{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.37735849056603776,
"eval_steps": 500,
"global_step": 600,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 332.5,
"completions/max_terminated_length": 332.5,
"completions/mean_length": 279.3291778564453,
"completions/mean_terminated_length": 279.3291778564453,
"completions/min_length": 220.8,
"completions/min_terminated_length": 220.8,
"entropy": 0.24179386061926683,
"epoch": 0.006289308176100629,
"frac_reward_zero_std": 0.10000000223517418,
"grad_norm": 0.3676387071609497,
"items/detected_max": 32.0,
"items/detected_mean": 26.7125,
"items/detected_min": 21.0,
"kl": 0.00015798515381296359,
"learning_rate": 1e-06,
"loss": 0.021,
"num_tokens": 279138.0,
"reward": 0.03066593613475561,
"reward_std": 0.024809306021779776,
"reward_total": 0.19474640637636184,
"rewards/pos_1/mean": 0.19474640265107154,
"rewards/pos_1/std": 0.07326587401330471,
"rewards/pos_10/mean": 0.011071646504569798,
"rewards/pos_10/std": 0.02179219771642238,
"rewards/pos_11/mean": 0.007418489165138454,
"rewards/pos_11/std": 0.01720283292233944,
"rewards/pos_12/mean": 0.006210686359554529,
"rewards/pos_12/std": 0.0179185833549127,
"rewards/pos_13/mean": 0.0039982261369004846,
"rewards/pos_13/std": 0.009272939292714,
"rewards/pos_14/mean": 0.0012971448712050914,
"rewards/pos_14/std": 0.007776029966771602,
"rewards/pos_15/mean": -0.0006621864042244852,
"rewards/pos_15/std": 0.005226330552250147,
"rewards/pos_16/mean": 0.0002548443153500557,
"rewards/pos_16/std": 0.0015290660317987203,
"rewards/pos_17/mean": -0.00012490232475101994,
"rewards/pos_17/std": 0.0002498046495020399,
"rewards/pos_18/mean": -4.625929407134921e-19,
"rewards/pos_18/std": 9.251858814269843e-19,
"rewards/pos_19/mean": -4.625929407134921e-19,
"rewards/pos_19/std": 9.251858814269843e-19,
"rewards/pos_2/mean": 0.11575367655605077,
"rewards/pos_2/std": 0.07039788216352463,
"rewards/pos_20/mean": -4.625929407134921e-19,
"rewards/pos_20/std": 9.251858814269843e-19,
"rewards/pos_3/mean": 0.08213621266186237,
"rewards/pos_3/std": 0.05401332974433899,
"rewards/pos_4/mean": 0.059673574194312094,
"rewards/pos_4/std": 0.05429626442492008,
"rewards/pos_5/mean": 0.04313204605132341,
"rewards/pos_5/std": 0.04542485494166613,
"rewards/pos_6/mean": 0.03173917992971838,
"rewards/pos_6/std": 0.0341308044269681,
"rewards/pos_7/mean": 0.025256002554669976,
"rewards/pos_7/std": 0.03226408157497644,
"rewards/pos_8/mean": 0.016526030853856354,
"rewards/pos_8/std": 0.027043108874931932,
"rewards/pos_9/mean": 0.014891598536632956,
"rewards/pos_9/std": 0.024382113991305233,
"step": 10
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 337.5,
"completions/max_terminated_length": 337.5,
"completions/mean_length": 279.1000091552734,
"completions/mean_terminated_length": 279.1000091552734,
"completions/min_length": 221.7,
"completions/min_terminated_length": 221.7,
"entropy": 0.24853383572772145,
"epoch": 0.012578616352201259,
"frac_reward_zero_std": 0.07500000223517418,
"grad_norm": 0.3557882308959961,
"items/detected_max": 33.0,
"items/detected_mean": 26.795833333333334,
"items/detected_min": 20.0,
"kl": 0.00038001918375509073,
"learning_rate": 1e-06,
"loss": 0.0214,
"num_tokens": 561062.0,
"reward": 0.027369741536676883,
"reward_std": 0.028840648476034403,
"reward_total": 0.1867876473814249,
"rewards/pos_1/mean": 0.1867876447737217,
"rewards/pos_1/std": 0.09987052232027054,
"rewards/pos_10/mean": 0.005311872425954789,
"rewards/pos_10/std": 0.0194315308239311,
"rewards/pos_11/mean": 0.002390401461161673,
"rewards/pos_11/std": 0.012136956956237555,
"rewards/pos_12/mean": 0.0009011006564833224,
"rewards/pos_12/std": 0.013748793490231037,
"rewards/pos_13/mean": 0.0011169454897753893,
"rewards/pos_13/std": 0.010598769155330956,
"rewards/pos_14/mean": 0.000874718651175499,
"rewards/pos_14/std": 0.006548650283366442,
"rewards/pos_15/mean": 0.0003694009268656373,
"rewards/pos_15/std": 0.004986041854135692,
"rewards/pos_16/mean": 0.00024476498365402224,
"rewards/pos_16/std": 0.001508907275274396,
"rewards/pos_17/mean": 0.00012490232475101948,
"rewards/pos_17/std": 0.0017486325465142726,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.0002410189714282751,
"rewards/pos_19/std": 0.0002783047268167138,
"rewards/pos_2/mean": 0.10891713332384825,
"rewards/pos_2/std": 0.08113598320633172,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.07684338670223952,
"rewards/pos_3/std": 0.0730495484545827,
"rewards/pos_4/mean": 0.05329529843293131,
"rewards/pos_4/std": 0.06378404032438993,
"rewards/pos_5/mean": 0.038301121164113286,
"rewards/pos_5/std": 0.05285612549632788,
"rewards/pos_6/mean": 0.026770615531131626,
"rewards/pos_6/std": 0.047124340385198596,
"rewards/pos_7/mean": 0.01945003040600568,
"rewards/pos_7/std": 0.034271925175562504,
"rewards/pos_8/mean": 0.015267963660880924,
"rewards/pos_8/std": 0.028611483983695507,
"rewards/pos_9/mean": 0.010668529532267712,
"rewards/pos_9/std": 0.02512239357456565,
"step": 20
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 327.5,
"completions/max_terminated_length": 327.5,
"completions/mean_length": 275.84167938232423,
"completions/mean_terminated_length": 275.84167938232423,
"completions/min_length": 216.4,
"completions/min_terminated_length": 216.4,
"entropy": 0.2397403221887847,
"epoch": 0.018867924528301886,
"frac_reward_zero_std": 0.15833333730697632,
"grad_norm": 0.3486528694629669,
"items/detected_max": 30.0,
"items/detected_mean": 26.129166666666666,
"items/detected_min": 19.0,
"kl": 0.0006169086136045128,
"learning_rate": 1e-06,
"loss": 0.0205,
"num_tokens": 846746.0,
"reward": 0.023525450099259614,
"reward_std": 0.02375112334266305,
"reward_total": 0.17002619691193105,
"rewards/pos_1/mean": 0.1700261954218149,
"rewards/pos_1/std": 0.06453945254907012,
"rewards/pos_10/mean": 0.007429246860556305,
"rewards/pos_10/std": 0.01979129179380834,
"rewards/pos_11/mean": 0.003910829390224535,
"rewards/pos_11/std": 0.013063314463943243,
"rewards/pos_12/mean": 0.0029709233087487517,
"rewards/pos_12/std": 0.012117477436549961,
"rewards/pos_13/mean": 0.001196519553195685,
"rewards/pos_13/std": 0.007730130874551833,
"rewards/pos_14/mean": 0.001231371099129319,
"rewards/pos_14/std": 0.006333466898649931,
"rewards/pos_15/mean": 0.0003137066261842847,
"rewards/pos_15/std": 0.00301144989207387,
"rewards/pos_16/mean": 0.00045505964662879707,
"rewards/pos_16/std": 0.0028057243209332226,
"rewards/pos_17/mean": 0.0008398459758609533,
"rewards/pos_17/std": 0.0021793012507259845,
"rewards/pos_18/mean": 0.0008421394741162658,
"rewards/pos_18/std": 0.002036854811012745,
"rewards/pos_19/mean": 0.0004743130411952734,
"rewards/pos_19/std": 0.0009486260823905468,
"rewards/pos_2/mean": 0.08502510460093618,
"rewards/pos_2/std": 0.058614648133516314,
"rewards/pos_20/mean": 0.0004743130411952734,
"rewards/pos_20/std": 0.0009486260823905468,
"rewards/pos_3/mean": 0.05779995219781995,
"rewards/pos_3/std": 0.04915156504139304,
"rewards/pos_4/mean": 0.046246381290256974,
"rewards/pos_4/std": 0.054492526780813934,
"rewards/pos_5/mean": 0.031023955554701388,
"rewards/pos_5/std": 0.04292720593512058,
"rewards/pos_6/mean": 0.023734602285549043,
"rewards/pos_6/std": 0.036466607637703416,
"rewards/pos_7/mean": 0.016808894340647383,
"rewards/pos_7/std": 0.03472186983563006,
"rewards/pos_8/mean": 0.012099898888845929,
"rewards/pos_8/std": 0.03379071457311511,
"rewards/pos_9/mean": 0.007605718693230301,
"rewards/pos_9/std": 0.029351587407290936,
"step": 30
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 339.4,
"completions/max_terminated_length": 339.4,
"completions/mean_length": 278.9750091552734,
"completions/mean_terminated_length": 278.9750091552734,
"completions/min_length": 217.7,
"completions/min_terminated_length": 217.7,
"entropy": 0.2447954393302401,
"epoch": 0.025157232704402517,
"frac_reward_zero_std": 0.11666666939854622,
"grad_norm": 0.22751903533935547,
"items/detected_max": 31.0,
"items/detected_mean": 26.841666666666665,
"items/detected_min": 18.0,
"kl": 0.0006590462109670626,
"learning_rate": 1e-06,
"loss": 0.0212,
"num_tokens": 1127066.0,
"reward": 0.01795863511506468,
"reward_std": 0.01903191963210702,
"reward_total": 0.19013034030795098,
"rewards/pos_1/mean": 0.19013033509254457,
"rewards/pos_1/std": 0.07505300045013427,
"rewards/pos_10/mean": 0.0008360685140360146,
"rewards/pos_10/std": 0.011655322345905005,
"rewards/pos_11/mean": -0.0004029428935609758,
"rewards/pos_11/std": 0.011080938763916493,
"rewards/pos_12/mean": -0.00019872152479365467,
"rewards/pos_12/std": 0.006120560830458999,
"rewards/pos_13/mean": -0.0004367430577985942,
"rewards/pos_13/std": 0.0051455656997859475,
"rewards/pos_14/mean": -0.0005456584040075541,
"rewards/pos_14/std": 0.0031746502034366133,
"rewards/pos_15/mean": -0.0002604166860692203,
"rewards/pos_15/std": 0.002604166674427688,
"rewards/pos_16/mean": -0.00038226647302508354,
"rewards/pos_16/std": 0.0007645330158993601,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.07478453110670671,
"rewards/pos_2/std": 0.05773529279977083,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.044285639701411125,
"rewards/pos_3/std": 0.04824868915602565,
"rewards/pos_4/mean": 0.024659152084495872,
"rewards/pos_4/std": 0.04134202003479004,
"rewards/pos_5/mean": 0.014307024469599128,
"rewards/pos_5/std": 0.03745766724459827,
"rewards/pos_6/mean": 0.007001709949690849,
"rewards/pos_6/std": 0.02652715602889657,
"rewards/pos_7/mean": 0.0030801351997070013,
"rewards/pos_7/std": 0.022172968951053917,
"rewards/pos_8/mean": 0.001393768424168229,
"rewards/pos_8/std": 0.017827122542075812,
"rewards/pos_9/mean": 0.0010436996119096876,
"rewards/pos_9/std": 0.01348349207546562,
"step": 40
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 336.2,
"completions/max_terminated_length": 336.2,
"completions/mean_length": 277.8895919799805,
"completions/mean_terminated_length": 277.8895919799805,
"completions/min_length": 220.9,
"completions/min_terminated_length": 220.9,
"entropy": 0.24365584701299667,
"epoch": 0.031446540880503145,
"frac_reward_zero_std": 0.1666666693985462,
"grad_norm": 0.2256021499633789,
"items/detected_max": 31.0,
"items/detected_mean": 26.358333333333334,
"items/detected_min": 19.0,
"kl": 0.000693480301076003,
"learning_rate": 1e-06,
"loss": 0.0207,
"num_tokens": 1402153.0,
"reward": 0.018253592820838094,
"reward_std": 0.024873400665819646,
"reward_total": 0.12365640453062951,
"rewards/pos_1/mean": 0.12365640229545534,
"rewards/pos_1/std": 0.08260059077292681,
"rewards/pos_10/mean": 0.003422046871855855,
"rewards/pos_10/std": 0.015368315065279603,
"rewards/pos_11/mean": 0.0008227722952142358,
"rewards/pos_11/std": 0.012859936244785786,
"rewards/pos_12/mean": 0.0011632090085186065,
"rewards/pos_12/std": 0.008362313848920166,
"rewards/pos_13/mean": -0.0005874803406186402,
"rewards/pos_13/std": 0.00946071040816605,
"rewards/pos_14/mean": 0.00012089891824871302,
"rewards/pos_14/std": 0.003749142773449421,
"rewards/pos_15/mean": -0.00013020834885537624,
"rewards/pos_15/std": 0.00234375,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.08051420627161861,
"rewards/pos_2/std": 0.06426868531852961,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05728820329532027,
"rewards/pos_3/std": 0.061889216490089896,
"rewards/pos_4/mean": 0.0330103222746402,
"rewards/pos_4/std": 0.054461258091032506,
"rewards/pos_5/mean": 0.021673589176498354,
"rewards/pos_5/std": 0.05066446699202061,
"rewards/pos_6/mean": 0.017316767123702448,
"rewards/pos_6/std": 0.04409534465521574,
"rewards/pos_7/mean": 0.013347537699155509,
"rewards/pos_7/std": 0.037946426495909694,
"rewards/pos_8/mean": 0.009339081990765408,
"rewards/pos_8/std": 0.027106760535389185,
"rewards/pos_9/mean": 0.004739647073438391,
"rewards/pos_9/std": 0.021040741237811744,
"step": 50
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 332.2,
"completions/max_terminated_length": 332.2,
"completions/mean_length": 275.77292785644534,
"completions/mean_terminated_length": 275.77292785644534,
"completions/min_length": 210.5,
"completions/min_terminated_length": 210.5,
"entropy": 0.24245100983728965,
"epoch": 0.03773584905660377,
"frac_reward_zero_std": 0.1833333380520344,
"grad_norm": 0.34496742486953735,
"items/detected_max": 30.0,
"items/detected_mean": 26.470833333333335,
"items/detected_min": 19.0,
"kl": 0.0010004034176139007,
"learning_rate": 1e-06,
"loss": 0.0211,
"num_tokens": 1695536.0,
"reward": 0.027998355263844133,
"reward_std": 0.025541644264012574,
"reward_total": 0.23674236759543418,
"rewards/pos_1/mean": 0.23674236685037614,
"rewards/pos_1/std": 0.0875159040093422,
"rewards/pos_10/mean": 0.008996098558418453,
"rewards/pos_10/std": 0.018989878077991306,
"rewards/pos_11/mean": 0.005316582974046469,
"rewards/pos_11/std": 0.015308488998562098,
"rewards/pos_12/mean": 0.004086111392825842,
"rewards/pos_12/std": 0.014277325919829309,
"rewards/pos_13/mean": 0.0035586868296377363,
"rewards/pos_13/std": 0.010105705843307078,
"rewards/pos_14/mean": 0.0022290571592748164,
"rewards/pos_14/std": 0.00829645860940218,
"rewards/pos_15/mean": 0.0011718750232830645,
"rewards/pos_15/std": 0.0039062500465661286,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.10075544621795415,
"rewards/pos_2/std": 0.07253548875451088,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.0666294975206256,
"rewards/pos_3/std": 0.05721334554255009,
"rewards/pos_4/mean": 0.04308908779639751,
"rewards/pos_4/std": 0.05217773541808128,
"rewards/pos_5/mean": 0.031524106953293086,
"rewards/pos_5/std": 0.04607770144939423,
"rewards/pos_6/mean": 0.019696346682030706,
"rewards/pos_6/std": 0.03438961319625378,
"rewards/pos_7/mean": 0.013458259997423738,
"rewards/pos_7/std": 0.03490288313478231,
"rewards/pos_8/mean": 0.013818001028266736,
"rewards/pos_8/std": 0.02888077814131975,
"rewards/pos_9/mean": 0.009397907659877091,
"rewards/pos_9/std": 0.02525058896280825,
"step": 60
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 330.6,
"completions/max_terminated_length": 330.6,
"completions/mean_length": 271.752099609375,
"completions/mean_terminated_length": 271.752099609375,
"completions/min_length": 213.8,
"completions/min_terminated_length": 213.8,
"entropy": 0.23684501089155674,
"epoch": 0.0440251572327044,
"frac_reward_zero_std": 0.1416666701436043,
"grad_norm": 0.30509424209594727,
"items/detected_max": 30.0,
"items/detected_mean": 26.395833333333332,
"items/detected_min": 20.0,
"kl": 0.00137258033840529,
"learning_rate": 1e-06,
"loss": 0.0193,
"num_tokens": 1990757.0,
"reward": 0.019408026617020367,
"reward_std": 0.019861249532550574,
"reward_total": 0.15183117985725403,
"rewards/pos_1/mean": 0.15183117538690566,
"rewards/pos_1/std": 0.06800224892795086,
"rewards/pos_10/mean": 0.0017800831352360546,
"rewards/pos_10/std": 0.01105083387810737,
"rewards/pos_11/mean": -5.587483756244183e-05,
"rewards/pos_11/std": 0.011234215367585421,
"rewards/pos_12/mean": 0.0008566255099140107,
"rewards/pos_12/std": 0.006780216773040593,
"rewards/pos_13/mean": 0.0007554006297141314,
"rewards/pos_13/std": 0.006829833285883069,
"rewards/pos_14/mean": -0.0002901801839470863,
"rewards/pos_14/std": 0.006283610546961427,
"rewards/pos_15/mean": -0.0006621863809414208,
"rewards/pos_15/std": 0.0026221639243885876,
"rewards/pos_16/mean": 0.00038226647302508354,
"rewards/pos_16/std": 0.0012742216931656003,
"rewards/pos_17/mean": -0.0004996092990040779,
"rewards/pos_17/std": 0.0009992185980081559,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.10222204849123954,
"rewards/pos_2/std": 0.06664478406310081,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.060004901699721815,
"rewards/pos_3/std": 0.06095823477953673,
"rewards/pos_4/mean": 0.0354589288122952,
"rewards/pos_4/std": 0.05521870870143175,
"rewards/pos_5/mean": 0.01933925971388817,
"rewards/pos_5/std": 0.03796455860137939,
"rewards/pos_6/mean": 0.008141869585961104,
"rewards/pos_6/std": 0.021356439776718618,
"rewards/pos_7/mean": 0.0011259438935667277,
"rewards/pos_7/std": 0.017314390907995404,
"rewards/pos_8/mean": 0.003906934190308675,
"rewards/pos_8/std": 0.012045483710244298,
"rewards/pos_9/mean": 0.003862919790359834,
"rewards/pos_9/std": 0.010645803553052247,
"step": 70
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 334.6,
"completions/max_terminated_length": 334.6,
"completions/mean_length": 270.6020904541016,
"completions/mean_terminated_length": 270.6020904541016,
"completions/min_length": 208.7,
"completions/min_terminated_length": 208.7,
"entropy": 0.2382718403513233,
"epoch": 0.050314465408805034,
"frac_reward_zero_std": 0.1333333358168602,
"grad_norm": 0.23278222978115082,
"items/detected_max": 30.0,
"items/detected_mean": 26.154166666666665,
"items/detected_min": 15.0,
"kl": 0.0016101757981838696,
"learning_rate": 1e-06,
"loss": 0.0192,
"num_tokens": 2274070.0,
"reward": 0.018261409434489905,
"reward_std": 0.019813417922705412,
"reward_total": 0.13820529966615142,
"rewards/pos_1/mean": 0.138205295545049,
"rewards/pos_1/std": 0.06373181324452162,
"rewards/pos_10/mean": 0.001570706581696868,
"rewards/pos_10/std": 0.01301275547593832,
"rewards/pos_11/mean": 0.0004506188444793224,
"rewards/pos_11/std": 0.008641806431114674,
"rewards/pos_12/mean": -0.0003394705709069967,
"rewards/pos_12/std": 0.00797676993533969,
"rewards/pos_13/mean": 0.00011044344864785671,
"rewards/pos_13/std": 0.0031491465866565703,
"rewards/pos_14/mean": -0.0005456584272906184,
"rewards/pos_14/std": 0.0024994839914143085,
"rewards/pos_15/mean": -0.00026041667442768814,
"rewards/pos_15/std": 0.0026041667442768812,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.07644948430825024,
"rewards/pos_2/std": 0.05021180454641581,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04756893061567098,
"rewards/pos_3/std": 0.047028692159801724,
"rewards/pos_4/mean": 0.03352718676906079,
"rewards/pos_4/std": 0.04555289912968874,
"rewards/pos_5/mean": 0.02097760115284473,
"rewards/pos_5/std": 0.039121273113414644,
"rewards/pos_6/mean": 0.018290499058639397,
"rewards/pos_6/std": 0.03251297641545534,
"rewards/pos_7/mean": 0.016266733058728278,
"rewards/pos_7/std": 0.03300811811350286,
"rewards/pos_8/mean": 0.00988356041489169,
"rewards/pos_8/std": 0.028713068063370883,
"rewards/pos_9/mean": 0.0033275068912189455,
"rewards/pos_9/std": 0.017993874754756688,
"step": 80
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 329.7,
"completions/max_terminated_length": 329.7,
"completions/mean_length": 267.65834045410156,
"completions/mean_terminated_length": 267.65834045410156,
"completions/min_length": 209.7,
"completions/min_terminated_length": 209.7,
"entropy": 0.2301488552863399,
"epoch": 0.05660377358490566,
"frac_reward_zero_std": 0.1916666731238365,
"grad_norm": 0.18648235499858856,
"items/detected_max": 31.0,
"items/detected_mean": 25.570833333333333,
"items/detected_min": 19.0,
"kl": 0.0020055538250744577,
"learning_rate": 1e-06,
"loss": 0.0183,
"num_tokens": 2554026.0,
"reward": 0.023970679030753672,
"reward_std": 0.021331347804516555,
"reward_total": 0.19938781536184252,
"rewards/pos_1/mean": 0.19938780865631997,
"rewards/pos_1/std": 0.0798557098954916,
"rewards/pos_10/mean": 0.0008782313205301761,
"rewards/pos_10/std": 0.011960454075597227,
"rewards/pos_11/mean": 0.00036563378816936167,
"rewards/pos_11/std": 0.01036488190293312,
"rewards/pos_12/mean": 0.0002656884375028312,
"rewards/pos_12/std": 0.005923388386145234,
"rewards/pos_13/mean": -0.0001367966295219958,
"rewards/pos_13/std": 0.002462339331395924,
"rewards/pos_14/mean": -0.0006665573455393314,
"rewards/pos_14/std": 0.0013331146910786629,
"rewards/pos_15/mean": -0.0006510416860692203,
"rewards/pos_15/std": 0.0013020833721384406,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.10163866644725203,
"rewards/pos_2/std": 0.06948479041457176,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06862257414031774,
"rewards/pos_3/std": 0.06218107980675995,
"rewards/pos_4/mean": 0.044460686412639915,
"rewards/pos_4/std": 0.050121272914111614,
"rewards/pos_5/mean": 0.02706362712197006,
"rewards/pos_5/std": 0.04326969822868705,
"rewards/pos_6/mean": 0.017091114865615965,
"rewards/pos_6/std": 0.031835716869682076,
"rewards/pos_7/mean": 0.010764542536344379,
"rewards/pos_7/std": 0.019727899390272797,
"rewards/pos_8/mean": 0.007698593634025505,
"rewards/pos_8/std": 0.02043189574033022,
"rewards/pos_9/mean": 0.003255945723503828,
"rewards/pos_9/std": 0.01512229349464178,
"step": 90
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 333.0,
"completions/max_terminated_length": 333.0,
"completions/mean_length": 272.6104248046875,
"completions/mean_terminated_length": 272.6104248046875,
"completions/min_length": 210.0,
"completions/min_terminated_length": 210.0,
"entropy": 0.24247641147424778,
"epoch": 0.06289308176100629,
"frac_reward_zero_std": 0.2333333395421505,
"grad_norm": 0.4801751375198364,
"items/detected_max": 34.0,
"items/detected_mean": 26.158333333333335,
"items/detected_min": 18.0,
"kl": 0.0021246019886651387,
"learning_rate": 1e-06,
"loss": 0.0187,
"num_tokens": 2834739.0,
"reward": 0.02137735360302031,
"reward_std": 0.022431656159460543,
"reward_total": 0.15205592066049575,
"rewards/pos_1/mean": 0.1520559150725603,
"rewards/pos_1/std": 0.08682643361389637,
"rewards/pos_10/mean": 0.002806135616265237,
"rewards/pos_10/std": 0.014712602784857153,
"rewards/pos_11/mean": 0.0010838294634595512,
"rewards/pos_11/std": 0.010319896694272756,
"rewards/pos_12/mean": 9.85862105153501e-05,
"rewards/pos_12/std": 0.00413814561907202,
"rewards/pos_13/mean": -0.0005735396873205901,
"rewards/pos_13/std": 0.003230412770062685,
"rewards/pos_14/mean": -0.0004123469581827521,
"rewards/pos_14/std": 0.002757887914776802,
"rewards/pos_15/mean": 0.00013020833721384407,
"rewards/pos_15/std": 0.001822916720993817,
"rewards/pos_16/mean": 0.0,
"rewards/pos_16/std": 0.0,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.09235873296856881,
"rewards/pos_2/std": 0.0637791832908988,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06637982809916139,
"rewards/pos_3/std": 0.05287711825221777,
"rewards/pos_4/mean": 0.03900537546724081,
"rewards/pos_4/std": 0.054096858203411105,
"rewards/pos_5/mean": 0.027999201137572528,
"rewards/pos_5/std": 0.04368511065840721,
"rewards/pos_6/mean": 0.021048975386656822,
"rewards/pos_6/std": 0.03628326542675495,
"rewards/pos_7/mean": 0.012417805101722478,
"rewards/pos_7/std": 0.03044475233182311,
"rewards/pos_8/mean": 0.009451015887316316,
"rewards/pos_8/std": 0.022707051225006582,
"rewards/pos_9/mean": 0.004067648795899004,
"rewards/pos_9/std": 0.02021083978470415,
"step": 100
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 330.6,
"completions/max_terminated_length": 330.6,
"completions/mean_length": 274.5708465576172,
"completions/mean_terminated_length": 274.5708465576172,
"completions/min_length": 214.3,
"completions/min_terminated_length": 214.3,
"entropy": 0.23819080513591567,
"epoch": 0.06918238993710692,
"frac_reward_zero_std": 0.0666666679084301,
"grad_norm": 0.2140370011329651,
"items/detected_max": 30.0,
"items/detected_mean": 26.3875,
"items/detected_min": 19.0,
"kl": 0.0024520867267710856,
"learning_rate": 1e-06,
"loss": 0.0186,
"num_tokens": 3123245.0,
"reward": 0.028679715376347302,
"reward_std": 0.025797955319285394,
"reward_total": 0.1972983568906784,
"rewards/pos_1/mean": 0.1972983494400978,
"rewards/pos_1/std": 0.08312824480235577,
"rewards/pos_10/mean": 0.006511424528434872,
"rewards/pos_10/std": 0.012539930804632605,
"rewards/pos_11/mean": 0.004311096295714379,
"rewards/pos_11/std": 0.009560632146894931,
"rewards/pos_12/mean": 0.003057955822441727,
"rewards/pos_12/std": 0.009056452894583344,
"rewards/pos_13/mean": 0.0026791852549649777,
"rewards/pos_13/std": 0.0089504691073671,
"rewards/pos_14/mean": 0.003384079411625862,
"rewards/pos_14/std": 0.00754068112000823,
"rewards/pos_15/mean": 0.001009297906421125,
"rewards/pos_15/std": 0.004131007054820657,
"rewards/pos_16/mean": 0.0010092979297041899,
"rewards/pos_16/std": 0.0020185960456728944,
"rewards/pos_17/mean": -0.00012490234803408338,
"rewards/pos_17/std": 0.0021070301532745373,
"rewards/pos_18/mean": 4.625929407134921e-19,
"rewards/pos_18/std": 9.251858814269843e-19,
"rewards/pos_19/mean": 4.625929407134921e-19,
"rewards/pos_19/std": 9.251858814269843e-19,
"rewards/pos_2/mean": 0.11366754285991192,
"rewards/pos_2/std": 0.07409449964761734,
"rewards/pos_20/mean": 4.625929407134921e-19,
"rewards/pos_20/std": 9.251858814269843e-19,
"rewards/pos_3/mean": 0.078475221991539,
"rewards/pos_3/std": 0.06386412046849728,
"rewards/pos_4/mean": 0.05367650641128421,
"rewards/pos_4/std": 0.05735928248614073,
"rewards/pos_5/mean": 0.0359677754342556,
"rewards/pos_5/std": 0.051013645529747007,
"rewards/pos_6/mean": 0.030178581224754454,
"rewards/pos_6/std": 0.042489323485642674,
"rewards/pos_7/mean": 0.017106734635308384,
"rewards/pos_7/std": 0.031865698378533126,
"rewards/pos_8/mean": 0.013932319730520249,
"rewards/pos_8/std": 0.03123240638524294,
"rewards/pos_9/mean": 0.011453809647355229,
"rewards/pos_9/std": 0.025007078982889654,
"step": 110
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 334.4,
"completions/max_terminated_length": 334.4,
"completions/mean_length": 277.0229248046875,
"completions/mean_terminated_length": 277.0229248046875,
"completions/min_length": 216.8,
"completions/min_terminated_length": 216.8,
"entropy": 0.2525980423515042,
"epoch": 0.07547169811320754,
"frac_reward_zero_std": 0.15833333805203437,
"grad_norm": 0.4217924177646637,
"items/detected_max": 31.0,
"items/detected_mean": 26.629166666666666,
"items/detected_min": 14.0,
"kl": 0.0021200735282036476,
"learning_rate": 1e-06,
"loss": 0.0208,
"num_tokens": 3414244.0,
"reward": 0.021100316557567566,
"reward_std": 0.02556178504601121,
"reward_total": 0.15421188264153898,
"rewards/pos_1/mean": 0.1542118826182559,
"rewards/pos_1/std": 0.08668690882623195,
"rewards/pos_10/mean": 0.001462233636993915,
"rewards/pos_10/std": 0.014794380008243024,
"rewards/pos_11/mean": 0.0009601796511560679,
"rewards/pos_11/std": 0.009470279794186354,
"rewards/pos_12/mean": 0.00016555656911805273,
"rewards/pos_12/std": 0.008588731940835714,
"rewards/pos_13/mean": 0.0011903240345418453,
"rewards/pos_13/std": 0.006918352888897061,
"rewards/pos_14/mean": 0.0006779891438782216,
"rewards/pos_14/std": 0.005625644978135824,
"rewards/pos_15/mean": 0.0012298537883907556,
"rewards/pos_15/std": 0.004600600665435195,
"rewards/pos_16/mean": 0.0018836816772818566,
"rewards/pos_16/std": 0.003439636155962944,
"rewards/pos_17/mean": -1.7571356147527693e-05,
"rewards/pos_17/std": 0.001963294483721256,
"rewards/pos_18/mean": 0.0004820379428565502,
"rewards/pos_18/std": 0.0009640758857131004,
"rewards/pos_19/mean": 0.0004820379428565502,
"rewards/pos_19/std": 0.0009640758857131004,
"rewards/pos_2/mean": 0.09640580550767482,
"rewards/pos_2/std": 0.0769618958234787,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05643308900762349,
"rewards/pos_3/std": 0.06533879688940943,
"rewards/pos_4/mean": 0.03552527402061969,
"rewards/pos_4/std": 0.0584185004234314,
"rewards/pos_5/mean": 0.02315435045547929,
"rewards/pos_5/std": 0.04358684956096113,
"rewards/pos_6/mean": 0.020684696536045523,
"rewards/pos_6/std": 0.03866914417594671,
"rewards/pos_7/mean": 0.012705407536122948,
"rewards/pos_7/std": 0.03549118069931865,
"rewards/pos_8/mean": 0.009155850886600092,
"rewards/pos_8/std": 0.02957823732867837,
"rewards/pos_9/mean": 0.005213634658139199,
"rewards/pos_9/std": 0.019175097346305847,
"step": 120
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 345.9,
"completions/max_terminated_length": 345.9,
"completions/mean_length": 271.9437591552734,
"completions/mean_terminated_length": 271.9437591552734,
"completions/min_length": 216.0,
"completions/min_terminated_length": 216.0,
"entropy": 0.22371312171841662,
"epoch": 0.08176100628930817,
"frac_reward_zero_std": 0.21666667237877846,
"grad_norm": 0.2381807267665863,
"items/detected_max": 31.0,
"items/detected_mean": 25.783333333333335,
"items/detected_min": 19.0,
"kl": 0.0023203926954010966,
"learning_rate": 1e-06,
"loss": 0.0193,
"num_tokens": 3687189.0,
"reward": 0.017484810762107374,
"reward_std": 0.019230466429144145,
"reward_total": 0.12171695046126843,
"rewards/pos_1/mean": 0.1217169489711523,
"rewards/pos_1/std": 0.06389217637479305,
"rewards/pos_10/mean": 0.0034612691204529257,
"rewards/pos_10/std": 0.012485883384943008,
"rewards/pos_11/mean": 0.0005503416177816689,
"rewards/pos_11/std": 0.010597484931349755,
"rewards/pos_12/mean": 0.00045946375466883184,
"rewards/pos_12/std": 0.006546087586320937,
"rewards/pos_13/mean": 0.000506892695557326,
"rewards/pos_13/std": 0.006360282865352928,
"rewards/pos_14/mean": -0.00014572404325008392,
"rewards/pos_14/std": 0.003616921044886112,
"rewards/pos_15/mean": 0.0,
"rewards/pos_15/std": 0.002083333395421505,
"rewards/pos_16/mean": -0.0007645329460501671,
"rewards/pos_16/std": 0.0015290660317987203,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.07369502820074558,
"rewards/pos_2/std": 0.0545389074832201,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.048825977463275196,
"rewards/pos_3/std": 0.04302537702023983,
"rewards/pos_4/mean": 0.03623074255883694,
"rewards/pos_4/std": 0.04152446836233139,
"rewards/pos_5/mean": 0.02578731467947364,
"rewards/pos_5/std": 0.039170566760003564,
"rewards/pos_6/mean": 0.01823263019323349,
"rewards/pos_6/std": 0.03099028319120407,
"rewards/pos_7/mean": 0.010217601037584245,
"rewards/pos_7/std": 0.02759239007718861,
"rewards/pos_8/mean": 0.006804429623298347,
"rewards/pos_8/std": 0.02312811641022563,
"rewards/pos_9/mean": 0.004490223817992955,
"rewards/pos_9/std": 0.016783139877952637,
"step": 130
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 326.6,
"completions/max_terminated_length": 326.6,
"completions/mean_length": 269.00001220703126,
"completions/mean_terminated_length": 269.00001220703126,
"completions/min_length": 212.8,
"completions/min_terminated_length": 212.8,
"entropy": 0.23122907076030969,
"epoch": 0.0880503144654088,
"frac_reward_zero_std": 0.1833333380520344,
"grad_norm": 0.15572181344032288,
"items/detected_max": 32.0,
"items/detected_mean": 25.520833333333332,
"items/detected_min": 19.0,
"kl": 0.0024950778378600566,
"learning_rate": 1e-06,
"loss": 0.0175,
"num_tokens": 3974029.0,
"reward": 0.014080746471881867,
"reward_std": 0.016236986499279737,
"reward_total": 0.1274184116628021,
"rewards/pos_1/mean": 0.12741840942762792,
"rewards/pos_1/std": 0.05708431825041771,
"rewards/pos_10/mean": 0.00020685511990450322,
"rewards/pos_10/std": 0.007606808561831713,
"rewards/pos_11/mean": -0.002086038514971733,
"rewards/pos_11/std": 0.007140537723898888,
"rewards/pos_12/mean": 0.00026568842586129906,
"rewards/pos_12/std": 0.003909353911876678,
"rewards/pos_13/mean": -0.0001367966178804636,
"rewards/pos_13/std": 0.002462339401245117,
"rewards/pos_14/mean": -0.0011998032219707965,
"rewards/pos_14/std": 0.002399606443941593,
"rewards/pos_15/mean": -0.00039062501164153216,
"rewards/pos_15/std": 0.0007812500232830643,
"rewards/pos_16/mean": -0.0005096886307001114,
"rewards/pos_16/std": 0.0010193773545324803,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.05554983280599117,
"rewards/pos_2/std": 0.04518025889992714,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04318028772249818,
"rewards/pos_3/std": 0.044956205785274504,
"rewards/pos_4/mean": 0.030316960613708942,
"rewards/pos_4/std": 0.04388560689985752,
"rewards/pos_5/mean": 0.01796492321882397,
"rewards/pos_5/std": 0.03195359939709306,
"rewards/pos_6/mean": 0.008910082012880594,
"rewards/pos_6/std": 0.03093706537038088,
"rewards/pos_7/mean": 0.0019248059805249795,
"rewards/pos_7/std": 0.015874549094587565,
"rewards/pos_8/mean": 0.0006415006675524638,
"rewards/pos_8/std": 0.01841237680055201,
"rewards/pos_9/mean": -0.00019396404968574643,
"rewards/pos_9/std": 0.010641431924887002,
"step": 140
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 328.8,
"completions/max_terminated_length": 328.8,
"completions/mean_length": 269.02501068115237,
"completions/mean_terminated_length": 269.02501068115237,
"completions/min_length": 206.8,
"completions/min_terminated_length": 206.8,
"entropy": 0.23320260525991518,
"epoch": 0.09433962264150944,
"frac_reward_zero_std": 0.1666666716337204,
"grad_norm": 0.2597532272338867,
"items/detected_max": 32.0,
"items/detected_mean": 25.716666666666665,
"items/detected_min": 17.0,
"kl": 0.0030229106078574356,
"learning_rate": 1e-06,
"loss": 0.0171,
"num_tokens": 4257405.0,
"reward": 0.017440968845039605,
"reward_std": 0.014302454702556134,
"reward_total": 0.1550670871976763,
"rewards/pos_1/mean": 0.15506708160974086,
"rewards/pos_1/std": 0.04401266416534781,
"rewards/pos_10/mean": 0.0006411751499399542,
"rewards/pos_10/std": 0.00836768765002489,
"rewards/pos_11/mean": 0.0011508289724588394,
"rewards/pos_11/std": 0.007357292901724577,
"rewards/pos_12/mean": -0.0009468055446632207,
"rewards/pos_12/std": 0.0066732271574437615,
"rewards/pos_13/mean": -0.0007546751294285059,
"rewards/pos_13/std": 0.004247732646763325,
"rewards/pos_14/mean": -4.281066358089447e-05,
"rewards/pos_14/std": 0.0029963889624923467,
"rewards/pos_15/mean": 9.981024777516723e-05,
"rewards/pos_15/std": 0.001762120542116463,
"rewards/pos_16/mean": 0.00010816878639161587,
"rewards/pos_16/std": 0.001529983733780682,
"rewards/pos_17/mean": 0.00036553293466567993,
"rewards/pos_17/std": 0.0012306751683354377,
"rewards/pos_18/mean": 0.0002452176297083497,
"rewards/pos_18/std": 0.0014713057782500983,
"rewards/pos_19/mean": -0.0001205094857141394,
"rewards/pos_19/std": 0.00024101897142827724,
"rewards/pos_2/mean": 0.07839932842180133,
"rewards/pos_2/std": 0.03501760168001056,
"rewards/pos_20/mean": -1.8503717628539686e-18,
"rewards/pos_20/std": 2.136625223889579e-18,
"rewards/pos_3/mean": 0.04932655282318592,
"rewards/pos_3/std": 0.03529196488671005,
"rewards/pos_4/mean": 0.035052822227589785,
"rewards/pos_4/std": 0.03147873505949974,
"rewards/pos_5/mean": 0.014226565160788596,
"rewards/pos_5/std": 0.03353233290836215,
"rewards/pos_6/mean": 0.009943538601510227,
"rewards/pos_6/std": 0.02568354532122612,
"rewards/pos_7/mean": 0.0037480150582268833,
"rewards/pos_7/std": 0.01887619134504348,
"rewards/pos_8/mean": 0.00222595312516205,
"rewards/pos_8/std": 0.015316987968981266,
"rewards/pos_9/mean": 8.356973994523287e-05,
"rewards/pos_9/std": 0.010961612733080984,
"step": 150
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 331.5,
"completions/max_terminated_length": 331.5,
"completions/mean_length": 273.31251220703126,
"completions/mean_terminated_length": 273.31251220703126,
"completions/min_length": 217.2,
"completions/min_terminated_length": 217.2,
"entropy": 0.23213966513673465,
"epoch": 0.10062893081761007,
"frac_reward_zero_std": 0.14166666939854622,
"grad_norm": 0.3334786593914032,
"items/detected_max": 30.0,
"items/detected_mean": 25.591666666666665,
"items/detected_min": 19.0,
"kl": 0.0035994505208994574,
"learning_rate": 1e-06,
"loss": 0.0189,
"num_tokens": 4548299.0,
"reward": 0.015756109496578574,
"reward_std": 0.017546102358028292,
"reward_total": 0.14928646758198738,
"rewards/pos_1/mean": 0.14928646497428416,
"rewards/pos_1/std": 0.0711632726714015,
"rewards/pos_10/mean": 0.002952051186002791,
"rewards/pos_10/std": 0.011590327834710479,
"rewards/pos_11/mean": 4.112366586923599e-05,
"rewards/pos_11/std": 0.009777330886572599,
"rewards/pos_12/mean": 0.0006721259327605366,
"rewards/pos_12/std": 0.005886415066197514,
"rewards/pos_13/mean": -0.0004103898769244552,
"rewards/pos_13/std": 0.004967003781348467,
"rewards/pos_14/mean": -0.00039993440732359885,
"rewards/pos_14/std": 0.0007998688146471977,
"rewards/pos_15/mean": -0.0006510416860692203,
"rewards/pos_15/std": 0.0013020833721384406,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0002548443386331201,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.06409315429627896,
"rewards/pos_2/std": 0.05032593030482531,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04519490986131132,
"rewards/pos_3/std": 0.041902261599898336,
"rewards/pos_4/mean": 0.022920475325372537,
"rewards/pos_4/std": 0.036204461753368375,
"rewards/pos_5/mean": 0.013419941076426767,
"rewards/pos_5/std": 0.03253512904047966,
"rewards/pos_6/mean": 0.007387386483605951,
"rewards/pos_6/std": 0.025701201800256968,
"rewards/pos_7/mean": 0.005470841613714583,
"rewards/pos_7/std": 0.023291338980197907,
"rewards/pos_8/mean": 0.0032636421092320234,
"rewards/pos_8/std": 0.019500295352190732,
"rewards/pos_9/mean": 0.002256354584824294,
"rewards/pos_9/std": 0.01522524117026478,
"step": 160
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 336.9,
"completions/max_terminated_length": 336.9,
"completions/mean_length": 270.0791778564453,
"completions/mean_terminated_length": 270.0791778564453,
"completions/min_length": 207.9,
"completions/min_terminated_length": 207.9,
"entropy": 0.22610665323833626,
"epoch": 0.1069182389937107,
"frac_reward_zero_std": 0.15833333656191825,
"grad_norm": 0.2289900928735733,
"items/detected_max": 31.0,
"items/detected_mean": 25.416666666666668,
"items/detected_min": 18.0,
"kl": 0.003391667290755625,
"learning_rate": 1e-06,
"loss": 0.0179,
"num_tokens": 4831193.0,
"reward": 0.0186764198821038,
"reward_std": 0.01692580496892333,
"reward_total": 0.1566243339329958,
"rewards/pos_1/mean": 0.15662432909011842,
"rewards/pos_1/std": 0.05228233584202826,
"rewards/pos_10/mean": 0.0012737916316837073,
"rewards/pos_10/std": 0.009292289335280656,
"rewards/pos_11/mean": 0.0017623581690713764,
"rewards/pos_11/std": 0.006201159209012985,
"rewards/pos_12/mean": -0.0007756583509035408,
"rewards/pos_12/std": 0.00754371783696115,
"rewards/pos_13/mean": 0.000780485977884382,
"rewards/pos_13/std": 0.0026553449453786014,
"rewards/pos_14/mean": 0.0010540792252868413,
"rewards/pos_14/std": 0.0021081584505736827,
"rewards/pos_15/mean": -0.0006510416627861559,
"rewards/pos_15/std": 0.0031652866629883645,
"rewards/pos_16/mean": -0.00012742215767502877,
"rewards/pos_16/std": 0.0002548443386331219,
"rewards/pos_17/mean": -0.0003747069742530584,
"rewards/pos_17/std": 0.0007494139485061168,
"rewards/pos_18/mean": -0.0002452176297083506,
"rewards/pos_18/std": 0.0004904352594167012,
"rewards/pos_19/mean": -0.00012050948571413849,
"rewards/pos_19/std": 0.00024101897142827697,
"rewards/pos_2/mean": 0.09396323375403881,
"rewards/pos_2/std": 0.04918128461576998,
"rewards/pos_20/mean": -9.251858814269843e-19,
"rewards/pos_20/std": 1.8503717628539686e-18,
"rewards/pos_3/mean": 0.05363719742745161,
"rewards/pos_3/std": 0.038127432996407154,
"rewards/pos_4/mean": 0.028266194695606827,
"rewards/pos_4/std": 0.043761784210801125,
"rewards/pos_5/mean": 0.01885695867240429,
"rewards/pos_5/std": 0.04044422851875425,
"rewards/pos_6/mean": 0.010655944173777243,
"rewards/pos_6/std": 0.02998672341927886,
"rewards/pos_7/mean": 0.006501190612713497,
"rewards/pos_7/std": 0.022654754761606454,
"rewards/pos_8/mean": 0.002201350941322744,
"rewards/pos_8/std": 0.01561037302017212,
"rewards/pos_9/mean": 0.0002458267903421074,
"rewards/pos_9/std": 0.013765505608171224,
"step": 170
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 337.3,
"completions/max_terminated_length": 337.3,
"completions/mean_length": 268.4208419799805,
"completions/mean_terminated_length": 268.4208419799805,
"completions/min_length": 208.5,
"completions/min_terminated_length": 208.5,
"entropy": 0.22181829875335096,
"epoch": 0.11320754716981132,
"frac_reward_zero_std": 0.16666667088866233,
"grad_norm": 0.2745143473148346,
"items/detected_max": 30.0,
"items/detected_mean": 25.6,
"items/detected_min": 19.0,
"kl": 0.00353732905665917,
"learning_rate": 1e-06,
"loss": 0.0165,
"num_tokens": 5095139.0,
"reward": 0.01957973469980061,
"reward_std": 0.019774758955463767,
"reward_total": 0.17836526185274124,
"rewards/pos_1/mean": 0.17836525700986386,
"rewards/pos_1/std": 0.0697622362524271,
"rewards/pos_10/mean": 0.0003152468823827803,
"rewards/pos_10/std": 0.008289119158871473,
"rewards/pos_11/mean": -0.0011112218722701072,
"rewards/pos_11/std": 0.005890296772122383,
"rewards/pos_12/mean": -0.0004802196519449353,
"rewards/pos_12/std": 0.004994568601250649,
"rewards/pos_13/mean": 0.0005208333488553763,
"rewards/pos_13/std": 0.003230412770062685,
"rewards/pos_14/mean": 0.00012089894153177738,
"rewards/pos_14/std": 0.0018415355123579502,
"rewards/pos_15/mean": -0.0009114583372138441,
"rewards/pos_15/std": 0.0033854166977107525,
"rewards/pos_16/mean": 0.0,
"rewards/pos_16/std": 0.0,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08470222726464272,
"rewards/pos_2/std": 0.05917537584900856,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.0567516652867198,
"rewards/pos_3/std": 0.057415686547756195,
"rewards/pos_4/mean": 0.033095260383561254,
"rewards/pos_4/std": 0.050696741044521335,
"rewards/pos_5/mean": 0.020724335825070737,
"rewards/pos_5/std": 0.040101373288780454,
"rewards/pos_6/mean": 0.016150066489353777,
"rewards/pos_6/std": 0.0350839720107615,
"rewards/pos_7/mean": 0.005806612188462168,
"rewards/pos_7/std": 0.0280417391564697,
"rewards/pos_8/mean": -0.0006850265664979816,
"rewards/pos_8/std": 0.015849439846351742,
"rewards/pos_9/mean": -0.0016472050338052213,
"rewards/pos_9/std": 0.011492027598433197,
"step": 180
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 320.5,
"completions/max_terminated_length": 320.5,
"completions/mean_length": 266.6770935058594,
"completions/mean_terminated_length": 266.6770935058594,
"completions/min_length": 213.0,
"completions/min_terminated_length": 213.0,
"entropy": 0.2238691563097139,
"epoch": 0.11949685534591195,
"frac_reward_zero_std": 0.1333333358168602,
"grad_norm": 0.27348336577415466,
"items/detected_max": 30.0,
"items/detected_mean": 25.604166666666668,
"items/detected_min": 20.0,
"kl": 0.00398559235521437,
"learning_rate": 1e-06,
"loss": 0.0177,
"num_tokens": 5367136.0,
"reward": 0.019349526677979158,
"reward_std": 0.020760491117835046,
"reward_total": 0.14788751676678658,
"rewards/pos_1/mean": 0.1478875134140253,
"rewards/pos_1/std": 0.07359497845172883,
"rewards/pos_10/mean": -0.00036135842092335226,
"rewards/pos_10/std": 0.009093766659498214,
"rewards/pos_11/mean": 0.00014302353374660016,
"rewards/pos_11/std": 0.006097358372062445,
"rewards/pos_12/mean": 0.0004698589909821749,
"rewards/pos_12/std": 0.005443687411025166,
"rewards/pos_13/mean": -0.00047190774930641057,
"rewards/pos_13/std": 0.004843967990018428,
"rewards/pos_14/mean": 0.0004996092990040779,
"rewards/pos_14/std": 0.003132202010601759,
"rewards/pos_15/mean": -0.0005420573987066745,
"rewards/pos_15/std": 0.003082551993429661,
"rewards/pos_16/mean": 0.00024476498365402224,
"rewards/pos_16/std": 0.001508907275274396,
"rewards/pos_17/mean": 0.00024980464950203897,
"rewards/pos_17/std": 0.0014988278970122337,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.09178031878545881,
"rewards/pos_2/std": 0.05649234391748905,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.0645428148796782,
"rewards/pos_3/std": 0.04721329659223557,
"rewards/pos_4/mean": 0.034123249468393624,
"rewards/pos_4/std": 0.050931732170283794,
"rewards/pos_5/mean": 0.020877907634712756,
"rewards/pos_5/std": 0.04322007759474218,
"rewards/pos_6/mean": 0.014087293948978185,
"rewards/pos_6/std": 0.03900466645136476,
"rewards/pos_7/mean": 0.008370040811132639,
"rewards/pos_7/std": 0.03408770642708987,
"rewards/pos_8/mean": 0.004736725857947021,
"rewards/pos_8/std": 0.02168282950296998,
"rewards/pos_9/mean": 0.00047341859608422964,
"rewards/pos_9/std": 0.014039887744002045,
"step": 190
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 326.2,
"completions/max_terminated_length": 326.2,
"completions/mean_length": 265.7000045776367,
"completions/mean_terminated_length": 265.7000045776367,
"completions/min_length": 209.1,
"completions/min_terminated_length": 209.1,
"entropy": 0.22170269154012204,
"epoch": 0.12578616352201258,
"frac_reward_zero_std": 0.2000000074505806,
"grad_norm": 0.24164925515651703,
"items/detected_max": 31.0,
"items/detected_mean": 25.566666666666666,
"items/detected_min": 18.0,
"kl": 0.004068990563973784,
"learning_rate": 1e-06,
"loss": 0.0167,
"num_tokens": 5619256.0,
"reward": 0.017621663835598156,
"reward_std": 0.018194446712732314,
"reward_total": 0.13456704604905098,
"rewards/pos_1/mean": 0.1345670434413478,
"rewards/pos_1/std": 0.046814880799502134,
"rewards/pos_10/mean": 0.001287172012962401,
"rewards/pos_10/std": 0.010988512635231018,
"rewards/pos_11/mean": 0.0032654690090566873,
"rewards/pos_11/std": 0.007643049582839012,
"rewards/pos_12/mean": 0.0005549672991037369,
"rewards/pos_12/std": 0.0068269922398030754,
"rewards/pos_13/mean": 0.0007194308214820922,
"rewards/pos_13/std": 0.004568580654449761,
"rewards/pos_14/mean": 0.000740341772325337,
"rewards/pos_14/std": 0.004380421992391348,
"rewards/pos_15/mean": 0.000758960610255599,
"rewards/pos_15/std": 0.004496230185031891,
"rewards/pos_16/mean": 0.0006371107883751392,
"rewards/pos_16/std": 0.002659227070398629,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.07656875005923211,
"rewards/pos_2/std": 0.05108541529625654,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04204599694348872,
"rewards/pos_3/std": 0.04441921235993505,
"rewards/pos_4/mean": 0.03253965424373746,
"rewards/pos_4/std": 0.043719809129834176,
"rewards/pos_5/mean": 0.02265897230245173,
"rewards/pos_5/std": 0.039311152510344984,
"rewards/pos_6/mean": 0.016424931492656468,
"rewards/pos_6/std": 0.03444936061277985,
"rewards/pos_7/mean": 0.011618296918459236,
"rewards/pos_7/std": 0.026631731120869518,
"rewards/pos_8/mean": 0.005892342084553093,
"rewards/pos_8/std": 0.01966498149558902,
"rewards/pos_9/mean": 0.002278735424624756,
"rewards/pos_9/std": 0.015979555086232722,
"step": 200
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 328.1,
"completions/max_terminated_length": 328.1,
"completions/mean_length": 265.3000061035156,
"completions/mean_terminated_length": 265.3000061035156,
"completions/min_length": 211.1,
"completions/min_terminated_length": 211.1,
"entropy": 0.22155435737222434,
"epoch": 0.1320754716981132,
"frac_reward_zero_std": 0.20000000149011612,
"grad_norm": 0.2742804288864136,
"items/detected_max": 31.0,
"items/detected_mean": 25.2,
"items/detected_min": 17.0,
"kl": 0.004075667206052458,
"learning_rate": 1e-06,
"loss": 0.0175,
"num_tokens": 5876148.0,
"reward": 0.0260989835485816,
"reward_std": 0.02093745619058609,
"reward_total": 0.2333697572350502,
"rewards/pos_1/mean": 0.23336975127458573,
"rewards/pos_1/std": 0.08852581698447466,
"rewards/pos_10/mean": 0.0008219308452680707,
"rewards/pos_10/std": 0.00651337243616581,
"rewards/pos_11/mean": 0.00041771340183913706,
"rewards/pos_11/std": 0.004837532434612513,
"rewards/pos_12/mean": -0.0018297375878319143,
"rewards/pos_12/std": 0.005276431818492711,
"rewards/pos_13/mean": -0.0008207797771319747,
"rewards/pos_13/std": 0.001410291320644319,
"rewards/pos_14/mean": -0.0007998688146471977,
"rewards/pos_14/std": 0.0015997376292943954,
"rewards/pos_15/mean": -0.0006510416860692203,
"rewards/pos_15/std": 0.001081953290849924,
"rewards/pos_16/mean": -0.0007645329460501671,
"rewards/pos_16/std": 0.0013136462308466434,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.11015560580417513,
"rewards/pos_2/std": 0.060006641410291196,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06867622348072473,
"rewards/pos_3/std": 0.06746885739266872,
"rewards/pos_4/mean": 0.04123723595403135,
"rewards/pos_4/std": 0.04667948558926582,
"rewards/pos_5/mean": 0.030098050978267565,
"rewards/pos_5/std": 0.041850338224321604,
"rewards/pos_6/mean": 0.023275516787543894,
"rewards/pos_6/std": 0.0351828612620011,
"rewards/pos_7/mean": 0.011897218006197363,
"rewards/pos_7/std": 0.025269487011246382,
"rewards/pos_8/mean": 0.0065091788419522345,
"rewards/pos_8/std": 0.018076164787635206,
"rewards/pos_9/mean": 0.0007596121751703322,
"rewards/pos_9/std": 0.012911638477817178,
"step": 210
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 322.8,
"completions/max_terminated_length": 322.8,
"completions/mean_length": 260.1895919799805,
"completions/mean_terminated_length": 260.1895919799805,
"completions/min_length": 204.2,
"completions/min_terminated_length": 204.2,
"entropy": 0.2140471934651335,
"epoch": 0.13836477987421383,
"frac_reward_zero_std": 0.20000000447034835,
"grad_norm": 0.2579440176486969,
"items/detected_max": 30.0,
"items/detected_mean": 24.854166666666668,
"items/detected_min": 18.0,
"kl": 0.005129828681189489,
"learning_rate": 1e-06,
"loss": 0.0144,
"num_tokens": 6144383.0,
"reward": 0.019254540232941507,
"reward_std": 0.01791708227247,
"reward_total": 0.17324519529938698,
"rewards/pos_1/mean": 0.1732451893389225,
"rewards/pos_1/std": 0.06843348871916533,
"rewards/pos_10/mean": 0.0009092353051528334,
"rewards/pos_10/std": 0.009819345409050583,
"rewards/pos_11/mean": 0.0003755505429580808,
"rewards/pos_11/std": 0.006670900527387858,
"rewards/pos_12/mean": 9.85862105153501e-05,
"rewards/pos_12/std": 0.0018861609743908048,
"rewards/pos_13/mean": -2.6353169232606888e-05,
"rewards/pos_13/std": 0.002136039733886719,
"rewards/pos_14/mean": 0.00012089894153177738,
"rewards/pos_14/std": 0.0018415355123579502,
"rewards/pos_15/mean": 0.00026041667442768814,
"rewards/pos_15/std": 0.0015625000465661286,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0002548443386331201,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08983773421496152,
"rewards/pos_2/std": 0.061134057305753234,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.051768475910648704,
"rewards/pos_3/std": 0.048210570914670826,
"rewards/pos_4/mean": 0.03269893038086593,
"rewards/pos_4/std": 0.04562271069735289,
"rewards/pos_5/mean": 0.013085525191854686,
"rewards/pos_5/std": 0.03348905979655683,
"rewards/pos_6/mean": 0.009051869111135601,
"rewards/pos_6/std": 0.023240961600095034,
"rewards/pos_7/mean": 0.007865509623661638,
"rewards/pos_7/std": 0.02395617887377739,
"rewards/pos_8/mean": 0.00524594159796834,
"rewards/pos_8/std": 0.01783889103680849,
"rewards/pos_9/mean": 0.0008032934507355094,
"rewards/pos_9/std": 0.01199915090110153,
"step": 220
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 316.4,
"completions/max_terminated_length": 316.4,
"completions/mean_length": 256.67084045410155,
"completions/mean_terminated_length": 256.67084045410155,
"completions/min_length": 199.4,
"completions/min_terminated_length": 199.4,
"entropy": 0.21115152711669605,
"epoch": 0.14465408805031446,
"frac_reward_zero_std": 0.2250000074505806,
"grad_norm": 0.21807080507278442,
"items/detected_max": 33.0,
"items/detected_mean": 24.458333333333332,
"items/detected_min": 18.0,
"kl": 0.00473162657605523,
"learning_rate": 1e-06,
"loss": 0.0154,
"num_tokens": 6408505.0,
"reward": 0.013173915771767497,
"reward_std": 0.017613324616104365,
"reward_total": 0.11636085771024227,
"rewards/pos_1/mean": 0.116360854357481,
"rewards/pos_1/std": 0.06958029698580503,
"rewards/pos_10/mean": -0.0015289061702787876,
"rewards/pos_10/std": 0.006670845299959182,
"rewards/pos_11/mean": -0.0004223058931529522,
"rewards/pos_11/std": 0.004721084609627724,
"rewards/pos_12/mean": 0.00032677670242264867,
"rewards/pos_12/std": 0.0034685343271121383,
"rewards/pos_13/mean": -6.385107990354299e-05,
"rewards/pos_13/std": 0.004249790078029036,
"rewards/pos_14/mean": 0.00063058752566576,
"rewards/pos_14/std": 0.0028609128668904303,
"rewards/pos_15/mean": 0.0003794802469201386,
"rewards/pos_15/std": 0.0033631272381171584,
"rewards/pos_16/mean": 0.0005096886307001114,
"rewards/pos_16/std": 0.0010193773545324803,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.05666366685181856,
"rewards/pos_2/std": 0.05859119575470686,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.03987346403300762,
"rewards/pos_3/std": 0.05046229064464569,
"rewards/pos_4/mean": 0.02973797037266195,
"rewards/pos_4/std": 0.04398397598415613,
"rewards/pos_5/mean": 0.011401956726331264,
"rewards/pos_5/std": 0.033489659521728755,
"rewards/pos_6/mean": 0.007691465364769101,
"rewards/pos_6/std": 0.02227671444416046,
"rewards/pos_7/mean": 0.001962298701982945,
"rewards/pos_7/std": 0.021468895603902638,
"rewards/pos_8/mean": 0.0008060719235800206,
"rewards/pos_8/std": 0.016375967115163804,
"rewards/pos_9/mean": -0.0008509156294167041,
"rewards/pos_9/std": 0.009683813899755478,
"step": 230
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 344.9,
"completions/max_terminated_length": 344.9,
"completions/mean_length": 262.8958404541016,
"completions/mean_terminated_length": 262.8958404541016,
"completions/min_length": 202.1,
"completions/min_terminated_length": 202.1,
"entropy": 0.22451901606594524,
"epoch": 0.1509433962264151,
"frac_reward_zero_std": 0.20833333879709243,
"grad_norm": 0.3170786201953888,
"items/detected_max": 41.0,
"items/detected_mean": 25.341666666666665,
"items/detected_min": 19.0,
"kl": 0.005417148351746922,
"learning_rate": 1e-06,
"loss": 0.016,
"num_tokens": 6710187.0,
"reward": 0.026339445263147354,
"reward_std": 0.019665919244289398,
"reward_total": 0.1830954909324646,
"rewards/pos_1/mean": 0.1830954894423485,
"rewards/pos_1/std": 0.05954671716317535,
"rewards/pos_10/mean": 0.002383789012674242,
"rewards/pos_10/std": 0.0101193432463333,
"rewards/pos_11/mean": -0.0002576601458713412,
"rewards/pos_11/std": 0.007039384357631207,
"rewards/pos_12/mean": 0.0005050237057730555,
"rewards/pos_12/std": 0.00495102065615356,
"rewards/pos_13/mean": 0.0006576299783773683,
"rewards/pos_13/std": 0.0029568195110186934,
"rewards/pos_14/mean": 0.00012089891824870931,
"rewards/pos_14/std": 0.0016913962550461291,
"rewards/pos_15/mean": 0.00013020832557230816,
"rewards/pos_15/std": 0.0018229166511446238,
"rewards/pos_16/mean": -0.00038226647302508354,
"rewards/pos_16/std": 0.0007645330158993601,
"rewards/pos_17/mean": -0.0003747069742530621,
"rewards/pos_17/std": 0.0007494139485061168,
"rewards/pos_18/mean": -0.00036782644456252825,
"rewards/pos_18/std": 0.0007356528891250491,
"rewards/pos_19/mean": -0.0002410189714282751,
"rewards/pos_19/std": 0.0004820379428565502,
"rewards/pos_2/mean": 0.11684439983218908,
"rewards/pos_2/std": 0.05454435236752033,
"rewards/pos_20/mean": -3.700743525707937e-18,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.07016311129555106,
"rewards/pos_3/std": 0.05461613442748785,
"rewards/pos_4/mean": 0.05105745694600046,
"rewards/pos_4/std": 0.049194585625082256,
"rewards/pos_5/mean": 0.04083440322428942,
"rewards/pos_5/std": 0.04271569908596575,
"rewards/pos_6/mean": 0.02642920259386301,
"rewards/pos_6/std": 0.033980270475149156,
"rewards/pos_7/mean": 0.01962945028208196,
"rewards/pos_7/std": 0.028517392510548235,
"rewards/pos_8/mean": 0.010012811232202998,
"rewards/pos_8/std": 0.022616432514041662,
"rewards/pos_9/mean": 0.006548473122529685,
"rewards/pos_9/std": 0.016274257795885205,
"step": 240
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 335.2,
"completions/max_terminated_length": 335.2,
"completions/mean_length": 265.06250915527346,
"completions/mean_terminated_length": 265.06250915527346,
"completions/min_length": 206.5,
"completions/min_terminated_length": 206.5,
"entropy": 0.23201529448851943,
"epoch": 0.15723270440251572,
"frac_reward_zero_std": 0.15833333879709244,
"grad_norm": 0.18924497067928314,
"items/detected_max": 30.0,
"items/detected_mean": 25.55,
"items/detected_min": 16.0,
"kl": 0.005678520092139176,
"learning_rate": 1e-06,
"loss": 0.0164,
"num_tokens": 6991369.0,
"reward": 0.025188619922846555,
"reward_std": 0.02295760540291667,
"reward_total": 0.189043590426445,
"rewards/pos_1/mean": 0.18904358446598052,
"rewards/pos_1/std": 0.07828422243474051,
"rewards/pos_10/mean": 0.005751247936859727,
"rewards/pos_10/std": 0.014085922576487064,
"rewards/pos_11/mean": 0.003690930129960179,
"rewards/pos_11/std": 0.01120929727330804,
"rewards/pos_12/mean": 0.0018294563284143805,
"rewards/pos_12/std": 0.006877762172371149,
"rewards/pos_13/mean": 0.0014269712381064891,
"rewards/pos_13/std": 0.005435216333717107,
"rewards/pos_14/mean": 0.0004972761031240225,
"rewards/pos_14/std": 0.005260519497096538,
"rewards/pos_15/mean": 0.0010305219795554868,
"rewards/pos_15/std": 0.0020610440522432346,
"rewards/pos_16/mean": 0.00012742215767502693,
"rewards/pos_16/std": 0.0017839103471487777,
"rewards/pos_17/mean": -0.0001249023247510204,
"rewards/pos_17/std": 0.0002498046495020408,
"rewards/pos_18/mean": -0.0001226088148541758,
"rewards/pos_18/std": 0.0002452176297083516,
"rewards/pos_19/mean": -0.00024101897142827602,
"rewards/pos_19/std": 0.0002783047268167157,
"rewards/pos_2/mean": 0.10324889775365591,
"rewards/pos_2/std": 0.06747091393917799,
"rewards/pos_20/mean": -9.251858814269843e-19,
"rewards/pos_20/std": 1.8503717628539686e-18,
"rewards/pos_3/mean": 0.06996003771200776,
"rewards/pos_3/std": 0.05566513775847852,
"rewards/pos_4/mean": 0.0460869993083179,
"rewards/pos_4/std": 0.05539451166987419,
"rewards/pos_5/mean": 0.02949588217306882,
"rewards/pos_5/std": 0.04718504603952169,
"rewards/pos_6/mean": 0.020764206012245268,
"rewards/pos_6/std": 0.03319612480700016,
"rewards/pos_7/mean": 0.012898960616439581,
"rewards/pos_7/std": 0.031006166944280267,
"rewards/pos_8/mean": 0.011339673865586519,
"rewards/pos_8/std": 0.025132232066243886,
"rewards/pos_9/mean": 0.007068848155904561,
"rewards/pos_9/std": 0.01833072875160724,
"step": 250
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 325.5,
"completions/max_terminated_length": 325.5,
"completions/mean_length": 264.6687637329102,
"completions/mean_terminated_length": 264.6687637329102,
"completions/min_length": 212.8,
"completions/min_terminated_length": 212.8,
"entropy": 0.21104721585288644,
"epoch": 0.16352201257861634,
"frac_reward_zero_std": 0.12500000298023223,
"grad_norm": 0.1994742453098297,
"items/detected_max": 31.0,
"items/detected_mean": 25.033333333333335,
"items/detected_min": 19.0,
"kl": 0.005256364249119846,
"learning_rate": 1e-06,
"loss": 0.0156,
"num_tokens": 7230110.0,
"reward": 0.01918751737102866,
"reward_std": 0.018093076907098292,
"reward_total": 0.17405339246615767,
"rewards/pos_1/mean": 0.17405338347889482,
"rewards/pos_1/std": 0.06521443482488394,
"rewards/pos_10/mean": -0.0009455020190216601,
"rewards/pos_10/std": 0.005048538488335907,
"rewards/pos_11/mean": -0.0004358483478426933,
"rewards/pos_11/std": 0.004538286104798317,
"rewards/pos_12/mean": -0.0009852433227933942,
"rewards/pos_12/std": 0.0019704866455867885,
"rewards/pos_13/mean": -0.0004103898885659873,
"rewards/pos_13/std": 0.0008207797771319747,
"rewards/pos_14/mean": -0.0009331803303211927,
"rewards/pos_14/std": 0.0016409843228757381,
"rewards/pos_15/mean": -0.00026041667442768814,
"rewards/pos_15/std": 0.0005208333488553763,
"rewards/pos_16/mean": -0.0005096886307001114,
"rewards/pos_16/std": 0.0010193773545324803,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09146424243226647,
"rewards/pos_2/std": 0.05889338552951813,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.049712089775130155,
"rewards/pos_3/std": 0.0506577804684639,
"rewards/pos_4/mean": 0.03075085999444127,
"rewards/pos_4/std": 0.04471208192408085,
"rewards/pos_5/mean": 0.017505517741665245,
"rewards/pos_5/std": 0.03927737530320883,
"rewards/pos_6/mean": 0.011149797862162813,
"rewards/pos_6/std": 0.03163430448621511,
"rewards/pos_7/mean": 0.007539704500231892,
"rewards/pos_7/std": 0.02429402128327638,
"rewards/pos_8/mean": 0.005642501241527498,
"rewards/pos_8/std": 0.019538127072155475,
"rewards/pos_9/mean": 0.0005351163446903229,
"rewards/pos_9/std": 0.011835501319728792,
"step": 260
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 321.6,
"completions/max_terminated_length": 321.6,
"completions/mean_length": 266.3895889282227,
"completions/mean_terminated_length": 266.3895889282227,
"completions/min_length": 212.2,
"completions/min_terminated_length": 212.2,
"entropy": 0.22741040562589962,
"epoch": 0.16981132075471697,
"frac_reward_zero_std": 0.19166667237877846,
"grad_norm": 0.11161798983812332,
"items/detected_max": 31.0,
"items/detected_mean": 25.679166666666667,
"items/detected_min": 19.0,
"kl": 0.006663059488103803,
"learning_rate": 1e-06,
"loss": 0.0178,
"num_tokens": 7504485.0,
"reward": 0.02319504979532212,
"reward_std": 0.021584799420088528,
"reward_total": 0.2078800647519529,
"rewards/pos_1/mean": 0.20788006028160452,
"rewards/pos_1/std": 0.08386271893978119,
"rewards/pos_10/mean": 0.0002914302225690335,
"rewards/pos_10/std": 0.008567684050649404,
"rewards/pos_11/mean": 0.0008116275537759065,
"rewards/pos_11/std": 0.007888751104474068,
"rewards/pos_12/mean": 0.0014335623593069613,
"rewards/pos_12/std": 0.006815162324346602,
"rewards/pos_13/mean": 0.0010429344838485122,
"rewards/pos_13/std": 0.005427724355831743,
"rewards/pos_14/mean": 0.0003763771615922451,
"rewards/pos_14/std": 0.002538005914539099,
"rewards/pos_15/mean": 0.0003794802934862673,
"rewards/pos_15/std": 0.0012797940289601684,
"rewards/pos_16/mean": 0.00038226647302508354,
"rewards/pos_16/std": 0.0012742216931656003,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08499452217947692,
"rewards/pos_2/std": 0.06867573540657759,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04896511500701308,
"rewards/pos_3/std": 0.05700467079877854,
"rewards/pos_4/mean": 0.03908236059360206,
"rewards/pos_4/std": 0.04660152085125446,
"rewards/pos_5/mean": 0.02731589375374218,
"rewards/pos_5/std": 0.03737078960984945,
"rewards/pos_6/mean": 0.02225884739673347,
"rewards/pos_6/std": 0.03675803560763598,
"rewards/pos_7/mean": 0.01433403476839885,
"rewards/pos_7/std": 0.0255989148048684,
"rewards/pos_8/mean": 0.008673226041719318,
"rewards/pos_8/std": 0.024009060859680176,
"rewards/pos_9/mean": 0.005679246410727501,
"rewards/pos_9/std": 0.018023168295621873,
"step": 270
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 321.7,
"completions/max_terminated_length": 321.7,
"completions/mean_length": 260.93334045410154,
"completions/mean_terminated_length": 260.93334045410154,
"completions/min_length": 206.3,
"completions/min_terminated_length": 206.3,
"entropy": 0.22694507626195748,
"epoch": 0.1761006289308176,
"frac_reward_zero_std": 0.06666666865348816,
"grad_norm": 0.231883704662323,
"items/detected_max": 31.0,
"items/detected_mean": 24.904166666666665,
"items/detected_min": 19.0,
"kl": 0.00740518799527005,
"learning_rate": 1e-06,
"loss": 0.0152,
"num_tokens": 7805425.0,
"reward": 0.02050045975483954,
"reward_std": 0.020080117974430323,
"reward_total": 0.1723884368315339,
"rewards/pos_1/mean": 0.17238843599334358,
"rewards/pos_1/std": 0.08533855974674225,
"rewards/pos_10/mean": 0.0010272695566527546,
"rewards/pos_10/std": 0.011435505677945912,
"rewards/pos_11/mean": 0.00032194284722208977,
"rewards/pos_11/std": 0.0058740658685565,
"rewards/pos_12/mean": 0.0010664916946552694,
"rewards/pos_12/std": 0.004227267880924046,
"rewards/pos_13/mean": 0.0005193052464164793,
"rewards/pos_13/std": 0.0032273565186187623,
"rewards/pos_14/mean": 0.0005332458764314652,
"rewards/pos_14/std": 0.0031994753517210484,
"rewards/pos_15/mean": -0.0005208333488553763,
"rewards/pos_15/std": 0.0010416666977107526,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0002548443386331201,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.0002452176297083497,
"rewards/pos_18/std": 0.0004904352594166994,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09616097155958414,
"rewards/pos_2/std": 0.05851127915084362,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05270993052981794,
"rewards/pos_3/std": 0.05299665480852127,
"rewards/pos_4/mean": 0.02972645848058164,
"rewards/pos_4/std": 0.04188747890293598,
"rewards/pos_5/mean": 0.023145902354735882,
"rewards/pos_5/std": 0.0399086520075798,
"rewards/pos_6/mean": 0.014170866855420173,
"rewards/pos_6/std": 0.02930837543681264,
"rewards/pos_7/mean": 0.008681701379828155,
"rewards/pos_7/std": 0.022060082969255747,
"rewards/pos_8/mean": 0.006511726812459528,
"rewards/pos_8/std": 0.02124700667336583,
"rewards/pos_9/mean": 0.004063288937322796,
"rewards/pos_9/std": 0.02034384049475193,
"step": 280
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 318.9,
"completions/max_terminated_length": 318.9,
"completions/mean_length": 259.00834503173826,
"completions/mean_terminated_length": 259.00834503173826,
"completions/min_length": 205.2,
"completions/min_terminated_length": 205.2,
"entropy": 0.21961077268545826,
"epoch": 0.18238993710691823,
"frac_reward_zero_std": 0.2250000037252903,
"grad_norm": 0.3105533719062805,
"items/detected_max": 30.0,
"items/detected_mean": 24.458333333333332,
"items/detected_min": 18.0,
"kl": 0.005909793743679378,
"learning_rate": 1e-06,
"loss": 0.0153,
"num_tokens": 8051641.0,
"reward": 0.01701394822448492,
"reward_std": 0.01878683748655021,
"reward_total": 0.12582458220422268,
"rewards/pos_1/mean": 0.12582458183169365,
"rewards/pos_1/std": 0.06831286940723658,
"rewards/pos_10/mean": 0.0021640648017637433,
"rewards/pos_10/std": 0.008456035633571446,
"rewards/pos_11/mean": -0.00034264513524249194,
"rewards/pos_11/std": 0.004963404871523381,
"rewards/pos_12/mean": -0.00015655870083719492,
"rewards/pos_12/std": 0.004190851980820298,
"rewards/pos_13/mean": -0.0004103898885659873,
"rewards/pos_13/std": 0.002778257662430406,
"rewards/pos_14/mean": -0.0006665573455393314,
"rewards/pos_14/std": 0.0013331146910786629,
"rewards/pos_15/mean": -0.00026041667442768814,
"rewards/pos_15/std": 0.0005208333488553763,
"rewards/pos_16/mean": -0.00038226647302508354,
"rewards/pos_16/std": 0.0007645330158993601,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.080518504884094,
"rewards/pos_2/std": 0.06397041976451874,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05210294816643,
"rewards/pos_3/std": 0.05435645366087556,
"rewards/pos_4/mean": 0.03605008292943239,
"rewards/pos_4/std": 0.04513352606445551,
"rewards/pos_5/mean": 0.023162062186747788,
"rewards/pos_5/std": 0.04228055775165558,
"rewards/pos_6/mean": 0.011037050012964756,
"rewards/pos_6/std": 0.026784436777234077,
"rewards/pos_7/mean": 0.0049487472046166655,
"rewards/pos_7/std": 0.022359331441111862,
"rewards/pos_8/mean": 0.003727375087328255,
"rewards/pos_8/std": 0.013905423972755671,
"rewards/pos_9/mean": 0.0032054831506684424,
"rewards/pos_9/std": 0.015140429511666298,
"step": 290
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 329.0,
"completions/max_terminated_length": 329.0,
"completions/mean_length": 262.5062622070312,
"completions/mean_terminated_length": 262.5062622070312,
"completions/min_length": 210.3,
"completions/min_terminated_length": 210.3,
"entropy": 0.2162858149347206,
"epoch": 0.18867924528301888,
"frac_reward_zero_std": 0.17500000223517417,
"grad_norm": 0.2965948283672333,
"items/detected_max": 30.0,
"items/detected_mean": 24.929166666666667,
"items/detected_min": 18.0,
"kl": 0.006186072509444784,
"learning_rate": 1e-06,
"loss": 0.0166,
"num_tokens": 8319180.0,
"reward": 0.021316327748354524,
"reward_std": 0.0197633249219507,
"reward_total": 0.17199704954400657,
"rewards/pos_1/mean": 0.17199704800732435,
"rewards/pos_1/std": 0.07101438688114285,
"rewards/pos_10/mean": 0.0023221148061566056,
"rewards/pos_10/std": 0.006007013586349785,
"rewards/pos_11/mean": 0.0007187323644757271,
"rewards/pos_11/std": 0.0063957332633435724,
"rewards/pos_12/mean": 0.0003055526991374791,
"rewards/pos_12/std": 0.0025730435503646732,
"rewards/pos_13/mean": 0.0004621113766916096,
"rewards/pos_13/std": 0.0022599261021241544,
"rewards/pos_14/mean": 0.0003427405841648579,
"rewards/pos_14/std": 0.0020978176034986973,
"rewards/pos_15/mean": 0.0007488812552765012,
"rewards/pos_15/std": 0.0016863864380866288,
"rewards/pos_16/mean": 0.0003721871413290501,
"rewards/pos_16/std": 0.0020621417788788676,
"rewards/pos_17/mean": 0.0002498046262189746,
"rewards/pos_17/std": 0.0014988278970122337,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.0910015320405364,
"rewards/pos_2/std": 0.07098599858582019,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05390574834309518,
"rewards/pos_3/std": 0.05022534327581525,
"rewards/pos_4/mean": 0.035313252126798034,
"rewards/pos_4/std": 0.045174385979771615,
"rewards/pos_5/mean": 0.02623851306270808,
"rewards/pos_5/std": 0.03764389446005225,
"rewards/pos_6/mean": 0.017836013311170972,
"rewards/pos_6/std": 0.03342308038845658,
"rewards/pos_7/mean": 0.015512678562663496,
"rewards/pos_7/std": 0.02850998127833009,
"rewards/pos_8/mean": 0.006407566787675023,
"rewards/pos_8/std": 0.021089862566441298,
"rewards/pos_9/mean": 0.002592064579948783,
"rewards/pos_9/std": 0.012618646910414099,
"step": 300
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 315.3,
"completions/max_terminated_length": 315.3,
"completions/mean_length": 258.1125122070313,
"completions/mean_terminated_length": 258.1125122070313,
"completions/min_length": 201.0,
"completions/min_terminated_length": 201.0,
"entropy": 0.22230708294858534,
"epoch": 0.1949685534591195,
"frac_reward_zero_std": 0.16666667014360428,
"grad_norm": 0.21009129285812378,
"items/detected_max": 30.0,
"items/detected_mean": 24.758333333333333,
"items/detected_min": 18.0,
"kl": 0.00682944758979526,
"learning_rate": 1e-06,
"loss": 0.0156,
"num_tokens": 8583422.0,
"reward": 0.020729794539511202,
"reward_std": 0.019597215484827755,
"reward_total": 0.13572603426873683,
"rewards/pos_1/mean": 0.1357260297983885,
"rewards/pos_1/std": 0.06282865833491087,
"rewards/pos_10/mean": 0.002956245210953057,
"rewards/pos_10/std": 0.012046903045848012,
"rewards/pos_11/mean": 0.0010622974252328277,
"rewards/pos_11/std": 0.007764216605573892,
"rewards/pos_12/mean": 0.0009396833018399775,
"rewards/pos_12/std": 0.003768626763485372,
"rewards/pos_13/mean": 0.0003964492352679372,
"rewards/pos_13/std": 0.003297562850639224,
"rewards/pos_14/mean": 0.0,
"rewards/pos_14/std": 0.0021329835057258607,
"rewards/pos_15/mean": 0.0,
"rewards/pos_15/std": 0.0,
"rewards/pos_16/mean": -0.00038226647302508354,
"rewards/pos_16/std": 0.0007645330158993601,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09809606708586216,
"rewards/pos_2/std": 0.05868534594774246,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06796926595270633,
"rewards/pos_3/std": 0.05614562537521124,
"rewards/pos_4/mean": 0.04019765388220549,
"rewards/pos_4/std": 0.050240844232030214,
"rewards/pos_5/mean": 0.02712703449651599,
"rewards/pos_5/std": 0.038400023430585864,
"rewards/pos_6/mean": 0.016242858988698573,
"rewards/pos_6/std": 0.031802404299378396,
"rewards/pos_7/mean": 0.011441315151751041,
"rewards/pos_7/std": 0.028564795944839717,
"rewards/pos_8/mean": 0.009426339023048059,
"rewards/pos_8/std": 0.020744784269481897,
"rewards/pos_9/mean": 0.0036466998630203308,
"rewards/pos_9/std": 0.01425737168174237,
"step": 310
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 327.3,
"completions/max_terminated_length": 327.3,
"completions/mean_length": 265.0479278564453,
"completions/mean_terminated_length": 265.0479278564453,
"completions/min_length": 207.5,
"completions/min_terminated_length": 207.5,
"entropy": 0.23266545981168746,
"epoch": 0.20125786163522014,
"frac_reward_zero_std": 0.21666666865348816,
"grad_norm": 0.3723245859146118,
"items/detected_max": 30.0,
"items/detected_mean": 25.533333333333335,
"items/detected_min": 18.0,
"kl": 0.015437694112673246,
"learning_rate": 1e-06,
"loss": 0.018,
"num_tokens": 8874261.0,
"reward": 0.01982590677216649,
"reward_std": 0.022089423984289168,
"reward_total": 0.15025124587118627,
"rewards/pos_1/mean": 0.15025124410167337,
"rewards/pos_1/std": 0.07577884886413813,
"rewards/pos_10/mean": 0.001985669624991715,
"rewards/pos_10/std": 0.011629964830353856,
"rewards/pos_11/mean": 0.0015761803835630417,
"rewards/pos_11/std": 0.007801409997045994,
"rewards/pos_12/mean": 0.00045018800301477313,
"rewards/pos_12/std": 0.004841349297203123,
"rewards/pos_13/mean": 0.0010289938771165907,
"rewards/pos_13/std": 0.004246733780018985,
"rewards/pos_14/mean": 0.0003763771615922451,
"rewards/pos_14/std": 0.005101214628666639,
"rewards/pos_15/mean": -0.0002715613809414208,
"rewards/pos_15/std": 0.0023617473198100924,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0022935990244150163,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.07865543810185045,
"rewards/pos_2/std": 0.061682610772550106,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05971371060004458,
"rewards/pos_3/std": 0.057435553334653375,
"rewards/pos_4/mean": 0.03755526882596314,
"rewards/pos_4/std": 0.05022326456382871,
"rewards/pos_5/mean": 0.024758547963574528,
"rewards/pos_5/std": 0.05079233031719923,
"rewards/pos_6/mean": 0.021620591334067286,
"rewards/pos_6/std": 0.04191543236374855,
"rewards/pos_7/mean": 0.012061798262099426,
"rewards/pos_7/std": 0.028514722548425197,
"rewards/pos_8/mean": 0.005660012143198401,
"rewards/pos_8/std": 0.020607514400035144,
"rewards/pos_9/mean": 0.0015910826507024467,
"rewards/pos_9/std": 0.0158261253265664,
"step": 320
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 322.7,
"completions/max_terminated_length": 322.7,
"completions/mean_length": 260.2916717529297,
"completions/mean_terminated_length": 260.2916717529297,
"completions/min_length": 203.1,
"completions/min_terminated_length": 203.1,
"entropy": 0.2289639735283951,
"epoch": 0.20754716981132076,
"frac_reward_zero_std": 0.1666666693985462,
"grad_norm": 0.36900240182876587,
"items/detected_max": 30.0,
"items/detected_mean": 24.883333333333333,
"items/detected_min": 19.0,
"kl": 0.007082089926310194,
"learning_rate": 1e-06,
"loss": 0.0139,
"num_tokens": 9170361.0,
"reward": 0.015596002503298224,
"reward_std": 0.018469544406980277,
"reward_total": 0.1345834530889988,
"rewards/pos_1/mean": 0.13458345010876654,
"rewards/pos_1/std": 0.05662547051906586,
"rewards/pos_10/mean": 0.0029571169754490255,
"rewards/pos_10/std": 0.011131436098366976,
"rewards/pos_11/mean": 0.0010631690733134747,
"rewards/pos_11/std": 0.007871837262064219,
"rewards/pos_12/mean": -8.095824159681797e-05,
"rewards/pos_12/std": 0.0019351853523403406,
"rewards/pos_13/mean": -0.00047553846379742024,
"rewards/pos_13/std": 0.002647960395552218,
"rewards/pos_14/mean": 8.210353553295135e-05,
"rewards/pos_14/std": 0.0016157859936356544,
"rewards/pos_15/mean": -0.0001690037315711379,
"rewards/pos_15/std": 0.002266159188002348,
"rewards/pos_16/mean": -0.0002824950031936169,
"rewards/pos_16/std": 0.0024931418942287565,
"rewards/pos_17/mean": 0.00023223331663757563,
"rewards/pos_17/std": 0.0014636851847171783,
"rewards/pos_18/mean": 0.0002368203247897327,
"rewards/pos_18/std": 0.0009640759089961648,
"rewards/pos_19/mean": 0.0004820379428565502,
"rewards/pos_19/std": 0.0009640758857131004,
"rewards/pos_2/mean": 0.06383709013462066,
"rewards/pos_2/std": 0.05119516551494598,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04193787656258792,
"rewards/pos_3/std": 0.05090108048170805,
"rewards/pos_4/mean": 0.02564535040874034,
"rewards/pos_4/std": 0.04551187176257372,
"rewards/pos_5/mean": 0.01711053295293823,
"rewards/pos_5/std": 0.037988297548145054,
"rewards/pos_6/mean": 0.010198239896756908,
"rewards/pos_6/std": 0.03463809639215469,
"rewards/pos_7/mean": 0.006971109099686146,
"rewards/pos_7/std": 0.02396038032602519,
"rewards/pos_8/mean": 0.004388767573982477,
"rewards/pos_8/std": 0.01984125506132841,
"rewards/pos_9/mean": 0.003202139458153397,
"rewards/pos_9/std": 0.015375914028845727,
"step": 330
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 316.0,
"completions/max_terminated_length": 316.0,
"completions/mean_length": 257.50000915527346,
"completions/mean_terminated_length": 257.50000915527346,
"completions/min_length": 207.4,
"completions/min_terminated_length": 207.4,
"entropy": 0.22065137525399525,
"epoch": 0.2138364779874214,
"frac_reward_zero_std": 0.1916666716337204,
"grad_norm": 0.28421109914779663,
"items/detected_max": 30.0,
"items/detected_mean": 24.35,
"items/detected_min": 19.0,
"kl": 0.007017532674944959,
"learning_rate": 1e-06,
"loss": 0.0149,
"num_tokens": 9446513.0,
"reward": 0.019425443094223737,
"reward_std": 0.018430884974077345,
"reward_total": 0.1281277360394597,
"rewards/pos_1/mean": 0.12812773007899522,
"rewards/pos_1/std": 0.05213469823356718,
"rewards/pos_10/mean": 0.002732968854252249,
"rewards/pos_10/std": 0.011932172696106136,
"rewards/pos_11/mean": 0.002629860525485128,
"rewards/pos_11/std": 0.009663155302405357,
"rewards/pos_12/mean": 0.0012087689246982336,
"rewards/pos_12/std": 0.005961433122865856,
"rewards/pos_13/mean": 0.0003840367076918483,
"rewards/pos_13/std": 0.002956819487735629,
"rewards/pos_14/mean": -1.241255085915327e-05,
"rewards/pos_14/std": 0.0021081584505736827,
"rewards/pos_15/mean": 0.00026041667442768814,
"rewards/pos_15/std": 0.0015625000465661286,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09696470415219664,
"rewards/pos_2/std": 0.05546027736272663,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.060085851489566265,
"rewards/pos_3/std": 0.04964439510367811,
"rewards/pos_4/mean": 0.03614059896208346,
"rewards/pos_4/std": 0.048409922327846286,
"rewards/pos_5/mean": 0.022982617770321668,
"rewards/pos_5/std": 0.039155343035236004,
"rewards/pos_6/mean": 0.014983089175075293,
"rewards/pos_6/std": 0.030907634552568196,
"rewards/pos_7/mean": 0.01095429282868281,
"rewards/pos_7/std": 0.021905113221146168,
"rewards/pos_8/mean": 0.007376816321630031,
"rewards/pos_8/std": 0.020483076642267405,
"rewards/pos_9/mean": 0.004069264605641365,
"rewards/pos_9/std": 0.015573488874360919,
"step": 340
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 331.2,
"completions/max_terminated_length": 331.2,
"completions/mean_length": 259.6979278564453,
"completions/mean_terminated_length": 259.6979278564453,
"completions/min_length": 198.2,
"completions/min_terminated_length": 198.2,
"entropy": 0.22916316849490007,
"epoch": 0.22012578616352202,
"frac_reward_zero_std": 0.20000000447034835,
"grad_norm": 0.3008638918399811,
"items/detected_max": 34.0,
"items/detected_mean": 25.133333333333333,
"items/detected_min": 19.0,
"kl": 0.007831003968021832,
"learning_rate": 1e-06,
"loss": 0.0162,
"num_tokens": 9722756.0,
"reward": 0.02412949312129058,
"reward_std": 0.01957730036228895,
"reward_total": 0.19526247452013196,
"rewards/pos_1/mean": 0.1952624715398997,
"rewards/pos_1/std": 0.07063707131892442,
"rewards/pos_10/mean": 0.004335769626777619,
"rewards/pos_10/std": 0.007979208161123097,
"rewards/pos_11/mean": 0.002140712761320174,
"rewards/pos_11/std": 0.00789201883599162,
"rewards/pos_12/mean": 0.002158847451210022,
"rewards/pos_12/std": 0.004358783643692732,
"rewards/pos_13/mean": -6.151784909889103e-05,
"rewards/pos_13/std": 0.004023188166320324,
"rewards/pos_14/mean": 0.0004996092990040779,
"rewards/pos_14/std": 0.002981483284384012,
"rewards/pos_15/mean": -0.0001514323754236102,
"rewards/pos_15/std": 0.002301301993429661,
"rewards/pos_16/mean": 0.00011734282597899436,
"rewards/pos_16/std": 0.0012540629599243403,
"rewards/pos_17/mean": 0.0003747069742530584,
"rewards/pos_17/std": 0.0012490232475101948,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09324139822274446,
"rewards/pos_2/std": 0.06644056802615524,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.0615955856628716,
"rewards/pos_3/std": 0.051313380664214495,
"rewards/pos_4/mean": 0.04688858595909551,
"rewards/pos_4/std": 0.04371719118207693,
"rewards/pos_5/mean": 0.03025575850624591,
"rewards/pos_5/std": 0.04001493654213846,
"rewards/pos_6/mean": 0.02127471216954291,
"rewards/pos_6/std": 0.030604945309460164,
"rewards/pos_7/mean": 0.01315246180165559,
"rewards/pos_7/std": 0.023588302149437367,
"rewards/pos_8/mean": 0.006093456991948187,
"rewards/pos_8/std": 0.019216742552816867,
"rewards/pos_9/mean": 0.005533974547870457,
"rewards/pos_9/std": 0.0137285475153476,
"step": 350
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 317.6,
"completions/max_terminated_length": 317.6,
"completions/mean_length": 260.34375762939453,
"completions/mean_terminated_length": 260.34375762939453,
"completions/min_length": 213.4,
"completions/min_terminated_length": 213.4,
"entropy": 0.21896101338788868,
"epoch": 0.22641509433962265,
"frac_reward_zero_std": 0.11666666939854622,
"grad_norm": 0.27910512685775757,
"items/detected_max": 30.0,
"items/detected_mean": 24.866666666666667,
"items/detected_min": 17.0,
"kl": 0.006924989688074371,
"learning_rate": 1e-06,
"loss": 0.0149,
"num_tokens": 9980249.0,
"reward": 0.028933631628751753,
"reward_std": 0.019675091747194527,
"reward_total": 0.22213271036744117,
"rewards/pos_1/mean": 0.22213271036744117,
"rewards/pos_1/std": 0.07323975777253508,
"rewards/pos_10/mean": 0.0012367580784484744,
"rewards/pos_10/std": 0.007548078242689371,
"rewards/pos_11/mean": 0.0018706073053181172,
"rewards/pos_11/std": 0.004126033373177051,
"rewards/pos_12/mean": 0.00045404935954138634,
"rewards/pos_12/std": 0.004291756404563784,
"rewards/pos_13/mean": 0.0011696518515236676,
"rewards/pos_13/std": 0.003098501614294946,
"rewards/pos_14/mean": 0.0006329207681119442,
"rewards/pos_14/std": 0.002865579165518284,
"rewards/pos_15/mean": 0.00010898428736254573,
"rewards/pos_15/std": 0.0017804686212912202,
"rewards/pos_16/mean": 0.00011734282597899436,
"rewards/pos_16/std": 0.001763751613907516,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0017872774740681052,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 1.5108222277709536e-18,
"rewards/pos_2/mean": 0.1439829993993044,
"rewards/pos_2/std": 0.06127358330413699,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 1.5108222277709536e-18,
"rewards/pos_3/mean": 0.08225387446582318,
"rewards/pos_3/std": 0.05610301522538066,
"rewards/pos_4/mean": 0.04696628153324127,
"rewards/pos_4/std": 0.04185368986800313,
"rewards/pos_5/mean": 0.030603414541110395,
"rewards/pos_5/std": 0.04589109029620886,
"rewards/pos_6/mean": 0.022831282811239362,
"rewards/pos_6/std": 0.02951878057792783,
"rewards/pos_7/mean": 0.014235854789149016,
"rewards/pos_7/std": 0.024564327974803745,
"rewards/pos_8/mean": 0.007167542690876871,
"rewards/pos_8/std": 0.019931788416579367,
"rewards/pos_9/mean": 0.0030309490975923836,
"rewards/pos_9/std": 0.013619128172285855,
"step": 360
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 315.6,
"completions/max_terminated_length": 315.6,
"completions/mean_length": 257.2437576293945,
"completions/mean_terminated_length": 257.2437576293945,
"completions/min_length": 209.5,
"completions/min_terminated_length": 209.5,
"entropy": 0.225166599607716,
"epoch": 0.23270440251572327,
"frac_reward_zero_std": 0.1666666693985462,
"grad_norm": 0.1677815020084381,
"items/detected_max": 31.0,
"items/detected_mean": 24.379166666666666,
"items/detected_min": 19.0,
"kl": 0.007016470714976701,
"learning_rate": 1e-06,
"loss": 0.014,
"num_tokens": 10236114.0,
"reward": 0.017631425987929106,
"reward_std": 0.022127764578908683,
"reward_total": 0.14379139244556427,
"rewards/pos_1/mean": 0.14379138946533204,
"rewards/pos_1/std": 0.0882169634103775,
"rewards/pos_10/mean": 0.0022585043567232787,
"rewards/pos_10/std": 0.009334756038151681,
"rewards/pos_11/mean": 0.001532090362161398,
"rewards/pos_11/std": 0.005388705246150494,
"rewards/pos_12/mean": 0.0005332458647899329,
"rewards/pos_12/std": 0.0033184764208272098,
"rewards/pos_13/mean": 0.0002596526173874736,
"rewards/pos_13/std": 0.0016136782709509135,
"rewards/pos_14/mean": 0.0001333114691078663,
"rewards/pos_14/std": 0.001866360567510128,
"rewards/pos_15/mean": -0.0005208333488553763,
"rewards/pos_15/std": 0.0010416666977107526,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08060946250334382,
"rewards/pos_2/std": 0.06794189997017383,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.054246106650680304,
"rewards/pos_3/std": 0.06542806439101696,
"rewards/pos_4/mean": 0.03207998964935541,
"rewards/pos_4/std": 0.05759974382817745,
"rewards/pos_5/mean": 0.017739857686683534,
"rewards/pos_5/std": 0.049878083541989325,
"rewards/pos_6/mean": 0.010747757053468376,
"rewards/pos_6/std": 0.03253991464152932,
"rewards/pos_7/mean": 0.0061981274979189035,
"rewards/pos_7/std": 0.023129416280426085,
"rewards/pos_8/mean": 0.0031042592250742017,
"rewards/pos_8/std": 0.019796630647033454,
"rewards/pos_9/mean": 0.00017042522376868873,
"rewards/pos_9/std": 0.0149512107251212,
"step": 370
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 325.4,
"completions/max_terminated_length": 325.4,
"completions/mean_length": 256.8833480834961,
"completions/mean_terminated_length": 256.8833480834961,
"completions/min_length": 202.0,
"completions/min_terminated_length": 202.0,
"entropy": 0.22051387261599303,
"epoch": 0.2389937106918239,
"frac_reward_zero_std": 0.19166667088866235,
"grad_norm": 0.40871769189834595,
"items/detected_max": 30.0,
"items/detected_mean": 24.458333333333332,
"items/detected_min": 17.0,
"kl": 0.007386551384115592,
"learning_rate": 1e-06,
"loss": 0.0145,
"num_tokens": 10495002.0,
"reward": 0.023296015989035367,
"reward_std": 0.021536412183195353,
"reward_total": 0.18946771025657655,
"rewards/pos_1/mean": 0.18946770280599595,
"rewards/pos_1/std": 0.0835272241383791,
"rewards/pos_10/mean": 0.002924169832840562,
"rewards/pos_10/std": 0.007994202524423599,
"rewards/pos_11/mean": 0.0019019185565412045,
"rewards/pos_11/std": 0.006062317825853825,
"rewards/pos_12/mean": 0.0017702369252219797,
"rewards/pos_12/std": 0.004168829647824168,
"rewards/pos_13/mean": 0.0006561018875800073,
"rewards/pos_13/std": 0.001643702038563788,
"rewards/pos_14/mean": 0.0007998688146471977,
"rewards/pos_14/std": 0.0017647244036197661,
"rewards/pos_15/mean": -0.00039062501164153216,
"rewards/pos_15/std": 0.0005611199419945478,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": -0.0003747069742530584,
"rewards/pos_17/std": 0.0007494139485061168,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09561244631186128,
"rewards/pos_2/std": 0.06435443423688411,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06603119405917823,
"rewards/pos_3/std": 0.06552036330103875,
"rewards/pos_4/mean": 0.04784354259027168,
"rewards/pos_4/std": 0.05719647705554962,
"rewards/pos_5/mean": 0.024211431900039316,
"rewards/pos_5/std": 0.042845807410776614,
"rewards/pos_6/mean": 0.017558457027189434,
"rewards/pos_6/std": 0.035367506369948384,
"rewards/pos_7/mean": 0.010088088770862669,
"rewards/pos_7/std": 0.028223303635604678,
"rewards/pos_8/mean": 0.005037966254167259,
"rewards/pos_8/std": 0.01750833559781313,
"rewards/pos_9/mean": 0.0030373329878784715,
"rewards/pos_9/std": 0.01273077039513737,
"step": 380
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 326.7,
"completions/max_terminated_length": 326.7,
"completions/mean_length": 260.4479248046875,
"completions/mean_terminated_length": 260.4479248046875,
"completions/min_length": 209.6,
"completions/min_terminated_length": 209.6,
"entropy": 0.23070885874330999,
"epoch": 0.24528301886792453,
"frac_reward_zero_std": 0.14166666939854622,
"grad_norm": 0.1970636546611786,
"items/detected_max": 30.0,
"items/detected_mean": 24.816666666666666,
"items/detected_min": 18.0,
"kl": 0.008075704663254631,
"learning_rate": 1e-06,
"loss": 0.015,
"num_tokens": 10768141.0,
"reward": 0.0202477783896029,
"reward_std": 0.020860963594168426,
"reward_total": 0.15246383249759674,
"rewards/pos_1/mean": 0.15246383231133223,
"rewards/pos_1/std": 0.07292928658425808,
"rewards/pos_10/mean": -0.0004727510269731283,
"rewards/pos_10/std": 0.007833246374502778,
"rewards/pos_11/mean": -0.0005811311304569245,
"rewards/pos_11/std": 0.00324117187410593,
"rewards/pos_12/mean": -0.0005629961844533682,
"rewards/pos_12/std": 0.0011259923689067365,
"rewards/pos_13/mean": -0.0009575764182955026,
"rewards/pos_13/std": 0.0019151528365910053,
"rewards/pos_14/mean": -0.0007998688146471972,
"rewards/pos_14/std": 0.0015997376292943965,
"rewards/pos_15/mean": -0.00026041667442768766,
"rewards/pos_15/std": 0.0005208333488553772,
"rewards/pos_16/mean": -0.00038226647302508305,
"rewards/pos_16/std": 0.0007645330158993611,
"rewards/pos_17/mean": -0.0004996092990040775,
"rewards/pos_17/std": 0.0009992185980081567,
"rewards/pos_18/mean": -0.00012260881485417439,
"rewards/pos_18/std": 0.0002452176297083506,
"rewards/pos_19/mean": 4.625929407134921e-19,
"rewards/pos_19/std": 9.251858814269843e-19,
"rewards/pos_2/mean": 0.10268718618899583,
"rewards/pos_2/std": 0.07012189291417599,
"rewards/pos_20/mean": 4.625929407134921e-19,
"rewards/pos_20/std": 9.251858814269843e-19,
"rewards/pos_3/mean": 0.06723444974049926,
"rewards/pos_3/std": 0.06676209792494774,
"rewards/pos_4/mean": 0.03524471241980791,
"rewards/pos_4/std": 0.055582412891089915,
"rewards/pos_5/mean": 0.022470815673295876,
"rewards/pos_5/std": 0.04258656706660986,
"rewards/pos_6/mean": 0.01445532594419395,
"rewards/pos_6/std": 0.031042478699237108,
"rewards/pos_7/mean": 0.007936406962107867,
"rewards/pos_7/std": 0.026095511973835528,
"rewards/pos_8/mean": 0.004532542885863222,
"rewards/pos_8/std": 0.01912364289164543,
"rewards/pos_9/mean": 0.002569500356912613,
"rewards/pos_9/std": 0.014730246970430017,
"step": 390
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 328.5,
"completions/max_terminated_length": 328.5,
"completions/mean_length": 256.43959350585936,
"completions/mean_terminated_length": 256.43959350585936,
"completions/min_length": 206.3,
"completions/min_terminated_length": 206.3,
"entropy": 0.21603431180119514,
"epoch": 0.25157232704402516,
"frac_reward_zero_std": 0.15833333730697632,
"grad_norm": 0.30788999795913696,
"items/detected_max": 30.0,
"items/detected_mean": 24.45,
"items/detected_min": 19.0,
"kl": 0.008252161745622289,
"learning_rate": 1e-06,
"loss": 0.0147,
"num_tokens": 11031828.0,
"reward": 0.016329724296701896,
"reward_std": 0.016512975608929992,
"reward_total": 0.14265724457800388,
"rewards/pos_1/mean": 0.14265724280849099,
"rewards/pos_1/std": 0.0651577839627862,
"rewards/pos_10/mean": 0.00025893463753163816,
"rewards/pos_10/std": 0.005557593540288508,
"rewards/pos_11/mean": 0.0004358483478426933,
"rewards/pos_11/std": 0.002794892713427544,
"rewards/pos_12/mean": -0.0007037452305667102,
"rewards/pos_12/std": 0.0014074904611334205,
"rewards/pos_13/mean": -0.0010943730361759663,
"rewards/pos_13/std": 0.0021887460723519327,
"rewards/pos_14/mean": -0.000933180283755064,
"rewards/pos_14/std": 0.0016409843228757381,
"rewards/pos_15/mean": -0.00013020833721384407,
"rewards/pos_15/std": 0.00026041667442768814,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0002548443386331201,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08800087512936443,
"rewards/pos_2/std": 0.05617033746093512,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.03881473926594481,
"rewards/pos_3/std": 0.04208221547305584,
"rewards/pos_4/mean": 0.023159040114842357,
"rewards/pos_4/std": 0.03938644733279943,
"rewards/pos_5/mean": 0.012225279369158671,
"rewards/pos_5/std": 0.03276767339557409,
"rewards/pos_6/mean": 0.009199051139876246,
"rewards/pos_6/std": 0.025476095825433732,
"rewards/pos_7/mean": 0.007665468950290233,
"rewards/pos_7/std": 0.023648051917552947,
"rewards/pos_8/mean": 0.004900209908373654,
"rewards/pos_8/std": 0.017847701627761125,
"rewards/pos_9/mean": 0.002391626697499305,
"rewards/pos_9/std": 0.013368407892994583,
"step": 400
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 316.3,
"completions/max_terminated_length": 316.3,
"completions/mean_length": 256.9125091552734,
"completions/mean_terminated_length": 256.9125091552734,
"completions/min_length": 198.5,
"completions/min_terminated_length": 198.5,
"entropy": 0.22275403228898843,
"epoch": 0.2578616352201258,
"frac_reward_zero_std": 0.18333333507180213,
"grad_norm": 0.2904108762741089,
"items/detected_max": 30.0,
"items/detected_mean": 24.691666666666666,
"items/detected_min": 18.0,
"kl": 0.007832498365799741,
"learning_rate": 1e-06,
"loss": 0.0133,
"num_tokens": 11296478.0,
"reward": 0.028976143896579744,
"reward_std": 0.022660434059798718,
"reward_total": 0.2048182100057602,
"rewards/pos_1/mean": 0.20481819957494735,
"rewards/pos_1/std": 0.07572797238826752,
"rewards/pos_10/mean": 0.0031875847955234347,
"rewards/pos_10/std": 0.009897435712628067,
"rewards/pos_11/mean": 0.0017400289420038463,
"rewards/pos_11/std": 0.008549216762185096,
"rewards/pos_12/mean": 0.0019125141203403472,
"rewards/pos_12/std": 0.006288737198337913,
"rewards/pos_13/mean": 0.0005208333488553763,
"rewards/pos_13/std": 0.0029991445364430545,
"rewards/pos_14/mean": 0.00012089891824871302,
"rewards/pos_14/std": 0.0018415355123579502,
"rewards/pos_15/mean": -1.1641532182693482e-11,
"rewards/pos_15/std": 0.001936122798360884,
"rewards/pos_16/mean": -0.0006371107883751392,
"rewards/pos_16/std": 0.0012742216931656003,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.11721294932067394,
"rewards/pos_2/std": 0.06804916113615037,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.08078673807904124,
"rewards/pos_3/std": 0.055386250652372834,
"rewards/pos_4/mean": 0.06849569948390126,
"rewards/pos_4/std": 0.05045593734830618,
"rewards/pos_5/mean": 0.0406662215013057,
"rewards/pos_5/std": 0.04149941392242908,
"rewards/pos_6/mean": 0.02355194711126387,
"rewards/pos_6/std": 0.040676692873239516,
"rewards/pos_7/mean": 0.019361452711746097,
"rewards/pos_7/std": 0.03839451922103763,
"rewards/pos_8/mean": 0.011890080338343979,
"rewards/pos_8/std": 0.030131167080253363,
"rewards/pos_9/mean": 0.006267227104399353,
"rewards/pos_9/std": 0.01935631341766566,
"step": 410
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 314.6,
"completions/max_terminated_length": 314.6,
"completions/mean_length": 254.96250762939454,
"completions/mean_terminated_length": 254.96250762939454,
"completions/min_length": 202.4,
"completions/min_terminated_length": 202.4,
"entropy": 0.21695980379978816,
"epoch": 0.2641509433962264,
"frac_reward_zero_std": 0.2083333395421505,
"grad_norm": 0.30839505791664124,
"items/detected_max": 30.0,
"items/detected_mean": 24.95,
"items/detected_min": 20.0,
"kl": 0.01072122643769641,
"learning_rate": 1e-06,
"loss": 0.0122,
"num_tokens": 11561600.0,
"reward": 0.01675855405628681,
"reward_std": 0.01783207766711712,
"reward_total": 0.14728330001235007,
"rewards/pos_1/mean": 0.14728329554200173,
"rewards/pos_1/std": 0.06452072151005268,
"rewards/pos_10/mean": 0.0008343550143763423,
"rewards/pos_10/std": 0.007690893951803446,
"rewards/pos_11/mean": -0.0001931680366396904,
"rewards/pos_11/std": 0.0025193195790052412,
"rewards/pos_12/mean": -0.00031124838860705493,
"rewards/pos_12/std": 0.0027554802829399703,
"rewards/pos_13/mean": -0.00028753390070050954,
"rewards/pos_13/std": 0.002476783096790314,
"rewards/pos_14/mean": 0.0001333114691078663,
"rewards/pos_14/std": 0.001866360567510128,
"rewards/pos_15/mean": -0.00026041667442768814,
"rewards/pos_15/std": 0.0005208333488553763,
"rewards/pos_16/mean": 0.0,
"rewards/pos_16/std": 0.0,
"rewards/pos_17/mean": -0.0003747069742530584,
"rewards/pos_17/std": 0.0007494139485061168,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.07640054933726788,
"rewards/pos_2/std": 0.06290208585560322,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.052256912970915434,
"rewards/pos_3/std": 0.05409074127674103,
"rewards/pos_4/mean": 0.027486630342900754,
"rewards/pos_4/std": 0.043330289982259275,
"rewards/pos_5/mean": 0.018951812526211143,
"rewards/pos_5/std": 0.03892630822956562,
"rewards/pos_6/mean": 0.00856637657561805,
"rewards/pos_6/std": 0.027429566346108915,
"rewards/pos_7/mean": 0.0033461297823426618,
"rewards/pos_7/std": 0.022302147187292575,
"rewards/pos_8/mean": 0.0017868415918201209,
"rewards/pos_8/std": 0.01229224568232894,
"rewards/pos_9/mean": -0.0003254689043387771,
"rewards/pos_9/std": 0.012023128243163228,
"step": 420
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 322.3,
"completions/max_terminated_length": 322.3,
"completions/mean_length": 257.87501068115233,
"completions/mean_terminated_length": 257.87501068115233,
"completions/min_length": 202.9,
"completions/min_terminated_length": 202.9,
"entropy": 0.217508593822519,
"epoch": 0.27044025157232704,
"frac_reward_zero_std": 0.2083333373069763,
"grad_norm": 0.2622852623462677,
"items/detected_max": 35.0,
"items/detected_mean": 24.483333333333334,
"items/detected_min": 18.0,
"kl": 0.009693127268595466,
"learning_rate": 1e-06,
"loss": 0.0139,
"num_tokens": 11848076.0,
"reward": 0.023260173643939196,
"reward_std": 0.01729170950129628,
"reward_total": 0.2099197213537991,
"rewards/pos_1/mean": 0.20991971981711685,
"rewards/pos_1/std": 0.08556453036144376,
"rewards/pos_10/mean": -0.0009033275535330185,
"rewards/pos_10/std": 0.001806655107066037,
"rewards/pos_11/mean": -0.0007264139130711555,
"rewards/pos_11/std": 0.001452827826142311,
"rewards/pos_12/mean": -0.000985243311151862,
"rewards/pos_12/std": 0.001970486622303724,
"rewards/pos_13/mean": -0.0001367966295219958,
"rewards/pos_13/std": 0.0002735932590439916,
"rewards/pos_14/mean": -0.00039993440732359885,
"rewards/pos_14/std": 0.0007998688146471977,
"rewards/pos_15/mean": -0.0006510416860692203,
"rewards/pos_15/std": 0.0013020833721384406,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0002548443386331201,
"rewards/pos_17/mean": -0.0001249023247510204,
"rewards/pos_17/std": 0.0002498046495020408,
"rewards/pos_18/mean": -9.251858814269843e-19,
"rewards/pos_18/std": 1.8503717628539686e-18,
"rewards/pos_19/mean": -9.251858814269843e-19,
"rewards/pos_19/std": 1.8503717628539686e-18,
"rewards/pos_2/mean": 0.10043946434743703,
"rewards/pos_2/std": 0.05938351722434163,
"rewards/pos_20/mean": -9.251858814269843e-19,
"rewards/pos_20/std": 1.8503717628539686e-18,
"rewards/pos_3/mean": 0.05927633896353655,
"rewards/pos_3/std": 0.04616741510108113,
"rewards/pos_4/mean": 0.042810957436449824,
"rewards/pos_4/std": 0.0379714198410511,
"rewards/pos_5/mean": 0.030907541140913965,
"rewards/pos_5/std": 0.033647235855460166,
"rewards/pos_6/mean": 0.01658214651979506,
"rewards/pos_6/std": 0.031235989369452,
"rewards/pos_7/mean": 0.008262457395903765,
"rewards/pos_7/std": 0.01830614060163498,
"rewards/pos_8/mean": 0.0016870602848939598,
"rewards/pos_8/std": 0.01442420994862914,
"rewards/pos_9/mean": -0.0006271459045819938,
"rewards/pos_9/std": 0.01102356247138232,
"step": 430
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 319.4,
"completions/max_terminated_length": 319.4,
"completions/mean_length": 257.2979248046875,
"completions/mean_terminated_length": 257.2979248046875,
"completions/min_length": 202.6,
"completions/min_terminated_length": 202.6,
"entropy": 0.213536003852884,
"epoch": 0.27672955974842767,
"frac_reward_zero_std": 0.18333333879709243,
"grad_norm": 0.5048796534538269,
"items/detected_max": 30.0,
"items/detected_mean": 24.325,
"items/detected_min": 17.0,
"kl": 0.010342476896767039,
"learning_rate": 1e-06,
"loss": 0.015,
"num_tokens": 12102227.0,
"reward": 0.025915137585252525,
"reward_std": 0.02463643942028284,
"reward_total": 0.17452566474676132,
"rewards/pos_1/mean": 0.1745256520807743,
"rewards/pos_1/std": 0.08961139153689146,
"rewards/pos_10/mean": 0.005741290561854839,
"rewards/pos_10/std": 0.01339336596429348,
"rewards/pos_11/mean": 0.002170154917985201,
"rewards/pos_11/std": 0.008989358879625797,
"rewards/pos_12/mean": 0.0031916681327857077,
"rewards/pos_12/std": 0.0069463324500247834,
"rewards/pos_13/mean": 0.0002596526290290058,
"rewards/pos_13/std": 0.0038024243200197815,
"rewards/pos_14/mean": 0.00039993440732359885,
"rewards/pos_14/std": 0.0013331146910786629,
"rewards/pos_15/mean": -0.00039062501164153173,
"rewards/pos_15/std": 0.0007812500232830652,
"rewards/pos_16/mean": 4.625929407134921e-19,
"rewards/pos_16/std": 9.251858814269843e-19,
"rewards/pos_17/mean": -0.0002498046495020385,
"rewards/pos_17/std": 0.0004996092990040789,
"rewards/pos_18/mean": -0.00012260881485417439,
"rewards/pos_18/std": 0.0002452176297083506,
"rewards/pos_19/mean": -0.0001205094857141371,
"rewards/pos_19/std": 0.00024101897142827602,
"rewards/pos_2/mean": 0.121816236525774,
"rewards/pos_2/std": 0.07342976573854684,
"rewards/pos_20/mean": -0.00011857826029881789,
"rewards/pos_20/std": 0.00023715652059763762,
"rewards/pos_3/mean": 0.07539536720141768,
"rewards/pos_3/std": 0.07011822983622551,
"rewards/pos_4/mean": 0.04657440833980218,
"rewards/pos_4/std": 0.06269463896751404,
"rewards/pos_5/mean": 0.03148499419912696,
"rewards/pos_5/std": 0.04455104125663638,
"rewards/pos_6/mean": 0.021406761987600477,
"rewards/pos_6/std": 0.041934129782021046,
"rewards/pos_7/mean": 0.016635867313016205,
"rewards/pos_7/std": 0.031167579744942487,
"rewards/pos_8/mean": 0.011998111993307248,
"rewards/pos_8/std": 0.025716074835509063,
"rewards/pos_9/mean": 0.0077047318452969195,
"rewards/pos_9/std": 0.01703706863336265,
"step": 440
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 333.4,
"completions/max_terminated_length": 333.4,
"completions/mean_length": 258.3291778564453,
"completions/mean_terminated_length": 258.3291778564453,
"completions/min_length": 205.5,
"completions/min_terminated_length": 205.5,
"entropy": 0.21201196995874247,
"epoch": 0.2830188679245283,
"frac_reward_zero_std": 0.24166667386889457,
"grad_norm": 0.28489571809768677,
"items/detected_max": 30.0,
"items/detected_mean": 24.383333333333333,
"items/detected_min": 19.0,
"kl": 0.013525528619356918,
"learning_rate": 1e-06,
"loss": 0.0147,
"num_tokens": 12361545.0,
"reward": 0.02282193535938859,
"reward_std": 0.021278287377208473,
"reward_total": 0.15365880057215692,
"rewards/pos_1/mean": 0.15365879833698273,
"rewards/pos_1/std": 0.06613776870071889,
"rewards/pos_10/mean": 0.0027370264171622696,
"rewards/pos_10/std": 0.008689730637706818,
"rewards/pos_11/mean": 0.0024675481021404267,
"rewards/pos_11/std": 0.007933381665498018,
"rewards/pos_12/mean": 0.001032855163794011,
"rewards/pos_12/std": 0.0041585674742236735,
"rewards/pos_13/mean": 0.00034887201618403243,
"rewards/pos_13/std": 0.00343367662280798,
"rewards/pos_14/mean": 0.0008995437063276773,
"rewards/pos_14/std": 0.0023323331959545623,
"rewards/pos_15/mean": -2.1224049851298334e-05,
"rewards/pos_15/std": 0.001894681667909026,
"rewards/pos_16/mean": 0.00011734282597899482,
"rewards/pos_16/std": 0.0017637516837567101,
"rewards/pos_17/mean": 0.00012490232475101994,
"rewards/pos_17/std": 0.0017486325465142737,
"rewards/pos_18/mean": -0.00012260881485417439,
"rewards/pos_18/std": 0.0002452176297083506,
"rewards/pos_19/mean": 4.625929407134921e-19,
"rewards/pos_19/std": 9.251858814269843e-19,
"rewards/pos_2/mean": 0.1019910465925932,
"rewards/pos_2/std": 0.057884824462234974,
"rewards/pos_20/mean": -0.0002371565205976362,
"rewards/pos_20/std": 0.0004743130411952744,
"rewards/pos_3/mean": 0.07082258388400078,
"rewards/pos_3/std": 0.061851803213357925,
"rewards/pos_4/mean": 0.05265027731657028,
"rewards/pos_4/std": 0.05185747928917408,
"rewards/pos_5/mean": 0.029668276105076073,
"rewards/pos_5/std": 0.049659001640975475,
"rewards/pos_6/mean": 0.020517667289823294,
"rewards/pos_6/std": 0.03811166062951088,
"rewards/pos_7/mean": 0.008061961038038135,
"rewards/pos_7/std": 0.027415334060788155,
"rewards/pos_8/mean": 0.008093093032948672,
"rewards/pos_8/std": 0.024487523501738905,
"rewards/pos_9/mean": 0.0036278903018683193,
"rewards/pos_9/std": 0.015486051235347987,
"step": 450
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 336.9,
"completions/max_terminated_length": 336.9,
"completions/mean_length": 259.5166748046875,
"completions/mean_terminated_length": 259.5166748046875,
"completions/min_length": 206.9,
"completions/min_terminated_length": 206.9,
"entropy": 0.21642931795989473,
"epoch": 0.2893081761006289,
"frac_reward_zero_std": 0.2083333373069763,
"grad_norm": 0.2088499367237091,
"items/detected_max": 30.0,
"items/detected_mean": 24.8875,
"items/detected_min": 19.0,
"kl": 0.01066165748294831,
"learning_rate": 1e-06,
"loss": 0.0149,
"num_tokens": 12617377.0,
"reward": 0.015607737097889185,
"reward_std": 0.023038433864712715,
"reward_total": 0.10255183950066567,
"rewards/pos_1/mean": 0.10255183652043343,
"rewards/pos_1/std": 0.07729223407804967,
"rewards/pos_10/mean": 0.0008214917848818004,
"rewards/pos_10/std": 0.006460730661638081,
"rewards/pos_11/mean": 0.0004120025085285306,
"rewards/pos_11/std": 0.004891922930255532,
"rewards/pos_12/mean": 0.0007297705044038594,
"rewards/pos_12/std": 0.00483751802239567,
"rewards/pos_13/mean": 0.00046408207854256036,
"rewards/pos_13/std": 0.0025697237113490702,
"rewards/pos_14/mean": 0.0004780227085575461,
"rewards/pos_14/std": 0.0025557830464094876,
"rewards/pos_15/mean": 0.0007508519338443875,
"rewards/pos_15/std": 0.0025433705653995276,
"rewards/pos_16/mean": 0.0001081687631085515,
"rewards/pos_16/std": 0.0016022538300603628,
"rewards/pos_17/mean": 0.00011572828516364098,
"rewards/pos_17/std": 0.0017302843742072582,
"rewards/pos_18/mean": 0.00036782644456252457,
"rewards/pos_18/std": 0.0012260881485417485,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.07489825356751681,
"rewards/pos_2/std": 0.07403903417289256,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04760490842163563,
"rewards/pos_3/std": 0.0639339717105031,
"rewards/pos_4/mean": 0.031833214685320854,
"rewards/pos_4/std": 0.054455436579883096,
"rewards/pos_5/mean": 0.02260264167562127,
"rewards/pos_5/std": 0.04626600733026862,
"rewards/pos_6/mean": 0.01071704727364704,
"rewards/pos_6/std": 0.04206645265221596,
"rewards/pos_7/mean": 0.009195378096774221,
"rewards/pos_7/std": 0.03155197752639651,
"rewards/pos_8/mean": 0.00475915347924456,
"rewards/pos_8/std": 0.02537555042654276,
"rewards/pos_9/mean": 0.0037443477544002236,
"rewards/pos_9/std": 0.01737030434887856,
"step": 460
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 318.4,
"completions/max_terminated_length": 318.4,
"completions/mean_length": 253.72084045410156,
"completions/mean_terminated_length": 253.72084045410156,
"completions/min_length": 202.4,
"completions/min_terminated_length": 202.4,
"entropy": 0.20708943608527383,
"epoch": 0.29559748427672955,
"frac_reward_zero_std": 0.15833333656191825,
"grad_norm": 0.2726491093635559,
"items/detected_max": 30.0,
"items/detected_mean": 24.233333333333334,
"items/detected_min": 18.0,
"kl": 0.010009914560456915,
"learning_rate": 1e-06,
"loss": 0.0141,
"num_tokens": 12867423.0,
"reward": 0.023438004404306413,
"reward_std": 0.022939678560942412,
"reward_total": 0.20833126753568648,
"rewards/pos_1/mean": 0.20833126306533814,
"rewards/pos_1/std": 0.08295661797747016,
"rewards/pos_10/mean": 0.003840826149098575,
"rewards/pos_10/std": 0.01358814868144691,
"rewards/pos_11/mean": 0.004028283571824431,
"rewards/pos_11/std": 0.011297739017754794,
"rewards/pos_12/mean": 0.0031792555819265544,
"rewards/pos_12/std": 0.008610495855100453,
"rewards/pos_13/mean": 0.0009312232374213636,
"rewards/pos_13/std": 0.0024096329929307104,
"rewards/pos_14/mean": 0.00038752187974750997,
"rewards/pos_14/std": 0.001308289635926485,
"rewards/pos_15/mean": 0.00039062501164153216,
"rewards/pos_15/std": 0.0013020833721384406,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.09070989573374391,
"rewards/pos_2/std": 0.07176251346245408,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05962197887711227,
"rewards/pos_3/std": 0.0604378919582814,
"rewards/pos_4/mean": 0.040848955418914555,
"rewards/pos_4/std": 0.05593613050878048,
"rewards/pos_5/mean": 0.025873664231039584,
"rewards/pos_5/std": 0.04659653939306736,
"rewards/pos_6/mean": 0.013208038732409477,
"rewards/pos_6/std": 0.03292058128863573,
"rewards/pos_7/mean": 0.009250723465811461,
"rewards/pos_7/std": 0.02974128625355661,
"rewards/pos_8/mean": 0.004960190316584583,
"rewards/pos_8/std": 0.023272993229329585,
"rewards/pos_9/mean": 0.0034524708811659367,
"rewards/pos_9/std": 0.016142918425612152,
"step": 470
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 314.6,
"completions/max_terminated_length": 314.6,
"completions/mean_length": 256.15625762939453,
"completions/mean_terminated_length": 256.15625762939453,
"completions/min_length": 202.4,
"completions/min_terminated_length": 202.4,
"entropy": 0.22198334643617273,
"epoch": 0.3018867924528302,
"frac_reward_zero_std": 0.17500000298023224,
"grad_norm": 0.2422315627336502,
"items/detected_max": 30.0,
"items/detected_mean": 24.533333333333335,
"items/detected_min": 19.0,
"kl": 0.012731472958694211,
"learning_rate": 1e-06,
"loss": 0.0159,
"num_tokens": 13147890.0,
"reward": 0.019851746130734683,
"reward_std": 0.021506846975535156,
"reward_total": 0.13185822628438473,
"rewards/pos_1/mean": 0.13185822516679763,
"rewards/pos_1/std": 0.06748964563012123,
"rewards/pos_10/mean": 0.004497004637960345,
"rewards/pos_10/std": 0.010416024201549589,
"rewards/pos_11/mean": 0.0031894596293568613,
"rewards/pos_11/std": 0.008589645940810442,
"rewards/pos_12/mean": 0.002049866074230522,
"rewards/pos_12/std": 0.0058181500295177106,
"rewards/pos_13/mean": 0.0006700425175949932,
"rewards/pos_13/std": 0.004630456329323351,
"rewards/pos_14/mean": -0.0001333114691078663,
"rewards/pos_14/std": 0.002069247793406248,
"rewards/pos_15/mean": -0.0005208333488553763,
"rewards/pos_15/std": 0.0010416666977107526,
"rewards/pos_16/mean": -0.0006371107883751392,
"rewards/pos_16/std": 0.0012742216931656003,
"rewards/pos_17/mean": -0.0003747069742530584,
"rewards/pos_17/std": 0.0007494139485061168,
"rewards/pos_18/mean": -0.00036782644456252457,
"rewards/pos_18/std": 0.0007356528891250491,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.0955985370092094,
"rewards/pos_2/std": 0.06465087793767452,
"rewards/pos_20/mean": -0.00011857826029881835,
"rewards/pos_20/std": 0.0002371565205976367,
"rewards/pos_3/mean": 0.05621484536677599,
"rewards/pos_3/std": 0.056038703257218006,
"rewards/pos_4/mean": 0.03342725089751184,
"rewards/pos_4/std": 0.04756846679374575,
"rewards/pos_5/mean": 0.022668213793076576,
"rewards/pos_5/std": 0.043966184742748736,
"rewards/pos_6/mean": 0.01648205746896565,
"rewards/pos_6/std": 0.03921035025268793,
"rewards/pos_7/mean": 0.015421654691454022,
"rewards/pos_7/std": 0.032973642507568,
"rewards/pos_8/mean": 0.01127681282814592,
"rewards/pos_8/std": 0.026968108676373958,
"rewards/pos_9/mean": 0.00583330043591559,
"rewards/pos_9/std": 0.01570929940789938,
"step": 480
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 328.6,
"completions/max_terminated_length": 328.6,
"completions/mean_length": 259.42708892822264,
"completions/mean_terminated_length": 259.42708892822264,
"completions/min_length": 210.7,
"completions/min_terminated_length": 210.7,
"entropy": 0.2256676376486818,
"epoch": 0.3081761006289308,
"frac_reward_zero_std": 0.1750000037252903,
"grad_norm": 0.29480305314064026,
"items/detected_max": 30.0,
"items/detected_mean": 25.2875,
"items/detected_min": 19.0,
"kl": 0.011438539253625398,
"learning_rate": 1e-06,
"loss": 0.0163,
"num_tokens": 13409963.0,
"reward": 0.021010580559959635,
"reward_std": 0.020790635049343108,
"reward_total": 0.16183093758299946,
"rewards/pos_1/mean": 0.1618309345562011,
"rewards/pos_1/std": 0.05938454018905759,
"rewards/pos_10/mean": 0.0036000264110043643,
"rewards/pos_10/std": 0.01220149858854711,
"rewards/pos_11/mean": 0.0015766112133860587,
"rewards/pos_11/std": 0.009965774789452552,
"rewards/pos_12/mean": 0.0016310159582644701,
"rewards/pos_12/std": 0.008706640219315886,
"rewards/pos_13/mean": -2.635318087413861e-05,
"rewards/pos_13/std": 0.0038622493389993914,
"rewards/pos_14/mean": -1.2412527576088906e-05,
"rewards/pos_14/std": 0.0021081584505736827,
"rewards/pos_15/mean": -1.164153172010054e-11,
"rewards/pos_15/std": 0.0020833333255723127,
"rewards/pos_16/mean": -0.0007645329460501671,
"rewards/pos_16/std": 0.0015290660317987203,
"rewards/pos_17/mean": -0.0004996092990040775,
"rewards/pos_17/std": 0.0007880588760599503,
"rewards/pos_18/mean": -0.0004904352594166994,
"rewards/pos_18/std": 0.0009808705188333988,
"rewards/pos_19/mean": 4.625929407134921e-19,
"rewards/pos_19/std": 9.251858814269843e-19,
"rewards/pos_2/mean": 0.0853801101911813,
"rewards/pos_2/std": 0.055204817373305556,
"rewards/pos_20/mean": 4.625929407134921e-19,
"rewards/pos_20/std": 9.251858814269843e-19,
"rewards/pos_3/mean": 0.06138521353714168,
"rewards/pos_3/std": 0.06057459069415927,
"rewards/pos_4/mean": 0.04186704335734248,
"rewards/pos_4/std": 0.05362486112862826,
"rewards/pos_5/mean": 0.02881924912217073,
"rewards/pos_5/std": 0.04141271626576781,
"rewards/pos_6/mean": 0.018322084681130946,
"rewards/pos_6/std": 0.03851036261767149,
"rewards/pos_7/mean": 0.010271314741112291,
"rewards/pos_7/std": 0.030041778273880482,
"rewards/pos_8/mean": 0.004253976827021689,
"rewards/pos_8/std": 0.02209998187609017,
"rewards/pos_9/mean": 0.003067348466720432,
"rewards/pos_9/std": 0.012733368459157646,
"step": 490
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 319.2,
"completions/max_terminated_length": 319.2,
"completions/mean_length": 257.12708892822263,
"completions/mean_terminated_length": 257.12708892822263,
"completions/min_length": 205.3,
"completions/min_terminated_length": 205.3,
"entropy": 0.2168938455792765,
"epoch": 0.31446540880503143,
"frac_reward_zero_std": 0.1916666693985462,
"grad_norm": 0.3637041747570038,
"items/detected_max": 30.0,
"items/detected_mean": 24.4,
"items/detected_min": 19.0,
"kl": 0.011324638859756911,
"learning_rate": 1e-06,
"loss": 0.0145,
"num_tokens": 13669780.0,
"reward": 0.028080825228244066,
"reward_std": 0.026594148110598327,
"reward_total": 0.19896745309233665,
"rewards/pos_1/mean": 0.19896744936704636,
"rewards/pos_1/std": 0.09368667528033256,
"rewards/pos_10/mean": 0.0030387546285055578,
"rewards/pos_10/std": 0.011216811486519873,
"rewards/pos_11/mean": 0.0017206659773364662,
"rewards/pos_11/std": 0.00750925000756979,
"rewards/pos_12/mean": 0.001734267210122198,
"rewards/pos_12/std": 0.005157523066736757,
"rewards/pos_13/mean": 0.0003465388319455087,
"rewards/pos_13/std": 0.0034290103474631907,
"rewards/pos_14/mean": 0.0004972761031240225,
"rewards/pos_14/std": 0.003127535805106163,
"rewards/pos_15/mean": 0.0007701053051277995,
"rewards/pos_15/std": 0.002581877401098609,
"rewards/pos_16/mean": 0.00012742215767502785,
"rewards/pos_16/std": 0.001783910347148776,
"rewards/pos_17/mean": -0.0004996092990040779,
"rewards/pos_17/std": 0.0009992185980081559,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.11547301951795816,
"rewards/pos_2/std": 0.08222815282642841,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.08356678988784552,
"rewards/pos_3/std": 0.07231327965855598,
"rewards/pos_4/mean": 0.05913681124026576,
"rewards/pos_4/std": 0.06760933250188828,
"rewards/pos_5/mean": 0.03961076997220516,
"rewards/pos_5/std": 0.0505744569003582,
"rewards/pos_6/mean": 0.025688346242532134,
"rewards/pos_6/std": 0.04620512621477246,
"rewards/pos_7/mean": 0.01850563199259341,
"rewards/pos_7/std": 0.03958346555009484,
"rewards/pos_8/mean": 0.008969539403915405,
"rewards/pos_8/std": 0.026932531129568817,
"rewards/pos_9/mean": 0.004205800127238035,
"rewards/pos_9/std": 0.016458538291044534,
"step": 500
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 317.6,
"completions/max_terminated_length": 317.6,
"completions/mean_length": 257.35000762939455,
"completions/mean_terminated_length": 257.35000762939455,
"completions/min_length": 206.9,
"completions/min_terminated_length": 206.9,
"entropy": 0.2289095859353741,
"epoch": 0.32075471698113206,
"frac_reward_zero_std": 0.20000000298023224,
"grad_norm": 0.3724060654640198,
"items/detected_max": 30.0,
"items/detected_mean": 24.945833333333333,
"items/detected_min": 19.0,
"kl": 0.012526460485726905,
"learning_rate": 1e-06,
"loss": 0.0161,
"num_tokens": 13941636.0,
"reward": 0.013600436504930258,
"reward_std": 0.013879508594982326,
"reward_total": 0.12234384417533875,
"rewards/pos_1/mean": 0.12234384194016457,
"rewards/pos_1/std": 0.06252944990992546,
"rewards/pos_10/mean": 0.0011139262584038078,
"rewards/pos_10/std": 0.005974171520210803,
"rewards/pos_11/mean": 0.0007149806478992105,
"rewards/pos_11/std": 0.004180813068524003,
"rewards/pos_12/mean": 0.0007421830203384161,
"rewards/pos_12/std": 0.0037812422029674052,
"rewards/pos_13/mean": 0.00033969798823818564,
"rewards/pos_13/std": 0.0032613215735182166,
"rewards/pos_14/mean": 0.0008903696667402983,
"rewards/pos_14/std": 0.002313985209912062,
"rewards/pos_15/mean": -0.00016060643829405333,
"rewards/pos_15/std": 0.0022829538211226468,
"rewards/pos_16/mean": -0.00027409770991653225,
"rewards/pos_16/std": 0.002509936527349055,
"rewards/pos_17/mean": 0.00024063060991466023,
"rewards/pos_17/std": 0.0014804798178374773,
"rewards/pos_18/mean": 0.0004904352594166992,
"rewards/pos_18/std": 0.0009808705188333994,
"rewards/pos_19/mean": -0.0001205094857141378,
"rewards/pos_19/std": 0.0002410189714282756,
"rewards/pos_2/mean": 0.05304830120876432,
"rewards/pos_2/std": 0.04234171658754349,
"rewards/pos_20/mean": -2.3129647035674607e-19,
"rewards/pos_20/std": 4.625929407134921e-19,
"rewards/pos_3/mean": 0.03204201725311577,
"rewards/pos_3/std": 0.02614194988273084,
"rewards/pos_4/mean": 0.024951637379126625,
"rewards/pos_4/std": 0.02837535673752427,
"rewards/pos_5/mean": 0.01782722012139857,
"rewards/pos_5/std": 0.030218360666185617,
"rewards/pos_6/mean": 0.008628727751784027,
"rewards/pos_6/std": 0.022143259178847075,
"rewards/pos_7/mean": 0.00442632056074217,
"rewards/pos_7/std": 0.017501901322975754,
"rewards/pos_8/mean": 0.0035614772117696703,
"rewards/pos_8/std": 0.013929071929305792,
"rewards/pos_9/mean": 0.001202161784749478,
"rewards/pos_9/std": 0.00740228877402842,
"step": 510
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.008333333333333337,
"completions/max_length": 345.0,
"completions/max_terminated_length": 325.2,
"completions/mean_length": 259.12084197998047,
"completions/mean_terminated_length": 257.01781158447267,
"completions/min_length": 203.6,
"completions/min_terminated_length": 203.6,
"entropy": 0.21795971781636278,
"epoch": 0.3270440251572327,
"frac_reward_zero_std": 0.24166667610406875,
"grad_norm": 0.44123509526252747,
"items/detected_max": 47.0,
"items/detected_mean": 24.404166666666665,
"items/detected_min": 19.0,
"kl": 0.009536893166174802,
"learning_rate": 1e-06,
"loss": 0.0151,
"num_tokens": 14200682.0,
"reward": 0.018912672251462936,
"reward_std": 0.020285651087760925,
"reward_total": 0.16123745497316122,
"rewards/pos_1/mean": 0.1612374499440193,
"rewards/pos_1/std": 0.07130067069083453,
"rewards/pos_10/mean": 0.002000838378444314,
"rewards/pos_10/std": 0.009915737924166024,
"rewards/pos_11/mean": 0.0014460663311183452,
"rewards/pos_11/std": 0.006357286497950554,
"rewards/pos_12/mean": 0.002331364236306399,
"rewards/pos_12/std": 0.0051678218180313705,
"rewards/pos_13/mean": 0.0012172291171737015,
"rewards/pos_13/std": 0.0028921225806698204,
"rewards/pos_14/mean": 0.0001333114691078663,
"rewards/pos_14/std": 0.001866360567510128,
"rewards/pos_15/mean": -0.00026041667442768814,
"rewards/pos_15/std": 0.0005208333488553763,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": -0.0004996092990040779,
"rewards/pos_17/std": 0.0009992185980081559,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08295135851949453,
"rewards/pos_2/std": 0.06343129687011242,
"rewards/pos_20/mean": -0.0004743130411952734,
"rewards/pos_20/std": 0.0009486260823905468,
"rewards/pos_3/mean": 0.05000346293672919,
"rewards/pos_3/std": 0.06582604087889195,
"rewards/pos_4/mean": 0.0349484842736274,
"rewards/pos_4/std": 0.051855390705168244,
"rewards/pos_5/mean": 0.01802287108730525,
"rewards/pos_5/std": 0.038701156992465256,
"rewards/pos_6/mean": 0.012096035139256856,
"rewards/pos_6/std": 0.02850966826081276,
"rewards/pos_7/mean": 0.006792393838986754,
"rewards/pos_7/std": 0.02349440399557352,
"rewards/pos_8/mean": 0.005078107689041644,
"rewards/pos_8/std": 0.019577187206596135,
"rewards/pos_9/mean": 0.0016062510956544429,
"rewards/pos_9/std": 0.01359426868148148,
"step": 520
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 335.6,
"completions/max_terminated_length": 335.6,
"completions/mean_length": 257.13334503173826,
"completions/mean_terminated_length": 257.13334503173826,
"completions/min_length": 201.4,
"completions/min_terminated_length": 201.4,
"entropy": 0.2276484333910048,
"epoch": 0.3333333333333333,
"frac_reward_zero_std": 0.16666667088866233,
"grad_norm": 0.39736422896385193,
"items/detected_max": 31.0,
"items/detected_mean": 24.495833333333334,
"items/detected_min": 17.0,
"kl": 0.011707671616750303,
"learning_rate": 1e-06,
"loss": 0.0136,
"num_tokens": 14494766.0,
"reward": 0.021821795543655754,
"reward_std": 0.018616222217679022,
"reward_total": 0.1836221881210804,
"rewards/pos_1/mean": 0.18362218141555786,
"rewards/pos_1/std": 0.06593040227890015,
"rewards/pos_10/mean": -8.139674318954348e-05,
"rewards/pos_10/std": 0.0068946087034419175,
"rewards/pos_11/mean": 0.00041771335527300834,
"rewards/pos_11/std": 0.005574377067387104,
"rewards/pos_12/mean": 0.0,
"rewards/pos_12/std": 0.0020140345441177487,
"rewards/pos_13/mean": -0.0008207797771319747,
"rewards/pos_13/std": 0.001410291320644319,
"rewards/pos_14/mean": -0.0002666229382157326,
"rewards/pos_14/std": 0.0005332458764314652,
"rewards/pos_15/mean": -0.0009114583604969084,
"rewards/pos_15/std": 0.001822916720993817,
"rewards/pos_16/mean": -0.00012742215767502785,
"rewards/pos_16/std": 0.0002548443386331201,
"rewards/pos_17/mean": -0.0003747069742530584,
"rewards/pos_17/std": 0.0007494139485061168,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.09686637232080102,
"rewards/pos_2/std": 0.058121581189334394,
"rewards/pos_20/mean": -0.00011857826029881835,
"rewards/pos_20/std": 0.0002371565205976367,
"rewards/pos_3/mean": 0.06536929956637323,
"rewards/pos_3/std": 0.05751038622111082,
"rewards/pos_4/mean": 0.03915250832214952,
"rewards/pos_4/std": 0.048235630802810195,
"rewards/pos_5/mean": 0.02973933550529182,
"rewards/pos_5/std": 0.03755170316435397,
"rewards/pos_6/mean": 0.012826546526048333,
"rewards/pos_6/std": 0.030991469509899616,
"rewards/pos_7/mean": 0.008809663902502507,
"rewards/pos_7/std": 0.023566680843941867,
"rewards/pos_8/mean": 0.00226218614843674,
"rewards/pos_8/std": 0.016075075697153808,
"rewards/pos_9/mean": 0.00031418022699654103,
"rewards/pos_9/std": 0.01436436150688678,
"step": 530
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.002083333333333337,
"completions/max_length": 348.2,
"completions/max_terminated_length": 328.0,
"completions/mean_length": 259.75834197998046,
"completions/mean_terminated_length": 259.19579620361327,
"completions/min_length": 203.2,
"completions/min_terminated_length": 203.2,
"entropy": 0.23366477421174447,
"epoch": 0.33962264150943394,
"frac_reward_zero_std": 0.18333333730697632,
"grad_norm": 0.2925439774990082,
"items/detected_max": 50.0,
"items/detected_mean": 24.9375,
"items/detected_min": 19.0,
"kl": 0.010972671315539629,
"learning_rate": 1e-06,
"loss": 0.016,
"num_tokens": 14773898.0,
"reward": 0.02019151346758008,
"reward_std": 0.018987514078617096,
"reward_total": 0.17083178609609603,
"rewards/pos_1/mean": 0.17083178013563155,
"rewards/pos_1/std": 0.06076769046485424,
"rewards/pos_10/mean": 0.0035903325304389,
"rewards/pos_10/std": 0.011628140788525343,
"rewards/pos_11/mean": 0.0027449949644505977,
"rewards/pos_11/std": 0.009627606812864542,
"rewards/pos_12/mean": 0.002477097907103598,
"rewards/pos_12/std": 0.0066403827862814065,
"rewards/pos_13/mean": 0.0019338638870976866,
"rewards/pos_13/std": 0.0057128167012706395,
"rewards/pos_14/mean": 0.002217912580817938,
"rewards/pos_14/std": 0.004919342882931233,
"rewards/pos_15/mean": 0.0012909386656247079,
"rewards/pos_15/std": 0.0036235440289601683,
"rewards/pos_16/mean": 0.00012742215767502785,
"rewards/pos_16/std": 0.0017839103704318405,
"rewards/pos_17/mean": -0.00024980464950203897,
"rewards/pos_17/std": 0.0004996092990040779,
"rewards/pos_18/mean": -0.0002452176297083497,
"rewards/pos_18/std": 0.0004904352594166994,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.08002098277211189,
"rewards/pos_2/std": 0.0625394044443965,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.050222796481102706,
"rewards/pos_3/std": 0.047927550598979,
"rewards/pos_4/mean": 0.03301227022893727,
"rewards/pos_4/std": 0.045436117053031924,
"rewards/pos_5/mean": 0.01862254913430661,
"rewards/pos_5/std": 0.036664932314306495,
"rewards/pos_6/mean": 0.013114646542817354,
"rewards/pos_6/std": 0.024240608420223,
"rewards/pos_7/mean": 0.010791311541106552,
"rewards/pos_7/std": 0.022284213569946587,
"rewards/pos_8/mean": 0.00868327144999057,
"rewards/pos_8/std": 0.01720749761443585,
"rewards/pos_9/mean": 0.00476360940374434,
"rewards/pos_9/std": 0.017515426012687386,
"step": 540
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 326.3,
"completions/max_terminated_length": 326.3,
"completions/mean_length": 257.66875610351565,
"completions/mean_terminated_length": 257.66875610351565,
"completions/min_length": 205.4,
"completions/min_terminated_length": 205.4,
"entropy": 0.2219298164670666,
"epoch": 0.34591194968553457,
"frac_reward_zero_std": 0.16666667014360428,
"grad_norm": 0.18160945177078247,
"items/detected_max": 30.0,
"items/detected_mean": 24.6875,
"items/detected_min": 19.0,
"kl": 0.013025570806348696,
"learning_rate": 1e-06,
"loss": 0.0135,
"num_tokens": 15045635.0,
"reward": 0.02094852374866605,
"reward_std": 0.018986396910622715,
"reward_total": 0.1696576789021492,
"rewards/pos_1/mean": 0.16965767368674278,
"rewards/pos_1/std": 0.06637014597654342,
"rewards/pos_10/mean": -0.0004375260672532022,
"rewards/pos_10/std": 0.004892604262568057,
"rewards/pos_11/mean": 0.0004868887132033706,
"rewards/pos_11/std": 0.0027171708177775146,
"rewards/pos_12/mean": 0.000505023670848459,
"rewards/pos_12/std": 0.003104969975538552,
"rewards/pos_13/mean": 0.0003840367076918483,
"rewards/pos_13/std": 0.0032727377722039817,
"rewards/pos_14/mean": -0.00014572401996701956,
"rewards/pos_14/std": 0.0023747813887894154,
"rewards/pos_15/mean": 0.00026041667442768814,
"rewards/pos_15/std": 0.0015625000465661286,
"rewards/pos_16/mean": -0.0005096886307001114,
"rewards/pos_16/std": 0.0008039575535804033,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.10143493013456464,
"rewards/pos_2/std": 0.06443125363439321,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06565358490915969,
"rewards/pos_3/std": 0.0606054674834013,
"rewards/pos_4/mean": 0.04483333106618374,
"rewards/pos_4/std": 0.05223893783986568,
"rewards/pos_5/mean": 0.017703550728037955,
"rewards/pos_5/std": 0.040874285716563465,
"rewards/pos_6/mean": 0.013590086996555329,
"rewards/pos_6/std": 0.030414795130491258,
"rewards/pos_7/mean": 0.006060151034034789,
"rewards/pos_7/std": 0.022422701702453196,
"rewards/pos_8/mean": 0.0007093379856087268,
"rewards/pos_8/std": 0.01532193636521697,
"rewards/pos_9/mean": -0.0009702042210847139,
"rewards/pos_9/std": 0.00782886240631342,
"step": 550
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 318.8,
"completions/max_terminated_length": 318.8,
"completions/mean_length": 255.1791748046875,
"completions/mean_terminated_length": 255.1791748046875,
"completions/min_length": 206.8,
"completions/min_terminated_length": 206.8,
"entropy": 0.2227155176612238,
"epoch": 0.3522012578616352,
"frac_reward_zero_std": 0.20833333507180213,
"grad_norm": 0.4057275652885437,
"items/detected_max": 30.0,
"items/detected_mean": 24.0875,
"items/detected_min": 18.0,
"kl": 0.013039527038442127,
"learning_rate": 1e-06,
"loss": 0.0158,
"num_tokens": 15341969.0,
"reward": 0.02573305545374751,
"reward_std": 0.025089992582798003,
"reward_total": 0.15084369219839572,
"rewards/pos_1/mean": 0.1508436843752861,
"rewards/pos_1/std": 0.08285234421491623,
"rewards/pos_10/mean": 0.0003826912841759622,
"rewards/pos_10/std": 0.00601864627096802,
"rewards/pos_11/mean": -0.0006290163844823837,
"rewards/pos_11/std": 0.0025642702355980875,
"rewards/pos_12/mean": -2.975030802190304e-05,
"rewards/pos_12/std": 0.0021924841217696666,
"rewards/pos_13/mean": -0.00015073727117851377,
"rewards/pos_13/std": 0.002434458048082888,
"rewards/pos_14/mean": -0.0002666229382157326,
"rewards/pos_14/std": 0.0026662293821573257,
"rewards/pos_15/mean": -0.00026041667442768814,
"rewards/pos_15/std": 0.0005208333488553763,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.11551399044692516,
"rewards/pos_2/std": 0.08242853209376336,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.0830869296565652,
"rewards/pos_3/std": 0.0705016914755106,
"rewards/pos_4/mean": 0.058664625696837905,
"rewards/pos_4/std": 0.06315769106149674,
"rewards/pos_5/mean": 0.04447245486080646,
"rewards/pos_5/std": 0.057740448787808415,
"rewards/pos_6/mean": 0.03110660398378968,
"rewards/pos_6/std": 0.046270192787051204,
"rewards/pos_7/mean": 0.020051701087504627,
"rewards/pos_7/std": 0.0408051636070013,
"rewards/pos_8/mean": 0.008981027461898824,
"rewards/pos_8/std": 0.02832955224439502,
"rewards/pos_9/mean": 0.003148760786280036,
"rewards/pos_9/std": 0.012807594193145633,
"step": 560
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 315.7,
"completions/max_terminated_length": 315.7,
"completions/mean_length": 255.00833740234376,
"completions/mean_terminated_length": 255.00833740234376,
"completions/min_length": 200.4,
"completions/min_terminated_length": 200.4,
"entropy": 0.2198466069375475,
"epoch": 0.3584905660377358,
"frac_reward_zero_std": 0.18333333879709243,
"grad_norm": 0.2640518844127655,
"items/detected_max": 30.0,
"items/detected_mean": 24.033333333333335,
"items/detected_min": 17.0,
"kl": 0.011254959223151673,
"learning_rate": 1e-06,
"loss": 0.014,
"num_tokens": 15608789.0,
"reward": 0.023907679412513972,
"reward_std": 0.02189104985445738,
"reward_total": 0.19489617720246316,
"rewards/pos_1/mean": 0.19489617235958576,
"rewards/pos_1/std": 0.08845191560685635,
"rewards/pos_10/mean": 0.002839635347481817,
"rewards/pos_10/std": 0.008555516763590277,
"rewards/pos_11/mean": 0.002113221539184451,
"rewards/pos_11/std": 0.006886483822017908,
"rewards/pos_12/mean": 0.0005332458764314652,
"rewards/pos_12/std": 0.003080526296980679,
"rewards/pos_13/mean": -0.00015073727117851377,
"rewards/pos_13/std": 0.002434458048082888,
"rewards/pos_14/mean": 0.00039993440732359885,
"rewards/pos_14/std": 0.0013331146910786629,
"rewards/pos_15/mean": -0.00013020833721384407,
"rewards/pos_15/std": 0.00026041667442768814,
"rewards/pos_16/mean": -0.0002548443153500557,
"rewards/pos_16/std": 0.0005096886772662402,
"rewards/pos_17/mean": 0.0,
"rewards/pos_17/std": 0.0,
"rewards/pos_18/mean": 0.0,
"rewards/pos_18/std": 0.0,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.09868481699377299,
"rewards/pos_2/std": 0.0687787925824523,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.06857036959845572,
"rewards/pos_3/std": 0.06220374419353902,
"rewards/pos_4/mean": 0.04410428605042398,
"rewards/pos_4/std": 0.05569797847419977,
"rewards/pos_5/mean": 0.024646718322765084,
"rewards/pos_5/std": 0.04680635128170252,
"rewards/pos_6/mean": 0.01727755949832499,
"rewards/pos_6/std": 0.03226131461560726,
"rewards/pos_7/mean": 0.010465893859509379,
"rewards/pos_7/std": 0.02676454761531204,
"rewards/pos_8/mean": 0.008822847646661103,
"rewards/pos_8/std": 0.018864544108510017,
"rewards/pos_9/mean": 0.0054551504668779675,
"rewards/pos_9/std": 0.014690567925572395,
"step": 570
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 318.2,
"completions/max_terminated_length": 318.2,
"completions/mean_length": 252.0562545776367,
"completions/mean_terminated_length": 252.0562545776367,
"completions/min_length": 200.3,
"completions/min_terminated_length": 200.3,
"entropy": 0.22020154868563016,
"epoch": 0.36477987421383645,
"frac_reward_zero_std": 0.20833333805203438,
"grad_norm": 0.19639542698860168,
"items/detected_max": 30.0,
"items/detected_mean": 23.941666666666666,
"items/detected_min": 18.0,
"kl": 0.012878992815482585,
"learning_rate": 1e-06,
"loss": 0.0142,
"num_tokens": 15882596.0,
"reward": 0.01625635549426079,
"reward_std": 0.020171427354216575,
"reward_total": 0.13932952359318734,
"rewards/pos_1/mean": 0.13932951912283897,
"rewards/pos_1/std": 0.06653505899012088,
"rewards/pos_10/mean": 0.0011549281771294773,
"rewards/pos_10/std": 0.007822985318489373,
"rewards/pos_11/mean": 0.0019340602215379477,
"rewards/pos_11/std": 0.006773776095360518,
"rewards/pos_12/mean": 0.0022382270311936737,
"rewards/pos_12/std": 0.005602446431294083,
"rewards/pos_13/mean": 0.0021132876281626523,
"rewards/pos_13/std": 0.006415321328677237,
"rewards/pos_14/mean": 0.0011661666445434093,
"rewards/pos_14/std": 0.003782981541007757,
"rewards/pos_15/mean": 0.00010898428736254573,
"rewards/pos_15/std": 0.0017804686212912202,
"rewards/pos_16/mean": 0.00024476498365402224,
"rewards/pos_16/std": 0.0013653157511726022,
"rewards/pos_17/mean": 0.0003747069742530584,
"rewards/pos_17/std": 0.0012490232475101948,
"rewards/pos_18/mean": -0.00012260881485417485,
"rewards/pos_18/std": 0.0002452176297083497,
"rewards/pos_19/mean": -0.00036152845714241264,
"rewards/pos_19/std": 0.0005193236982449889,
"rewards/pos_2/mean": 0.07585011683404445,
"rewards/pos_2/std": 0.06930038519203663,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.04002528414130211,
"rewards/pos_3/std": 0.057030481100082395,
"rewards/pos_4/mean": 0.026532804872840644,
"rewards/pos_4/std": 0.04809789825230837,
"rewards/pos_5/mean": 0.015283433720469475,
"rewards/pos_5/std": 0.04078430999070406,
"rewards/pos_6/mean": 0.008291334100067615,
"rewards/pos_6/std": 0.03033105507493019,
"rewards/pos_7/mean": 0.006710097620089073,
"rewards/pos_7/std": 0.024835748365148902,
"rewards/pos_8/mean": 0.003944838512688875,
"rewards/pos_8/std": 0.019218799006193877,
"rewards/pos_9/mean": 0.00030867705354467037,
"rewards/pos_9/std": 0.011737936455756427,
"step": 580
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 315.5,
"completions/max_terminated_length": 315.5,
"completions/mean_length": 256.16459197998046,
"completions/mean_terminated_length": 256.16459197998046,
"completions/min_length": 208.2,
"completions/min_terminated_length": 208.2,
"entropy": 0.20928714523712794,
"epoch": 0.3710691823899371,
"frac_reward_zero_std": 0.1916666716337204,
"grad_norm": 0.19928312301635742,
"items/detected_max": 30.0,
"items/detected_mean": 24.333333333333332,
"items/detected_min": 20.0,
"kl": 0.010333529350949297,
"learning_rate": 1e-06,
"loss": 0.0129,
"num_tokens": 16140451.0,
"reward": 0.018881635484285654,
"reward_std": 0.01549559785053134,
"reward_total": 0.15202718488872052,
"rewards/pos_1/mean": 0.1520271772518754,
"rewards/pos_1/std": 0.045391360158100726,
"rewards/pos_10/mean": -0.00017164183082059027,
"rewards/pos_10/std": 0.004313099361024797,
"rewards/pos_11/mean": -0.0007264139130711555,
"rewards/pos_11/std": 0.0036130989901721478,
"rewards/pos_12/mean": -0.0009852433227933942,
"rewards/pos_12/std": 0.0017325364518910647,
"rewards/pos_13/mean": -0.0008207797771319747,
"rewards/pos_13/std": 0.001410291320644319,
"rewards/pos_14/mean": -0.0007998688146471977,
"rewards/pos_14/std": 0.0013743613380938768,
"rewards/pos_15/mean": -0.0006510416860692203,
"rewards/pos_15/std": 0.0013020833721384406,
"rewards/pos_16/mean": -0.00038226647302508354,
"rewards/pos_16/std": 0.0007645330158993601,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.0002452176297083497,
"rewards/pos_18/std": 0.0004904352594166994,
"rewards/pos_19/mean": 0.0,
"rewards/pos_19/std": 0.0,
"rewards/pos_2/mean": 0.08341998513787985,
"rewards/pos_2/std": 0.052314735017716885,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.0601381441578269,
"rewards/pos_3/std": 0.04900498539209366,
"rewards/pos_4/mean": 0.03645331035368145,
"rewards/pos_4/std": 0.03896988406777382,
"rewards/pos_5/mean": 0.024979627598077057,
"rewards/pos_5/std": 0.03800388928502798,
"rewards/pos_6/mean": 0.01494928018655628,
"rewards/pos_6/std": 0.026825455389916895,
"rewards/pos_7/mean": 0.007466998533345759,
"rewards/pos_7/std": 0.022271700343117116,
"rewards/pos_8/mean": 0.0020789593312656508,
"rewards/pos_8/std": 0.01322312243282795,
"rewards/pos_9/mean": 0.0010265630786307155,
"rewards/pos_9/std": 0.008656562282703817,
"step": 590
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 316.6,
"completions/max_terminated_length": 316.6,
"completions/mean_length": 256.9187576293945,
"completions/mean_terminated_length": 256.9187576293945,
"completions/min_length": 200.8,
"completions/min_terminated_length": 200.8,
"entropy": 0.2289535324089229,
"epoch": 0.37735849056603776,
"frac_reward_zero_std": 0.2166666716337204,
"grad_norm": 0.11284702271223068,
"items/detected_max": 30.0,
"items/detected_mean": 24.408333333333335,
"items/detected_min": 20.0,
"kl": 0.014122579808463343,
"learning_rate": 1e-06,
"loss": 0.014,
"num_tokens": 16420668.0,
"reward": 0.020397144777234644,
"reward_std": 0.01687405025586486,
"reward_total": 0.18865224489709362,
"rewards/pos_1/mean": 0.18865224173059686,
"rewards/pos_1/std": 0.06502700969576836,
"rewards/pos_10/mean": -0.00011006370186805725,
"rewards/pos_10/std": 0.007006493420340121,
"rewards/pos_11/mean": 0.0002226763404905796,
"rewards/pos_11/std": 0.0039321394637227055,
"rewards/pos_12/mean": 0.00024987878277897834,
"rewards/pos_12/std": 0.0034018342383205892,
"rewards/pos_13/mean": -0.0002735932474024594,
"rewards/pos_13/std": 0.004693410592153668,
"rewards/pos_14/mean": -0.0006665573455393314,
"rewards/pos_14/std": 0.0013331146910786629,
"rewards/pos_15/mean": -0.0005208333488553763,
"rewards/pos_15/std": 0.0010416666977107526,
"rewards/pos_16/mean": 0.0,
"rewards/pos_16/std": 0.0,
"rewards/pos_17/mean": -0.00012490232475101948,
"rewards/pos_17/std": 0.00024980464950203897,
"rewards/pos_18/mean": -0.0002452176297083497,
"rewards/pos_18/std": 0.0004904352594166994,
"rewards/pos_19/mean": -0.00012050948571413755,
"rewards/pos_19/std": 0.0002410189714282751,
"rewards/pos_2/mean": 0.0862714304268593,
"rewards/pos_2/std": 0.05415545515716076,
"rewards/pos_20/mean": 0.0,
"rewards/pos_20/std": 0.0,
"rewards/pos_3/mean": 0.05922614808077924,
"rewards/pos_3/std": 0.0429378055036068,
"rewards/pos_4/mean": 0.04153857089113444,
"rewards/pos_4/std": 0.03988835141062737,
"rewards/pos_5/mean": 0.018643930507823824,
"rewards/pos_5/std": 0.038961370615288614,
"rewards/pos_6/mean": 0.01132265494670719,
"rewards/pos_6/std": 0.030610042996704577,
"rewards/pos_7/mean": 0.003978243132587522,
"rewards/pos_7/std": 0.021603101422078906,
"rewards/pos_8/mean": 0.0007107637298759073,
"rewards/pos_8/std": 0.01438979608938098,
"rewards/pos_9/mean": -0.0008119920123135671,
"rewards/pos_9/std": 0.007518141204491258,
"step": 600
}
],
"logging_steps": 10,
"max_steps": 3180,
"num_input_tokens_seen": 16420668,
"num_train_epochs": 2,
"save_steps": 200,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.37735849056603776,
"eval_steps": 500,
"global_step": 600,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 332.5,
"completions/max_terminated_length": 332.5,
"completions/mean_length": 279.3291778564453,
"completions/mean_terminated_length": 279.3291778564453,
"completions/min_length": 220.8,
"completions/min_terminated_length": 220.8,
"entropy": 0.24179386061926683,
"epoch": 0.006289308176100629,
"frac_reward_zero_std": 0.10000000223517418,
"grad_norm": 0.3676387071609497,
"items/detected_max": 32.0,
"items/detected_mean": 26.7125,
"items/detected_min": 21.0,
"kl": 0.00015798515381296359,
"learning_rate": 1e-06,
"loss": 0.021,
"num_tokens": 279138.0,
"reward": 0.03066593613475561,
"reward_std": 0.024809306021779776,
"reward_total": 0.19474640637636184,
"rewards/pos_1/mean": 0.19474640265107154,
"rewards/pos_1/std": 0.07326587401330471,
"rewards/pos_10/mean": 0.011071646504569798,
"rewards/pos_10/std": 0.02179219771642238,
"rewards/pos_11/mean": 0.007418489165138454,
"rewards/pos_11/std": 0.01720283292233944,
"rewards/pos_12/mean": 0.006210686359554529,
"rewards/pos_12/std": 0.0179185833549127,
"rewards/pos_13/mean": 0.0039982261369004846,
"rewards/pos_13/std": 0.009272939292714,
"rewards/pos_14/mean": 0.0012971448712050914,
"rewards/pos_14/std": 0.007776029966771602,
"rewards/pos_15/mean": -0.0006621864042244852,
"rewards/pos_15/std": 0.005226330552250147,
"rewards/pos_16/mean": 0.0002548443153500557,
"rewards/pos_16/std": 0.0015290660317987203,
"rewards/pos_17/mean": -0.00012490232475101994,
"rewards/pos_17/std": 0.0002498046495020399,
"rewards/pos_18/mean": -4.625929407134921e-19,
"rewards/pos_18/std": 9.251858814269843e-19,
"rewards/pos_19/mean": -4.625929407134921e-19,
"rewards/pos_19/std": 9.251858814269843e-19,
"rewards/pos_2/mean": 0.11575367655605077,
"rewards/pos_2/std": 0.07039788216352463,
"rewards/pos_20/mean": -4.625929407134921e-19,
"rewards/pos_20/std": 9.251858814269843e-19,
"rewards/pos_3/mean": 0.08213621266186237,
"rewards/pos_3/std": 0.05401332974433899,
"rewards/pos_4/mean": 0.059673574194312094,
"rewards/pos_4/std": 0.05429626442492008,
"rewards/pos_5/mean": 0.04313204605132341,
"rewards/pos_5/std": 0.04542485494166613,
"rewards/pos_6/mean": 0.03173917992971838,
"rewards/pos_6/std": 0.0341308044269681,
"rewards/pos_7/mean": 0.025256002554669976,
"rewards/pos_7/std": 0.03226408157497644,
"rewards/pos_8/mean": 0.016526030853856354,
"rewards/pos_8/std": 0.027043108874931932,
"rewards/pos_9/mean": 0.014891598536632956,
"rewards/pos_9/std": 0.024382113991305233,
"step": 10
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.0,
"completions/max_length": 337.5,
"completions/max_terminated_length": 337.5,
"completions/mean_length": 279.1000091552734,
"completions/mean_terminated_length": 279.1000091552734,
"completions/min_length": 221.7,
"completions/min_terminated_length": 221.7,
"entropy": 0.24853383572772145,
"epoch": 0.012578616352201259,
"frac_reward_zero_std": 0.07500000223517418,
"grad_norm": 0.3557882308959961,
"items/detected_max": 33.0,
"items/detected_mean": 26.795833333333334,
"items/detected_min": 20.0,