2526.5208587646484,
"epoch": 0.11428571428571428,
"grad_norm": 0.04443981498479843,
"kl": 3.993883728981018e-05,
"learning_rate": 9.739258537542835e-07,
"loss": -0.0003,
"reward": 0.4763328084954992,
"reward_std": 0.8469073474407196,
"rewards/cosine_scaled_reward": -0.0014169311616569757,
"rewards/format_reward": 0.4791666679084301,
"step": 100
},
{
"completion_length": 2756.645866394043,
"epoch": 0.11542857142857142,
"grad_norm": 0.03890707716345787,
"kl": 1.1265277862548828e-05,
"learning_rate": 9.728616793536587e-07,
"loss": 0.0634,
"reward": 0.35849976912140846,
"reward_std": 0.6256566159427166,
"rewards/cosine_scaled_reward": -0.01866680383682251,
"rewards/format_reward": 0.3958333358168602,
"step": 101
},
{
"completion_length": 2056.3125762939453,
"epoch": 0.11657142857142858,
"grad_norm": 0.03498723357915878,
"kl": 4.578381776809692e-06,
"learning_rate": 9.717768952713511e-07,
"loss": 0.04,
"reward": 0.6048059780150652,
"reward_std": 0.6034604609012604,
"rewards/cosine_scaled_reward": -0.0517636826261878,
"rewards/format_reward": 0.7083333488553762,
"step": 102
},
{
"completion_length": 2677.979202270508,
"epoch": 0.11771428571428572,
"grad_norm": 0.033229317516088486,
"kl": 4.5493245124816895e-05,
"learning_rate": 9.706715543782064e-07,
"loss": 0.0244,
"reward": 0.22921494115144014,
"reward_std": 0.5147040607407689,
"rewards/cosine_scaled_reward": -0.09372587502002716,
"rewards/format_reward": 0.4166666679084301,
"step": 103
},
{
"completion_length": 2607.9375,
"epoch": 0.11885714285714286,
"grad_norm": 0.04574336111545563,
"kl": 2.5503337383270264e-05,
"learning_rate": 9.695457105469804e-07,
"loss": 0.0431,
"reward": 0.19741635338868946,
"reward_std": 0.7826841361820698,
"rewards/cosine_scaled_reward": -0.14087516069412231,
"rewards/format_reward": 0.4791666753590107,
"step": 104
},
{
"completion_length": 2667.354217529297,
"epoch": 0.12,
"grad_norm": 0.06220825016498566,
"kl": 2.7738511562347412e-05,
"learning_rate": 9.683994186497132e-07,
"loss": 0.11,
"reward": 0.3817732520401478,
"reward_std": 1.0063364561647177,
"rewards/cosine_scaled_reward": -0.007030060514807701,
"rewards/format_reward": 0.39583334140479565,
"step": 105
},
{
"completion_length": 2350.1667137145996,
"epoch": 0.12114285714285715,
"grad_norm": 0.038587577641010284,
"kl": -3.159046173095703e-06,
"learning_rate": 9.672327345550543e-07,
"loss": 0.0546,
"reward": 0.8650688156485558,
"reward_std": 0.8709494546055794,
"rewards/cosine_scaled_reward": 0.14086772687733173,
"rewards/format_reward": 0.5833333395421505,
"step": 106
},
{
"completion_length": 2880.0833435058594,
"epoch": 0.12228571428571429,
"grad_norm": 0.03471750020980835,
"kl": 5.945563316345215e-06,
"learning_rate": 9.66045715125541e-07,
"loss": 0.0146,
"reward": 0.45387474820017815,
"reward_std": 0.527577817440033,
"rewards/cosine_scaled_reward": -0.0022292807698249817,
"rewards/format_reward": 0.4583333432674408,
"step": 107
},
{
"completion_length": 2783.312530517578,
"epoch": 0.12342857142857143,
"grad_norm": 0.031732089817523956,
"kl": 9.931158274412155e-06,
"learning_rate": 9.648384182148252e-07,
"loss": 0.0456,
"reward": 0.1532750353217125,
"reward_std": 0.5677235350012779,
"rewards/cosine_scaled_reward": -0.10044581908732653,
"rewards/format_reward": 0.35416667722165585,
"step": 108
},
{
"completion_length": 3024.354179382324,
"epoch": 0.12457142857142857,
"grad_norm": 0.023847682401537895,
"kl": -6.922753527760506e-06,
"learning_rate": 9.636109026648554e-07,
"loss": 0.0082,
"reward": 0.10752910934388638,
"reward_std": 0.4340343438088894,
"rewards/cosine_scaled_reward": -0.09206879511475563,
"rewards/format_reward": 0.29166667349636555,
"step": 109
},
{
"completion_length": 2805.270851135254,
"epoch": 0.12571428571428572,
"grad_norm": 0.037369612604379654,
"kl": -4.994682967662811e-06,
"learning_rate": 9.623632283030077e-07,
"loss": -0.0008,
"reward": 0.11598340794444084,
"reward_std": 0.7331918552517891,
"rewards/cosine_scaled_reward": -0.16075829323381186,
"rewards/format_reward": 0.4375000074505806,
"step": 110
},
{
"completion_length": 3019.500045776367,
"epoch": 0.12685714285714286,
"grad_norm": 0.03501195088028908,
"kl": 1.4557619579136372e-05,
"learning_rate": 9.610954559391704e-07,
"loss": 0.0307,
"reward": 0.0815199427306652,
"reward_std": 0.6842854991555214,
"rewards/cosine_scaled_reward": -0.1050733660813421,
"rewards/format_reward": 0.29166666977107525,
"step": 111
},
{
"completion_length": 3347.0833740234375,
"epoch": 0.128,
"grad_norm": 0.04136764630675316,
"kl": 1.260777935385704e-05,
"learning_rate": 9.598076473627796e-07,
"loss": 0.0226,
"reward": 0.20088393986225128,
"reward_std": 0.6008856520056725,
"rewards/cosine_scaled_reward": -0.04539137287065387,
"rewards/format_reward": 0.29166666977107525,
"step": 112
},
{
"completion_length": 2591.145854949951,
"epoch": 0.12914285714285714,
"grad_norm": 0.028484085574746132,
"kl": 6.703287363052368e-05,
"learning_rate": 9.58499865339809e-07,
"loss": 0.0377,
"reward": 0.1814913973212242,
"reward_std": 0.5701524242758751,
"rewards/cosine_scaled_reward": -0.1280043087899685,
"rewards/format_reward": 0.4375000111758709,
"step": 113
},
{
"completion_length": 2411.8958587646484,
"epoch": 0.13028571428571428,
"grad_norm": 0.03289726749062538,
"kl": 8.968636393547058e-06,
"learning_rate": 9.571721736097088e-07,
"loss": 0.0053,
"reward": 0.12762277480214834,
"reward_std": 0.43860352924093604,
"rewards/cosine_scaled_reward": -0.22785530053079128,
"rewards/format_reward": 0.5833333358168602,
"step": 114
},
{
"completion_length": 2959.895835876465,
"epoch": 0.13142857142857142,
"grad_norm": 0.0381602868437767,
"kl": 2.3990869522094727e-05,
"learning_rate": 9.55824636882301e-07,
"loss": 0.0392,
"reward": 0.28831441630609334,
"reward_std": 0.6521844463422894,
"rewards/cosine_scaled_reward": -0.01209279743488878,
"rewards/format_reward": 0.3125,
"step": 115
},
{
"completion_length": 3210.875,
"epoch": 0.13257142857142856,
"grad_norm": 0.027132300660014153,
"kl": 1.683831214904785e-05,
"learning_rate": 9.54457320834625e-07,
"loss": 0.002,
"reward": -0.266488716006279,
"reward_std": 0.5367148518562317,
"rewards/cosine_scaled_reward": -0.21657769149169326,
"rewards/format_reward": 0.1666666679084301,
"step": 116
},
{
"completion_length": 3263.1666717529297,
"epoch": 0.1337142857142857,
"grad_norm": 0.026704585179686546,
"kl": 8.605420589447021e-06,
"learning_rate": 9.530702921077358e-07,
"loss": 0.005,
"reward": -0.2799858786165714,
"reward_std": 0.43628356233239174,
"rewards/cosine_scaled_reward": -0.2233262713998556,
"rewards/format_reward": 0.1666666679084301,
"step": 117
},
{
"completion_length": 2891.6875610351562,
"epoch": 0.13485714285714287,
"grad_norm": 0.04840896278619766,
"kl": -8.52346420288086e-06,
"learning_rate": 9.516636183034564e-07,
"loss": 0.0269,
"reward": 0.8288209615275264,
"reward_std": 0.9317138865590096,
"rewards/cosine_scaled_reward": 0.15399381378665566,
"rewards/format_reward": 0.5208333414047956,
"step": 118
},
{
"completion_length": 1989.208366394043,
"epoch": 0.136,
"grad_norm": 0.03215627372264862,
"kl": 6.823241710662842e-05,
"learning_rate": 9.502373679810839e-07,
"loss": 0.0413,
"reward": 0.5099336765706539,
"reward_std": 0.5711662992835045,
"rewards/cosine_scaled_reward": -0.06794983521103859,
"rewards/format_reward": 0.6458333358168602,
"step": 119
},
{
"completion_length": 2788.00004196167,
"epoch": 0.13714285714285715,
"grad_norm": 0.04669712111353874,
"kl": 1.7192214727401733e-05,
"learning_rate": 9.487916106540465e-07,
"loss": 0.0839,
"reward": 0.23134121485054493,
"reward_std": 0.6627340568229556,
"rewards/cosine_scaled_reward": -0.06141273118555546,
"rewards/format_reward": 0.3541666716337204,
"step": 120
},
{
"completion_length": 2010.562515258789,
"epoch": 0.1382857142857143,
"grad_norm": 0.03950565680861473,
"kl": 4.994124174118042e-05,
"learning_rate": 9.473264167865171e-07,
"loss": 0.0413,
"reward": 0.5943057034164667,
"reward_std": 0.6402759049087763,
"rewards/cosine_scaled_reward": -0.0049304962158203125,
"rewards/format_reward": 0.6041666697710752,
"step": 121
},
{
"completion_length": 3024.6666870117188,
"epoch": 0.13942857142857143,
"grad_norm": 0.05417582765221596,
"kl": 3.859400749206543e-06,
"learning_rate": 9.458418577899774e-07,
"loss": 0.0436,
"reward": 0.3850366286933422,
"reward_std": 0.8862559981644154,
"rewards/cosine_scaled_reward": 0.005018303170800209,
"rewards/format_reward": 0.3750000037252903,
"step": 122
},
{
"completion_length": 2911.104232788086,
"epoch": 0.14057142857142857,
"grad_norm": 0.05250710994005203,
"kl": 7.831677794456482e-05,
"learning_rate": 9.443380060197385e-07,
"loss": 0.1311,
"reward": -0.02938712202012539,
"reward_std": 0.7504731137305498,
"rewards/cosine_scaled_reward": -0.18136022426187992,
"rewards/format_reward": 0.33333333395421505,
"step": 123
},
{
"completion_length": 2103.000026702881,
"epoch": 0.1417142857142857,
"grad_norm": 0.03832955285906792,
"kl": 2.69375741481781e-05,
"learning_rate": 9.428149347714143e-07,
"loss": -0.0021,
"reward": 0.6034899442456663,
"reward_std": 0.6659752391278744,
"rewards/cosine_scaled_reward": -0.00033837370574474335,
"rewards/format_reward": 0.6041666679084301,
"step": 124
},
{
"completion_length": 2813.000015258789,
"epoch": 0.14285714285714285,
"grad_norm": 0.033055417239665985,
"kl": 4.522502422332764e-06,
"learning_rate": 9.412727182773486e-07,
"loss": 0.0306,
"reward": 0.4082658104598522,
"reward_std": 0.5916998572647572,
"rewards/cosine_scaled_reward": 0.047882895451039076,
"rewards/format_reward": 0.31250000186264515,
"step": 125
},
{
"completion_length": 2842.0416870117188,
"epoch": 0.144,
"grad_norm": 0.04313673824071884,
"kl": 2.8833746910095215e-06,
"learning_rate": 9.397114317029974e-07,
"loss": 0.0148,
"reward": 0.1651114380802028,
"reward_std": 0.7616937188431621,
"rewards/cosine_scaled_reward": -0.12577760918065906,
"rewards/format_reward": 0.4166666716337204,
"step": 126
},
{
"completion_length": 3460.750030517578,
"epoch": 0.14514285714285713,
"grad_norm": 0.02390168234705925,
"kl": 3.158301115036011e-05,
"learning_rate": 9.381311511432658e-07,
"loss": 0.022,
"reward": -0.43700045719742775,
"reward_std": 0.405607839114964,
"rewards/cosine_scaled_reward": -0.2705835606902838,
"rewards/format_reward": 0.1041666679084301,
"step": 127
},
{
"completion_length": 2771.4583740234375,
"epoch": 0.1462857142857143,
"grad_norm": 0.04748675227165222,
"kl": 1.5042722225189209e-05,
"learning_rate": 9.36531953618799e-07,
"loss": 0.0512,
"reward": 0.8968718275427818,
"reward_std": 0.8156895972788334,
"rewards/cosine_scaled_reward": 0.1671859212219715,
"rewards/format_reward": 0.5625000055879354,
"step": 128
},
{
"completion_length": 3298.250030517578,
"epoch": 0.14742857142857144,
"grad_norm": 0.04449458047747612,
"kl": 1.4479272067546844e-05,
"learning_rate": 9.34913917072228e-07,
"loss": 0.009,
"reward": -0.10969683527946472,
"reward_std": 0.7850774228572845,
"rewards/cosine_scaled_reward": -0.17984841065481305,
"rewards/format_reward": 0.2500000037252903,
"step": 129
},
{
"completion_length": 3292.7083587646484,
"epoch": 0.14857142857142858,
"grad_norm": 0.04384580999612808,
"kl": 2.863258123397827e-05,
"learning_rate": 9.332771203643714e-07,
"loss": 0.0654,
"reward": -0.18806715682148933,
"reward_std": 0.7124089896678925,
"rewards/cosine_scaled_reward": -0.16695024794898927,
"rewards/format_reward": 0.14583333767950535,
"step": 130
},
{
"completion_length": 2807.4792289733887,
"epoch": 0.14971428571428572,
"grad_norm": 0.05430476740002632,
"kl": 4.511140286922455e-05,
"learning_rate": 9.316216432703916e-07,
"loss": 0.1049,
"reward": 0.31562044098973274,
"reward_std": 0.7439009035006166,
"rewards/cosine_scaled_reward": -0.0401064483448863,
"rewards/format_reward": 0.3958333395421505,
"step": 131
},
{
"completion_length": 2769.083339691162,
"epoch": 0.15085714285714286,
"grad_norm": 0.0387246310710907,
"kl": -2.1746382117271423e-05,
"learning_rate": 9.299475664759068e-07,
"loss": -0.0041,
"reward": 0.15846600383520126,
"reward_std": 0.6973005346953869,
"rewards/cosine_scaled_reward": -0.09785035438835621,
"rewards/format_reward": 0.35416666977107525,
"step": 132
},
{
"completion_length": 3294.4583740234375,
"epoch": 0.152,
"grad_norm": 0.043031562119722366,
"kl": 1.8611550331115723e-05,
"learning_rate": 9.282549715730579e-07,
"loss": 0.0035,
"reward": 0.021218243055045605,
"reward_std": 0.645587831735611,
"rewards/cosine_scaled_reward": -0.12480754964053631,
"rewards/format_reward": 0.27083333767950535,
"step": 133
},
{
"completion_length": 2709.437526702881,
"epoch": 0.15314285714285714,
"grad_norm": 0.049496814608573914,
"kl": 9.149685502052307e-05,
"learning_rate": 9.265439410565328e-07,
"loss": 0.039,
"reward": 0.39167931117117405,
"reward_std": 0.6326965596526861,
"rewards/cosine_scaled_reward": -0.0020770253613591194,
"rewards/format_reward": 0.3958333358168602,
"step": 134
},
{
"completion_length": 2031.5417022705078,
"epoch": 0.15428571428571428,
"grad_norm": 0.05865013226866722,
"kl": 8.575990796089172e-05,
"learning_rate": 9.248145583195447e-07,
"loss": 0.1061,
"reward": 1.1254341015592217,
"reward_std": 0.8684800090268254,
"rewards/cosine_scaled_reward": 0.25021704845130444,
"rewards/format_reward": 0.6250000074505806,
"step": 135
},
{
"completion_length": 2618.687515258789,
"epoch": 0.15542857142857142,
"grad_norm": 0.034709468483924866,
"kl": 4.778057336807251e-05,
"learning_rate": 9.230669076497687e-07,
"loss": 0.023,
"reward": 0.48167144507169724,
"reward_std": 0.6561040449887514,
"rewards/cosine_scaled_reward": 0.02208572020754218,
"rewards/format_reward": 0.43750000186264515,
"step": 136
},
{
"completion_length": 3057.8958435058594,
"epoch": 0.15657142857142858,
"grad_norm": 0.021808914840221405,
"kl": 1.4843419194221497e-05,
"learning_rate": 9.213010742252327e-07,
"loss": 0.0142,
"reward": -0.08836949244141579,
"reward_std": 0.4218329731374979,
"rewards/cosine_scaled_reward": -0.1796014215797186,
"rewards/format_reward": 0.2708333395421505,
"step": 137
},
{
"completion_length": 2785.1875228881836,
"epoch": 0.15771428571428572,
"grad_norm": 0.0407271683216095,
"kl": 1.3926997780799866e-05,
"learning_rate": 9.195171441101668e-07,
"loss": 0.003,
"reward": 0.33298980072140694,
"reward_std": 0.7713625803589821,
"rewards/cosine_scaled_reward": -0.06267178524285555,
"rewards/format_reward": 0.45833334140479565,
"step": 138
},
{
"completion_length": 2804.0833435058594,
"epoch": 0.15885714285714286,
"grad_norm": 0.04251856729388237,
"kl": 0.00016787275671958923,
"learning_rate": 9.177152042508077e-07,
"loss": 0.0123,
"reward": 0.13880036072805524,
"reward_std": 0.7100770510733128,
"rewards/cosine_scaled_reward": -0.13893315801396966,
"rewards/format_reward": 0.41666667349636555,
"step": 139
},
{
"completion_length": 3116.958396911621,
"epoch": 0.16,
"grad_norm": 0.03125656023621559,
"kl": 8.677691221237183e-05,
"learning_rate": 9.158953424711624e-07,
"loss": 0.0472,
"reward": 0.06694573163986206,
"reward_std": 0.5762135218828917,
"rewards/cosine_scaled_reward": -0.07069380325265229,
"rewards/format_reward": 0.2083333358168602,
"step": 140
},
{
"completion_length": 2992.2708892822266,
"epoch": 0.16114285714285714,
"grad_norm": 0.041330691426992416,
"kl": 9.230338037014008e-06,
"learning_rate": 9.140576474687263e-07,
"loss": 0.0029,
"reward": 0.19635527953505516,
"reward_std": 0.7302406523376703,
"rewards/cosine_scaled_reward": -0.12057237513363361,
"rewards/format_reward": 0.43750000558793545,
"step": 141
},
{
"completion_length": 2647.187545776367,
"epoch": 0.16228571428571428,
"grad_norm": 0.037891265004873276,
"kl": 3.4689903259277344e-05,
"learning_rate": 9.122022088101613e-07,
"loss": 0.0244,
"reward": 0.33850586879998446,
"reward_std": 0.7777349427342415,
"rewards/cosine_scaled_reward": -0.12241373397409916,
"rewards/format_reward": 0.5833333395421505,
"step": 142
},
{
"completion_length": 2517.7917098999023,
"epoch": 0.16342857142857142,
"grad_norm": 0.038469068706035614,
"kl": 4.398822784423828e-05,
"learning_rate": 9.103291169269299e-07,
"loss": 0.0177,
"reward": 0.10013854561839253,
"reward_std": 0.6364383529871702,
"rewards/cosine_scaled_reward": -0.17909739445894957,
"rewards/format_reward": 0.4583333358168602,
"step": 143
},
{
"completion_length": 3101.375030517578,
"epoch": 0.16457142857142856,
"grad_norm": 0.047153256833553314,
"kl": 2.6047229766845703e-05,
"learning_rate": 9.084384631108882e-07,
"loss": 0.0287,
"reward": 0.22925031930208206,
"reward_std": 0.7248009741306305,
"rewards/cosine_scaled_reward": -0.020791523158550262,
"rewards/format_reward": 0.2708333395421505,
"step": 144
},
{
"completion_length": 2370.1875228881836,
"epoch": 0.1657142857142857,
"grad_norm": 0.0311027429997921,
"kl": 4.681199789047241e-05,
"learning_rate": 9.065303395098358e-07,
"loss": 0.0081,
"reward": 0.4629867151379585,
"reward_std": 0.7178505845367908,
"rewards/cosine_scaled_reward": -0.028923317790031433,
"rewards/format_reward": 0.520833333954215,
"step": 145
},
{
"completion_length": 2521.020866394043,
"epoch": 0.16685714285714287,
"grad_norm": 0.022678615525364876,
"kl": 6.041303277015686e-05,
"learning_rate": 9.046048391230247e-07,
"loss": 0.0343,
"reward": 0.04914231039583683,
"reward_std": 0.4269477091729641,
"rewards/cosine_scaled_reward": -0.2775121796876192,
"rewards/format_reward": 0.604166679084301,
"step": 146
},
{
"completion_length": 3500.4166870117188,
"epoch": 0.168,
"grad_norm": 0.04476216807961464,
"kl": 4.99039888381958e-05,
"learning_rate": 9.026620557966279e-07,
"loss": 0.0183,
"reward": -0.014710072427988052,
"reward_std": 0.8408091142773628,
"rewards/cosine_scaled_reward": -0.09068836877122521,
"rewards/format_reward": 0.1666666716337204,
"step": 147
},
{
"completion_length": 2774.5416717529297,
"epoch": 0.16914285714285715,
"grad_norm": 0.028264645487070084,
"kl": 2.7420930564403534e-05,
"learning_rate": 9.007020842191634e-07,
"loss": 0.0294,
"reward": 0.31881215423345566,
"reward_std": 0.535629041492939,
"rewards/cosine_scaled_reward": -0.06976059079170227,
"rewards/format_reward": 0.45833334140479565,
"step": 148
},
{
"completion_length": 2830.875015258789,
"epoch": 0.1702857142857143,
"grad_norm": 0.037253376096487045,
"kl": 3.8120895624160767e-05,
"learning_rate": 8.987250199168808e-07,
"loss": 0.047,
"reward": 0.2539705242961645,
"reward_std": 0.6319136489182711,
"rewards/cosine_scaled_reward": -0.08134806924499571,
"rewards/format_reward": 0.4166666716337204,
"step": 149
},
{
"completion_length": 2714.9167098999023,
"epoch": 0.17142857142857143,
"grad_norm": 0.03853278234601021,
"kl": 0.00010446656960994005,
"learning_rate": 8.967309592491052e-07,
"loss": 0.0443,
"reward": -0.02947249379940331,
"reward_std": 0.6823416836559772,
"rewards/cosine_scaled_reward": -0.19181958585977554,
"rewards/format_reward": 0.35416666977107525,
"step": 150
},
{
"completion_length": 2669.3334045410156,
"epoch": 0.17257142857142857,
"grad_norm": 0.05166688933968544,
"kl": 0.0001301094889640808,
"learning_rate": 8.9471999940354e-07,
"loss": 0.0583,
"reward": 0.3849489726126194,
"reward_std": 0.9069008305668831,
"rewards/cosine_scaled_reward": -0.04710885416716337,
"rewards/format_reward": 0.47916667349636555,
"step": 151
},
{
"completion_length": 3023.062530517578,
"epoch": 0.1737142857142857,
"grad_norm": 0.03680495545268059,
"kl": 4.774332046508789e-05,
"learning_rate": 8.926922383915315e-07,
"loss": 0.0118,
"reward": -0.15647598542273045,
"reward_std": 0.4975847424939275,
"rewards/cosine_scaled_reward": -0.2240713369101286,
"rewards/format_reward": 0.2916666679084301,
"step": 152
},
{
"completion_length": 2933.187545776367,
"epoch": 0.17485714285714285,
"grad_norm": 0.04766285791993141,
"kl": 4.67933714389801e-05,
"learning_rate": 8.906477750432903e-07,
"loss": 0.0522,
"reward": 0.08462437242269516,
"reward_std": 0.8217967357486486,
"rewards/cosine_scaled_reward": -0.12435449287295341,
"rewards/format_reward": 0.3333333395421505,
"step": 153
},
{
"completion_length": 3438.8333740234375,
"epoch": 0.176,
"grad_norm": 0.06172926723957062,
"kl": 5.605979822576046e-06,
"learning_rate": 8.88586709003076e-07,
"loss": -0.0067,
"reward": 0.6275491081178188,
"reward_std": 1.102588940411806,
"rewards/cosine_scaled_reward": 0.10544120147824287,
"rewards/format_reward": 0.416666679084301,
"step": 154
},
{
"completion_length": 2463.5208435058594,
"epoch": 0.17714285714285713,
"grad_norm": 0.026917118579149246,
"kl": 6.724148988723755e-05,
"learning_rate": 8.865091407243394e-07,
"loss": 0.0085,
"reward": 0.40650070272386074,
"reward_std": 0.5913192741572857,
"rewards/cosine_scaled_reward": 0.005333678796887398,
"rewards/format_reward": 0.39583333395421505,
"step": 155
},
{
"completion_length": 2991.145835876465,
"epoch": 0.1782857142857143,
"grad_norm": 0.05746571347117424,
"kl": 2.1982938051223755e-05,
"learning_rate": 8.844151714648274e-07,
"loss": 0.0599,
"reward": 0.1549793779850006,
"reward_std": 0.6098005510866642,
"rewards/cosine_scaled_reward": -0.06834365800023079,
"rewards/format_reward": 0.2916666716337204,
"step": 156
},
{
"completion_length": 3096.5208435058594,
"epoch": 0.17942857142857144,
"grad_norm": 0.037943121045827866,
"kl": 5.745887756347656e-05,
"learning_rate": 8.823049032816478e-07,
"loss": 0.0273,
"reward": 0.06849924847483635,
"reward_std": 0.5304312091320753,
"rewards/cosine_scaled_reward": -0.13241703808307648,
"rewards/format_reward": 0.3333333432674408,
"step": 157
},
{
"completion_length": 2825.979232788086,
"epoch": 0.18057142857142858,
"grad_norm": 0.04266660660505295,
"kl": 0.00021417438983917236,
"learning_rate": 8.801784390262943e-07,
"loss": 0.063,
"reward": 0.6505898106843233,
"reward_std": 0.7261264026165009,
"rewards/cosine_scaled_reward": 0.0961282174102962,
"rewards/format_reward": 0.45833334140479565,
"step": 158
},
{
"completion_length": 3209.979202270508,
"epoch": 0.18171428571428572,
"grad_norm": 0.0403718538582325,
"kl": 3.383122384548187e-05,
"learning_rate": 8.780358823396352e-07,
"loss": 0.0048,
"reward": -0.04634229780640453,
"reward_std": 0.7521483562886715,
"rewards/cosine_scaled_reward": -0.15858781896531582,
"rewards/format_reward": 0.27083333767950535,
"step": 159
},
{
"completion_length": 3112.020866394043,
"epoch": 0.18285714285714286,
"grad_norm": 0.06450565159320831,
"kl": 0.00014490634202957153,
"learning_rate": 8.758773376468604e-07,
"loss": 0.0331,
"reward": 0.17703042179346085,
"reward_std": 0.8230392001569271,
"rewards/cosine_scaled_reward": -0.0364847918972373,
"rewards/format_reward": 0.25000000186264515,
"step": 160
},
{
"completion_length": 1949.7291870117188,
"epoch": 0.184,
"grad_norm": 0.04517294839024544,
"kl": 0.00019407272338867188,
"learning_rate": 8.737029101523929e-07,
"loss": 0.0781,
"reward": 0.4693678207695484,
"reward_std": 0.6783907152712345,
"rewards/cosine_scaled_reward": -0.11948275938630104,
"rewards/format_reward": 0.7083333507180214,
"step": 161
},
{
"completion_length": 3424.2084350585938,
"epoch": 0.18514285714285714,
"grad_norm": 0.05282009020447731,
"kl": 0.00010896846652030945,
"learning_rate": 8.715127058347614e-07,
"loss": 0.0252,
"reward": 0.03611734416335821,
"reward_std": 0.8644157852977514,
"rewards/cosine_scaled_reward": -0.09652466559782624,
"rewards/format_reward": 0.22916667349636555,
"step": 162
},
{
"completion_length": 2668.750045776367,
"epoch": 0.18628571428571428,
"grad_norm": 0.030406512320041656,
"kl": 9.234249591827393e-05,
"learning_rate": 8.693068314414344e-07,
"loss": 0.0055,
"reward": 0.8739382587373257,
"reward_std": 0.6097183618694544,
"rewards/cosine_scaled_reward": 0.14530246332287788,
"rewards/format_reward": 0.5833333358168602,
"step": 163
},
{
"completion_length": 2629.8125228881836,
"epoch": 0.18742857142857142,
"grad_norm": 0.04248577728867531,
"kl": 8.138641715049744e-05,
"learning_rate": 8.670853944836176e-07,
"loss": 0.0345,
"reward": 0.42669382283929735,
"reward_std": 0.7748677339404821,
"rewards/cosine_scaled_reward": -0.0054031130857765675,
"rewards/format_reward": 0.43750000558793545,
"step": 164
},
{
"completion_length": 3354.8541870117188,
"epoch": 0.18857142857142858,
"grad_norm": 0.051660098135471344,
"kl": 9.429454803466797e-05,
"learning_rate": 8.648485032310144e-07,
"loss": 0.0124,
"reward": -0.10903086792677641,
"reward_std": 0.8036016914993525,
"rewards/cosine_scaled_reward": -0.14826543792150915,
"rewards/format_reward": 0.18750000558793545,
"step": 165
},
{
"completion_length": 2896.4583740234375,
"epoch": 0.18971428571428572,
"grad_norm": 0.03565189614892006,
"kl": 0.00016220659017562866,
"learning_rate": 8.625962667065487e-07,
"loss": 0.0169,
"reward": -0.00203709676861763,
"reward_std": 0.6630012057721615,
"rewards/cosine_scaled_reward": -0.1468518888577819,
"rewards/format_reward": 0.29166667349636555,
"step": 166
},
{
"completion_length": 2433.2292251586914,
"epoch": 0.19085714285714286,
"grad_norm": 0.07079535722732544,
"kl": 0.0008552521467208862,
"learning_rate": 8.603287946810513e-07,
"loss": -0.0086,
"reward": 0.42900961078703403,
"reward_std": 0.7506549386307597,
"rewards/cosine_scaled_reward": -0.0771618754370138,
"rewards/format_reward": 0.5833333395421505,
"step": 167
},
{
"completion_length": 3103.7708740234375,
"epoch": 0.192,
"grad_norm": 0.053577445447444916,
"kl": 8.881837129592896e-05,
"learning_rate": 8.580461976679099e-07,
"loss": 0.0283,
"reward": 0.275710541754961,
"reward_std": 0.9691016413271427,
"rewards/cosine_scaled_reward": -0.0288114077411592,
"rewards/format_reward": 0.3333333395421505,
"step": 168
},
{
"completion_length": 2718.1041870117188,
"epoch": 0.19314285714285714,
"grad_norm": 0.028121525421738625,
"kl": 5.266070365905762e-05,
"learning_rate": 8.557485869176825e-07,
"loss": 0.0111,
"reward": 0.6644045971333981,
"reward_std": 0.6455990113317966,
"rewards/cosine_scaled_reward": 0.08220228180289268,
"rewards/format_reward": 0.5000000055879354,
"step": 169
},
{
"completion_length": 2760.3125228881836,
"epoch": 0.19428571428571428,
"grad_norm": 0.02649330534040928,
"kl": 7.764250040054321e-05,
"learning_rate": 8.534360744126753e-07,
"loss": 0.008,
"reward": 0.22724516410380602,
"reward_std": 0.5219327211380005,
"rewards/cosine_scaled_reward": -0.06346075981855392,
"rewards/format_reward": 0.3541666679084301,
"step": 170
},
{
"completion_length": 2697.1250228881836,
"epoch": 0.19542857142857142,
"grad_norm": 0.030017036944627762,
"kl": 0.00019634515047073364,
"learning_rate": 8.511087728614862e-07,
"loss": 0.0456,
"reward": 0.48459129920229316,
"reward_std": 0.6084076017141342,
"rewards/cosine_scaled_reward": 0.0339623149484396,
"rewards/format_reward": 0.4166666716337204,
"step": 171
},
{
"completion_length": 3093.937530517578,
"epoch": 0.19657142857142856,
"grad_norm": 0.037974875420331955,
"kl": 0.0001561082899570465,
"learning_rate": 8.487667956935087e-07,
"loss": 0.0175,
"reward": 0.4516556728631258,
"reward_std": 0.6006606463342905,
"rewards/cosine_scaled_reward": 0.059161149663850665,
"rewards/format_reward": 0.33333333395421505,
"step": 172
},
{
"completion_length": 2046.0000343322754,
"epoch": 0.1977142857142857,
"grad_norm": 0.029817163944244385,
"kl": 6.169034168124199e-05,
"learning_rate": 8.464102570534061e-07,
"loss": -0.008,
"reward": 0.14771767449565232,
"reward_std": 0.6277614803984761,
"rewards/cosine_scaled_reward": -0.19697449961677194,
"rewards/format_reward": 0.5416666679084301,
"step": 173
},
{
"completion_length": 2523.833381652832,
"epoch": 0.19885714285714284,
"grad_norm": 0.02026381902396679,
"kl": 0.0006473064422607422,
"learning_rate": 8.440392717955475e-07,
"loss": 0.0157,
"reward": 0.3172628991305828,
"reward_std": 0.5097487587481737,
"rewards/cosine_scaled_reward": -0.10178520064800978,
"rewards/format_reward": 0.5208333376795053,
"step": 174
},
{
"completion_length": 2944.562545776367,
"epoch": 0.2,
"grad_norm": 0.04808598756790161,
"kl": 9.900331497192383e-05,
"learning_rate": 8.416539554784089e-07,
"loss": 0.0413,
"reward": 0.3816917687654495,
"reward_std": 0.6620169542729855,
"rewards/cosine_scaled_reward": 0.0033458725083619356,
"rewards/format_reward": 0.3750000074505806,
"step": 175
},
{
"completion_length": 2707.2500228881836,
"epoch": 0.20114285714285715,
"grad_norm": 0.048613887280225754,
"kl": 8.374452590942383e-05,
"learning_rate": 8.392544243589427e-07,
"loss": 0.0113,
"reward": 0.5291388533078134,
"reward_std": 0.7917350004427135,
"rewards/cosine_scaled_reward": 0.0458194212988019,
"rewards/format_reward": 0.4375000037252903,
"step": 176
},
{
"completion_length": 3056.229248046875,
"epoch": 0.2022857142857143,
"grad_norm": 0.04693533480167389,
"kl": 0.00028914958238601685,
"learning_rate": 8.368407953869103e-07,
"loss": 0.038,
"reward": 0.3346361154690385,
"reward_std": 0.8135349042713642,
"rewards/cosine_scaled_reward": -0.030598610173910856,
"rewards/format_reward": 0.39583334140479565,
"step": 177
},
{
"completion_length": 2260.6250076293945,
"epoch": 0.20342857142857143,
"grad_norm": 0.02989472635090351,
"kl": 0.0001444593071937561,
"learning_rate": 8.344131861991828e-07,
"loss": 0.0159,
"reward": 0.4688864201307297,
"reward_std": 0.6323110768571496,
"rewards/cosine_scaled_reward": -0.04680680809542537,
"rewards/format_reward": 0.5625,
"step": 178
},
{
"completion_length": 2971.125015258789,
"epoch": 0.20457142857142857,
"grad_norm": 0.03039778396487236,
"kl": 0.0001380816102027893,
"learning_rate": 8.319717151140072e-07,
"loss": 0.007,
"reward": -0.12367597967386246,
"reward_std": 0.5375447850674391,
"rewards/cosine_scaled_reward": -0.21808799169957638,
"rewards/format_reward": 0.3125000074505806,
"step": 179
},
{
"completion_length": 2201.8125343322754,
"epoch": 0.2057142857142857,
"grad_norm": 0.019160043448209763,
"kl": 0.00012689828872680664,
"learning_rate": 8.295165011252396e-07,
"loss": -0.0167,
"reward": 0.7198411021381617,
"reward_std": 0.4092927183955908,
"rewards/cosine_scaled_reward": 0.047420548275113106,
"rewards/format_reward": 0.625,
"step": 180
},
{
"completion_length": 3133.6875228881836,
"epoch": 0.20685714285714285,
"grad_norm": 0.0425766259431839,
"kl": 8.432567119598389e-05,
"learning_rate": 8.270476638965461e-07,
"loss": 0.031,
"reward": 0.2028632378205657,
"reward_std": 0.7808081433176994,
"rewards/cosine_scaled_reward": -0.03398505225777626,
"rewards/format_reward": 0.27083333395421505,
"step": 181
},
{
"completion_length": 2342.3750381469727,
"epoch": 0.208,
"grad_norm": 0.04425498843193054,
"kl": 0.00025501102209091187,
"learning_rate": 8.245653237555705e-07,
"loss": 0.004,
"reward": 0.31899992749094963,
"reward_std": 0.8116986956447363,
"rewards/cosine_scaled_reward": -0.11133337735373061,
"rewards/format_reward": 0.5416666716337204,
"step": 182
},
{
"completion_length": 1913.8541870117188,
"epoch": 0.20914285714285713,
"grad_norm": 0.029876578599214554,
"kl": 0.00026257336139678955,
"learning_rate": 8.220696016880687e-07,
"loss": 0.0399,
"reward": 0.5424154847860336,
"reward_std": 0.636641725897789,
"rewards/cosine_scaled_reward": -0.07254226750228554,
"rewards/format_reward": 0.6875,
"step": 183
},
{
"completion_length": 3032.395854949951,
"epoch": 0.2102857142857143,
"grad_norm": 0.020256591960787773,
"kl": 0.00012014806270599365,
"learning_rate": 8.195606193320136e-07,
"loss": 0.011,
"reward": -0.12789930403232574,
"reward_std": 0.4264537561684847,
"rewards/cosine_scaled_reward": -0.17853298410773277,
"rewards/format_reward": 0.22916666977107525,
"step": 184
},
{
"completion_length": 2749.9375343322754,
"epoch": 0.21142857142857144,
"grad_norm": 0.03656504303216934,
"kl": 0.00017721951007843018,
"learning_rate": 8.170384989716657e-07,
"loss": 0.0238,
"reward": 0.023411770351231098,
"reward_std": 0.640469016507268,
"rewards/cosine_scaled_reward": -0.1549607841297984,
"rewards/format_reward": 0.33333333395421505,
"step": 185
},
{
"completion_length": 2899.375015258789,
"epoch": 0.21257142857142858,
"grad_norm": 0.03554477170109749,
"kl": 9.587407112121582e-05,
"learning_rate": 8.145033635316128e-07,
"loss": 0.0213,
"reward": 0.3093617632985115,
"reward_std": 0.6292181555181742,
"rewards/cosine_scaled_reward": -0.022402465343475342,
"rewards/format_reward": 0.35416666977107525,
"step": 186
},
{
"completion_length": 2228.0416870117188,
"epoch": 0.21371428571428572,
"grad_norm": 0.023758431896567345,
"kl": 0.00029593706130981445,
"learning_rate": 8.119553365707802e-07,
"loss": 0.0325,
"reward": 0.09883421659469604,
"reward_std": 0.5377654749900103,
"rewards/cosine_scaled_reward": -0.2109995698556304,
"rewards/format_reward": 0.5208333395421505,
"step": 187
},
{
"completion_length": 3543.2916870117188,
"epoch": 0.21485714285714286,
"grad_norm": 0.03328506276011467,
"kl": 5.07943332195282e-05,
"learning_rate": 8.093945422764069e-07,
"loss": 0.0109,
"reward": -0.2927169185131788,
"reward_std": 0.5551654025912285,
"rewards/cosine_scaled_reward": -0.1880251243710518,
"rewards/format_reward": 0.0833333358168602,
"step": 188
},
{
"completion_length": 2309.729202270508,
"epoch": 0.216,
"grad_norm": 0.032304707914590836,
"kl": 0.0003388524055480957,
"learning_rate": 8.068211054579943e-07,
"loss": 0.0375,
"reward": 0.05538259819149971,
"reward_std": 0.521863205358386,
"rewards/cosine_scaled_reward": -0.24314204324036837,
"rewards/format_reward": 0.541666679084301,
"step": 189
},
{
"completion_length": 2878.229202270508,
"epoch": 0.21714285714285714,
"grad_norm": 0.026849228888750076,
"kl": 0.0002093091607093811,
"learning_rate": 8.04235151541222e-07,
"loss": 0.007,
"reward": 0.20538577064871788,
"reward_std": 0.4814210291951895,
"rewards/cosine_scaled_reward": -0.06397378008114174,
"rewards/format_reward": 0.3333333358168602,
"step": 190
},
{
"completion_length": 2574.1041870117188,
"epoch": 0.21828571428571428,
"grad_norm": 0.04074694588780403,
"kl": 0.00031504780054092407,
"learning_rate": 8.01636806561836e-07,
"loss": -0.0004,
"reward": 0.32336311414837837,
"reward_std": 0.6287105903029442,
"rewards/cosine_scaled_reward": -0.06748512433841825,
"rewards/format_reward": 0.45833333395421505,
"step": 191
},
{
"completion_length": 3383.8958435058594,
"epoch": 0.21942857142857142,
"grad_norm": 0.03389768674969673,
"kl": 0.00012552272528409958,
"learning_rate": 7.990261971595048e-07,
"loss": 0.0249,
"reward": -0.01775958389043808,
"reward_std": 0.6146044600754976,
"rewards/cosine_scaled_reward": -0.14429646357893944,
"rewards/format_reward": 0.2708333395421505,
"step": 192
},
{
"completion_length": 2860.687530517578,
"epoch": 0.22057142857142858,
"grad_norm": 0.04098944365978241,
"kl": 0.00015905871987342834,
"learning_rate": 7.964034505716476e-07,
"loss": 0.0101,
"reward": 0.28354677464812994,
"reward_std": 0.7337615005671978,
"rewards/cosine_scaled_reward": -0.08739327266812325,
"rewards/format_reward": 0.4583333395421505,
"step": 193
},
{
"completion_length": 3281.0000610351562,
"epoch": 0.22171428571428572,
"grad_norm": 0.05677646026015282,
"kl": 7.194280624389648e-05,
"learning_rate": 7.93768694627233e-07,
"loss": 0.0276,
"reward": 0.6990666054189205,
"reward_std": 0.836215304210782,
"rewards/cosine_scaled_reward": 0.14119997806847095,
"rewards/format_reward": 0.41666666977107525,
"step": 194
},
{
"completion_length": 2777.250030517578,
"epoch": 0.22285714285714286,
"grad_norm": 0.025516223162412643,
"kl": 0.0005084574222564697,
"learning_rate": 7.911220577405484e-07,
"loss": 0.0117,
"reward": -0.06681104190647602,
"reward_std": 0.5002289786934853,
"rewards/cosine_scaled_reward": -0.23132219538092613,
"rewards/format_reward": 0.3958333358168602,
"step": 195
},
{
"completion_length": 3459.750030517578,
"epoch": 0.224,
"grad_norm": 0.029978394508361816,
"kl": 0.00012940913438796997,
"learning_rate": 7.884636689049422e-07,
"loss": 0.0098,
"reward": 0.272402954287827,
"reward_std": 0.5220753028988838,
"rewards/cosine_scaled_reward": -0.020048514008522034,
"rewards/format_reward": 0.31250000558793545,
"step": 196
},
{
"completion_length": 2773.645851135254,
"epoch": 0.22514285714285714,
"grad_norm": 0.048345357179641724,
"kl": 0.00043816864490509033,
"learning_rate": 7.857936576865356e-07,
"loss": 0.02,
"reward": 0.21063009148929268,
"reward_std": 0.9322120416909456,
"rewards/cosine_scaled_reward": -0.08218496525660157,
"rewards/format_reward": 0.37500000186264515,
"step": 197
},
{
"completion_length": 2772.1875076293945,
"epoch": 0.22628571428571428,
"grad_norm": 0.03043530508875847,
"kl": 0.0005050599575042725,
"learning_rate": 7.831121542179086e-07,
"loss": 0.0052,
"reward": 0.29729510098695755,
"reward_std": 0.5535988472402096,
"rewards/cosine_scaled_reward": -0.028435785323381424,
"rewards/format_reward": 0.3541666716337204,
"step": 198
},
{
"completion_length": 3581.5625,
"epoch": 0.22742857142857142,
"grad_norm": 0.02494131401181221,
"kl": 8.360063657164574e-05,
"learning_rate": 7.804192891917571e-07,
"loss": 0.0005,
"reward": -0.43263228237628937,
"reward_std": 0.47081154212355614,
"rewards/cosine_scaled_reward": -0.25798280723392963,
"rewards/format_reward": 0.0833333358168602,
"step": 199
},
{
"completion_length": 2370.666717529297,
"epoch": 0.22857142857142856,
"grad_norm": 0.03381456062197685,
"kl": 0.0002932697534561157,
"learning_rate": 7.777151938545235e-07,
"loss": 0.0061,
"reward": 1.13365722540766,
"reward_std": 0.56923015601933,
"rewards/cosine_scaled_reward": 0.24391193874180317,
"rewards/format_reward": 0.6458333358168602,
"step": 200
},
{
"completion_length": 2703.2708587646484,
"epoch": 0.2297142857142857,
"grad_norm": 0.04875003919005394,
"kl": 9.842216968536377e-05,
"learning_rate": 7.75e-07,
"loss": 0.0188,
"reward": 0.6816193205304444,
"reward_std": 0.9491816088557243,
"rewards/cosine_scaled_reward": 0.09080967318732291,
"rewards/format_reward": 0.5000000074505806,
"step": 201
},
{
"completion_length": 2428.520866394043,
"epoch": 0.23085714285714284,
"grad_norm": 0.029104772955179214,
"kl": 0.0004838407039642334,
"learning_rate": 7.72273839962904e-07,
"loss": 0.0076,
"reward": 0.8386933319270611,
"reward_std": 0.4117593439295888,
"rewards/cosine_scaled_reward": 0.1901800061459653,
"rewards/format_reward": 0.4583333358168602,
"step": 202
},
{
"completion_length": 3084.0416717529297,
"epoch": 0.232,
"grad_norm": 0.01628260500729084,
"kl": 0.00013211369514465332,
"learning_rate": 7.695368466124296e-07,
"loss": 0.0065,
"reward": -0.11864904500544071,
"reward_std": 0.2944770138710737,
"rewards/cosine_scaled_reward": -0.17390786111354828,
"rewards/format_reward": 0.2291666716337204,
"step": 203
},
{
"completion_length": 2296.666717529297,
"epoch": 0.23314285714285715,
"grad_norm": 0.028948936611413956,
"kl": 0.0003097057342529297,
"learning_rate": 7.667891533457718e-07,
"loss": 0.0015,
"reward": 0.3847005255520344,
"reward_std": 0.6527448669075966,
"rewards/cosine_scaled_reward": -0.09931641444563866,
"rewards/format_reward": 0.5833333395421505,
"step": 204
},
{
"completion_length": 2873.229232788086,
"epoch": 0.2342857142857143,
"grad_norm": 0.04777153953909874,
"kl": 0.00016327202320098877,
"learning_rate": 7.640308940816239e-07,
"loss": 0.0657,
"reward": 0.5415339916944504,
"reward_std": 0.8121820725500584,
"rewards/cosine_scaled_reward": 0.041600316762924194,
"rewards/format_reward": 0.4583333395421505,
"step": 205
},
{
"completion_length": 2879.979202270508,
"epoch": 0.23542857142857143,
"grad_norm": 0.03969888389110565,
"kl": 5.835294723510742e-05,
"learning_rate": 7.612622032536507e-07,
"loss": 0.0206,
"reward": 0.03187836706638336,
"reward_std": 0.6640341132879257,
"rewards/cosine_scaled_reward": -0.16114415228366852,
"rewards/format_reward": 0.3541666753590107,
"step": 206
},
{
"completion_length": 3115.6666870117188,
"epoch": 0.23657142857142857,
"grad_norm": 0.03586557134985924,
"kl": 0.00026457011699676514,
"learning_rate": 7.584832158039378e-07,
"loss": 0.0158,
"reward": -0.057422760874032974,
"reward_std": 0.601855780929327,
"rewards/cosine_scaled_reward": -0.20579471392557025,
"rewards/format_reward": 0.35416668094694614,
"step": 207
},
{
"completion_length": 2807.125015258789,
"epoch": 0.2377142857142857,
"grad_norm": 0.028317734599113464,
"kl": 0.0002864748239517212,
"learning_rate": 7.556940671764124e-07,
"loss": 0.0223,
"reward": 0.241438128054142,
"reward_std": 0.5762086287140846,
"rewards/cosine_scaled_reward": -0.066780935972929,
"rewards/format_reward": 0.3750000111758709,
"step": 208
},
{
"completion_length": 2613.0417098999023,
"epoch": 0.23885714285714285,
"grad_norm": 0.042905014008283615,
"kl": 0.00037148594856262207,
"learning_rate": 7.528948933102438e-07,
"loss": 0.0532,
"reward": 0.4992841109633446,
"reward_std": 0.6773411780595779,
"rewards/cosine_scaled_reward": 0.01005872106179595,
"rewards/format_reward": 0.4791666753590107,
"step": 209
},
{
"completion_length": 2671.1458435058594,
"epoch": 0.24,
"grad_norm": 0.017170142382383347,
"kl": 0.00027191638946533203,
"learning_rate": 7.500858306332172e-07,
"loss": -0.0012,
"reward": 0.31039364635944366,
"reward_std": 0.4712617564946413,
"rewards/cosine_scaled_reward": -0.03230317682027817,
"rewards/format_reward": 0.375,
"step": 210
},
{
"completion_length": 2630.500030517578,
"epoch": 0.24114285714285713,
"grad_norm": 0.03910549357533455,
"kl": 0.000303804874420166,
"learning_rate": 7.472670160550848e-07,
"loss": 0.0362,
"reward": 0.4297244977205992,
"reward_std": 0.6497951280325651,
"rewards/cosine_scaled_reward": -0.024721081601455808,
"rewards/format_reward": 0.47916666977107525,
"step": 211
},
{
"completion_length": 2433.3541870117188,
"epoch": 0.2422857142857143,
"grad_norm": 0.01393934153020382,
"kl": 0.00036037713289260864,
"learning_rate": 7.444385869608921e-07,
"loss": -0.0072,
"reward": 0.4994283504784107,
"reward_std": 0.31923972349613905,
"rewards/cosine_scaled_reward": -0.00028583407402038574,
"rewards/format_reward": 0.5,
"step": 212
},
{
"completion_length": 2127.208396911621,
"epoch": 0.24342857142857144,
"grad_norm": 0.03985314443707466,
"kl": 0.0005207061767578125,
"learning_rate": 7.416006812042827e-07,
"loss": 0.0386,
"reward": 0.9866954833269119,
"reward_std": 0.8575046211481094,
"rewards/cosine_scaled_reward": 0.17043107002973557,
"rewards/format_reward": 0.6458333395421505,
"step": 213
},
{
"completion_length": 2820.625045776367,
"epoch": 0.24457142857142858,
"grad_norm": 0.05272352322936058,
"kl": 0.00022479891777038574,
"learning_rate": 7.387534371007797e-07,
"loss": 0.0394,
"reward": 0.6067041866481304,
"reward_std": 0.8792510256171227,
"rewards/cosine_scaled_reward": 0.07418542727828026,
"rewards/format_reward": 0.45833334140479565,
"step": 214
},
{
"completion_length": 2600.395896911621,
"epoch": 0.24571428571428572,
"grad_norm": 0.02042786218225956,
"kl": 0.0003108680248260498,
"learning_rate": 7.358969934210438e-07,
"loss": 0.0222,
"reward": 0.038707589730620384,
"reward_std": 0.41035953164100647,
"rewards/cosine_scaled_reward": -0.2514795530587435,
"rewards/format_reward": 0.5416666679084301,
"step": 215
},
{
"completion_length": 2171.187526702881,
"epoch": 0.24685714285714286,
"grad_norm": 0.030419498682022095,
"kl": 0.0006196498870849609,
"learning_rate": 7.330314893841101e-07,
"loss": 0.0272,
"reward": 0.47143199387937784,
"reward_std": 0.6538783833384514,
"rewards/cosine_scaled_reward": -0.03511733375489712,
"rewards/format_reward": 0.5416666716337204,
"step": 216
},
{
"completion_length": 2755.916679382324,
"epoch": 0.248,
"grad_norm": 0.042481135576963425,
"kl": 0.0003572404384613037,
"learning_rate": 7.301570646506027e-07,
"loss": 0.0345,
"reward": 0.2117303553968668,
"reward_std": 0.9634007029235363,
"rewards/cosine_scaled_reward": -0.11288482137024403,
"rewards/format_reward": 0.4375000111758709,
"step": 217
},
{
"completion_length": 2861.312530517578,
"epoch": 0.24914285714285714,
"grad_norm": 0.03387823700904846,
"kl": 0.0003177523612976074,
"learning_rate": 7.27273859315928e-07,
"loss": 0.0076,
"reward": 0.04665176011621952,
"reward_std": 0.6162600554525852,
"rewards/cosine_scaled_reward": -0.12250744178891182,
"rewards/format_reward": 0.2916666679084301,
"step": 218
},
{
"completion_length": 2398.937515258789,
"epoch": 0.2502857142857143,
"grad_norm": 0.03438759222626686,
"kl": 0.00035077333450317383,
"learning_rate": 7.243820139034464e-07,
"loss": 0.0038,
"reward": 0.5890939775854349,
"reward_std": 0.7752751149237156,
"rewards/cosine_scaled_reward": 0.02371366578154266,
"rewards/format_reward": 0.5416666734963655,
"step": 219
},
{
"completion_length": 2606.1041946411133,
"epoch": 0.25142857142857145,
"grad_norm": 0.012811764143407345,
"kl": 0.0001837611198425293,
"learning_rate": 7.214816693576234e-07,
"loss": 0.0032,
"reward": -0.1477857008576393,
"reward_std": 0.2786343488842249,
"rewards/cosine_scaled_reward": -0.282226188108325,
"rewards/format_reward": 0.4166666679084301,
"step": 220
},
{
"completion_length": 2278.2708587646484,
"epoch": 0.25257142857142856,
"grad_norm": 0.026400132104754448,
"kl": 0.00043514370918273926,
"learning_rate": 7.185729670371604e-07,
"loss": 0.003,
"reward": 0.5976126529276371,
"reward_std": 0.723968205973506,
"rewards/cosine_scaled_reward": 0.027972997166216373,
"rewards/format_reward": 0.5416666734963655,
"step": 221
},
{
"completion_length": 2221.833366394043,
"epoch": 0.2537142857142857,
"grad_norm": 0.032811444252729416,
"kl": 0.000776827335357666,
"learning_rate": 7.156560487081051e-07,
"loss": 0.0661,
"reward": 0.5920759253203869,
"reward_std": 0.5432851929217577,
"rewards/cosine_scaled_reward": 0.01478794775903225,
"rewards/format_reward": 0.5625000055879354,
"step": 222
},
{
"completion_length": 2540.0208892822266,
"epoch": 0.25485714285714284,
"grad_norm": 0.03456755727529526,
"kl": 0.0004057884216308594,
"learning_rate": 7.127310565369415e-07,
"loss": 0.0262,
"reward": 0.34556883573532104,
"reward_std": 0.7227263785898685,
"rewards/cosine_scaled_reward": -0.077215576035087,
"rewards/format_reward": 0.5000000093132257,
"step": 223
},
{
"completion_length": 3318.875030517578,
"epoch": 0.256,
"grad_norm": 0.04437468200922012,
"kl": 0.00018399953842163086,
"learning_rate": 7.097981330836616e-07,
"loss": 0.0454,
"reward": 0.16465971246361732,
"reward_std": 0.7901786658912897,
"rewards/cosine_scaled_reward": -0.06350347894476727,
"rewards/format_reward": 0.29166667722165585,
"step": 224
},
{
"completion_length": 2950.5417404174805,
"epoch": 0.2571428571428571,
"grad_norm": 0.04053080826997757,
"kl": 0.00041607022285461426,
"learning_rate": 7.068574212948169e-07,
"loss": 0.0464,
"reward": 0.01986863650381565,
"reward_std": 0.649408801458776,
"rewards/cosine_scaled_reward": -0.1463156845420599,
"rewards/format_reward": 0.31250000558793545,
"step": 225
},
{
"completion_length": 2706.7500610351562,
"epoch": 0.2582857142857143,
"grad_norm": 0.0555625818669796,
"kl": 0.0008983612060546875,
"learning_rate": 7.039090644965509e-07,
"loss": 0.0504,
"reward": 0.7885134015232325,
"reward_std": 0.8663289472460747,
"rewards/cosine_scaled_reward": 0.14425668492913246,
"rewards/format_reward": 0.5000000055879354,
"step": 226
},
{
"completion_length": 1995.1041831970215,
"epoch": 0.25942857142857145,
"grad_norm": 0.03679594770073891,
"kl": 0.0006301403045654297,
"learning_rate": 7.009532063876148e-07,
"loss": 0.0354,
"reward": 0.3761005823034793,
"reward_std": 0.7344776913523674,
"rewards/cosine_scaled_reward": -0.12444971542572603,
"rewards/format_reward": 0.6250000037252903,
"step": 227
},
{
"completion_length": 2323.4583435058594,
"epoch": 0.26057142857142856,
"grad_norm": 0.03594999015331268,
"kl": 0.00032375752925872803,
"learning_rate": 6.979899910323624e-07,
"loss": 0.019,
"reward": 0.7633006721735001,
"reward_std": 0.7087479755282402,
"rewards/cosine_scaled_reward": 0.1316503193229437,
"rewards/format_reward": 0.5000000037252903,
"step": 228
},
{
"completion_length": 3394.75,
"epoch": 0.26171428571428573,
"grad_norm": 0.030971722677350044,
"kl": 0.00020450353622436523,
"learning_rate": 6.950195628537299e-07,
"loss": 0.0338,
"reward": 0.06433436647057533,
"reward_std": 0.4984890129417181,
"rewards/cosine_scaled_reward": -0.05116615444421768,
"rewards/format_reward": 0.16666667349636555,
"step": 229
},
{
"completion_length": 3045.0625534057617,
"epoch": 0.26285714285714284,
"grad_norm": 0.0360412634909153,
"kl": 0.00022391974925994873,
"learning_rate": 6.920420666261961e-07,
"loss": 0.038,
"reward": 0.02990809455513954,
"reward_std": 0.6434181109070778,
"rewards/cosine_scaled_reward": -0.14129596762359142,
"rewards/format_reward": 0.31250000558793545,
"step": 230
},
{
"completion_length": 2590.312515258789,
"epoch": 0.264,
"grad_norm": 0.025739798322319984,
"kl": 0.00043526291847229004,
"learning_rate": 6.890576474687263e-07,
"loss": -0.0002,
"reward": 0.34738841000944376,
"reward_std": 0.5048788581043482,
"rewards/cosine_scaled_reward": -0.06588914524763823,
"rewards/format_reward": 0.47916666977107525,
"step": 231
},
{
"completion_length": 3344.5625610351562,
"epoch": 0.2651428571428571,
"grad_norm": 0.034720636904239655,
"kl": 0.0002605915069580078,
"learning_rate": 6.860664508377001e-07,
"loss": 0.0408,
"reward": -0.1548066809773445,
"reward_std": 0.537173880264163,
"rewards/cosine_scaled_reward": -0.160736670717597,
"rewards/format_reward": 0.16666666977107525,
"step": 232
},
{
"completion_length": 2555.1666946411133,
"epoch": 0.2662857142857143,
"grad_norm": 0.04706970229744911,
"kl": 0.00042426586151123047,
"learning_rate": 6.83068622519821e-07,
"loss": 0.0466,
"reward": 0.36857184290420264,
"reward_std": 0.7978673111647367,
"rewards/cosine_scaled_reward": -0.08654742129147053,
"rewards/format_reward": 0.5416666753590107,
"step": 233
},
{
"completion_length": 2744.833335876465,
"epoch": 0.2674285714285714,
"grad_norm": 0.030540212988853455,
"kl": 0.0005177855491638184,
"learning_rate": 6.800643086250121e-07,
"loss": 0.0196,
"reward": 0.16414142958819866,
"reward_std": 0.5068646352738142,
"rewards/cosine_scaled_reward": -0.09501262754201889,
"rewards/format_reward": 0.3541666679084301,
"step": 234
},
{
"completion_length": 2465.2708625793457,
"epoch": 0.26857142857142857,
"grad_norm": 0.023667052388191223,
"kl": 0.00041544437408447266,
"learning_rate": 6.770536555792944e-07,
"loss": 0.0041,
"reward": 0.6389858163893223,
"reward_std": 0.5823408327996731,
"rewards/cosine_scaled_reward": 0.11115959542803466,
"rewards/format_reward": 0.4166666679084301,
"step": 235
},
{
"completion_length": 2748.5417098999023,
"epoch": 0.26971428571428574,
"grad_norm": 0.0339183434844017,
"kl": 0.0003077387809753418,
"learning_rate": 6.740368101176495e-07,
"loss": -0.0088,
"reward": 0.16632992029190063,
"reward_std": 0.7507955506443977,
"rewards/cosine_scaled_reward": -0.12516836973372847,
"rewards/format_reward": 0.41666667349636555,
"step": 236
},
{
"completion_length": 2843.1458740234375,
"epoch": 0.27085714285714285,
"grad_norm": 0.03987835720181465,
"kl": 0.0008280277252197266,
"learning_rate": 6.710139192768694e-07,
"loss": 0.0305,
"reward": 0.2756032687611878,
"reward_std": 0.7616374446079135,
"rewards/cosine_scaled_reward": -0.06011504400521517,
"rewards/format_reward": 0.3958333395421505,
"step": 237
},
{
"completion_length": 3365.2083740234375,
"epoch": 0.272,
"grad_norm": 0.051739439368247986,
"kl": 0.00021576881408691406,
"learning_rate": 6.679851303883891e-07,
"loss": 0.0456,
"reward": 0.186602633446455,
"reward_std": 0.8221456026658416,
"rewards/cosine_scaled_reward": -0.07336535991635174,
"rewards/format_reward": 0.33333334513008595,
"step": 238
},
{
"completion_length": 1754.3750076293945,
"epoch": 0.27314285714285713,
"grad_norm": 0.027313968166708946,
"kl": 0.0010634064674377441,
"learning_rate": 6.649505910711058e-07,
"loss": 0.0085,
"reward": 1.1690001524984837,
"reward_std": 0.7178722098469734,
"rewards/cosine_scaled_reward": 0.20950003527104855,
"rewards/format_reward": 0.7500000055879354,
"step": 239
},
{
"completion_length": 3099.708335876465,
"epoch": 0.2742857142857143,
"grad_norm": 0.023615289479494095,
"kl": 0.0005241632461547852,
"learning_rate": 6.619104492241847e-07,
"loss": 0.0084,
"reward": -0.22005018219351768,
"reward_std": 0.3783438242971897,
"rewards/cosine_scaled_reward": -0.2246084287762642,
"rewards/format_reward": 0.2291666716337204,
"step": 240
},
{
"completion_length": 3463.6041870117188,
"epoch": 0.2754285714285714,
"grad_norm": 0.03201238811016083,
"kl": 0.0002460479736328125,
"learning_rate": 6.588648530198504e-07,
"loss": 0.0075,
"reward": -0.10366834700107574,
"reward_std": 0.5739485621452332,
"rewards/cosine_scaled_reward": -0.18725084699690342,
"rewards/format_reward": 0.2708333432674408,
"step": 241
},
{
"completion_length": 2606.791717529297,
"epoch": 0.2765714285714286,
"grad_norm": 0.03816881403326988,
"kl": 0.0006372928619384766,
"learning_rate": 6.558139508961654e-07,
"loss": 0.0182,
"reward": 0.16975077614188194,
"reward_std": 0.6602888819179498,
"rewards/cosine_scaled_reward": -0.14429128426127136,
"rewards/format_reward": 0.4583333395421505,
"step": 242
},
{
"completion_length": 2902.8542098999023,
"epoch": 0.2777142857142857,
"grad_norm": 0.047052547335624695,
"kl": 0.0002627074718475342,
"learning_rate": 6.527578915497951e-07,
"loss": 0.0463,
"reward": 0.24082460720092058,
"reward_std": 0.658990764990449,
"rewards/cosine_scaled_reward": -0.05667102336883545,
"rewards/format_reward": 0.35416667722165585,
"step": 243
},
{
"completion_length": 3030.4375534057617,
"epoch": 0.27885714285714286,
"grad_norm": 0.04656903073191643,
"kl": 0.0015395283699035645,
"learning_rate": 6.496968239287603e-07,
"loss": 0.0069,
"reward": 0.38501549884676933,
"reward_std": 0.8437606617808342,
"rewards/cosine_scaled_reward": 0.01542440801858902,
"rewards/format_reward": 0.35416667349636555,
"step": 244
},
{
"completion_length": 2680.8541870117188,
"epoch": 0.28,
"grad_norm": 0.03648906201124191,
"kl": 0.000688149593770504,
"learning_rate": 6.466308972251785e-07,
"loss": 0.0536,
"reward": 0.19830208271741867,
"reward_std": 0.6960498318076134,
"rewards/cosine_scaled_reward": -0.09876563586294651,
"rewards/format_reward": 0.3958333358168602,
"step": 245
},
{
"completion_length": 2983.104202270508,
"epoch": 0.28114285714285714,
"grad_norm": 0.05028293654322624,
"kl": 0.00043523311614990234,
"learning_rate": 6.435602608679916e-07,
"loss": 0.0866,
"reward": 0.4245212096720934,
"reward_std": 0.8826902657747269,
"rewards/cosine_scaled_reward": -0.006489396095275879,
"rewards/format_reward": 0.43750000186264515,
"step": 246
},
{
"completion_length": 3257.187530517578,
"epoch": 0.2822857142857143,
"grad_norm": 0.02985483780503273,
"kl": 0.0003070831298828125,
"learning_rate": 6.404850645156841e-07,
"loss": 0.0182,
"reward": -0.2367397672496736,
"reward_std": 0.5003039184957743,
"rewards/cosine_scaled_reward": -0.2121198857203126,
"rewards/format_reward": 0.18750000186264515,
"step": 247
},
{
"completion_length": 2173.875030517578,
"epoch": 0.2834285714285714,
"grad_norm": 0.041356880217790604,
"kl": 0.0006842613220214844,
"learning_rate": 6.374054580489873e-07,
"loss": 0.0086,
"reward": 1.2376226875931025,
"reward_std": 0.6682583466172218,
"rewards/cosine_scaled_reward": 0.3271446730941534,
"rewards/format_reward": 0.583333333954215,
"step": 248
},
{
"completion_length": 2189.1666870117188,
"epoch": 0.2845714285714286,
"grad_norm": 0.04028550907969475,
"kl": 0.0006466060876846313,
"learning_rate": 6.343215915635761e-07,
"loss": 0.0563,
"reward": 0.7381115891039371,
"reward_std": 0.7054900173097849,
"rewards/cosine_scaled_reward": 0.08780577592551708,
"rewards/format_reward": 0.5625000055879354,
"step": 249
},
{
"completion_length": 2997.187515258789,
"epoch": 0.2857142857142857,
"grad_norm": 0.06118670478463173,
"kl": 0.0006084442138671875,
"learning_rate": 6.31233615362752e-07,
"loss": 0.0546,
"reward": 0.07891043275594711,
"reward_std": 0.8724489659070969,
"rewards/cosine_scaled_reward": -0.12721144780516624,
"rewards/format_reward": 0.33333333767950535,
"step": 250
},
{
"completion_length": 2369.50004196167,
"epoch": 0.28685714285714287,
"grad_norm": 0.03798350319266319,
"kl": 0.00027310848236083984,
"learning_rate": 6.281416799501187e-07,
"loss": 0.0488,
"reward": 0.7616634331643581,
"reward_std": 0.7558214440941811,
"rewards/cosine_scaled_reward": 0.0787483798339963,
"rewards/format_reward": 0.6041666772216558,
"step": 251
},
{
"completion_length": 2793.125015258789,
"epoch": 0.288,
"grad_norm": 0.026387983933091164,
"kl": 0.0006684064865112305,
"learning_rate": 6.25045936022246e-07,
"loss": 0.0151,
"reward": 0.24926170520484447,
"reward_std": 0.5719782337546349,
"rewards/cosine_scaled_reward": -0.07328582555055618,
"rewards/format_reward": 0.3958333358168602,
"step": 252
},
{
"completion_length": 2910.3125534057617,
"epoch": 0.28914285714285715,
"grad_norm": 0.03298799693584442,
"kl": 0.00051116943359375,
"learning_rate": 6.219465344613258e-07,
"loss": 0.0004,
"reward": 0.2467479296028614,
"reward_std": 0.5817495100200176,
"rewards/cosine_scaled_reward": -0.04329271428287029,
"rewards/format_reward": 0.3333333358168602,
"step": 253
},
{
"completion_length": 2709.437530517578,
"epoch": 0.29028571428571426,
"grad_norm": 0.040090933442115784,
"kl": 0.0008482933044433594,
"learning_rate": 6.188436263278172e-07,
"loss": 0.0401,
"reward": 0.19492281647399068,
"reward_std": 0.803937803953886,
"rewards/cosine_scaled_reward": -0.13170527666807175,
"rewards/format_reward": 0.4583333469927311,
"step": 254
},
{
"completion_length": 3190.4375228881836,
"epoch": 0.2914285714285714,
"grad_norm": 0.032341450452804565,
"kl": 0.0004298090934753418,
"learning_rate": 6.157373628530852e-07,
"loss": 0.0009,
"reward": -0.1900683492422104,
"reward_std": 0.6940745413303375,
"rewards/cosine_scaled_reward": -0.1992008574306965,
"rewards/format_reward": 0.2083333358168602,
"step": 255
},
{
"completion_length": 3150.5625610351562,
"epoch": 0.2925714285714286,
"grad_norm": 0.05708003789186478,
"kl": 0.00038230419158935547,
"learning_rate": 6.126278954320294e-07,
"loss": 0.079,
"reward": 0.40191663336008787,
"reward_std": 0.9853764846920967,
"rewards/cosine_scaled_reward": 0.023874984588474035,
"rewards/format_reward": 0.35416667349636555,
"step": 256
},
{
"completion_length": 3104.395896911621,
"epoch": 0.2937142857142857,
"grad_norm": 0.04910450056195259,
"kl": 0.00029408931732177734,
"learning_rate": 6.095153756157051e-07,
"loss": 0.0179,
"reward": 0.5222748387604952,
"reward_std": 0.8857487104833126,
"rewards/cosine_scaled_reward": 0.07363741181325167,
"rewards/format_reward": 0.37500000558793545,
"step": 257
},
{
"completion_length": 3502.9166870117188,
"epoch": 0.2948571428571429,
"grad_norm": 0.04182790219783783,
"kl": 0.0003981590270996094,
"learning_rate": 6.06399955103937e-07,
"loss": 0.0028,
"reward": -0.13565716426819563,
"reward_std": 0.6289124507457018,
"rewards/cosine_scaled_reward": -0.19282858772203326,
"rewards/format_reward": 0.25000000186264515,
"step": 258
},
{
"completion_length": 2970.1458740234375,
"epoch": 0.296,
"grad_norm": 0.04299155995249748,
"kl": 0.0008516311645507812,
"learning_rate": 6.032817857379256e-07,
"loss": 0.064,
"reward": 0.15694763511419296,
"reward_std": 0.5294061750173569,
"rewards/cosine_scaled_reward": -0.056942859664559364,
"rewards/format_reward": 0.27083333767950535,
"step": 259
},
{
"completion_length": 2157.0625228881836,
"epoch": 0.29714285714285715,
"grad_norm": 0.04957187548279762,
"kl": 0.0002192854881286621,
"learning_rate": 6.001610194928464e-07,
"loss": 0.0369,
"reward": 0.9001389928162098,
"reward_std": 0.9185776375234127,
"rewards/cosine_scaled_reward": 0.1584028152283281,
"rewards/format_reward": 0.5833333395421505,
"step": 260
},
{
"completion_length": 3176.3750228881836,
"epoch": 0.29828571428571427,
"grad_norm": 0.04083000496029854,
"kl": 0.0002732276916503906,
"learning_rate": 5.97037808470444e-07,
"loss": 0.0182,
"reward": 0.18310315907001495,
"reward_std": 0.6294377241283655,
"rewards/cosine_scaled_reward": -0.043865093030035496,
"rewards/format_reward": 0.2708333358168602,
"step": 261
},
{
"completion_length": 3245.2291717529297,
"epoch": 0.29942857142857143,
"grad_norm": 0.019805744290351868,
"kl": 0.00045013427734375,
"learning_rate": 5.939123048916173e-07,
"loss": 0.0029,
"reward": -0.25881996005773544,
"reward_std": 0.3580850623548031,
"rewards/cosine_scaled_reward": -0.20232664234936237,
"rewards/format_reward": 0.14583333395421505,
"step": 262
},
{
"completion_length": 2710.8125228881836,
"epoch": 0.30057142857142854,
"grad_norm": 0.030118314549326897,
"kl": 0.0011047124862670898,
"learning_rate": 5.907846610890011e-07,
"loss": 0.0096,
"reward": -0.03488955553621054,
"reward_std": 0.6062153112143278,
"rewards/cosine_scaled_reward": -0.19452812150120735,
"rewards/format_reward": 0.35416666977107525,
"step": 263
},
{
"completion_length": 2831.041702270508,
"epoch": 0.3017142857142857,
"grad_norm": 0.025041300803422928,
"kl": 0.0004094839096069336,
"learning_rate": 5.87655029499542e-07,
"loss": 0.0005,
"reward": 0.15325931343249977,
"reward_std": 0.5651900470256805,
"rewards/cosine_scaled_reward": -0.11087033338844776,
"rewards/format_reward": 0.3750000037252903,
"step": 264
},
{
"completion_length": 2064.687545776367,
"epoch": 0.3028571428571429,
"grad_norm": 0.03248697519302368,
"kl": 0.001496434211730957,
"learning_rate": 5.845235626570683e-07,
"loss": 0.0528,
"reward": 0.6930933259427547,
"reward_std": 0.6866586022078991,
"rewards/cosine_scaled_reward": 0.04446331039071083,
"rewards/format_reward": 0.6041666772216558,
"step": 265
},
{
"completion_length": 3208.5416717529297,
"epoch": 0.304,
"grad_norm": 0.0215072650462389,
"kl": 0.000579833984375,
"learning_rate": 5.813904131848564e-07,
"loss": -0.0004,
"reward": -0.09674983285367489,
"reward_std": 0.40319861471652985,
"rewards/cosine_scaled_reward": -0.11087492946535349,
"rewards/format_reward": 0.125,
"step": 266
},
{
"completion_length": 3249.3541870117188,
"epoch": 0.30514285714285716,
"grad_norm": 0.03392777964472771,
"kl": 0.0008037090301513672,
"learning_rate": 5.78255733788191e-07,
"loss": -0.0111,
"reward": -0.2167503908276558,
"reward_std": 0.5864087976515293,
"rewards/cosine_scaled_reward": -0.20212520100176334,
"rewards/format_reward": 0.1875000074505806,
"step": 267
},
{
"completion_length": 2485.8750228881836,
"epoch": 0.3062857142857143,
"grad_norm": 0.040282804518938065,
"kl": 0.0007268190383911133,
"learning_rate": 5.751196772469237e-07,
"loss": 0.0522,
"reward": 0.48096081614494324,
"reward_std": 0.8175039198249578,
"rewards/cosine_scaled_reward": 0.021730393171310425,
"rewards/format_reward": 0.43750000558793545,
"step": 268
},
{
"completion_length": 2987.9166870117188,
"epoch": 0.30742857142857144,
"grad_norm": 0.03961595892906189,
"kl": 0.0005930662155151367,
"learning_rate": 5.71982396408026e-07,
"loss": 0.0168,
"reward": 0.15192145481705666,
"reward_std": 0.7950946986675262,
"rewards/cosine_scaled_reward": -0.08028928004205227,
"rewards/format_reward": 0.31250000558793545,
"step": 269
},
{
"completion_length": 3080.625030517578,
"epoch": 0.30857142857142855,
"grad_norm": 0.04469272121787071,
"kl": 0.00038649141788482666,
"learning_rate": 5.688440441781398e-07,
"loss": 0.0194,
"reward": 0.43130027689039707,
"reward_std": 0.9086147882044315,
"rewards/cosine_scaled_reward": 0.03856680044555105,
"rewards/format_reward": 0.35416667349636555,
"step": 270
},
{
"completion_length": 2210.687530517578,
"epoch": 0.3097142857142857,
"grad_norm": 0.03942432254552841,
"kl": 0.0009759664535522461,
"learning_rate": 5.657047735161255e-07,
"loss": 0.0599,
"reward": 0.6358127221465111,
"reward_std": 0.803847155533731,
"rewards/cosine_scaled_reward": -0.02584364265203476,
"rewards/format_reward": 0.687500013038516,
"step": 271
},
{
"completion_length": 2895.625030517578,
"epoch": 0.31085714285714283,
"grad_norm": 0.029458731412887573,
"kl": 0.00041615962982177734,
"learning_rate": 5.625647374256061e-07,
"loss": 0.0218,
"reward": 0.2146900595689658,
"reward_std": 0.5223457869142294,
"rewards/cosine_scaled_reward": -0.06973831099458039,
"rewards/format_reward": 0.35416666977107525,
"step": 272
},
{
"completion_length": 2339.8958435058594,
"epoch": 0.312,
"grad_norm": 0.02511209435760975,
"kl": 0.0005612373352050781,
"learning_rate": 5.594240889475106e-07,
"loss": 0.0052,
"reward": 0.6176238246262074,
"reward_std": 0.6360815949738026,
"rewards/cosine_scaled_reward": 0.058811913593672216,
"rewards/format_reward": 0.5000000055879354,
"step": 273
},
{
"completion_length": 1936.6250305175781,
"epoch": 0.31314285714285717,
"grad_norm": 0.0384528823196888,
"kl": 0.00042247772216796875,
"learning_rate": 5.562829811526154e-07,
"loss": 0.0107,
"reward": 1.0110880993306637,
"reward_std": 0.8275207653641701,
"rewards/cosine_scaled_reward": 0.16179403848946095,
"rewards/format_reward": 0.6875000018626451,
"step": 274
},
{
"completion_length": 2277.541702270508,
"epoch": 0.3142857142857143,
"grad_norm": 0.03816717863082886,
"kl": 0.001892685890197754,
"learning_rate": 5.531415671340826e-07,
"loss": 0.048,
"reward": 0.8937440374866128,
"reward_std": 0.7764279395341873,
"rewards/cosine_scaled_reward": 0.1656220075674355,
"rewards/format_reward": 0.5625000055879354,
"step": 275
},
{
"completion_length": 2728.437530517578,
"epoch": 0.31542857142857145,
"grad_norm": 0.03653620183467865,
"kl": 0.0014216899871826172,
"learning_rate": 5.5e-07,
"loss": -0.0137,
"reward": 0.44955473113805056,
"reward_std": 0.6848132107406855,
"rewards/cosine_scaled_reward": 0.026860691141337156,
"rewards/format_reward": 0.3958333395421505,
"step": 276
},
{
"completion_length": 2431.7500381469727,
"epoch": 0.31657142857142856,
"grad_norm": 0.06454865634441376,
"kl": 0.0006378889083862305,
"learning_rate": 5.468584328659172e-07,
"loss": 0.0178,
"reward": 0.48344410210847855,
"reward_std": 0.7250744290649891,
"rewards/cosine_scaled_reward": 0.012555371969938278,
"rewards/format_reward": 0.45833333395421505,
"step": 277
},
{
"completion_length": 2068.0416870117188,
"epoch": 0.3177142857142857,
"grad_norm": 0.04215235263109207,
"kl": 0.0013701915740966797,
"learning_rate": 5.437170188473847e-07,
"loss": 0.0084,
"reward": 0.961402993183583,
"reward_std": 0.8674832321703434,
"rewards/cosine_scaled_reward": 0.12653482053428888,
"rewards/format_reward": 0.7083333395421505,
"step": 278
},
{
"completion_length": 3224.625,
"epoch": 0.31885714285714284,
"grad_norm": 0.022943608462810516,
"kl": 0.0012241601943969727,
"learning_rate": 5.405759110524894e-07,
"loss": -0.0093,
"reward": -0.18678050953894854,
"reward_std": 0.4048001943156123,
"rewards/cosine_scaled_reward": -0.17672359757125378,
"rewards/format_reward": 0.1666666679084301,
"step": 279
},
{
"completion_length": 2174.9583625793457,
"epoch": 0.32,
"grad_norm": 0.039122022688388824,
"kl": 0.0006496906280517578,
"learning_rate": 5.37435262574394e-07,
"loss": 0.0228,
"reward": 0.7332258597016335,
"reward_std": 0.9436057284474373,
"rewards/cosine_scaled_reward": 0.08536292100325227,
"rewards/format_reward": 0.5625000018626451,
"step": 280
},
{
"completion_length": 3451.625,
"epoch": 0.3211428571428571,
"grad_norm": 0.02864747680723667,
"kl": 0.00041675567626953125,
"learning_rate": 5.342952264838747e-07,
"loss": 0.0233,
"reward": -0.32699717301875353,
"reward_std": 0.46905585937201977,
"rewards/cosine_scaled_reward": -0.21558191592339426,
"rewards/format_reward": 0.10416666977107525,
"step": 281
},
{
"completion_length": 2733.9583740234375,
"epoch": 0.3222857142857143,
"grad_norm": 0.04495016112923622,
"kl": 0.0015126466751098633,
"learning_rate": 5.311559558218603e-07,
"loss": -0.0161,
"reward": 0.39272309094667435,
"reward_std": 0.6825822591781616,
"rewards/cosine_scaled_reward": -0.03280512057244778,
"rewards/format_reward": 0.4583333395421505,
"step": 282
},
{
"completion_length": 2954.0208587646484,
"epoch": 0.32342857142857145,
"grad_norm": 0.03212931752204895,
"kl": 0.0006183385848999023,
"learning_rate": 5.28017603591974e-07,
"loss": 0.0181,
"reward": 0.7553600296378136,
"reward_std": 0.6611499711871147,
"rewards/cosine_scaled_reward": 0.13809666596353054,
"rewards/format_reward": 0.4791666679084301,
"step": 283
},
{
"completion_length": 2272.000045776367,
"epoch": 0.32457142857142857,
"grad_norm": 0.04936413839459419,
"kl": 0.0004665851593017578,
"learning_rate": 5.248803227530763e-07,
"loss": 0.0378,
"reward": 0.5260394886136055,
"reward_std": 0.8827262446284294,
"rewards/cosine_scaled_reward": -0.05989692127332091,
"rewards/format_reward": 0.6458333432674408,
"step": 284
},
{
"completion_length": 2675.916732788086,
"epoch": 0.32571428571428573,
"grad_norm": 0.021733392030000687,
"kl": 0.0008116960525512695,
"learning_rate": 5.21744266211809e-07,
"loss": 0.0038,
"reward": 0.5852530999109149,
"reward_std": 0.5077806441113353,
"rewards/cosine_scaled_reward": -0.030290118884295225,
"rewards/format_reward": 0.6458333395421505,
"step": 285
},
{
"completion_length": 2508.9583892822266,
"epoch": 0.32685714285714285,
"grad_norm": 0.04166712611913681,
"kl": 0.0005065202713012695,
"learning_rate": 5.186095868151436e-07,
"loss": 0.0488,
"reward": 0.4353667050600052,
"reward_std": 0.8739612437784672,
"rewards/cosine_scaled_reward": -0.03231665026396513,
"rewards/format_reward": 0.5,
"step": 286
},
{
"completion_length": 1978.8125343322754,
"epoch": 0.328,
"grad_norm": 0.03577245771884918,
"kl": 0.0008919239044189453,
"learning_rate": 5.154764373429315e-07,
"loss": 0.0558,
"reward": 0.5270357728004456,
"reward_std": 0.6659693606197834,
"rewards/cosine_scaled_reward": -0.01773212105035782,
"rewards/format_reward": 0.5625,
"step": 287
},
{
"completion_length": 3146.0208587646484,
"epoch": 0.3291428571428571,
"grad_norm": 0.04421306028962135,
"kl": 0.0004572868347167969,
"learning_rate": 5.123449705004581e-07,
"loss": 0.0147,
"reward": 0.1639587520621717,
"reward_std": 0.741458848118782,
"rewards/cosine_scaled_reward": -0.0742706386372447,
"rewards/format_reward": 0.3125000074505806,
"step": 288
},
{
"completion_length": 2259.9375038146973,
"epoch": 0.3302857142857143,
"grad_norm": 0.030598286539316177,
"kl": 0.0006687641143798828,
"learning_rate": 5.09215338910999e-07,
"loss": -0.0,
"reward": 0.5977206081151962,
"reward_std": 0.6630996689200401,
"rewards/cosine_scaled_reward": 0.03844361938536167,
"rewards/format_reward": 0.520833333954215,
"step": 289
},
{
"completion_length": 1871.6875305175781,
"epoch": 0.3314285714285714,
"grad_norm": 0.041915781795978546,
"kl": 0.0013911724090576172,
"learning_rate": 5.060876951083828e-07,
"loss": 0.0246,
"reward": 0.7659082859754562,
"reward_std": 0.9081554226577282,
"rewards/cosine_scaled_reward": 0.007954126223921776,
"rewards/format_reward": 0.7500000074505806,
"step": 290
},
{
"completion_length": 2547.1250534057617,
"epoch": 0.3325714285714286,
"grad_norm": 0.0511026568710804,
"kl": 0.0010018348693847656,
"learning_rate": 5.02962191529556e-07,
"loss": 0.0576,
"reward": 0.62677151709795,
"reward_std": 0.9465496614575386,
"rewards/cosine_scaled_reward": 0.0529690757393837,
"rewards/format_reward": 0.5208333414047956,
"step": 291
},
{
"completion_length": 3320.250030517578,
"epoch": 0.33371428571428574,
"grad_norm": 0.04419681057333946,
"kl": 0.0006957054138183594,
"learning_rate": 4.998389805071536e-07,
"loss": 0.0262,
"reward": -0.012292207218706608,
"reward_std": 0.8140805494040251,
"rewards/cosine_scaled_reward": -0.17281277664005756,
"rewards/format_reward": 0.3333333395421505,
"step": 292
},
{
"completion_length": 2636.7916870117188,
"epoch": 0.33485714285714285,
"grad_norm": 0.017586873844265938,
"kl": 0.0006794929504394531,
"learning_rate": 4.967182142620745e-07,
"loss": 0.0078,
"reward": 0.3901491202414036,
"reward_std": 0.326943714171648,
"rewards/cosine_scaled_reward": -0.054925426840782166,
"rewards/format_reward": 0.5000000055879354,
"step": 293
},
{
"completion_length": 3162.4583587646484,
"epoch": 0.336,
"grad_norm": 0.027655253186821938,
"kl": 0.0004392862319946289,
"learning_rate": 4.93600044896063e-07,
"loss": 0.0199,
"reward": 0.11368493363261223,
"reward_std": 0.44556378945708275,
"rewards/cosine_scaled_reward": -0.0994075471535325,
"rewards/format_reward": 0.31250000558793545,
"step": 294
},
{
"completion_length": 3195.6458435058594,
"epoch": 0.33714285714285713,
"grad_norm": 0.03356190025806427,
"kl": 0.0003917217254638672,
"learning_rate": 4.904846243842949e-07,
"loss": 0.0145,
"reward": -0.07907827943563461,
"reward_std": 0.5164209427312016,
"rewards/cosine_scaled_reward": -0.12287248979555443,
"rewards/format_reward": 0.1666666679084301,
"step": 295
},
{
"completion_length": 3008.5625534057617,
"epoch": 0.3382857142857143,
"grad_norm": 0.03211989253759384,
"kl": 0.0007897615432739258,
"learning_rate": 4.873721045679706e-07,
"loss": 0.0291,
"reward": -0.17747718811733648,
"reward_std": 0.5453593544661999,
"rewards/cosine_scaled_reward": -0.23457193188369274,
"rewards/format_reward": 0.2916666716337204,
"step": 296
},
{
"completion_length": 3558.3958740234375,
"epoch": 0.3394285714285714,
"grad_norm": 0.0319380946457386,
"kl": 0.00032269954681396484,
"learning_rate": 4.842626371469149e-07,
"loss": 0.0192,
"reward": -0.3608126677572727,
"reward_std": 0.5530665349215269,
"rewards/cosine_scaled_reward": -0.20123967458494008,
"rewards/format_reward": 0.0416666679084301,
"step": 297
},
{
"completion_length": 2848.7708740234375,
"epoch": 0.3405714285714286,
"grad_norm": 0.03126157075166702,
"kl": 0.00034427642822265625,
"learning_rate": 4.811563736721829e-07,
"loss": 0.0216,
"reward": 0.294229757739231,
"reward_std": 0.6536255031824112,
"rewards/cosine_scaled_reward": -0.06121844658628106,
"rewards/format_reward": 0.41666666977107525,
"step": 298
},
{
"completion_length": 3238.6666717529297,
"epoch": 0.3417142857142857,
"grad_norm": 0.03061261959373951,
"kl": 0.0004975795745849609,
"learning_rate": 4.780534655386743e-07,
"loss": 0.0007,
"reward": 0.2360800839960575,
"reward_std": 0.48118265718221664,
"rewards/cosine_scaled_reward": -0.01737664802931249,
"rewards/format_reward": 0.27083333395421505,
"step": 299
},
{
"completion_length": 3446.6041870117188,
"epoch": 0.34285714285714286,
"grad_norm": 0.0574151873588562,
"kl": 0.0003980398178100586,
"learning_rate": 4.749540639777539e-07,
"loss": 0.061,
"reward": -0.1822480447590351,
"reward_std": 0.7430878691375256,
"rewards/cosine_scaled_reward": -0.1744573526084423,
"rewards/format_reward": 0.16666666977107525,
"step": 300
},
{
"completion_length": 2296.041679382324,
"epoch": 0.344,
"grad_norm": 0.036582205444574356,
"kl": 0.0013877153396606445,
"learning_rate": 4.7185832004988133e-07,
"loss": 0.0257,
"reward": 0.3718048664741218,
"reward_std": 0.7346825525164604,
"rewards/cosine_scaled_reward": -0.1161809004843235,
"rewards/format_reward": 0.6041666809469461,
"step": 301
},
{
"completion_length": 2405.125026702881,
"epoch": 0.34514285714285714,
"grad_norm": 0.032144494354724884,
"kl": 0.0007233619689941406,
"learning_rate": 4.68766384637248e-07,
"loss": 0.0155,
"reward": 0.7637341096997261,
"reward_std": 0.5499859433621168,
"rewards/cosine_scaled_reward": 0.12145039439201355,
"rewards/format_reward": 0.5208333395421505,
"step": 302
},
{
"completion_length": 2446.1250381469727,
"epoch": 0.3462857142857143,
"grad_norm": 0.04364825040102005,
"kl": 0.0014946460723876953,
"learning_rate": 4.656784084364238e-07,
"loss": 0.062,
"reward": 0.37057532370090485,
"reward_std": 0.7990845926105976,
"rewards/cosine_scaled_reward": -0.08554568700492382,
"rewards/format_reward": 0.5416666734963655,
"step": 303
},
{
"completion_length": 2854.937530517578,
"epoch": 0.3474285714285714,
"grad_norm": 0.03800087422132492,
"kl": 0.0013117790222167969,
"learning_rate": 4.6259454195101267e-07,
"loss": 0.0191,
"reward": -0.04802834615111351,
"reward_std": 0.5804362390190363,
"rewards/cosine_scaled_reward": -0.19068085309118032,
"rewards/format_reward": 0.33333333395421505,
"step": 304
},
{
"completion_length": 2861.166679382324,
"epoch": 0.3485714285714286,
"grad_norm": 0.03797345981001854,
"kl": 0.001414179801940918,
"learning_rate": 4.59514935484316e-07,
"loss": -0.001,
"reward": 0.20239191502332687,
"reward_std": 0.6676407773047686,
"rewards/cosine_scaled_reward": -0.07588737271726131,
"rewards/format_reward": 0.3541666679084301,
"step": 305
},
{
"completion_length": 2406.4791870117188,
"epoch": 0.3497142857142857,
"grad_norm": 0.041493017226457596,
"kl": 0.0019769668579101562,
"learning_rate": 4.5643973913200837e-07,
"loss": 0.0715,
"reward": 0.4393942207098007,
"reward_std": 0.6489860704168677,
"rewards/cosine_scaled_reward": -0.03030291385948658,
"rewards/format_reward": 0.5000000037252903,
"step": 306
},
{
"completion_length": 2538.68754196167,
"epoch": 0.35085714285714287,
"grad_norm": 0.03819847106933594,
"kl": 0.0011295080184936523,
"learning_rate": 4.5336910277482155e-07,
"loss": 0.0372,
"reward": 0.3050793409347534,
"reward_std": 0.7908993605524302,
"rewards/cosine_scaled_reward": -0.09746033605188131,
"rewards/format_reward": 0.5000000093132257,
"step": 307
},
{
"completion_length": 3305.375,
"epoch": 0.352,
"grad_norm": 0.025576284155249596,
"kl": 0.0006071329116821289,
"learning_rate": 4.503031760712397e-07,
"loss": -0.0213,
"reward": -0.008270788937807083,
"reward_std": 0.5344270393252373,
"rewards/cosine_scaled_reward": -0.10830206237733364,
"rewards/format_reward": 0.2083333358168602,
"step": 308
},
{
"completion_length": 3398.062530517578,
"epoch": 0.35314285714285715,
"grad_norm": 0.036179594695568085,
"kl": 0.0006154775619506836,
"learning_rate": 4.4724210845020494e-07,
"loss": 0.0303,
"reward": 0.17174796015024185,
"reward_std": 0.718455646187067,
"rewards/cosine_scaled_reward": -0.039126007817685604,
"rewards/format_reward": 0.2500000037252903,
"step": 309
},
{
"completion_length": 2386.5208892822266,
"epoch": 0.35428571428571426,
"grad_norm": 0.019351672381162643,
"kl": 0.0006570816040039062,
"learning_rate": 4.441860491038345e-07,
"loss": 0.0162,
"reward": 0.014205506071448326,
"reward_std": 0.39019110426306725,
"rewards/cosine_scaled_reward": -0.23248059302568436,
"rewards/format_reward": 0.47916666977107525,
"step": 310
},
{
"completion_length": 2644.9791984558105,
"epoch": 0.3554285714285714,
"grad_norm": 0.03210190311074257,
"kl": 0.0006589889526367188,
"learning_rate": 4.4113514698014953e-07,
"loss": 0.0523,
"reward": 0.45767392963171005,
"reward_std": 0.5884373337030411,
"rewards/cosine_scaled_reward": 0.010086966678500175,
"rewards/format_reward": 0.4375000149011612,
"step": 311
},
{
"completion_length": 2269.062530517578,
"epoch": 0.3565714285714286,
"grad_norm": 0.054676324129104614,
"kl": 0.0008935928344726562,
"learning_rate": 4.3808955077581546e-07,
"loss": 0.1022,
"reward": 0.5176713224500418,
"reward_std": 0.6970217786729336,
"rewards/cosine_scaled_reward": 0.02966897562146187,
"rewards/format_reward": 0.45833334885537624,
"step": 312
},
{
"completion_length": 3095.1250534057617,
"epoch": 0.3577142857142857,
"grad_norm": 0.051631879061460495,
"kl": 0.0010917186737060547,
"learning_rate": 4.350494089288943e-07,
"loss": 0.0596,
"reward": 0.1791437454521656,
"reward_std": 0.832375283818692,
"rewards/cosine_scaled_reward": -0.03542814147658646,
"rewards/format_reward": 0.2500000037252903,
"step": 313
},
{
"completion_length": 2249.2291946411133,
"epoch": 0.3588571428571429,
"grad_norm": 0.0347190760076046,
"kl": 0.0015243291854858398,
"learning_rate": 4.3201486961161093e-07,
"loss": 0.0493,
"reward": 0.6420404259115458,
"reward_std": 0.6056624613702297,
"rewards/cosine_scaled_reward": 0.029353542253375053,
"rewards/format_reward": 0.5833333432674408,
"step": 314
},
{
"completion_length": 2830.104217529297,
"epoch": 0.36,
"grad_norm": 0.029440823942422867,
"kl": 0.0004910826683044434,
"learning_rate": 4.2898608072313045e-07,
"loss": 0.0184,
"reward": 0.5590658411383629,
"reward_std": 0.5381645411252975,
"rewards/cosine_scaled_reward": 0.060782887041568756,
"rewards/format_reward": 0.43750000186264515,
"step": 315
},
{
"completion_length": 3479.0625,
"epoch": 0.36114285714285715,
"grad_norm": 0.039480630308389664,
"kl": 0.00043773651123046875,
"learning_rate": 4.2596318988235037e-07,
"loss": 0.0194,
"reward": -0.34035787358880043,
"reward_std": 0.6184021793305874,
"rewards/cosine_scaled_reward": -0.22226226888597012,
"rewards/format_reward": 0.10416666977107525,
"step": 316
},
{
"completion_length": 3024.6458892822266,
"epoch": 0.36228571428571427,
"grad_norm": 0.04857990890741348,
"kl": 0.0009016990661621094,
"learning_rate": 4.2294634442070553e-07,
"loss": 0.0733,
"reward": 0.15108232758939266,
"reward_std": 0.8656900152564049,
"rewards/cosine_scaled_reward": -0.09112551622092724,
"rewards/format_reward": 0.33333333767950535,
"step": 317
},
{
"completion_length": 2164.416732788086,
"epoch": 0.36342857142857143,
"grad_norm": 0.06030331179499626,
"kl": 0.002919435501098633,
"learning_rate": 4.1993569137498776e-07,
"loss": 0.0517,
"reward": 0.6099264821968973,
"reward_std": 0.9467958249151707,
"rewards/cosine_scaled_reward": -0.028370092622935772,
"rewards/format_reward": 0.6666666716337204,
"step": 318
},
{
"completion_length": 2844.4375228881836,
"epoch": 0.36457142857142855,
"grad_norm": 0.02526162751019001,
"kl": 0.0008511543273925781,
"learning_rate": 4.1693137748017915e-07,
"loss": 0.041,
"reward": -0.2095866478048265,
"reward_std": 0.45644592493772507,
"rewards/cosine_scaled_reward": -0.24020998924970627,
"rewards/format_reward": 0.27083333767950535,
"step": 319
},
{
"completion_length": 2073.020835876465,
"epoch": 0.3657142857142857,
"grad_norm": 0.05878617987036705,
"kl": 0.001214742660522461,
"learning_rate": 4.1393354916230005e-07,
"loss": 0.0585,
"reward": 0.3800775744020939,
"reward_std": 0.6021665297448635,
"rewards/cosine_scaled_reward": -0.1016278937458992,
"rewards/format_reward": 0.5833333432674408,
"step": 320
},
{
"completion_length": 1917.3750228881836,
"epoch": 0.3668571428571429,
"grad_norm": 0.03980216011404991,
"kl": 0.001294553279876709,
"learning_rate": 4.1094235253127374e-07,
"loss": 0.0804,
"reward": 0.7248329035937786,
"reward_std": 0.6771653518080711,
"rewards/cosine_scaled_reward": 0.008249786798842251,
"rewards/format_reward": 0.7083333395421505,
"step": 321
},
{
"completion_length": 2903.875011444092,
"epoch": 0.368,
"grad_norm": 0.0290102269500494,
"kl": 0.0006527900695800781,
"learning_rate": 4.079579333738039e-07,
"loss": 0.0155,
"reward": -0.1609306987375021,
"reward_std": 0.4162718243896961,
"rewards/cosine_scaled_reward": -0.1950486795976758,
"rewards/format_reward": 0.2291666716337204,
"step": 322
},
{
"completion_length": 2931.895851135254,
"epoch": 0.36914285714285716,
"grad_norm": 0.019431445747613907,
"kl": 0.0010836124420166016,
"learning_rate": 4.0498043714627006e-07,
"loss": 0.0199,
"reward": 0.35527924820780754,
"reward_std": 0.44612617418169975,
"rewards/cosine_scaled_reward": -0.020277027040719986,
"rewards/format_reward": 0.39583333395421505,
"step": 323
},
{
"completion_length": 2903.729217529297,
"epoch": 0.3702857142857143,
"grad_norm": 0.0290395338088274,
"kl": 0.0006427764892578125,
"learning_rate": 4.020100089676376e-07,
"loss": 0.019,
"reward": 0.27246633544564247,
"reward_std": 0.5725754294544458,
"rewards/cosine_scaled_reward": -0.05126684159040451,
"rewards/format_reward": 0.37500000558793545,
"step": 324
},
{
"completion_length": 2823.7500610351562,
"epoch": 0.37142857142857144,
"grad_norm": 0.027312133461236954,
"kl": 0.0011664628982543945,
"learning_rate": 3.9904679361238526e-07,
"loss": 0.0044,
"reward": 0.5526620149612427,
"reward_std": 0.578266765922308,
"rewards/cosine_scaled_reward": 0.05758099630475044,
"rewards/format_reward": 0.43750000186264515,
"step": 325
},
{
"completion_length": 2344.4792098999023,
"epoch": 0.37257142857142855,
"grad_norm": 0.025410566478967667,
"kl": 0.0010523796081542969,
"learning_rate": 3.9609093550344907e-07,
"loss": 0.036,
"reward": 0.6994030121713877,
"reward_std": 0.48659895546734333,
"rewards/cosine_scaled_reward": 0.06845148652791977,
"rewards/format_reward": 0.5625000018626451,
"step": 326
},
{
"completion_length": 2858.5833435058594,
"epoch": 0.3737142857142857,
"grad_norm": 0.033931881189346313,
"kl": 0.0009119510650634766,
"learning_rate": 3.931425787051832e-07,
"loss": -0.0144,
"reward": 0.597864555194974,
"reward_std": 0.6859371922910213,
"rewards/cosine_scaled_reward": 0.10101561388000846,
"rewards/format_reward": 0.3958333395421505,
"step": 327
},
{
"completion_length": 3584.0,
"epoch": 0.37485714285714283,
"grad_norm": 0.02540241740643978,
"kl": 0.0006210803985595703,
"learning_rate": 3.902018669163384e-07,
"loss": 0.0,
"reward": -0.4537691902369261,
"reward_std": 0.4180416464805603,
"rewards/cosine_scaled_reward": -0.237301261164248,
"rewards/format_reward": 0.02083333395421505,
"step": 328
},
{
"completion_length": 2313.520881652832,
"epoch": 0.376,
"grad_norm": 0.0391564816236496,
"kl": 0.0012059211730957031,
"learning_rate": 3.872689434630585e-07,
"loss": 0.007,
"reward": 0.6544864065945148,
"reward_std": 0.6928528789430857,
"rewards/cosine_scaled_reward": 0.056409850250929594,
"rewards/format_reward": 0.5416666716337204,
"step": 329
},
{
"completion_length": 2332.1250228881836,
"epoch": 0.37714285714285717,
"grad_norm": 0.04612577334046364,
"kl": 0.0016050338745117188,
"learning_rate": 3.843439512918949e-07,
"loss": 0.0728,
"reward": 0.27069247141480446,
"reward_std": 0.5735555719584227,
"rewards/cosine_scaled_reward": -0.1250704526901245,
"rewards/format_reward": 0.520833345130086,
"step": 330
},
{
"completion_length": 2372.687515258789,
"epoch": 0.3782857142857143,
"grad_norm": 0.02713729254901409,
"kl": 0.0010589361190795898,
"learning_rate": 3.8142703296283953e-07,
"loss": 0.0437,
"reward": 0.070272296667099,
"reward_std": 0.49196940287947655,
"rewards/cosine_scaled_reward": -0.18361386097967625,
"rewards/format_reward": 0.4375,
"step": 331
},
{
"completion_length": 2441.0833587646484,
"epoch": 0.37942857142857145,
"grad_norm": 0.31212538480758667,
"kl": 0.0019620656967163086,
"learning_rate": 3.785183306423767e-07,
"loss": 0.037,
"reward": 0.42729340121150017,
"reward_std": 0.5920264758169651,
"rewards/cosine_scaled_reward": -0.046769993379712105,
"rewards/format_reward": 0.5208333395421505,
"step": 332
},
{
"completion_length": 2342.8958435058594,
"epoch": 0.38057142857142856,
"grad_norm": 0.024794911965727806,
"kl": 0.0012290477752685547,
"learning_rate": 3.7561798609655373e-07,
"loss": 0.0199,
"reward": 0.3754307944327593,
"reward_std": 0.525186138227582,
"rewards/cosine_scaled_reward": -0.06228460434067529,
"rewards/format_reward": 0.5000000037252903,
"step": 333
},
{
"completion_length": 3427.000030517578,
"epoch": 0.38171428571428573,
"grad_norm": 0.03751340135931969,
"kl": 0.0007123947143554688,
"learning_rate": 3.72726140684072e-07,
"loss": 0.0284,
"reward": -0.16761411912739277,
"reward_std": 0.5903099803254008,
"rewards/cosine_scaled_reward": -0.1671403832733631,
"rewards/format_reward": 0.16666666977107525,
"step": 334
},
{
"completion_length": 2378.3958892822266,
"epoch": 0.38285714285714284,
"grad_norm": 0.03808344900608063,
"kl": 0.0016427040100097656,
"learning_rate": 3.6984293534939737e-07,
"loss": 0.0017,
"reward": 0.486670202575624,
"reward_std": 0.7270042449235916,
"rewards/cosine_scaled_reward": -0.03791489452123642,
"rewards/format_reward": 0.5625000074505806,
"step": 335
},
{
"completion_length": 3137.541717529297,
"epoch": 0.384,
"grad_norm": 0.05256379023194313,
"kl": 0.0007867813110351562,
"learning_rate": 3.6696851061588994e-07,
"loss": 0.0447,
"reward": 0.24836627580225468,
"reward_std": 0.8177415505051613,
"rewards/cosine_scaled_reward": -0.05290019465610385,
"rewards/format_reward": 0.35416668094694614,
"step": 336
},
{
"completion_length": 3336.8333435058594,
"epoch": 0.3851428571428571,
"grad_norm": 0.040195003151893616,
"kl": 0.0008945465087890625,
"learning_rate": 3.641030065789562e-07,
"loss": -0.0036,
"reward": -0.056929025799036026,
"reward_std": 0.6236543748527765,
"rewards/cosine_scaled_reward": -0.13263118267059326,
"rewards/format_reward": 0.2083333395421505,
"step": 337
},
{
"completion_length": 2392.1250534057617,
"epoch": 0.3862857142857143,
"grad_norm": 0.06893110275268555,
"kl": 0.0015697479248046875,
"learning_rate": 3.612465628992203e-07,
"loss": 0.1161,
"reward": 0.6895103082060814,
"reward_std": 0.8643823079764843,
"rewards/cosine_scaled_reward": 0.07392181595787406,
"rewards/format_reward": 0.5416666716337204,
"step": 338
},
{
"completion_length": 3073.0625228881836,
"epoch": 0.38742857142857146,
"grad_norm": 0.03818909078836441,
"kl": 0.0008697509765625,
"learning_rate": 3.5839931879571725e-07,
"loss": 0.0467,
"reward": -0.09941295813769102,
"reward_std": 0.49821673333644867,
"rewards/cosine_scaled_reward": -0.1538731474429369,
"rewards/format_reward": 0.2083333358168602,
"step": 339
},
{
"completion_length": 2484.5625381469727,
"epoch": 0.38857142857142857,
"grad_norm": 0.040925703942775726,
"kl": 0.0010941028594970703,
"learning_rate": 3.555614130391079e-07,
"loss": 0.0031,
"reward": 0.39815794909372926,
"reward_std": 0.7465380132198334,
"rewards/cosine_scaled_reward": -0.030087683349847794,
"rewards/format_reward": 0.4583333358168602,
"step": 340
},
{
"completion_length": 2501.6041946411133,
"epoch": 0.38971428571428574,
"grad_norm": 0.03836272656917572,
"kl": 0.0011022090911865234,
"learning_rate": 3.5273298394491515e-07,
"loss": 0.036,
"reward": 0.49983580166008323,
"reward_std": 0.7316713891923428,
"rewards/cosine_scaled_reward": -8.210213854908943e-05,
"rewards/format_reward": 0.5000000055879354,
"step": 341
},
{
"completion_length": 3069.2291870117188,
"epoch": 0.39085714285714285,
"grad_norm": 0.03620956465601921,
"kl": 0.0009083747863769531,
"learning_rate": 3.4991416936678276e-07,
"loss": 0.0476,
"reward": 0.07453913427889347,
"reward_std": 0.5634984392672777,
"rewards/cosine_scaled_reward": -0.0668970886617899,
"rewards/format_reward": 0.2083333358168602,
"step": 342
},
{
"completion_length": 3340.1875,
"epoch": 0.392,
"grad_norm": 0.0366065539419651,
"kl": 0.0004994869232177734,
"learning_rate": 3.471051066897562e-07,
"loss": 0.048,
"reward": 0.05115175619721413,
"reward_std": 0.7214457457885146,
"rewards/cosine_scaled_reward": -0.09942414052784443,
"rewards/format_reward": 0.2500000074505806,
"step": 343
},
{
"completion_length": 2400.5000915527344,
"epoch": 0.3931428571428571,
"grad_norm": 0.026572655886411667,
"kl": 0.0007510185241699219,
"learning_rate": 3.4430593282358777e-07,
"loss": 0.0261,
"reward": 0.9928230717778206,
"reward_std": 0.5358506329357624,
"rewards/cosine_scaled_reward": 0.18391155824065208,
"rewards/format_reward": 0.6250000055879354,
"step": 344
},
{
"completion_length": 2765.3541870117188,
"epoch": 0.3942857142857143,
"grad_norm": 0.04380709305405617,
"kl": 0.0007009506225585938,
"learning_rate": 3.4151678419606233e-07,
"loss": 0.0862,
"reward": 0.2515047416090965,
"reward_std": 0.5667482540011406,
"rewards/cosine_scaled_reward": -0.051330966874957085,
"rewards/format_reward": 0.35416666977107525,
"step": 345
},
{
"completion_length": 3413.5833740234375,
"epoch": 0.3954285714285714,
"grad_norm": 0.04162891209125519,
"kl": 0.0006406307220458984,
"learning_rate": 3.387377967463493e-07,
"loss": 0.0233,
"reward": -0.05835883319377899,
"reward_std": 0.6794736217707396,
"rewards/cosine_scaled_reward": -0.17501276172697544,
"rewards/format_reward": 0.29166667349636555,
"step": 346
},
{
"completion_length": 3015.500030517578,
"epoch": 0.3965714285714286,
"grad_norm": 0.04500904306769371,
"kl": 0.0009529590606689453,
"learning_rate": 3.359691059183761e-07,
"loss": 0.0428,
"reward": -0.06104391813278198,
"reward_std": 0.7086496539413929,
"rewards/cosine_scaled_reward": -0.20760529045946896,
"rewards/format_reward": 0.3541666716337204,
"step": 347
},
{
"completion_length": 2832.8541946411133,
"epoch": 0.3977142857142857,
"grad_norm": 0.019439518451690674,
"kl": 0.0007596015930175781,
"learning_rate": 3.3321084665422803e-07,
"loss": 0.0,
"reward": 0.2556960296933539,
"reward_std": 0.40971279237419367,
"rewards/cosine_scaled_reward": -0.0596519922837615,
"rewards/format_reward": 0.375,
"step": 348
},
{
"completion_length": 2908.812515258789,
"epoch": 0.39885714285714285,
"grad_norm": 0.03599149361252785,
"kl": 0.0007562637329101562,
"learning_rate": 3.3046315338757026e-07,
"loss": 0.0131,
"reward": 0.2589077763259411,
"reward_std": 0.626338753849268,
"rewards/cosine_scaled_reward": -0.08929613418877125,
"rewards/format_reward": 0.4375000149011612,
"step": 349
},
{
"completion_length": 2553.083339691162,
"epoch": 0.4,
"grad_norm": 0.027066772803664207,
"kl": 0.0008411407470703125,
"learning_rate": 3.2772616003709616e-07,
"loss": 0.0154,
"reward": 0.4995659068226814,
"reward_std": 0.4548289757221937,
"rewards/cosine_scaled_reward": 0.010199605487287045,
"rewards/format_reward": 0.4791666716337204,
"step": 350
},
{
"completion_length": 3268.625030517578,
"epoch": 0.40114285714285713,
"grad_norm": 0.03507920354604721,
"kl": 0.0008692741394042969,
"learning_rate": 3.250000000000001e-07,
"loss": 0.0301,
"reward": -0.001983426511287689,
"reward_std": 0.5756698325276375,
"rewards/cosine_scaled_reward": -0.13640837743878365,
"rewards/format_reward": 0.2708333395421505,
"step": 351
},
{
"completion_length": 2692.9583587646484,
"epoch": 0.4022857142857143,
"grad_norm": 0.03933379054069519,
"kl": 0.0012662410736083984,
"learning_rate": 3.222848061454764e-07,
"loss": 0.0339,
"reward": 0.39516958221793175,
"reward_std": 0.7729976400732994,
"rewards/cosine_scaled_reward": -0.010748542845249176,
"rewards/format_reward": 0.41666667349636555,
"step": 352
},
{
"completion_length": 2451.9375,
"epoch": 0.4034285714285714,
"grad_norm": 0.02567175030708313,
"kl": 0.0020258426666259766,
"learning_rate": 3.195807108082429e-07,
"loss": -0.0063,
"reward": 0.22795281372964382,
"reward_std": 0.49539407063275576,
"rewards/cosine_scaled_reward": -0.12560690939426422,
"rewards/format_reward": 0.4791666716337204,
"step": 353
},
{
"completion_length": 1874.9791793823242,
"epoch": 0.4045714285714286,
"grad_norm": 0.029671069234609604,
"kl": 0.0017531514167785645,
"learning_rate": 3.168878457820915e-07,
"loss": 0.0081,
"reward": 0.940945464768447,
"reward_std": 0.7216029353439808,
"rewards/cosine_scaled_reward": 0.11630604974925518,
"rewards/format_reward": 0.7083333358168602,
"step": 354
},
{
"completion_length": 2448.7500381469727,
"epoch": 0.4057142857142857,
"grad_norm": 0.03424989804625511,
"kl": 0.0019562244415283203,
"learning_rate": 3.142063423134644e-07,
"loss": -0.0181,
"reward": 0.7008613422513008,
"reward_std": 0.6350522413849831,
"rewards/cosine_scaled_reward": 0.06918067391961813,
"rewards/format_reward": 0.5625000074505806,
"step": 355
},
{
"completion_length": 2569.250015258789,
"epoch": 0.40685714285714286,
"grad_norm": 0.032561905682086945,
"kl": 0.0011096000671386719,
"learning_rate": 3.115363310950578e-07,
"loss": 0.0258,
"reward": 0.5701954085379839,
"reward_std": 0.7334288004785776,
"rewards/cosine_scaled_reward": -0.006568958284333348,
"rewards/format_reward": 0.5833333432674408,
"step": 356
},
{
"completion_length": 3241.1250610351562,
"epoch": 0.408,
"grad_norm": 0.03412913531064987,
"kl": 0.0009722709655761719,
"learning_rate": 3.0887794225945143e-07,
"loss": 0.0399,
"reward": -0.1981919500976801,
"reward_std": 0.6031934916973114,
"rewards/cosine_scaled_reward": -0.22409598156809807,
"rewards/format_reward": 0.25000000558793545,
"step": 357
},
{
"completion_length": 2790.041748046875,
"epoch": 0.40914285714285714,
"grad_norm": 0.05226265639066696,
"kl": 0.0027484893798828125,
"learning_rate": 3.062313053727671e-07,
"loss": 0.0425,
"reward": 0.5322411619126797,
"reward_std": 0.9997622594237328,
"rewards/cosine_scaled_reward": 0.03695392981171608,
"rewards/format_reward": 0.4583333358168602,
"step": 358
},
{
"completion_length": 2785.3958587646484,
"epoch": 0.4102857142857143,
"grad_norm": 0.03838330879807472,
"kl": 0.0009531974792480469,
"learning_rate": 3.0359654942835247e-07,
"loss": 0.0174,
"reward": 0.24072397220879793,
"reward_std": 0.742347358725965,
"rewards/cosine_scaled_reward": -0.08797135204076767,
"rewards/format_reward": 0.41666667722165585,
"step": 359
},
{
"completion_length": 2664.7083892822266,
"epoch": 0.4114285714285714,
"grad_norm": 0.04071582481265068,
"kl": 0.0034384727478027344,
"learning_rate": 3.0097380284049523e-07,
"loss": 0.0468,
"reward": 0.322513684630394,
"reward_std": 0.7571201398968697,
"rewards/cosine_scaled_reward": -0.0366598404943943,
"rewards/format_reward": 0.3958333432674408,
"step": 360
},
{
"completion_length": 3133.4167098999023,
"epoch": 0.4125714285714286,
"grad_norm": 0.04097485542297363,
"kl": 0.0013079643249511719,
"learning_rate": 2.9836319343816397e-07,
"loss": 0.0059,
"reward": 0.24762597493827343,
"reward_std": 0.7427336126565933,
"rewards/cosine_scaled_reward": -0.032437026500701904,
"rewards/format_reward": 0.3125000074505806,
"step": 361
},
{
"completion_length": 1420.1875114440918,
"epoch": 0.4137142857142857,
"grad_norm": 0.03498203307390213,
"kl": 0.0024487972259521484,
"learning_rate": 2.9576484845877793e-07,
"loss": 0.0363,
"reward": 1.1095450557768345,
"reward_std": 0.6055663973093033,
"rewards/cosine_scaled_reward": 0.127689179033041,
"rewards/format_reward": 0.8541666753590107,
"step": 362
},
{
"completion_length": 2218.708381652832,
"epoch": 0.41485714285714287,
"grad_norm": 0.03273200988769531,
"kl": 0.001424551010131836,
"learning_rate": 2.931788945420058e-07,
"loss": 0.0456,
"reward": 0.5941947225946933,
"reward_std": 0.6653445195406675,
"rewards/cosine_scaled_reward": 0.015847355127334595,
"rewards/format_reward": 0.5625000037252903,
"step": 363
},
{
"completion_length": 2686.854202270508,
"epoch": 0.416,
"grad_norm": 0.033127423375844955,
"kl": 0.0012614727020263672,
"learning_rate": 2.9060545772359305e-07,
"loss": 0.0182,
"reward": 0.13789374008774757,
"reward_std": 0.517003508284688,
"rewards/cosine_scaled_reward": -0.16021979413926601,
"rewards/format_reward": 0.4583333358168602,
"step": 364
},
{
"completion_length": 2956.166702270508,
"epoch": 0.41714285714285715,
"grad_norm": 0.048571985214948654,
"kl": 0.0009555816650390625,
"learning_rate": 2.8804466342921987e-07,
"loss": 0.0338,
"reward": 0.0561866071075201,
"reward_std": 0.7425644807517529,
"rewards/cosine_scaled_reward": -0.1177400229498744,
"rewards/format_reward": 0.29166667349636555,
"step": 365
},
{
"completion_length": 1896.750036239624,
"epoch": 0.41828571428571426,
"grad_norm": 0.030456840991973877,
"kl": 0.0005739927291870117,
"learning_rate": 2.854966364683872e-07,
"loss": 0.0198,
"reward": 0.781156275421381,
"reward_std": 0.6628811918199062,
"rewards/cosine_scaled_reward": 0.046828120946884155,
"rewards/format_reward": 0.6875,
"step": 366
},
{
"completion_length": 2907.187530517578,
"epoch": 0.41942857142857143,
"grad_norm": 0.04903361201286316,
"kl": 0.0008039474487304688,
"learning_rate": 2.829615010283344e-07,
"loss": -0.001,
"reward": 0.41779946256428957,
"reward_std": 0.8784012943506241,
"rewards/cosine_scaled_reward": -0.030683615244925022,
"rewards/format_reward": 0.47916666977107525,
"step": 367
},
{
"completion_length": 2909.0000076293945,
"epoch": 0.4205714285714286,
"grad_norm": 0.03092154674232006,
"kl": 0.0025424957275390625,
"learning_rate": 2.8043938066798645e-07,
"loss": -0.0078,
"reward": 0.0029743313789367676,
"reward_std": 0.4080376699566841,
"rewards/cosine_scaled_reward": -0.12351284455507994,
"rewards/format_reward": 0.25,
"step": 368
},
{
"completion_length": 3040.6042098999023,
"epoch": 0.4217142857142857,
"grad_norm": 0.05602690950036049,
"kl": 0.0014162063598632812,
"learning_rate": 2.7793039831193133e-07,
"loss": 0.0144,
"reward": 0.3254664018750191,
"reward_std": 0.957383282482624,
"rewards/cosine_scaled_reward": -0.035183459520339966,
"rewards/format_reward": 0.3958333358168602,
"step": 369
},
{
"completion_length": 3252.25,
"epoch": 0.4228571428571429,
"grad_norm": 0.02123953402042389,
"kl": 0.0016574859619140625,
"learning_rate": 2.7543467624442956e-07,
"loss": 0.0164,
"reward": 0.04274958372116089,
"reward_std": 0.37660638615489006,
"rewards/cosine_scaled_reward": -0.051541879773139954,
"rewards/format_reward": 0.14583333395421505,
"step": 370
},
{
"completion_length": 1920.6250305175781,
"epoch": 0.424,
"grad_norm": 0.013661705888807774,
"kl": 0.0009822845458984375,
"learning_rate": 2.729523361034538e-07,
"loss": 0.0059,
"reward": 0.5551425702869892,
"reward_std": 0.278125814627856,
"rewards/cosine_scaled_reward": 0.01715458557009697,
"rewards/format_reward": 0.520833333954215,
"step": 371
},
{
"completion_length": 3028.520896911621,
"epoch": 0.42514285714285716,
"grad_norm": 0.0525103434920311,
"kl": 0.00101470947265625,
"learning_rate": 2.7048349887476037e-07,
"loss": 0.0585,
"reward": 0.7306250557303429,
"reward_std": 0.8802761994302273,
"rewards/cosine_scaled_reward": 0.167395843192935,
"rewards/format_reward": 0.3958333358168602,
"step": 372
},
{
"completion_length": 1821.583351135254,
"epoch": 0.42628571428571427,
"grad_norm": 0.028777100145816803,
"kl": 0.0010874271392822266,
"learning_rate": 2.6802828488599294e-07,
"loss": 0.039,
"reward": 0.31937169190496206,
"reward_std": 0.6239332426339388,
"rewards/cosine_scaled_reward": -0.15281416662037373,
"rewards/format_reward": 0.6250000037252903,
"step": 373
},
{
"completion_length": 2500.604202270508,
"epoch": 0.42742857142857144,
"grad_norm": 0.0391635037958622,
"kl": 0.0013954639434814453,
"learning_rate": 2.655868138008171e-07,
"loss": 0.0541,
"reward": 0.37847861149930395,
"reward_std": 0.7538540698587894,
"rewards/cosine_scaled_reward": -0.039927379228174686,
"rewards/format_reward": 0.4583333358168602,
"step": 374
},
{
"completion_length": 2978.7291717529297,
"epoch": 0.42857142857142855,
"grad_norm": 0.03049525059759617,
"kl": 0.0025815963745117188,
"learning_rate": 2.631592046130896e-07,
"loss": 0.0099,
"reward": 0.37722547352313995,
"reward_std": 0.5434530153870583,
"rewards/cosine_scaled_reward": 0.0011127367615699768,
"rewards/format_reward": 0.37500000558793545,
"step": 375
},
{
"completion_length": 2328.083396911621,
"epoch": 0.4297142857142857,
"grad_norm": 0.04338321462273598,
"kl": 0.0022089481353759766,
"learning_rate": 2.6074557564105724e-07,
"loss": 0.0388,
"reward": 0.33715473487973213,
"reward_std": 0.6841284055262804,
"rewards/cosine_scaled_reward": -0.1126726483926177,
"rewards/format_reward": 0.5625000111758709,
"step": 376
},
{
"completion_length": 3434.3541870117188,
"epoch": 0.4308571428571429,
"grad_norm": 0.04104272648692131,
"kl": 0.0009138584136962891,
"learning_rate": 2.583460445215911e-07,
"loss": 0.0306,
"reward": -0.12041943147778511,
"reward_std": 0.6869102958589792,
"rewards/cosine_scaled_reward": -0.17479303665459156,
"rewards/format_reward": 0.22916667349636555,
"step": 377
},
{
"completion_length": 2245.437545776367,
"epoch": 0.432,
"grad_norm": 0.02392558939754963,
"kl": 0.0010918378829956055,
"learning_rate": 2.5596072820445254e-07,
"loss": 0.021,
"reward": 0.6035024983575568,
"reward_std": 0.6711528412997723,
"rewards/cosine_scaled_reward": 0.010084576904773712,
"rewards/format_reward": 0.583333333954215,
"step": 378
},
{
"completion_length": 3069.1041870117188,
"epoch": 0.43314285714285716,
"grad_norm": 0.02428600750863552,
"kl": 0.0008909702301025391,
"learning_rate": 2.5358974294659373e-07,
"loss": 0.0264,
"reward": -0.1455503827892244,
"reward_std": 0.5249228626489639,
"rewards/cosine_scaled_reward": -0.16652518138289452,
"rewards/format_reward": 0.18750000186264515,
"step": 379
},
{
"completion_length": 2903.437515258789,
"epoch": 0.4342857142857143,
"grad_norm": 0.05160444974899292,
"kl": 0.0016338825225830078,
"learning_rate": 2.512332043064913e-07,
"loss": 0.0391,
"reward": 0.152958738617599,
"reward_std": 0.8814124129712582,
"rewards/cosine_scaled_reward": -0.11102063581347466,
"rewards/format_reward": 0.37500000931322575,
"step": 380
},
{
"completion_length": 2945.3125534057617,
"epoch": 0.43542857142857144,
"grad_norm": 0.04407874122262001,
"kl": 0.0018510818481445312,
"learning_rate": 2.488912271385139e-07,
"loss": 0.0489,
"reward": -0.054892294108867645,
"reward_std": 0.6566630490124226,
"rewards/cosine_scaled_reward": -0.18369614984840155,
"rewards/format_reward": 0.3125,
"step": 381
},
{
"completion_length": 2357.4375534057617,
"epoch": 0.43657142857142855,
"grad_norm": 0.03399001434445381,
"kl": 0.0019459724426269531,
"learning_rate": 2.465639255873246e-07,
"loss": 0.0484,
"reward": 0.3038070909678936,
"reward_std": 0.4995873998850584,
"rewards/cosine_scaled_reward": -0.13976313127204776,
"rewards/format_reward": 0.5833333432674408,
"step": 382
},
{
"completion_length": 2986.9583435058594,
"epoch": 0.4377142857142857,
"grad_norm": 0.06184740364551544,
"kl": 0.0022149085998535156,
"learning_rate": 2.4425141308231765e-07,
"loss": 0.0698,
"reward": 0.3715968355536461,
"reward_std": 0.6343178395181894,
"rewards/cosine_scaled_reward": -0.0017015933990478516,
"rewards/format_reward": 0.3750000074505806,
"step": 383
},
{
"completion_length": 2208.4583892822266,
"epoch": 0.43885714285714283,
"grad_norm": 0.04220281541347504,
"kl": 0.002426624298095703,
"learning_rate": 2.4195380233209006e-07,
"loss": -0.0163,
"reward": 1.1835149601101875,
"reward_std": 0.8550020232796669,
"rewards/cosine_scaled_reward": 0.2584241358563304,
"rewards/format_reward": 0.6666666716337204,
"step": 384
},
{
"completion_length": 2592.8542098999023,
"epoch": 0.44,
"grad_norm": 0.02935204468667507,
"kl": 0.0010533332824707031,
"learning_rate": 2.3967120531894857e-07,
"loss": -0.0031,
"reward": 0.2314518727362156,
"reward_std": 0.5928132049739361,
"rewards/cosine_scaled_reward": -0.15510739805176854,
"rewards/format_reward": 0.5416666734963655,
"step": 385
},
{
"completion_length": 2843.5416870117188,
"epoch": 0.44114285714285717,
"grad_norm": 0.03138664364814758,
"kl": 0.0016608238220214844,
"learning_rate": 2.374037332934512e-07,
"loss": 0.0227,
"reward": 0.29218395706266165,
"reward_std": 0.5787023734301329,
"rewards/cosine_scaled_reward": -0.07265802705660462,
"rewards/format_reward": 0.4375000074505806,
"step": 386
},
{
"completion_length": 2934.5625,
"epoch": 0.4422857142857143,
"grad_norm": 0.020530929788947105,
"kl": 0.0008387565612792969,
"learning_rate": 2.3515149676898552e-07,
"loss": 0.013,
"reward": -0.04479955695569515,
"reward_std": 0.36785753816366196,
"rewards/cosine_scaled_reward": -0.147399777546525,
"rewards/format_reward": 0.25000000558793545,
"step": 387
},
{
"completion_length": 2610.062515258789,
"epoch": 0.44342857142857145,
"grad_norm": 0.040762562304735184,
"kl": 0.0011615753173828125,
"learning_rate": 2.3291460551638237e-07,
"loss": 0.0708,
"reward": 0.23008868657052517,
"reward_std": 0.6174043659120798,
"rewards/cosine_scaled_reward": -0.10370565578341484,
"rewards/format_reward": 0.43750000558793545,
"step": 388
},
{
"completion_length": 2543.354202270508,
"epoch": 0.44457142857142856,
"grad_norm": 0.04136168956756592,
"kl": 0.0022974014282226562,
"learning_rate": 2.306931685585657e-07,
"loss": 0.0045,
"reward": 0.37671348359435797,
"reward_std": 0.7336033247411251,
"rewards/cosine_scaled_reward": -0.07205992937088013,
"rewards/format_reward": 0.5208333414047956,
"step": 389
},
{
"completion_length": 3020.4791870117188,
"epoch": 0.44571428571428573,
"grad_norm": 0.05496221408247948,
"kl": 0.0008085966110229492,
"learning_rate": 2.2848729416523859e-07,
"loss": 0.0478,
"reward": 0.24858180806040764,
"reward_std": 0.8255434706807137,
"rewards/cosine_scaled_reward": -0.07362576154991984,
"rewards/format_reward": 0.39583334140479565,
"step": 390
},
{
"completion_length": 2627.8542098999023,
"epoch": 0.44685714285714284,
"grad_norm": 0.03822245076298714,
"kl": 0.0011529922485351562,
"learning_rate": 2.2629708984760706e-07,
"loss": -0.0056,
"reward": 1.1691500842571259,
"reward_std": 0.6973480880260468,
"rewards/cosine_scaled_reward": 0.30332503467798233,
"rewards/format_reward": 0.5625000074505806,
"step": 391
},
{
"completion_length": 2101.8750038146973,
"epoch": 0.448,
"grad_norm": 0.04114387929439545,
"kl": 0.0023064613342285156,
"learning_rate": 2.2412266235313973e-07,
"loss": 0.0151,
"reward": 0.4915292300283909,
"reward_std": 0.7246231567114592,
"rewards/cosine_scaled_reward": -0.07715206500142813,
"rewards/format_reward": 0.6458333358168602,
"step": 392
},
{
"completion_length": 2701.479232788086,
"epoch": 0.4491428571428571,
"grad_norm": 0.05024963617324829,
"kl": 0.0015339851379394531,
"learning_rate": 2.2196411766036487e-07,
"loss": 0.0441,
"reward": 0.4650035873055458,
"reward_std": 0.9533145241439342,
"rewards/cosine_scaled_reward": -0.04874821309931576,
"rewards/format_reward": 0.5625000167638063,
"step": 393
},
{
"completion_length": 3071.2083435058594,
"epoch": 0.4502857142857143,
"grad_norm": 0.03059108555316925,
"kl": 0.002701997756958008,
"learning_rate": 2.1982156097370557e-07,
"loss": 0.0348,
"reward": 0.016890098340809345,
"reward_std": 0.5630042045377195,
"rewards/cosine_scaled_reward": -0.10613830108195543,
"rewards/format_reward": 0.2291666679084301,
"step": 394
},
{
"completion_length": 2223.4375076293945,
"epoch": 0.4514285714285714,
"grad_norm": 0.016346530988812447,
"kl": 0.0012898445129394531,
"learning_rate": 2.1769509671835223e-07,
"loss": -0.0074,
"reward": 0.1286599598824978,
"reward_std": 0.4426659531891346,
"rewards/cosine_scaled_reward": -0.18567002192139626,
"rewards/format_reward": 0.5,
"step": 395
},
{
"completion_length": 3036.125030517578,
"epoch": 0.45257142857142857,
"grad_norm": 0.078156977891922,
"kl": 0.002143383026123047,
"learning_rate": 2.1558482853517253e-07,
"loss": 0.1641,
"reward": 0.2860925877466798,
"reward_std": 1.1900311894714832,
"rewards/cosine_scaled_reward": -0.02362038509454578,
"rewards/format_reward": 0.3333333395421505,
"step": 396
},
{
"completion_length": 2902.562545776367,
"epoch": 0.45371428571428574,
"grad_norm": 0.05982230231165886,
"kl": 0.0010967254638671875,
"learning_rate": 2.134908592756607e-07,
"loss": 0.0246,
"reward": 0.3248216570354998,
"reward_std": 1.0227275677025318,
"rewards/cosine_scaled_reward": -0.07717251125723124,
"rewards/format_reward": 0.4791666716337204,
"step": 397
},
{
"completion_length": 2611.916702270508,
"epoch": 0.45485714285714285,
"grad_norm": 0.05539969727396965,
"kl": 0.0018358230590820312,
"learning_rate": 2.1141329099692406e-07,
"loss": 0.0695,
"reward": 0.2927081920206547,
"reward_std": 0.8153704311698675,
"rewards/cosine_scaled_reward": -0.10364590771496296,
"rewards/format_reward": 0.5000000037252903,
"step": 398
},
{
"completion_length": 2259.0625610351562,
"epoch": 0.456,
"grad_norm": 0.04596502333879471,
"kl": 0.0021970272064208984,
"learning_rate": 2.0935222495670968e-07,
"loss": 0.0574,
"reward": 0.6550063313916326,
"reward_std": 0.7109723538160324,
"rewards/cosine_scaled_reward": 0.004586491733789444,
"rewards/format_reward": 0.6458333414047956,
"step": 399
},
{
"completion_length": 1735.2917175292969,
"epoch": 0.45714285714285713,
"grad_norm": 0.048301637172698975,
"kl": 0.0027017593383789062,
"learning_rate": 2.0730776160846853e-07,
"loss": 0.0258,
"reward": 0.9397249203175306,
"reward_std": 0.7951749358326197,
"rewards/cosine_scaled_reward": 0.09486244898289442,
"rewards/format_reward": 0.7500000093132257,
"step": 400
},
{
"completion_length": 3075.8333435058594,
"epoch": 0.4582857142857143,
"grad_norm": 0.02250814624130726,
"kl": 0.0019578933715820312,
"learning_rate": 2.0528000059645995e-07,
"loss": 0.0031,
"reward": 0.09915575897321105,
"reward_std": 0.5018129777163267,
"rewards/cosine_scaled_reward": -0.1066721323877573,
"rewards/format_reward": 0.3125000074505806,
"step": 401
},
{
"completion_length": 2486.0000076293945,
"epoch": 0.4594285714285714,
"grad_norm": 0.031174227595329285,
"kl": 0.0019636154174804688,
"learning_rate": 2.032690407508949e-07,
"loss": 0.0438,
"reward": 0.2905198944499716,
"reward_std": 0.5358554609119892,
"rewards/cosine_scaled_reward": -0.0839067182969302,
"rewards/format_reward": 0.45833333395421505,
"step": 402
},
{
"completion_length": 1904.0625,
"epoch": 0.4605714285714286,
"grad_norm": 0.03299061581492424,
"kl": 0.002137422561645508,
"learning_rate": 2.0127498008311922e-07,
"loss": 0.0207,
"reward": 0.615192599594593,
"reward_std": 0.5438333973288536,
"rewards/cosine_scaled_reward": -0.03615369647741318,
"rewards/format_reward": 0.6875,
"step": 403
},
{
"completion_length": 2334.2291870117188,
"epoch": 0.4617142857142857,
"grad_norm": 0.046507738530635834,
"kl": 0.0019326210021972656,
"learning_rate": 1.9929791578083655e-07,
"loss": 0.0536,
"reward": 0.22189591638743877,
"reward_std": 0.7510978206992149,
"rewards/cosine_scaled_reward": -0.11821871809661388,
"rewards/format_reward": 0.45833333395421505,
"step": 404
},
{
"completion_length": 2323.229206085205,
"epoch": 0.46285714285714286,
"grad_norm": 0.049295179545879364,
"kl": 0.0019037723541259766,
"learning_rate": 1.9733794420337213e-07,
"loss": 0.0334,
"reward": 0.8066165037453175,
"reward_std": 0.84293257817626,
"rewards/cosine_scaled_reward": 0.12205823619660805,
"rewards/format_reward": 0.5625000074505806,
"step": 405
},
{
"completion_length": 2272.812545776367,
"epoch": 0.464,
"grad_norm": 0.040780920535326004,
"kl": 0.0019860267639160156,
"learning_rate": 1.9539516087697517e-07,
"loss": 0.0374,
"reward": 0.6606193948537111,
"reward_std": 0.8417980894446373,
"rewards/cosine_scaled_reward": 0.007393034175038338,
"rewards/format_reward": 0.645833345130086,
"step": 406
},
{
"completion_length": 2579.8333778381348,
"epoch": 0.46514285714285714,
"grad_norm": 0.03767210990190506,
"kl": 0.0013513565063476562,
"learning_rate": 1.934696604901642e-07,
"loss": 0.0199,
"reward": 0.5471592950634658,
"reward_std": 0.5273555251769722,
"rewards/cosine_scaled_reward": 0.04441297799348831,
"rewards/format_reward": 0.4583333358168602,
"step": 407
},
{
"completion_length": 2723.500030517578,
"epoch": 0.4662857142857143,
"grad_norm": 0.03993203490972519,
"kl": 0.0015883445739746094,
"learning_rate": 1.915615368891117e-07,
"loss": 0.0578,
"reward": 0.43359660543501377,
"reward_std": 0.5000373655930161,
"rewards/cosine_scaled_reward": 0.008464958518743515,
"rewards/format_reward": 0.4166666716337204,
"step": 408
},
{
"completion_length": 3507.3958740234375,
"epoch": 0.4674285714285714,
"grad_norm": 0.026279030367732048,
"kl": 0.0009474754333496094,
"learning_rate": 1.8967088307307e-07,
"loss": 0.022,
"reward": -0.263646949082613,
"reward_std": 0.45825990103185177,
"rewards/cosine_scaled_reward": -0.18390680570155382,
"rewards/format_reward": 0.1041666679084301,
"step": 409
},
{
"completion_length": 2405.93754196167,
"epoch": 0.4685714285714286,
"grad_norm": 0.04245587810873985,
"kl": 0.0011708736419677734,
"learning_rate": 1.8779779118983867e-07,
"loss": 0.035,
"reward": 0.4530321164056659,
"reward_std": 0.8689349945634604,
"rewards/cosine_scaled_reward": -0.033900603419169784,
"rewards/format_reward": 0.5208333395421505,
"step": 410
},
{
"completion_length": 3133.0,
"epoch": 0.4697142857142857,
"grad_norm": 0.045392949134111404,
"kl": 0.001537322998046875,
"learning_rate": 1.8594235253127372e-07,
"loss": 0.0303,
"reward": 0.1537340972572565,
"reward_std": 0.6930182911455631,
"rewards/cosine_scaled_reward": -0.08979963045567274,
"rewards/format_reward": 0.3333333395421505,
"step": 411
},
{
"completion_length": 2966.1666870117188,
"epoch": 0.47085714285714286,
"grad_norm": 0.03469331935048103,
"kl": 0.001659393310546875,
"learning_rate": 1.8410465752883758e-07,
"loss": -0.0277,
"reward": 0.5148756094276905,
"reward_std": 0.6638762913644314,
"rewards/cosine_scaled_reward": 0.049104465171694756,
"rewards/format_reward": 0.4166666679084301,
"step": 412
},
{
"completion_length": 2527.187511444092,
"epoch": 0.472,
"grad_norm": 0.0342419259250164,
"kl": 0.002641916275024414,
"learning_rate": 1.822847957491922e-07,
"loss": 0.0405,
"reward": 0.49921591579914093,
"reward_std": 0.7078933194279671,
"rewards/cosine_scaled_reward": 0.010024622781202197,
"rewards/format_reward": 0.4791666716337204,
"step": 413
},
{
"completion_length": 3006.729202270508,
"epoch": 0.47314285714285714,
"grad_norm": 0.03782186284661293,
"kl": 0.0012745857238769531,
"learning_rate": 1.804828558898332e-07,
"loss": 0.0095,
"reward": -0.013905536383390427,
"reward_std": 0.6790427416563034,
"rewards/cosine_scaled_reward": -0.1423694370314479,
"rewards/format_reward": 0.2708333358168602,
"step": 414
},
{
"completion_length": 3249.333335876465,
"epoch": 0.4742857142857143,
"grad_norm": 0.017366327345371246,
"kl": 0.0010700225830078125,
"learning_rate": 1.7869892577476722e-07,
"loss": 0.0009,
"reward": -0.3696242421865463,
"reward_std": 0.3092060061171651,
"rewards/cosine_scaled_reward": -0.25772880017757416,
"rewards/format_reward": 0.14583333395421505,
"step": 415
},
{
"completion_length": 1827.9375267028809,
"epoch": 0.4754285714285714,
"grad_norm": 0.030641866847872734,
"kl": 0.0006153583526611328,
"learning_rate": 1.7693309235023127e-07,
"loss": 0.0127,
"reward": 0.7272525690495968,
"reward_std": 0.704660689458251,
"rewards/cosine_scaled_reward": -0.021790377446450293,
"rewards/format_reward": 0.7708333395421505,
"step": 416
},
{
"completion_length": 3335.5833435058594,
"epoch": 0.4765714285714286,
"grad_norm": 0.033628471195697784,
"kl": 0.0013918876647949219,
"learning_rate": 1.7518544168045524e-07,
"loss": 0.0198,
"reward": -0.24187161587178707,
"reward_std": 0.571142709814012,
"rewards/cosine_scaled_reward": -0.23551913630217314,
"rewards/format_reward": 0.22916667349636555,
"step": 417
},
{
"completion_length": 2151.937526702881,
"epoch": 0.4777142857142857,
"grad_norm": 0.02097875066101551,
"kl": 0.0009946823120117188,
"learning_rate": 1.7345605894346726e-07,
"loss": 0.0207,
"reward": 0.5706320106983185,
"reward_std": 0.5004640761762857,
"rewards/cosine_scaled_reward": -0.027184022590517998,
"rewards/format_reward": 0.6250000055879354,
"step": 418
},
{
"completion_length": 2525.7708587646484,
"epoch": 0.47885714285714287,
"grad_norm": 0.036434732377529144,
"kl": 0.0015592575073242188,
"learning_rate": 1.7174502842694212e-07,
"loss": 0.0261,
"reward": 0.16563719511032104,
"reward_std": 0.6850580945611,
"rewards/cosine_scaled_reward": -0.13593140104785562,
"rewards/format_reward": 0.4375,
"step": 419
},
{
"completion_length": 1836.8542098999023,
"epoch": 0.48,
"grad_norm": 0.0419175811111927,
"kl": 0.0034418106079101562,
"learning_rate": 1.7005243352409333e-07,
"loss": 0.0254,
"reward": 0.5675348229706287,
"reward_std": 0.7641635444015265,
"rewards/cosine_scaled_reward": -0.0808159353910014,
"rewards/format_reward": 0.7291666883975267,
"step": 420
},
{
"completion_length": 3365.437530517578,
"epoch": 0.48114285714285715,
"grad_norm": 0.04015079140663147,
"kl": 0.004153251647949219,
"learning_rate": 1.6837835672960831e-07,
"loss": 0.0357,
"reward": -0.140715591609478,
"reward_std": 0.6398463062942028,
"rewards/cosine_scaled_reward": -0.13285779720172286,
"rewards/format_reward": 0.12500000186264515,
"step": 421
},
{
"completion_length": 2927.6666870117188,
"epoch": 0.48228571428571426,
"grad_norm": 0.02774575911462307,
"kl": 0.0009286403656005859,
"learning_rate": 1.6672287963562852e-07,
"loss": 0.0137,
"reward": 0.16687930934131145,
"reward_std": 0.49705400690436363,
"rewards/cosine_scaled_reward": -0.08322702161967754,
"rewards/format_reward": 0.3333333358168602,
"step": 422
},
{
"completion_length": 2936.25,
"epoch": 0.48342857142857143,
"grad_norm": 0.031355079263448715,
"kl": 0.0010771751403808594,
"learning_rate": 1.6508608292777203e-07,
"loss": 0.005,
"reward": 0.07443832606077194,
"reward_std": 0.5912903603166342,
"rewards/cosine_scaled_reward": -0.13986417837440968,
"rewards/format_reward": 0.3541666679084301,
"step": 423
},
{
"completion_length": 3266.125030517578,
"epoch": 0.4845714285714286,
"grad_norm": 0.04135151579976082,
"kl": 0.0009894371032714844,
"learning_rate": 1.6346804638120098e-07,
"loss": 0.0246,
"reward": -0.2541657374240458,
"reward_std": 0.6840569227933884,
"rewards/cosine_scaled_reward": -0.220832874532789,
"rewards/format_reward": 0.1875000074505806,
"step": 424
},
{
"completion_length": 2052.7500610351562,
"epoch": 0.4857142857142857,
"grad_norm": 0.051734067499637604,
"kl": 0.0017888545989990234,
"learning_rate": 1.6186884885673413e-07,
"loss": 0.0244,
"reward": 1.310227697074879,
"reward_std": 1.0274488306604326,
"rewards/cosine_scaled_reward": 0.2696971520781517,
"rewards/format_reward": 0.770833345130086,
"step": 425
},
{
"completion_length": 2170.4583740234375,
"epoch": 0.4868571428571429,
"grad_norm": 0.03390904888510704,
"kl": 0.0038836002349853516,
"learning_rate": 1.6028856829700258e-07,
"loss": -0.0068,
"reward": 0.5739936158061028,
"reward_std": 0.8549008993431926,
"rewards/cosine_scaled_reward": 0.005746812559664249,
"rewards/format_reward": 0.5625,
"step": 426
},
{
"completion_length": 3058.1458740234375,
"epoch": 0.488,
"grad_norm": 0.04581299424171448,
"kl": 0.0008256435394287109,
"learning_rate": 1.5872728172265146e-07,
"loss": 0.0541,
"reward": 0.21424740552902222,
"reward_std": 0.5585381351411343,
"rewards/cosine_scaled_reward": -0.049126310274004936,
"rewards/format_reward": 0.3125000074505806,
"step": 427
},
{
"completion_length": 2518.250030517578,
"epoch": 0.48914285714285716,
"grad_norm": 0.03789118677377701,
"kl": 0.0015597343444824219,
"learning_rate": 1.5718506522858572e-07,
"loss": 0.0503,
"reward": 0.22319819265976548,
"reward_std": 0.7634806632995605,
"rewards/cosine_scaled_reward": -0.14881757437251508,
"rewards/format_reward": 0.5208333395421505,
"step": 428
},
{
"completion_length": 2405.25008392334,
"epoch": 0.49028571428571427,
"grad_norm": 0.03913323953747749,
"kl": 0.002506732940673828,
"learning_rate": 1.5566199398026147e-07,
"loss": 0.0521,
"reward": 0.43615793297067285,
"reward_std": 0.7961943745613098,
"rewards/cosine_scaled_reward": -0.06317103141918778,
"rewards/format_reward": 0.5625000186264515,
"step": 429
},
{
"completion_length": 2516.6041984558105,
"epoch": 0.49142857142857144,
"grad_norm": 0.03834352642297745,
"kl": 0.001491546630859375,
"learning_rate": 1.5415814221002265e-07,
"loss": 0.0167,
"reward": 0.698354821652174,
"reward_std": 0.5752123966813087,
"rewards/cosine_scaled_reward": 0.057510748505592346,
"rewards/format_reward": 0.5833333432674408,
"step": 430
},
{
"completion_length": 2556.229179382324,
"epoch": 0.49257142857142855,
"grad_norm": 0.021933771669864655,
"kl": 0.0019750595092773438,
"learning_rate": 1.5267358321348285e-07,
"loss": 0.027,
"reward": 0.1585888583213091,
"reward_std": 0.39748182776384056,
"rewards/cosine_scaled_reward": -0.16028890456072986,
"rewards/format_reward": 0.47916667722165585,
"step": 431
},
{
"completion_length": 2948.791717529297,
"epoch": 0.4937142857142857,
"grad_norm": 0.03498787805438042,
"kl": 0.0016126632690429688,
"learning_rate": 1.5120838934595337e-07,
"loss": 0.0353,
"reward": 0.1526618276257068,
"reward_std": 0.5853009335696697,
"rewards/cosine_scaled_reward": -0.1007524230517447,
"rewards/format_reward": 0.35416668094694614,
"step": 432
},
{
"completion_length": 2779.395866394043,
"epoch": 0.4948571428571429,
"grad_norm": 0.046449173241853714,
"kl": 0.001285552978515625,
"learning_rate": 1.4976263201891613e-07,
"loss": 0.0361,
"reward": 0.5470792800188065,
"reward_std": 0.7514309994876385,
"rewards/cosine_scaled_reward": 0.04437295813113451,
"rewards/format_reward": 0.4583333395421505,
"step": 433
},
{
"completion_length": 2833.0208740234375,
"epoch": 0.496,
"grad_norm": 0.03701100870966911,
"kl": 0.002968311309814453,
"learning_rate": 1.483363816965435e-07,
"loss": 0.0405,
"reward": -0.023492522537708282,
"reward_std": 0.5920409262180328,
"rewards/cosine_scaled_reward": -0.17841292917728424,
"rewards/format_reward": 0.33333334140479565,
"step": 434
},
{
"completion_length": 2294.895866394043,
"epoch": 0.49714285714285716,
"grad_norm": 0.03758431226015091,
"kl": 0.0011827945709228516,
"learning_rate": 1.469297078922642e-07,
"loss": 0.041,
"reward": 0.41764480620622635,
"reward_std": 0.6425671465694904,
"rewards/cosine_scaled_reward": -0.051594268530607224,
"rewards/format_reward": 0.5208333395421505,
"step": 435
},
{
"completion_length": 2106.375015258789,
"epoch": 0.4982857142857143,
"grad_norm": 0.021261492744088173,
"kl": 0.002359628677368164,
"learning_rate": 1.4554267916537495e-07,
"loss": -0.0056,
"reward": 0.863133005797863,
"reward_std": 0.5500740725547075,
"rewards/cosine_scaled_reward": 0.15031648427248,
"rewards/format_reward": 0.5625000018626451,
"step": 436
},
{
"completion_length": 2651.645866394043,
"epoch": 0.49942857142857144,
"grad_norm": 0.040734272450208664,
"kl": 0.0017135143280029297,
"learning_rate": 1.4417536311769885e-07,
"loss": 0.09,
"reward": 0.01604996738024056,
"reward_std": 0.665538102388382,
"rewards/cosine_scaled_reward": -0.22114169038832188,
"rewards/format_reward": 0.4583333395421505,
"step": 437
},
{
"completion_length": 2993.9791870117188,
"epoch": 0.5005714285714286,
"grad_norm": 0.05256485193967819,
"kl": 0.0008656978607177734,
"learning_rate": 1.4282782639029128e-07,
"loss": 0.0184,
"reward": 0.19979790039360523,
"reward_std": 0.8984194695949554,
"rewards/cosine_scaled_reward": -0.07718438468873501,
"rewards/format_reward": 0.35416667349636555,
"step": 438
},
{
"completion_length": 2457.937515258789,
"epoch": 0.5017142857142857,
"grad_norm": 0.026962121948599815,
"kl": 0.002255082130432129,
"learning_rate": 1.4150013466019114e-07,
"loss": 0.0319,
"reward": -0.06446752324700356,
"reward_std": 0.5604766383767128,
"rewards/cosine_scaled_reward": -0.2509837690740824,
"rewards/format_reward": 0.43750000558793545,
"step": 439
},
{
"completion_length": 2894.0625076293945,
"epoch": 0.5028571428571429,
"grad_norm": 0.02316948212683201,
"kl": 0.001163482666015625,
"learning_rate": 1.4019235263722034e-07,
"loss": 0.021,
"reward": -0.2982286214828491,
"reward_std": 0.4396653138101101,
"rewards/cosine_scaled_reward": -0.26369765028357506,
"rewards/format_reward": 0.2291666716337204,
"step": 440
},
{
"completion_length": 3043.270866394043,
"epoch": 0.504,
"grad_norm": 0.044711362570524216,
"kl": 0.0011162757873535156,
"learning_rate": 1.3890454406082956e-07,
"loss": 0.0672,
"reward": 0.10117581891245209,
"reward_std": 0.7147481255233288,
"rewards/cosine_scaled_reward": -0.07441211328841746,
"rewards/format_reward": 0.2500000037252903,
"step": 441
},
{
"completion_length": 2906.2083587646484,
"epoch": 0.5051428571428571,
"grad_norm": 0.03437203913927078,
"kl": 0.001821756362915039,
"learning_rate": 1.3763677169699217e-07,
"loss": 0.0265,
"reward": -0.09266237542033195,
"reward_std": 0.5656116344034672,
"rewards/cosine_scaled_reward": -0.19216451607644558,
"rewards/format_reward": 0.2916666716337204,
"step": 442
},
{
"completion_length": 3142.979179382324,
"epoch": 0.5062857142857143,
"grad_norm": 0.03172105550765991,
"kl": 0.0008325576782226562,
"learning_rate": 1.3638909733514452e-07,
"loss": 0.0029,
"reward": 0.1617426760494709,
"reward_std": 0.566162696108222,
"rewards/cosine_scaled_reward": -0.04412867687642574,
"rewards/format_reward": 0.2500000037252903,
"step": 443
},
{
"completion_length": 2948.208335876465,
"epoch": 0.5074285714285715,
"grad_norm": 0.031045345589518547,
"kl": 0.0011382102966308594,
"learning_rate": 1.351615817851748e-07,
"loss": 0.037,
"reward": 0.016486244276165962,
"reward_std": 0.395066162571311,
"rewards/cosine_scaled_reward": -0.11675689741969109,
"rewards/format_reward": 0.25,
"step": 444
},
{
"completion_length": 2974.375045776367,
"epoch": 0.5085714285714286,
"grad_norm": 0.0441366508603096,
"kl": 0.0014166831970214844,
"learning_rate": 1.3395428487445914e-07,
"loss": 0.0557,
"reward": 0.44049030914902687,
"reward_std": 0.8217526823282242,
"rewards/cosine_scaled_reward": 0.02232850785367191,
"rewards/format_reward": 0.39583334140479565,
"step": 445
},
{
"completion_length": 2984.479217529297,
"epoch": 0.5097142857142857,
"grad_norm": 0.04864772781729698,
"kl": 0.0013647079467773438,
"learning_rate": 1.3276726544494571e-07,
"loss": 0.0512,
"reward": 0.024046601727604866,
"reward_std": 0.828495018184185,
"rewards/cosine_scaled_reward": -0.14422670053318143,
"rewards/format_reward": 0.31250000186264515,
"step": 446
},
{
"completion_length": 2257.5625495910645,
"epoch": 0.5108571428571429,
"grad_norm": 0.0407555028796196,
"kl": 0.0022869110107421875,
"learning_rate": 1.316005813502869e-07,
"loss": 0.0436,
"reward": 0.5557453141082078,
"reward_std": 0.7650147341191769,
"rewards/cosine_scaled_reward": -0.045044018886983395,
"rewards/format_reward": 0.6458333414047956,
"step": 447
},
{
"completion_length": 2130.416690826416,
"epoch": 0.512,
"grad_norm": 0.018810981884598732,
"kl": 0.0023889541625976562,
"learning_rate": 1.3045428945301953e-07,
"loss": 0.0052,
"reward": 0.33343374729156494,
"reward_std": 0.37268934305757284,
"rewards/cosine_scaled_reward": -0.10411648452281952,
"rewards/format_reward": 0.5416666679084301,
"step": 448
},
{
"completion_length": 2702.791679382324,
"epoch": 0.5131428571428571,
"grad_norm": 0.027145305648446083,
"kl": 0.0036830902099609375,
"learning_rate": 1.2932844562179352e-07,
"loss": 0.0092,
"reward": -0.003954991698265076,
"reward_std": 0.48061698488891125,
"rewards/cosine_scaled_reward": -0.16864416131284088,
"rewards/format_reward": 0.3333333358168602,
"step": 449
},
{
"completion_length": 2385.8333587646484,
"epoch": 0.5142857142857142,
"grad_norm": 0.13469496369361877,
"kl": 0.0027618408203125,
"learning_rate": 1.2822310472864885e-07,
"loss": 0.0508,
"reward": 0.16872227331623435,
"reward_std": 0.7135936170816422,
"rewards/cosine_scaled_reward": -0.1343888696283102,
"rewards/format_reward": 0.43750000558793545,
"step": 450
},
{
"completion_length": 2780.7916946411133,
"epoch": 0.5154285714285715,
"grad_norm": 0.03198925033211708,
"kl": 0.0016019344329833984,
"learning_rate": 1.2713832064634125e-07,
"loss": 0.001,
"reward": -0.05699261091649532,
"reward_std": 0.5354622071608901,
"rewards/cosine_scaled_reward": -0.17432964034378529,
"rewards/format_reward": 0.2916666679084301,
"step": 451
},
{
"completion_length": 3177.8958740234375,
"epoch": 0.5165714285714286,
"grad_norm": 0.03804057464003563,
"kl": 0.0010409355163574219,
"learning_rate": 1.260741462457165e-07,
"loss": 0.0253,
"reward": 0.47885728627443314,
"reward_std": 0.6608881391584873,
"rewards/cosine_scaled_reward": 0.06234530918300152,
"rewards/format_reward": 0.3541666679084301,
"step": 452
},
{
"completion_length": 2606.3125381469727,
"epoch": 0.5177142857142857,
"grad_norm": 0.036739181727170944,
"kl": 0.0011627674102783203,
"learning_rate": 1.2503063339313356e-07,
"loss": 0.0038,
"reward": 0.17389470152556896,
"reward_std": 0.671304065734148,
"rewards/cosine_scaled_reward": -0.1422193255275488,
"rewards/format_reward": 0.45833333395421505,
"step": 453
},
{
"completion_length": 2695.1459045410156,
"epoch": 0.5188571428571429,
"grad_norm": 0.034578122198581696,
"kl": 0.0020804405212402344,
"learning_rate": 1.2400783294793668e-07,
"loss": 0.0247,
"reward": 0.44620663672685623,
"reward_std": 0.7098315693438053,
"rewards/cosine_scaled_reward": -0.06856334558688104,
"rewards/format_reward": 0.5833333395421505,
"step": 454
},
{
"completion_length": 2958.7708740234375,
"epoch": 0.52,
"grad_norm": 0.030018191784620285,
"kl": 0.0013599395751953125,
"learning_rate": 1.2300579475997657e-07,
"loss": 0.0325,
"reward": -0.15596182458102703,
"reward_std": 0.5225675888359547,
"rewards/cosine_scaled_reward": -0.25506424461491406,
"rewards/format_reward": 0.3541666753590107,
"step": 455
},
{
"completion_length": 3160.291679382324,
"epoch": 0.5211428571428571,
"grad_norm": 0.04109412059187889,
"kl": 0.0009658336639404297,
"learning_rate": 1.220245676671809e-07,
"loss": 0.0054,
"reward": 0.016125384718179703,
"reward_std": 0.6339304633438587,
"rewards/cosine_scaled_reward": -0.10652064997702837,
"rewards/format_reward": 0.22916666977107525,
"step": 456
},
{
"completion_length": 3051.416679382324,
"epoch": 0.5222857142857142,
"grad_norm": 0.046411171555519104,
"kl": 0.0009918212890625,
"learning_rate": 1.2106419949317388e-07,
"loss": 0.0284,
"reward": -0.02184194140136242,
"reward_std": 0.687283992767334,
"rewards/cosine_scaled_reward": -0.1255043102428317,
"rewards/format_reward": 0.2291666679084301,
"step": 457
},
{
"completion_length": 2320.4167098999023,
"epoch": 0.5234285714285715,
"grad_norm": 0.03369192034006119,
"kl": 0.0011844635009765625,
"learning_rate": 1.2012473704494537e-07,
"loss": -0.0035,
"reward": 0.22107653319835663,
"reward_std": 0.5749458260834217,
"rewards/cosine_scaled_reward": -0.13946174271404743,
"rewards/format_reward": 0.5000000055879354,
"step": 458
},
{
"completion_length": 1983.1667022705078,
"epoch": 0.5245714285714286,
"grad_norm": 0.04080894589424133,
"kl": 0.0028548240661621094,
"learning_rate": 1.1920622611056974e-07,
"loss": 0.0328,
"reward": 0.7140507474541664,
"reward_std": 0.9648182429373264,
"rewards/cosine_scaled_reward": 0.023692045360803604,
"rewards/format_reward": 0.6666666697710752,
"step": 459
},
{
"completion_length": 3069.437545776367,
"epoch": 0.5257142857142857,
"grad_norm": 0.05579410493373871,
"kl": 0.0021648406982421875,
"learning_rate": 1.1830871145697412e-07,
"loss": 0.0697,
"reward": 0.0923931347206235,
"reward_std": 0.8388851378113031,
"rewards/cosine_scaled_reward": -0.14130343310534954,
"rewards/format_reward": 0.37500000558793545,
"step": 460
},
{
"completion_length": 2985.0833435058594,
"epoch": 0.5268571428571428,
"grad_norm": 0.031334202736616135,
"kl": 0.0011229515075683594,
"learning_rate": 1.1743223682775649e-07,
"loss": 0.0152,
"reward": -0.03977155685424805,
"reward_std": 0.5789243541657925,
"rewards/cosine_scaled_reward": -0.15530244540423155,
"rewards/format_reward": 0.2708333395421505,
"step": 461
},
{
"completion_length": 2826.895835876465,
"epoch": 0.528,
"grad_norm": 0.01678531803190708,
"kl": 0.0008094310760498047,
"learning_rate": 1.1657684494105386e-07,
"loss": -0.0017,
"reward": -0.2524063400924206,
"reward_std": 0.340161694213748,
"rewards/cosine_scaled_reward": -0.2616198379546404,
"rewards/format_reward": 0.27083333395421505,
"step": 462
},
{
"completion_length": 2652.9791870117188,
"epoch": 0.5291428571428571,
"grad_norm": 0.030127134174108505,
"kl": 0.002656221389770508,
"learning_rate": 1.1574257748745986e-07,
"loss": 0.0208,
"reward": 0.33723626285791397,
"reward_std": 0.6605229675769806,
"rewards/cosine_scaled_reward": -0.050131882540881634,
"rewards/format_reward": 0.4375000074505806,
"step": 463
},
{
"completion_length": 1788.250015258789,
"epoch": 0.5302857142857142,
"grad_norm": 0.014693168923258781,
"kl": 0.0027604103088378906,
"learning_rate": 1.1492947512799328e-07,
"loss": -0.0001,
"reward": 0.8204499296844006,
"reward_std": 0.36152973771095276,
"rewards/cosine_scaled_reward": 0.09772495925426483,
"rewards/format_reward": 0.625,
"step": 464
},
{
"completion_length": 3023.291679382324,
"epoch": 0.5314285714285715,
"grad_norm": 0.02821505255997181,
"kl": 0.001667022705078125,
"learning_rate": 1.1413757749211602e-07,
"loss": 0.0299,
"reward": -0.02931244857609272,
"reward_std": 0.4763452261686325,
"rewards/cosine_scaled_reward": -0.1604895582422614,
"rewards/format_reward": 0.29166666977107525,
"step": 465
},
{
"completion_length": 2858.8958435058594,
"epoch": 0.5325714285714286,
"grad_norm": 0.016902461647987366,
"kl": 0.0013074874877929688,
"learning_rate": 1.1336692317580158e-07,
"loss": 0.0013,
"reward": 0.20501752942800522,
"reward_std": 0.389689764007926,
"rewards/cosine_scaled_reward": -0.012074556201696396,
"rewards/format_reward": 0.2291666716337204,
"step": 466
},
{
"completion_length": 3056.7291870117188,
"epoch": 0.5337142857142857,
"grad_norm": 0.024957498535513878,
"kl": 0.0012183189392089844,
"learning_rate": 1.1261754973965422e-07,
"loss": 0.0121,
"reward": -0.0974850058555603,
"reward_std": 0.47587360069155693,
"rewards/cosine_scaled_reward": -0.17374251037836075,
"rewards/format_reward": 0.2500000037252903,
"step": 467
},
{
"completion_length": 2902.0208740234375,
"epoch": 0.5348571428571428,
"grad_norm": 0.020729240030050278,
"kl": 0.0029973983764648438,
"learning_rate": 1.1188949370707787e-07,
"loss": 0.0248,
"reward": -0.3498317201156169,
"reward_std": 0.31344279972836375,
"rewards/cosine_scaled_reward": -0.3207492008805275,
"rewards/format_reward": 0.2916666716337204,
"step": 468
},
{
"completion_length": 2738.8958854675293,
"epoch": 0.536,
"grad_norm": 0.04061265289783478,
"kl": 0.0017533302307128906,
"learning_rate": 1.1118279056249653e-07,
"loss": 0.0035,
"reward": 0.16780934482812881,
"reward_std": 0.6997935622930527,
"rewards/cosine_scaled_reward": -0.10359533131122589,
"rewards/format_reward": 0.3750000074505806,
"step": 469
},
{
"completion_length": 3170.3334045410156,
"epoch": 0.5371428571428571,
"grad_norm": 0.05352408438920975,
"kl": 0.0013861656188964844,
"learning_rate": 1.1049747474962444e-07,
"loss": 0.0772,
"reward": -0.12879883614368737,
"reward_std": 0.7819614112377167,
"rewards/cosine_scaled_reward": -0.21023273840546608,
"rewards/format_reward": 0.2916666679084301,
"step": 470
},
{
"completion_length": 3140.2291870117188,
"epoch": 0.5382857142857143,
"grad_norm": 0.03319269046187401,
"kl": 0.0009810924530029297,
"learning_rate": 1.0983357966978745e-07,
"loss": -0.0023,
"reward": 0.078701913356781,
"reward_std": 0.5363645683974028,
"rewards/cosine_scaled_reward": -0.10648237727582455,
"rewards/format_reward": 0.2916666679084301,
"step": 471
},
{
"completion_length": 2882.3750610351562,
"epoch": 0.5394285714285715,
"grad_norm": 0.04058046266436577,
"kl": 0.0022449493408203125,
"learning_rate": 1.0919113768029517e-07,
"loss": -0.0056,
"reward": 0.23404515581205487,
"reward_std": 0.6296902745962143,
"rewards/cosine_scaled_reward": -0.11214408185333014,
"rewards/format_reward": 0.4583333358168602,
"step": 472
},
{
"completion_length": 3213.0625,
"epoch": 0.5405714285714286,
"grad_norm": 0.03943871706724167,
"kl": 0.0012938976287841797,
"learning_rate": 1.0857018009286381e-07,
"loss": 0.0228,
"reward": 0.07427317555993795,
"reward_std": 0.7803191654384136,
"rewards/cosine_scaled_reward": -0.06703009083867073,
"rewards/format_reward": 0.2083333358168602,
"step": 473
},
{
"completion_length": 2547.12504196167,
"epoch": 0.5417142857142857,
"grad_norm": 0.04717176407575607,
"kl": 0.0013964176177978516,
"learning_rate": 1.0797073717209013e-07,
"loss": 0.0238,
"reward": 0.7898019873537123,
"reward_std": 0.8049750197678804,
"rewards/cosine_scaled_reward": 0.15531766414642334,
"rewards/format_reward": 0.47916666977107525,
"step": 474
},
{
"completion_length": 2314.645881652832,
"epoch": 0.5428571428571428,
"grad_norm": 0.041776709258556366,
"kl": 0.0023360252380371094,
"learning_rate": 1.0739283813397639e-07,
"loss": 0.0467,
"reward": 0.635431744158268,
"reward_std": 0.7481589121744037,
"rewards/cosine_scaled_reward": 0.026049194857478142,
"rewards/format_reward": 0.5833333395421505,
"step": 475
},
{
"completion_length": 2867.8750610351562,
"epoch": 0.544,
"grad_norm": 0.053181666880846024,
"kl": 0.002093791961669922,
"learning_rate": 1.068365111445064e-07,
"loss": 0.0942,
"reward": 0.23149395920336246,
"reward_std": 0.7845707032829523,
"rewards/cosine_scaled_reward": -0.050919692032039165,
"rewards/format_reward": 0.33333334513008595,
"step": 476
},
{
"completion_length": 2275.062545776367,
"epoch": 0.5451428571428572,
"grad_norm": 0.0416170097887516,
"kl": 0.001888275146484375,
"learning_rate": 1.063017833182728e-07,
"loss": 0.031,
"reward": 1.290647640824318,
"reward_std": 0.759942751377821,
"rewards/cosine_scaled_reward": 0.22865712642669678,
"rewards/format_reward": 0.8333333488553762,
"step": 477
},
{
"completion_length": 3111.270896911621,
"epoch": 0.5462857142857143,
"grad_norm": 0.0378287248313427,
"kl": 0.0014688968658447266,
"learning_rate": 1.0578868071715544e-07,
"loss": 0.0332,
"reward": 0.21425859350711107,
"reward_std": 0.8052146192640066,
"rewards/cosine_scaled_reward": -0.06995403207838535,
"rewards/format_reward": 0.35416666977107525,
"step": 478
},
{
"completion_length": 2907.12508392334,
"epoch": 0.5474285714285714,
"grad_norm": 0.04446679353713989,
"kl": 0.0024013519287109375,
"learning_rate": 1.0529722834905125e-07,
"loss": 0.0509,
"reward": 0.1522473245859146,
"reward_std": 0.8119659163057804,
"rewards/cosine_scaled_reward": -0.13220967527013272,
"rewards/format_reward": 0.41666667722165585,
"step": 479
},
{
"completion_length": 2317.8125381469727,
"epoch": 0.5485714285714286,
"grad_norm": 0.040562789887189865,
"kl": 0.0023407936096191406,
"learning_rate": 1.0482745016665526e-07,
"loss": 0.0618,
"reward": 0.23905727022793144,
"reward_std": 0.696567952632904,
"rewards/cosine_scaled_reward": -0.14088802970945835,
"rewards/format_reward": 0.5208333395421505,
"step": 480
},
{
"completion_length": 3088.7291870117188,
"epoch": 0.5497142857142857,
"grad_norm": 0.04644802585244179,
"kl": 0.001194000244140625,
"learning_rate": 1.0437936906629334e-07,
"loss": 0.0554,
"reward": -0.07575955055654049,
"reward_std": 0.6837183889001608,
"rewards/cosine_scaled_reward": -0.22537978272885084,
"rewards/format_reward": 0.3750000111758709,
"step": 481
},
{
"completion_length": 2879.500030517578,
"epoch": 0.5508571428571428,
"grad_norm": 0.033522870391607285,
"kl": 0.0019969940185546875,
"learning_rate": 1.0395300688680625e-07,
"loss": 0.0197,
"reward": 0.3703960571438074,
"reward_std": 0.726120725274086,
"rewards/cosine_scaled_reward": 0.02894800715148449,
"rewards/format_reward": 0.31250000186264515,
"step": 482
},
{
"completion_length": 2807.6458587646484,
"epoch": 0.552,
"grad_norm": 0.029158450663089752,
"kl": 0.0014438629150390625,
"learning_rate": 1.0354838440848501e-07,
"loss": 0.0186,
"reward": 0.02249845489859581,
"reward_std": 0.5169437192380428,
"rewards/cosine_scaled_reward": -0.2179174479097128,
"rewards/format_reward": 0.4583333358168602,
"step": 483
},
{
"completion_length": 2524.9583778381348,
"epoch": 0.5531428571428572,
"grad_norm": 0.042325735092163086,
"kl": 0.0013332366943359375,
"learning_rate": 1.0316552135205837e-07,
"loss": 0.0593,
"reward": 0.270411791279912,
"reward_std": 0.7619899287819862,
"rewards/cosine_scaled_reward": -0.08354412391781807,
"rewards/format_reward": 0.4375000037252903,
"step": 484
},
{
"completion_length": 1929.500015258789,
"epoch": 0.5542857142857143,
"grad_norm": 0.03541100025177002,
"kl": 0.0027222633361816406,
"learning_rate": 1.0280443637773163e-07,
"loss": 0.0206,
"reward": 0.4553542360663414,
"reward_std": 0.7294214889407158,
"rewards/cosine_scaled_reward": -0.12648956006160006,
"rewards/format_reward": 0.7083333358168602,
"step": 485
},
{
"completion_length": 2066.2292137145996,
"epoch": 0.5554285714285714,
"grad_norm": 0.025865087285637856,
"kl": 0.0031957626342773438,
"learning_rate": 1.0246514708427701e-07,
"loss": 0.0397,
"reward": 0.31694683339446783,
"reward_std": 0.37571004405617714,
"rewards/cosine_scaled_reward": -0.11235993192531168,
"rewards/format_reward": 0.5416666697710752,
"step": 486
},
{
"completion_length": 1944.8750400543213,
"epoch": 0.5565714285714286,
"grad_norm": 0.028248826041817665,
"kl": 0.0013153553009033203,
"learning_rate": 1.0214767000817596e-07,
"loss": 0.0261,
"reward": 1.1763640493154526,
"reward_std": 0.6212100638076663,
"rewards/cosine_scaled_reward": 0.24443202713155188,
"rewards/format_reward": 0.6875,
"step": 487
},
{
"completion_length": 2371.125030517578,
"epoch": 0.5577142857142857,
"grad_norm": 0.01648075133562088,
"kl": 0.0020608901977539062,
"learning_rate": 1.0185202062281336e-07,
"loss": 0.0115,
"reward": 0.06456807442009449,
"reward_std": 0.45360255613923073,
"rewards/cosine_scaled_reward": -0.19688262417912483,
"rewards/format_reward": 0.4583333358168602,
"step": 488
},
{
"completion_length": 3052.4166717529297,
"epoch": 0.5588571428571428,
"grad_norm": 0.027904629707336426,
"kl": 0.0017666816711425781,
"learning_rate": 1.0157821333772304e-07,
"loss": 0.012,
"reward": -0.22679759562015533,
"reward_std": 0.4569165948778391,
"rewards/cosine_scaled_reward": -0.22798212803900242,
"rewards/format_reward": 0.22916666977107525,
"step": 489
},
{
"completion_length": 2675.854217529297,
"epoch": 0.56,
"grad_norm": 0.03642764315009117,
"kl": 0.0011510848999023438,
"learning_rate": 1.013262614978859e-07,
"loss": 0.0183,
"reward": 0.6243100725114346,
"reward_std": 0.6654925271868706,
"rewards/cosine_scaled_reward": 0.03090503066778183,
"rewards/format_reward": 0.5625000074505806,
"step": 490
},
{
"completion_length": 2528.312515258789,
"epoch": 0.5611428571428572,
"grad_norm": 0.053110428154468536,
"kl": 0.006933927536010742,
"learning_rate": 1.0109617738307911e-07,
"loss": 0.0718,
"reward": 0.5572945028543472,
"reward_std": 0.755212415009737,
"rewards/cosine_scaled_reward": 0.00781390443444252,
"rewards/format_reward": 0.5416666716337204,
"step": 491
},
{
"completion_length": 2572.291679382324,
"epoch": 0.5622857142857143,
"grad_norm": 0.028202729299664497,
"kl": 0.0012972354888916016,
"learning_rate": 1.0088797220727779e-07,
"loss": -0.0014,
"reward": 0.14036076702177525,
"reward_std": 0.5402057655155659,
"rewards/cosine_scaled_reward": -0.12773629277944565,
"rewards/format_reward": 0.39583333395421505,
"step": 492
},
{
"completion_length": 2267.0833740234375,
"epoch": 0.5634285714285714,
"grad_norm": 0.04113708436489105,
"kl": 0.001312255859375,
"learning_rate": 1.0070165611810855e-07,
"loss": 0.0483,
"reward": 0.8241308415308595,
"reward_std": 0.7886662483215332,
"rewards/cosine_scaled_reward": 0.08914875192567706,
"rewards/format_reward": 0.645833345130086,
"step": 493
},
{
"completion_length": 2292.354232788086,
"epoch": 0.5645714285714286,
"grad_norm": 0.05553987994790077,
"kl": 0.0015769004821777344,
"learning_rate": 1.005372381963547e-07,
"loss": 0.0064,
"reward": 0.927642391063273,
"reward_std": 1.0857042334973812,
"rewards/cosine_scaled_reward": 0.10965454811230302,
"rewards/format_reward": 0.7083333376795053,
"step": 494
},
{
"completion_length": 3287.5208740234375,
"epoch": 0.5657142857142857,
"grad_norm": 0.05122198164463043,
"kl": 0.0015921592712402344,
"learning_rate": 1.0039472645551372e-07,
"loss": 0.048,
"reward": 0.158145634457469,
"reward_std": 0.8010805640369654,
"rewards/cosine_scaled_reward": -0.07717718556523323,
"rewards/format_reward": 0.3125000074505806,
"step": 495
},
{
"completion_length": 1689.2500038146973,
"epoch": 0.5668571428571428,
"grad_norm": 0.03907603397965431,
"kl": 0.001201629638671875,
"learning_rate": 1.002741278414069e-07,
"loss": -0.003,
"reward": 1.0886746868491173,
"reward_std": 0.9113657027482986,
"rewards/cosine_scaled_reward": 0.16933735646307468,
"rewards/format_reward": 0.7500000111758709,
"step": 496
},
{
"completion_length": 2586.895851135254,
"epoch": 0.568,
"grad_norm": 0.039880234748125076,
"kl": 0.0023741722106933594,
"learning_rate": 1.0017544823184055e-07,
"loss": 0.0129,
"reward": 0.7896264218725264,
"reward_std": 0.6654196940362453,
"rewards/cosine_scaled_reward": 0.1656465344130993,
"rewards/format_reward": 0.45833333395421505,
"step": 497
},
{
"completion_length": 2759.7500228881836,
"epoch": 0.5691428571428572,
"grad_norm": 0.03684534505009651,
"kl": 0.0011496543884277344,
"learning_rate": 1.0009869243631952e-07,
"loss": -0.0041,
"reward": 0.49048804119229317,
"reward_std": 0.7747926935553551,
"rewards/cosine_scaled_reward": -0.004755992442369461,
"rewards/format_reward": 0.5000000018626451,
"step": 498
},
{
"completion_length": 2801.791702270508,
"epoch": 0.5702857142857143,
"grad_norm": 0.03754589334130287,
"kl": 0.0016314983367919922,
"learning_rate": 1.000438641958131e-07,
"loss": 0.0704,
"reward": 0.5553555823862553,
"reward_std": 0.9310576990246773,
"rewards/cosine_scaled_reward": 0.03809444326907396,
"rewards/format_reward": 0.47916666977107525,
"step": 499
},
{
"completion_length": 3048.500045776367,
"epoch": 0.5714285714285714,
"grad_norm": 0.03999260067939758,
"kl": 0.0013794898986816406,
"learning_rate": 1.0001096618257236e-07,
"loss": 0.0509,
"reward": -0.01913883164525032,
"reward_std": 0.7366000525653362,
"rewards/cosine_scaled_reward": -0.14498609118163586,
"rewards/format_reward": 0.27083333767950535,
"step": 500
},
{
"epoch": 0.5714285714285714,
"step": 500,
"total_flos": 0.0,
"train_loss": 0.027085746813583684,
"train_runtime": 165463.4808,
"train_samples_per_second": 0.145,
"train_steps_per_second": 0.003
}
],
"logging_steps": 1,
"max_steps": 500,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 50,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 6,
"trial_name": null,
"trial_params": null
}