{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.8, "eval_steps": 500, "global_step": 200, "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": 0.86328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1944.916748046875, "completions/mean_terminated_length": 1290.1565551757812, "completions/min_length": 592.5, "completions/min_terminated_length": 592.5, "epoch": 0.004, "format_reward_mean_0": 0.3541666865348816, "format_reward_mean_1": 0.4947916865348816, "format_reward_std_0": 0.48332110047340393, "format_reward_std_1": 0.5006251335144043, "frac_reward_zero_std_0": 0.8333333730697632, "frac_reward_zero_std_1": 0.7916666865348816, "grad_norm": 0.841902494430542, "kl": 7.798136371327725, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2224401.0, "problem_reward_mean_0": 0.09294871985912323, "problem_reward_mean_1": 0.08473558723926544, "problem_reward_std_0": 0.2337280809879303, "problem_reward_std_1": 0.2371104210615158, "reward_mean_0": 0.09294871985912323, "reward_mean_1": 0.08473558723926544, "reward_std_0": 0.04427918419241905, "reward_std_1": 0.06884881854057312, "rewards/compute_reward/mean": 0.08884214609861374, "rewards/compute_reward/std": 0.23541925102472305, "step": 1, "updated_reward_mean_0": 0.16346153616905212, "updated_reward_mean_1": 0.08473558723926544, "updated_reward_std_0": 0.13840968906879425, "updated_reward_std_1": 0.06884881854057312 }, { "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": 0.42578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1436.5950927734375, "completions/mean_terminated_length": 980.1548156738281, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.008, "format_reward_mean_0": 0.6041666865348816, "format_reward_mean_1": 0.8541666865348816, "format_reward_std_0": 0.49420401453971863, "format_reward_std_1": 0.35339951515197754, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3541666865348816, "grad_norm": 4.874773025512695, "kl": -7.350268983256936e+26, "learning_rate": 9.959999999999999e-07, "loss": 0.0005, "num_tokens": 4028280.0, "problem_reward_mean_0": 0.2753787934780121, "problem_reward_mean_1": 0.45638641715049744, "problem_reward_std_0": 0.4365801215171814, "problem_reward_std_1": 0.4678189158439636, "reward_mean_0": 0.2753787934780121, "reward_mean_1": 0.45638641715049744, "reward_std_0": 0.3564454913139343, "reward_std_1": 0.23965740203857422, "rewards/compute_reward/mean": 0.36588260531425476, "rewards/compute_reward/std": 0.4521995186805725, "step": 2, "updated_reward_mean_0": 0.6870039701461792, "updated_reward_mean_1": 0.31576141715049744, "updated_reward_std_0": 0.3474935293197632, "updated_reward_std_1": 0.18749010562896729 }, { "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": 0.28125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.5, "completions/mean_length": 1262.9193115234375, "completions/mean_terminated_length": 949.5337829589844, "completions/min_length": 248.5, "completions/min_terminated_length": 248.5, "epoch": 0.012, "format_reward_mean_0": 0.7083333730697632, "format_reward_mean_1": 0.9453125, "format_reward_std_0": 0.4593396484851837, "format_reward_std_1": 0.22766587138175964, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 3.13488507270813, "kl": -7.594635596034014e+32, "learning_rate": 9.92e-07, "loss": 0.0009, "num_tokens": 5895054.0, "problem_reward_mean_0": 0.49419647455215454, "problem_reward_mean_1": 0.5184133052825928, "problem_reward_std_0": 0.4466516077518463, "problem_reward_std_1": 0.418414443731308, "reward_mean_0": 0.49419641494750977, "reward_mean_1": 0.5184133052825928, "reward_std_0": 0.2824525833129883, "reward_std_1": 0.11427055299282074, "rewards/compute_reward/mean": 0.5063048750162125, "rewards/compute_reward/std": 0.43253301084041595, "step": 3, "updated_reward_mean_0": 0.7001208066940308, "updated_reward_mean_1": 0.21112164855003357, "updated_reward_std_0": 0.13358613848686218, "updated_reward_std_1": 0.09953916072845459 }, { "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": 0.203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1263.5182495117188, "completions/mean_terminated_length": 1085.0561828613281, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.016, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9505208730697632, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.21714918315410614, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.5, "grad_norm": 0.45377445220947266, "kl": 1.0528824551440139e+34, "learning_rate": 9.88e-07, "loss": 0.0001, "num_tokens": 7539133.0, "problem_reward_mean_0": 0.5214120149612427, "problem_reward_mean_1": 0.49952977895736694, "problem_reward_std_0": 0.462083101272583, "problem_reward_std_1": 0.4486338496208191, "reward_mean_0": 0.5214120149612427, "reward_mean_1": 0.49952980875968933, "reward_std_0": 0.193294495344162, "reward_std_1": 0.14607954025268555, "rewards/compute_reward/mean": 0.5104708820581436, "rewards/compute_reward/std": 0.45535847544670105, "step": 4, "updated_reward_mean_0": 0.7748842239379883, "updated_reward_mean_1": 0.1863064467906952, "updated_reward_std_0": 0.203306645154953, "updated_reward_std_1": 0.10103434324264526 }, { "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": 0.6002604166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1719.6940307617188, "completions/mean_terminated_length": 1259.7285766601562, "completions/min_length": 607.0, "completions/min_terminated_length": 607.0, "epoch": 0.02, "format_reward_mean_0": 0.5416666865348816, "format_reward_mean_1": 0.6927083730697632, "format_reward_std_0": 0.503533661365509, "format_reward_std_1": 0.46197324991226196, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 1.2209839820861816, "kl": 2.613428976628197e+27, "learning_rate": 9.84e-07, "loss": 0.0006, "num_tokens": 9515328.0, "problem_reward_mean_0": 0.2967236638069153, "problem_reward_mean_1": 0.3892689645290375, "problem_reward_std_0": 0.4210968315601349, "problem_reward_std_1": 0.45563551783561707, "reward_mean_0": 0.2967236638069153, "reward_mean_1": 0.3892689645290375, "reward_std_0": 0.12486772239208221, "reward_std_1": 0.06906403601169586, "rewards/compute_reward/mean": 0.3429962992668152, "rewards/compute_reward/std": 0.4383661597967148, "step": 5, "updated_reward_mean_0": 0.4855174422264099, "updated_reward_mean_1": 0.18093562126159668, "updated_reward_std_0": 0.13287250697612762, "updated_reward_std_1": 0.06906403601169586 }, { "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": 0.6041666666666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1683.2096557617188, "completions/mean_terminated_length": 1149.4609375, "completions/min_length": 557.5, "completions/min_terminated_length": 557.5, "epoch": 0.024, "format_reward_mean_0": 0.5, "format_reward_mean_1": 0.7708333730697632, "format_reward_std_0": 0.5052911639213562, "format_reward_std_1": 0.42084503173828125, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.75, "grad_norm": 0.7636227607727051, "kl": -6.430788328472071e+25, "learning_rate": 9.8e-07, "loss": 0.0015, "num_tokens": 11501095.0, "problem_reward_mean_0": 0.25887346267700195, "problem_reward_mean_1": 0.3063271641731262, "problem_reward_std_0": 0.4281981289386749, "problem_reward_std_1": 0.4312785267829895, "reward_mean_0": 0.25887346267700195, "reward_mean_1": 0.3063271641731262, "reward_std_0": 0.1053781658411026, "reward_std_1": 0.053087860345840454, "rewards/compute_reward/mean": 0.2826002985239029, "rewards/compute_reward/std": 0.4297383278608322, "step": 6, "updated_reward_mean_0": 0.3545524775981903, "updated_reward_mean_1": 0.07976465672254562, "updated_reward_std_0": 0.12988942861557007, "updated_reward_std_1": 0.04572216793894768 }, { "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": 0.578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1777.01171875, "completions/mean_terminated_length": 1482.2833862304688, "completions/min_length": 847.0, "completions/min_terminated_length": 847.0, "epoch": 0.028, "format_reward_mean_0": 0.625, "format_reward_mean_1": 0.890625, "format_reward_std_0": 0.48924607038497925, "format_reward_std_1": 0.31251633167266846, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.4375, "grad_norm": 2.3020262718200684, "kl": 9.215775846708343e+27, "learning_rate": 9.759999999999998e-07, "loss": 0.0004, "num_tokens": 13662972.0, "problem_reward_mean_0": 0.22083334624767303, "problem_reward_mean_1": 0.30930155515670776, "problem_reward_std_0": 0.3978443741798401, "problem_reward_std_1": 0.41395705938339233, "reward_mean_0": 0.22083334624767303, "reward_mean_1": 0.30930155515670776, "reward_std_0": 0.11206845939159393, "reward_std_1": 0.05749429389834404, "rewards/compute_reward/mean": 0.2650674358010292, "rewards/compute_reward/std": 0.4059007316827774, "step": 7, "updated_reward_mean_0": 0.3976784944534302, "updated_reward_mean_1": 0.17779111862182617, "updated_reward_std_0": 0.10732682049274445, "updated_reward_std_1": 0.04644574970006943 }, { "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": 0.5690104166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1746.0521240234375, "completions/mean_terminated_length": 1350.8394775390625, "completions/min_length": 531.5, "completions/min_terminated_length": 531.5, "epoch": 0.032, "format_reward_mean_0": 0.7708333730697632, "format_reward_mean_1": 0.8072916865348816, "format_reward_std_0": 0.4247443974018097, "format_reward_std_1": 0.3949405550956726, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1875, "grad_norm": 3.5024256706237793, "kl": 2.5970498696942426e+34, "learning_rate": 9.72e-07, "loss": 0.0005, "num_tokens": 15814270.0, "problem_reward_mean_0": 0.2699763774871826, "problem_reward_mean_1": 0.22859656810760498, "problem_reward_std_0": 0.2971177399158478, "problem_reward_std_1": 0.28961309790611267, "reward_mean_0": 0.269976407289505, "reward_mean_1": 0.22859656810760498, "reward_std_0": 0.228095144033432, "reward_std_1": 0.1364782452583313, "rewards/compute_reward/mean": 0.2492864727973938, "rewards/compute_reward/std": 0.2933654189109802, "step": 8, "updated_reward_mean_0": 0.47874945402145386, "updated_reward_mean_1": 0.20776323974132538, "updated_reward_std_0": 0.1828046292066574, "updated_reward_std_1": 0.1364782452583313 }, { "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": 0.36588541666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.5, "completions/mean_length": 1616.68359375, "completions/mean_terminated_length": 1434.0224609375, "completions/min_length": 745.0, "completions/min_terminated_length": 745.0, "epoch": 0.036, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.1016615629196167, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 7.923409938812256, "kl": 1.7192271820481986e+35, "learning_rate": 9.679999999999999e-07, "loss": 0.0014, "num_tokens": 17880161.0, "problem_reward_mean_0": 0.3193759620189667, "problem_reward_mean_1": 0.32553499937057495, "problem_reward_std_0": 0.3705745339393616, "problem_reward_std_1": 0.35076019167900085, "reward_mean_0": 0.3193759620189667, "reward_mean_1": 0.32553499937057495, "reward_std_0": 0.0820794329047203, "reward_std_1": 0.08465129137039185, "rewards/compute_reward/mean": 0.3224554657936096, "rewards/compute_reward/std": 0.36066734790802, "step": 9, "updated_reward_mean_0": 0.4528864324092865, "updated_reward_mean_1": 0.28126415610313416, "updated_reward_std_0": 0.12283589690923691, "updated_reward_std_1": 0.07728560268878937 }, { "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": 0.51953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1672.7369995117188, "completions/mean_terminated_length": 1285.803955078125, "completions/min_length": 589.5, "completions/min_terminated_length": 589.5, "epoch": 0.04, "format_reward_mean_0": 0.7083333730697632, "format_reward_mean_1": 0.9192708730697632, "format_reward_std_0": 0.4593396484851837, "format_reward_std_1": 0.2727741599082947, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4375, "grad_norm": 1.367235779762268, "kl": -6.165852519510532e+33, "learning_rate": 9.64e-07, "loss": 0.0004, "num_tokens": 19924204.0, "problem_reward_mean_0": 0.28016382455825806, "problem_reward_mean_1": 0.4327101409435272, "problem_reward_std_0": 0.40337231755256653, "problem_reward_std_1": 0.3946017920970917, "reward_mean_0": 0.28016382455825806, "reward_mean_1": 0.43271011114120483, "reward_std_0": 0.22735214233398438, "reward_std_1": 0.12387458980083466, "rewards/compute_reward/mean": 0.35643696784973145, "rewards/compute_reward/std": 0.3989870548248291, "step": 10, "updated_reward_mean_0": 0.5926549434661865, "updated_reward_mean_1": 0.2653743028640747, "updated_reward_std_0": 0.2570040225982666, "updated_reward_std_1": 0.1072855293750763 }, { "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": 0.30078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1515.5729370117188, "completions/mean_terminated_length": 1309.2459106445312, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.044, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.953125, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.21164682507514954, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.625, "grad_norm": 0.8776369094848633, "kl": 6.76082888364009e+31, "learning_rate": 9.6e-07, "loss": 0.0006, "num_tokens": 21897761.0, "problem_reward_mean_0": 0.4664939045906067, "problem_reward_mean_1": 0.45876893401145935, "problem_reward_std_0": 0.4135840833187103, "problem_reward_std_1": 0.39246559143066406, "reward_mean_0": 0.4664939045906067, "reward_mean_1": 0.45876893401145935, "reward_std_0": 0.12313437461853027, "reward_std_1": 0.09324336051940918, "rewards/compute_reward/mean": 0.4626314043998718, "rewards/compute_reward/std": 0.403024822473526, "step": 11, "updated_reward_mean_0": 0.5429162979125977, "updated_reward_mean_1": 0.24001893401145935, "updated_reward_std_0": 0.0960693508386612, "updated_reward_std_1": 0.052390895783901215 }, { "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": 0.3346354166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1484.530029296875, "completions/mean_terminated_length": 1275.1168823242188, "completions/min_length": 686.5, "completions/min_terminated_length": 686.5, "epoch": 0.048, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.1016615703701973, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 1.71711003780365, "kl": 1.0273182964901654e+27, "learning_rate": 9.559999999999998e-07, "loss": 0.0005, "num_tokens": 23901579.0, "problem_reward_mean_0": 0.46433913707733154, "problem_reward_mean_1": 0.5002963542938232, "problem_reward_std_0": 0.3845457434654236, "problem_reward_std_1": 0.36158451437950134, "reward_mean_0": 0.46433913707733154, "reward_mean_1": 0.5002963542938232, "reward_std_0": 0.310366690158844, "reward_std_1": 0.0930960401892662, "rewards/compute_reward/mean": 0.4823177456855774, "rewards/compute_reward/std": 0.37306512892246246, "step": 12, "updated_reward_mean_0": 0.6315736770629883, "updated_reward_mean_1": 0.3133750855922699, "updated_reward_std_0": 0.2558005452156067, "updated_reward_std_1": 0.07836464047431946 }, { "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": 0.22005208333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.5, "completions/mean_length": 1332.375, "completions/mean_terminated_length": 1137.5548095703125, "completions/min_length": 394.5, "completions/min_terminated_length": 394.5, "epoch": 0.052, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.0720745176076889, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 1.1444171667099, "kl": 3.7633376125184256e+29, "learning_rate": 9.52e-07, "loss": 0.0005, "num_tokens": 25768279.0, "problem_reward_mean_0": 0.555891215801239, "problem_reward_mean_1": 0.5619126558303833, "problem_reward_std_0": 0.3378259539604187, "problem_reward_std_1": 0.3490046262741089, "reward_mean_0": 0.555891215801239, "reward_mean_1": 0.5619126558303833, "reward_std_0": 0.12017383426427841, "reward_std_1": 0.06366707384586334, "rewards/compute_reward/mean": 0.5589019060134888, "rewards/compute_reward/std": 0.3434152752161026, "step": 13, "updated_reward_mean_0": 0.6513657569885254, "updated_reward_mean_1": 0.2910792827606201, "updated_reward_std_0": 0.1307113766670227, "updated_reward_std_1": 0.06366708129644394 }, { "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": 0.3658854166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1638.8125610351562, "completions/mean_terminated_length": 1461.5996704101562, "completions/min_length": 741.0, "completions/min_terminated_length": 741.0, "epoch": 0.056, "format_reward_mean_0": 1.0, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.5416666865348816, "grad_norm": 0.6270292401313782, "kl": 2.807394403501668e+25, "learning_rate": 9.479999999999999e-07, "loss": 0.0004, "num_tokens": 27797115.0, "problem_reward_mean_0": 0.5022467374801636, "problem_reward_mean_1": 0.550155758857727, "problem_reward_std_0": 0.4034174978733063, "problem_reward_std_1": 0.39027631282806396, "reward_mean_0": 0.5022467374801636, "reward_mean_1": 0.550155758857727, "reward_std_0": 0.2953236699104309, "reward_std_1": 0.07374735176563263, "rewards/compute_reward/mean": 0.5262012481689453, "rewards/compute_reward/std": 0.3968469351530075, "step": 14, "updated_reward_mean_0": 0.6360578536987305, "updated_reward_mean_1": 0.30308541655540466, "updated_reward_std_0": 0.22180214524269104, "updated_reward_std_1": 0.06546095013618469 }, { "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": 0.4036458333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.5, "completions/mean_length": 1566.979248046875, "completions/mean_terminated_length": 1291.3994750976562, "completions/min_length": 656.5, "completions/min_terminated_length": 656.5, "epoch": 0.06, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.15946698188781738, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 1.3812147378921509, "kl": 3.290063613994549e+32, "learning_rate": 9.439999999999999e-07, "loss": 0.0005, "num_tokens": 29841209.0, "problem_reward_mean_0": 0.46226954460144043, "problem_reward_mean_1": 0.4148913025856018, "problem_reward_std_0": 0.4359199106693268, "problem_reward_std_1": 0.4364767372608185, "reward_mean_0": 0.46226951479911804, "reward_mean_1": 0.4148913025856018, "reward_std_0": 0.1892150640487671, "reward_std_1": 0.14490483701229095, "rewards/compute_reward/mean": 0.4385804086923599, "rewards/compute_reward/std": 0.4361983388662338, "step": 15, "updated_reward_mean_0": 0.5978672504425049, "updated_reward_mean_1": 0.12886697053909302, "updated_reward_std_0": 0.1971031278371811, "updated_reward_std_1": 0.08642251789569855 }, { "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": 0.38151041666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.5, "completions/mean_length": 1501.9921875, "completions/mean_terminated_length": 1217.42529296875, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.064, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 2.072653293609619, "kl": 1.1824993414867342e+33, "learning_rate": 9.399999999999999e-07, "loss": 0.0008, "num_tokens": 31702697.0, "problem_reward_mean_0": 0.34278082847595215, "problem_reward_mean_1": 0.35962778329849243, "problem_reward_std_0": 0.40647101402282715, "problem_reward_std_1": 0.412203311920166, "reward_mean_0": 0.34278082847595215, "reward_mean_1": 0.35962778329849243, "reward_std_0": 0.1613975167274475, "reward_std_1": 0.07650136202573776, "rewards/compute_reward/mean": 0.3512042909860611, "rewards/compute_reward/std": 0.4093371480703354, "step": 16, "updated_reward_mean_0": 0.5239902138710022, "updated_reward_mean_1": 0.21379445493221283, "updated_reward_std_0": 0.17876403033733368, "updated_reward_std_1": 0.07650136202573776 }, { "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": 0.44010416666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.5, "completions/mean_length": 1619.1784057617188, "completions/mean_terminated_length": 1312.1194458007812, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.068, "format_reward_mean_0": 0.6458333730697632, "format_reward_mean_1": 0.9010416865348816, "format_reward_std_0": 0.48332110047340393, "format_reward_std_1": 0.29899564385414124, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.5208333730697632, "grad_norm": 2.0686347484588623, "kl": 1.2079281311203183e+33, "learning_rate": 9.36e-07, "loss": 0.0005, "num_tokens": 33675839.0, "problem_reward_mean_0": 0.2712808847427368, "problem_reward_mean_1": 0.37453925609588623, "problem_reward_std_0": 0.4479843080043793, "problem_reward_std_1": 0.44412103295326233, "reward_mean_0": 0.2712808847427368, "reward_mean_1": 0.37453925609588623, "reward_std_0": 0.0897064357995987, "reward_std_1": 0.11462261527776718, "rewards/compute_reward/mean": 0.32291005551815033, "rewards/compute_reward/std": 0.4460526704788208, "step": 17, "updated_reward_mean_0": 0.4944973886013031, "updated_reward_mean_1": 0.14006780087947845, "updated_reward_std_0": 0.17996369302272797, "updated_reward_std_1": 0.08087611198425293 }, { "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": 0.6731770833333334, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1753.0755615234375, "completions/mean_terminated_length": 1137.747314453125, "completions/min_length": 493.5, "completions/min_terminated_length": 493.5, "epoch": 0.072, "format_reward_mean_0": 0.5416666865348816, "format_reward_mean_1": 0.7890625, "format_reward_std_0": 0.503533661365509, "format_reward_std_1": 0.4085063934326172, "frac_reward_zero_std_0": 0.8333333730697632, "frac_reward_zero_std_1": 0.75, "grad_norm": 0.43813833594322205, "kl": 3.648555324525991e+26, "learning_rate": 9.32e-07, "loss": 0.0012, "num_tokens": 35694893.0, "problem_reward_mean_0": 0.18256579339504242, "problem_reward_mean_1": 0.19048607349395752, "problem_reward_std_0": 0.3732054829597473, "problem_reward_std_1": 0.3770767152309418, "reward_mean_0": 0.18256579339504242, "reward_mean_1": 0.19048607349395752, "reward_std_0": 0.017057400196790695, "reward_std_1": 0.03014661744236946, "rewards/compute_reward/mean": 0.18652593344449997, "rewards/compute_reward/std": 0.37514108419418335, "step": 18, "updated_reward_mean_0": 0.25450393557548523, "updated_reward_mean_1": 0.02381940186023712, "updated_reward_std_0": 0.12423785775899887, "updated_reward_std_1": 0.03014661744236946 }, { "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": 0.9127604166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 2000.7487182617188, "completions/mean_terminated_length": 1490.9755249023438, "completions/min_length": 806.0, "completions/min_terminated_length": 806.0, "epoch": 0.076, "format_reward_mean_0": 0.3333333432674408, "format_reward_mean_1": 0.6119791865348816, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.4879351258277893, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 1.5325229167938232, "kl": 3.853657821104819e+32, "learning_rate": 9.28e-07, "loss": 0.0006, "num_tokens": 37866463.0, "problem_reward_mean_0": 0.0364583358168602, "problem_reward_mean_1": 0.17470800876617432, "problem_reward_std_0": 0.1629970520734787, "problem_reward_std_1": 0.29387137293815613, "reward_mean_0": 0.0364583358168602, "reward_mean_1": 0.17470800876617432, "reward_std_0": 0.08992656320333481, "reward_std_1": 0.1284872591495514, "rewards/compute_reward/mean": 0.10558317042887211, "rewards/compute_reward/std": 0.228434219956398, "step": 19, "updated_reward_mean_0": 0.33308079838752747, "updated_reward_mean_1": 0.1551767736673355, "updated_reward_std_0": 0.14148420095443726, "updated_reward_std_1": 0.12669020891189575 }, { "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": 0.7604166666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1838.7513427734375, "completions/mean_terminated_length": 1162.0750122070312, "completions/min_length": 524.0, "completions/min_terminated_length": 524.0, "epoch": 0.08, "format_reward_mean_0": 0.4166666865348816, "format_reward_mean_1": 0.6171875, "format_reward_std_0": 0.49822381138801575, "format_reward_std_1": 0.48670727014541626, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.375, "grad_norm": 3.126662254333496, "kl": 4.034914321541098e+32, "learning_rate": 9.24e-07, "loss": 0.0009, "num_tokens": 39982519.0, "problem_reward_mean_0": 0.13638266921043396, "problem_reward_mean_1": 0.2074424922466278, "problem_reward_std_0": 0.31107404828071594, "problem_reward_std_1": 0.33535411953926086, "reward_mean_0": 0.13638266921043396, "reward_mean_1": 0.2074424922466278, "reward_std_0": 0.1358167976140976, "reward_std_1": 0.09339770674705505, "rewards/compute_reward/mean": 0.1719125658273697, "rewards/compute_reward/std": 0.3232140839099884, "step": 20, "updated_reward_mean_0": 0.35848915576934814, "updated_reward_mean_1": 0.10327579826116562, "updated_reward_std_0": 0.20424219965934753, "updated_reward_std_1": 0.09339770674705505 }, { "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": 0.4583333333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.5, "completions/mean_length": 1552.7591552734375, "completions/mean_terminated_length": 1168.71875, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.084, "format_reward_mean_0": 0.75, "format_reward_mean_1": 0.921875, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.26871830224990845, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 0.5133768916130066, "kl": -4.176670691251141e+35, "learning_rate": 9.2e-07, "loss": 0.0006, "num_tokens": 42047715.0, "problem_reward_mean_0": 0.317069411277771, "problem_reward_mean_1": 0.42149874567985535, "problem_reward_std_0": 0.4191499352455139, "problem_reward_std_1": 0.4369943141937256, "reward_mean_0": 0.317069411277771, "reward_mean_1": 0.42149877548217773, "reward_std_0": 0.22971075773239136, "reward_std_1": 0.11616530269384384, "rewards/compute_reward/mean": 0.36928407847881317, "rewards/compute_reward/std": 0.42807209491729736, "step": 21, "updated_reward_mean_0": 0.5768320560455322, "updated_reward_mean_1": 0.23877307772636414, "updated_reward_std_0": 0.24347564578056335, "updated_reward_std_1": 0.10727642476558685 }, { "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": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1969.0, "completions/mean_length": 1906.69921875, "completions/mean_terminated_length": 1500.2936401367188, "completions/min_length": 726.5, "completions/min_terminated_length": 726.5, "epoch": 0.088, "format_reward_mean_0": 0.5208333730697632, "format_reward_mean_1": 0.8411458730697632, "format_reward_std_0": 0.5048523545265198, "format_reward_std_1": 0.3660169243812561, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0833333358168602, "grad_norm": 1.3395600318908691, "kl": 3.0627492959926515e+34, "learning_rate": 9.16e-07, "loss": 0.0013, "num_tokens": 44243880.0, "problem_reward_mean_0": 0.16141358017921448, "problem_reward_mean_1": 0.2924758791923523, "problem_reward_std_0": 0.261919230222702, "problem_reward_std_1": 0.3009074330329895, "reward_mean_0": 0.16141358017921448, "reward_mean_1": 0.2924758791923523, "reward_std_0": 0.2085869014263153, "reward_std_1": 0.1719118058681488, "rewards/compute_reward/mean": 0.2269447222352028, "rewards/compute_reward/std": 0.28141331672668457, "step": 22, "updated_reward_mean_0": 0.5175771713256836, "updated_reward_mean_1": 0.2764788568019867, "updated_reward_std_0": 0.1667831391096115, "updated_reward_std_1": 0.1648084819316864 }, { "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": 0.4830729166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1646.608154296875, "completions/mean_terminated_length": 1277.0567016601562, "completions/min_length": 439.5, "completions/min_terminated_length": 439.5, "epoch": 0.092, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.8984375, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.3024662137031555, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.6041666865348816, "grad_norm": 1.1606178283691406, "kl": 9.865364031216154e+34, "learning_rate": 9.12e-07, "loss": 0.0006, "num_tokens": 46330032.0, "problem_reward_mean_0": 0.3260616958141327, "problem_reward_mean_1": 0.24213413894176483, "problem_reward_std_0": 0.45776891708374023, "problem_reward_std_1": 0.3715856075286865, "reward_mean_0": 0.3260616958141327, "reward_mean_1": 0.24213413894176483, "reward_std_0": 0.1304633766412735, "reward_std_1": 0.09449788182973862, "rewards/compute_reward/mean": 0.28409791737794876, "rewards/compute_reward/std": 0.4146772474050522, "step": 23, "updated_reward_mean_0": 0.3793705403804779, "updated_reward_mean_1": 0.09501218795776367, "updated_reward_std_0": 0.23230503499507904, "updated_reward_std_1": 0.05236729606986046 }, { "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": 0.48437499999999994, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1676.6797485351562, "completions/mean_terminated_length": 1358.755859375, "completions/min_length": 294.5, "completions/min_terminated_length": 294.5, "epoch": 0.096, "format_reward_mean_0": 0.7708333730697632, "format_reward_mean_1": 0.9140625, "format_reward_std_0": 0.4247443974018097, "format_reward_std_1": 0.28063738346099854, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 1.5191932916641235, "kl": -7.15233207451259e+24, "learning_rate": 9.08e-07, "loss": 0.0009, "num_tokens": 48454666.0, "problem_reward_mean_0": 0.3680555522441864, "problem_reward_mean_1": 0.3703894317150116, "problem_reward_std_0": 0.3973413109779358, "problem_reward_std_1": 0.3851271867752075, "reward_mean_0": 0.3680555522441864, "reward_mean_1": 0.3703894019126892, "reward_std_0": 0.3191315531730652, "reward_std_1": 0.16095389425754547, "rewards/compute_reward/mean": 0.369222491979599, "rewards/compute_reward/std": 0.39123424887657166, "step": 24, "updated_reward_mean_0": 0.6233074069023132, "updated_reward_mean_1": 0.20398545265197754, "updated_reward_std_0": 0.26643431186676025, "updated_reward_std_1": 0.12203611433506012 }, { "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": 0.6002604166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.5, "completions/mean_length": 1772.9544677734375, "completions/mean_terminated_length": 1380.35400390625, "completions/min_length": 759.5, "completions/min_terminated_length": 759.5, "epoch": 0.1, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9244791865348816, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.2645746171474457, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.6041666865348816, "grad_norm": 1.5432109832763672, "kl": -1.3902714556936954e+25, "learning_rate": 9.039999999999999e-07, "loss": 0.0015, "num_tokens": 50595187.0, "problem_reward_mean_0": 0.11336805671453476, "problem_reward_mean_1": 0.185546875, "problem_reward_std_0": 0.22890129685401917, "problem_reward_std_1": 0.308279812335968, "reward_mean_0": 0.11336805671453476, "reward_mean_1": 0.185546875, "reward_std_0": 0.16659682989120483, "reward_std_1": 0.08700822293758392, "rewards/compute_reward/mean": 0.14945746585726738, "rewards/compute_reward/std": 0.2685905396938324, "step": 25, "updated_reward_mean_0": 0.31388890743255615, "updated_reward_mean_1": 0.185546875, "updated_reward_std_0": 0.1517227590084076, "updated_reward_std_1": 0.08700822293758392 }, { "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": 0.41015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.5, "completions/mean_length": 1509.0091552734375, "completions/mean_terminated_length": 1150.7867431640625, "completions/min_length": 450.5, "completions/min_terminated_length": 450.5, "epoch": 0.104, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9348958730697632, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.24703142046928406, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.5416666865348816, "grad_norm": 1.7765531539916992, "kl": 2.7944653229914753e+33, "learning_rate": 9e-07, "loss": 0.0008, "num_tokens": 52481700.0, "problem_reward_mean_0": 0.3567708432674408, "problem_reward_mean_1": 0.4748264253139496, "problem_reward_std_0": 0.41603732109069824, "problem_reward_std_1": 0.37061572074890137, "reward_mean_0": 0.3567708432674408, "reward_mean_1": 0.4748264253139496, "reward_std_0": 0.1711132824420929, "reward_std_1": 0.08577477186918259, "rewards/compute_reward/mean": 0.4157986342906952, "rewards/compute_reward/std": 0.3933264911174774, "step": 26, "updated_reward_mean_0": 0.5477430820465088, "updated_reward_mean_1": 0.3394097685813904, "updated_reward_std_0": 0.1005217581987381, "updated_reward_std_1": 0.0613994225859642 }, { "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": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.5, "completions/mean_length": 1784.5690307617188, "completions/mean_terminated_length": 1412.2205200195312, "completions/min_length": 682.0, "completions/min_terminated_length": 682.0, "epoch": 0.108, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9036458730697632, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.2954613268375397, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 2.756042003631592, "kl": 4.932492143115746e+34, "learning_rate": 8.96e-07, "loss": 0.0011, "num_tokens": 54587469.0, "problem_reward_mean_0": 0.26516205072402954, "problem_reward_mean_1": 0.38288483023643494, "problem_reward_std_0": 0.4160582423210144, "problem_reward_std_1": 0.4509820342063904, "reward_mean_0": 0.26516205072402954, "reward_mean_1": 0.38288483023643494, "reward_std_0": 0.31572651863098145, "reward_std_1": 0.14094771444797516, "rewards/compute_reward/mean": 0.32402342557907104, "rewards/compute_reward/std": 0.4335201233625412, "step": 27, "updated_reward_mean_0": 0.5594907999038696, "updated_reward_mean_1": 0.19393810629844666, "updated_reward_std_0": 0.2731073200702667, "updated_reward_std_1": 0.11973695456981659 }, { "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": 0.50390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.5, "completions/mean_length": 1536.96484375, "completions/mean_terminated_length": 1030.1478881835938, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.112, "format_reward_mean_0": 0.6875, "format_reward_mean_1": 0.8046875, "format_reward_std_0": 0.4684174358844757, "format_reward_std_1": 0.39695829153060913, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.5, "grad_norm": 0.4476783275604248, "kl": 3.692239774899759e+35, "learning_rate": 8.92e-07, "loss": 0.0005, "num_tokens": 56469968.0, "problem_reward_mean_0": 0.41650885343551636, "problem_reward_mean_1": 0.5418977737426758, "problem_reward_std_0": 0.4598112106323242, "problem_reward_std_1": 0.44044870138168335, "reward_mean_0": 0.41650885343551636, "reward_mean_1": 0.5418977737426758, "reward_std_0": 0.1821129024028778, "reward_std_1": 0.10024255514144897, "rewards/compute_reward/mean": 0.4792032986879349, "rewards/compute_reward/std": 0.450129970908165, "step": 28, "updated_reward_mean_0": 0.6498016119003296, "updated_reward_mean_1": 0.1986212581396103, "updated_reward_std_0": 0.0973130315542221, "updated_reward_std_1": 0.08954077959060669 }, { "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": 0.3958333333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1653.7174682617188, "completions/mean_terminated_length": 1410.953857421875, "completions/min_length": 583.5, "completions/min_terminated_length": 583.5, "epoch": 0.116, "format_reward_mean_0": 0.8125, "format_reward_mean_1": 0.9036458730697632, "format_reward_std_0": 0.3944427967071533, "format_reward_std_1": 0.2954613268375397, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3125, "grad_norm": 0.8289758563041687, "kl": 1.9267819469220675e+32, "learning_rate": 8.88e-07, "loss": 0.0008, "num_tokens": 58415565.0, "problem_reward_mean_0": 0.3445204198360443, "problem_reward_mean_1": 0.3840733766555786, "problem_reward_std_0": 0.40211057662963867, "problem_reward_std_1": 0.4065639078617096, "reward_mean_0": 0.3445204198360443, "reward_mean_1": 0.3840733766555786, "reward_std_0": 0.16157305240631104, "reward_std_1": 0.16484598815441132, "rewards/compute_reward/mean": 0.36429689824581146, "rewards/compute_reward/std": 0.40433724224567413, "step": 29, "updated_reward_mean_0": 0.5434834361076355, "updated_reward_mean_1": 0.195271298289299, "updated_reward_std_0": 0.08039305359125137, "updated_reward_std_1": 0.09835918992757797 }, { "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": 0.5091145833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1639.6524047851562, "completions/mean_terminated_length": 1229.7548828125, "completions/min_length": 533.0, "completions/min_terminated_length": 533.0, "epoch": 0.12, "format_reward_mean_0": 0.75, "format_reward_mean_1": 0.9036458730697632, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.2954612970352173, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 2.1328725814819336, "kl": -41781411796284.59, "learning_rate": 8.839999999999999e-07, "loss": 0.001, "num_tokens": 60432918.0, "problem_reward_mean_0": 0.3567160367965698, "problem_reward_mean_1": 0.27230411767959595, "problem_reward_std_0": 0.4286772608757019, "problem_reward_std_1": 0.37302616238594055, "reward_mean_0": 0.3567160367965698, "reward_mean_1": 0.27230408787727356, "reward_std_0": 0.21640139818191528, "reward_std_1": 0.117093026638031, "rewards/compute_reward/mean": 0.3145100623369217, "rewards/compute_reward/std": 0.4008517265319824, "step": 30, "updated_reward_mean_0": 0.5628318786621094, "updated_reward_mean_1": 0.1467219889163971, "updated_reward_std_0": 0.1827128827571869, "updated_reward_std_1": 0.09581160545349121 }, { "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": 0.3671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1585.4570922851562, "completions/mean_terminated_length": 1329.8981323242188, "completions/min_length": 618.5, "completions/min_terminated_length": 618.5, "epoch": 0.124, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9453125, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.22766588628292084, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 2.7680063247680664, "kl": -1.5474250500488064e+27, "learning_rate": 8.799999999999999e-07, "loss": 0.0006, "num_tokens": 62490494.0, "problem_reward_mean_0": 0.4476562440395355, "problem_reward_mean_1": 0.49916452169418335, "problem_reward_std_0": 0.4208187758922577, "problem_reward_std_1": 0.41111353039741516, "reward_mean_0": 0.4476562440395355, "reward_mean_1": 0.49916452169418335, "reward_std_0": 0.11623167991638184, "reward_std_1": 0.15232595801353455, "rewards/compute_reward/mean": 0.47341036796569824, "rewards/compute_reward/std": 0.4159661680459976, "step": 31, "updated_reward_mean_0": 0.6457465887069702, "updated_reward_mean_1": 0.27954643964767456, "updated_reward_std_0": 0.08470186591148376, "updated_reward_std_1": 0.10934051126241684 }, { "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": 0.5533854166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.5, "completions/mean_length": 1654.4323120117188, "completions/mean_terminated_length": 1168.9890747070312, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.128, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.8567708730697632, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.3507634401321411, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.6348496675491333, "kl": -1.240475780637809e+34, "learning_rate": 8.76e-07, "loss": 0.0007, "num_tokens": 64461018.0, "problem_reward_mean_0": 0.2872109115123749, "problem_reward_mean_1": 0.26304107904434204, "problem_reward_std_0": 0.4071270525455475, "problem_reward_std_1": 0.3937910497188568, "reward_mean_0": 0.2872109115123749, "reward_mean_1": 0.26304107904434204, "reward_std_0": 0.1255507469177246, "reward_std_1": 0.07778667658567429, "rewards/compute_reward/mean": 0.27512598037719727, "rewards/compute_reward/std": 0.40045905113220215, "step": 32, "updated_reward_mean_0": 0.3842800259590149, "updated_reward_mean_1": 0.05731191858649254, "updated_reward_std_0": 0.17795154452323914, "updated_reward_std_1": 0.05568958446383476 }, { "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": 0.3854166666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.5, "completions/mean_length": 1540.2955932617188, "completions/mean_terminated_length": 1223.0338745117188, "completions/min_length": 520.5, "completions/min_terminated_length": 520.5, "epoch": 0.132, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9348958730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.24703142046928406, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4375, "grad_norm": 0.648619532585144, "kl": -9.702293772076356e+28, "learning_rate": 8.72e-07, "loss": 0.0008, "num_tokens": 66379693.0, "problem_reward_mean_0": 0.4518890976905823, "problem_reward_mean_1": 0.39619773626327515, "problem_reward_std_0": 0.4156479239463806, "problem_reward_std_1": 0.45329463481903076, "reward_mean_0": 0.4518890976905823, "reward_mean_1": 0.39619773626327515, "reward_std_0": 0.21837925910949707, "reward_std_1": 0.11740522086620331, "rewards/compute_reward/mean": 0.4240433871746063, "rewards/compute_reward/std": 0.4344712942838669, "step": 33, "updated_reward_mean_0": 0.5527971982955933, "updated_reward_mean_1": 0.15996262431144714, "updated_reward_std_0": 0.06121836602687836, "updated_reward_std_1": 0.0846874862909317 }, { "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": 0.4830729166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.5, "completions/mean_length": 1714.4036865234375, "completions/mean_terminated_length": 1404.8721923828125, "completions/min_length": 612.5, "completions/min_terminated_length": 612.5, "epoch": 0.136, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.890625, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.31251633167266846, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 0.7488571405410767, "kl": 4.670658636540978e+34, "learning_rate": 8.68e-07, "loss": 0.001, "num_tokens": 68597673.0, "problem_reward_mean_0": 0.1608167290687561, "problem_reward_mean_1": 0.12503518164157867, "problem_reward_std_0": 0.26600363850593567, "problem_reward_std_1": 0.24425163865089417, "reward_mean_0": 0.1608167290687561, "reward_mean_1": 0.12503518164157867, "reward_std_0": 0.15675964951515198, "reward_std_1": 0.041345302015542984, "rewards/compute_reward/mean": 0.1429259479045868, "rewards/compute_reward/std": 0.2551276460289955, "step": 34, "updated_reward_mean_0": 0.23412013053894043, "updated_reward_mean_1": 0.10458765923976898, "updated_reward_std_0": 0.15767638385295868, "updated_reward_std_1": 0.04093286022543907 }, { "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": 0.6341145833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.5, "completions/mean_length": 1819.6406860351562, "completions/mean_terminated_length": 1457.5316162109375, "completions/min_length": 832.0, "completions/min_terminated_length": 832.0, "epoch": 0.14, "format_reward_mean_0": 0.5833333730697632, "format_reward_mean_1": 0.9166666865348816, "format_reward_std_0": 0.49822381138801575, "format_reward_std_1": 0.27674600481987, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.5625, "grad_norm": 0.8160384297370911, "kl": 1.9686687939761257e+27, "learning_rate": 8.639999999999999e-07, "loss": 0.0009, "num_tokens": 70711050.0, "problem_reward_mean_0": 0.12534722685813904, "problem_reward_mean_1": 0.18177084624767303, "problem_reward_std_0": 0.2426619678735733, "problem_reward_std_1": 0.27367696166038513, "reward_mean_0": 0.12534722685813904, "reward_mean_1": 0.18177084624767303, "reward_std_0": 0.09210842102766037, "reward_std_1": 0.06698199361562729, "rewards/compute_reward/mean": 0.15355903655290604, "rewards/compute_reward/std": 0.258169449865818, "step": 35, "updated_reward_mean_0": 0.2414351999759674, "updated_reward_mean_1": 0.18177084624767303, "updated_reward_std_0": 0.037251394242048264, "updated_reward_std_1": 0.06698198616504669 }, { "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": 0.4140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1538.0430297851562, "completions/mean_terminated_length": 1192.973388671875, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.144, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9427083730697632, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.23270228505134583, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.6666666865348816, "grad_norm": 0.047207485884428024, "kl": -1.2331705540285776e+34, "learning_rate": 8.599999999999999e-07, "loss": 0.0007, "num_tokens": 72556234.0, "problem_reward_mean_0": 0.4305555820465088, "problem_reward_mean_1": 0.4240451455116272, "problem_reward_std_0": 0.468680202960968, "problem_reward_std_1": 0.46841391921043396, "reward_mean_0": 0.4305555820465088, "reward_mean_1": 0.4240451455116272, "reward_std_0": 0.15039034187793732, "reward_std_1": 0.09686611592769623, "rewards/compute_reward/mean": 0.4273003488779068, "rewards/compute_reward/std": 0.468547061085701, "step": 36, "updated_reward_mean_0": 0.5069444179534912, "updated_reward_mean_1": 0.1453993022441864, "updated_reward_std_0": 0.18892744183540344, "updated_reward_std_1": 0.06003763526678085 }, { "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": 0.45963541666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.5, "completions/mean_length": 1611.1575927734375, "completions/mean_terminated_length": 1271.5896606445312, "completions/min_length": 622.0, "completions/min_terminated_length": 622.0, "epoch": 0.148, "format_reward_mean_0": 0.7708333730697632, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.4247443974018097, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 3.5197649002075195, "kl": -2.9437839101250847e+31, "learning_rate": 8.559999999999999e-07, "loss": 0.001, "num_tokens": 74621331.0, "problem_reward_mean_0": 0.23007825016975403, "problem_reward_mean_1": 0.3218853175640106, "problem_reward_std_0": 0.34453606605529785, "problem_reward_std_1": 0.3551006019115448, "reward_mean_0": 0.23007825016975403, "reward_mean_1": 0.3218853175640106, "reward_std_0": 0.21284472942352295, "reward_std_1": 0.1172729879617691, "rewards/compute_reward/mean": 0.2759817764163017, "rewards/compute_reward/std": 0.3498183339834213, "step": 37, "updated_reward_mean_0": 0.4763135612010956, "updated_reward_mean_1": 0.20296171307563782, "updated_reward_std_0": 0.16361309587955475, "updated_reward_std_1": 0.10517378896474838 }, { "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": 0.4817708333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1652.2330932617188, "completions/mean_terminated_length": 1344.9456176757812, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.152, "format_reward_mean_0": 0.625, "format_reward_mean_1": 0.9817708730697632, "format_reward_std_0": 0.48924607038497925, "format_reward_std_1": 0.13395370543003082, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2291666716337204, "grad_norm": 1.3025144338607788, "kl": -7.253531947022663e+33, "learning_rate": 8.52e-07, "loss": 0.0011, "num_tokens": 76596521.0, "problem_reward_mean_0": 0.31863096356391907, "problem_reward_mean_1": 0.3997085988521576, "problem_reward_std_0": 0.35481682419776917, "problem_reward_std_1": 0.2967737317085266, "reward_mean_0": 0.31863096356391907, "reward_mean_1": 0.3997085988521576, "reward_std_0": 0.27501600980758667, "reward_std_1": 0.14819657802581787, "rewards/compute_reward/mean": 0.35916978120803833, "rewards/compute_reward/std": 0.3257952630519867, "step": 38, "updated_reward_mean_0": 0.5875396728515625, "updated_reward_mean_1": 0.393632173538208, "updated_reward_std_0": 0.13475559651851654, "updated_reward_std_1": 0.14131441712379456 }, { "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": 0.44010416666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.5, "completions/mean_length": 1621.8229370117188, "completions/mean_terminated_length": 1308.2723388671875, "completions/min_length": 617.5, "completions/min_terminated_length": 617.5, "epoch": 0.156, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9557291865348816, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.20596455037593842, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 0.6933267712593079, "kl": 4.431243574940188e+35, "learning_rate": 8.48e-07, "loss": 0.0008, "num_tokens": 78619232.0, "problem_reward_mean_0": 0.3569924831390381, "problem_reward_mean_1": 0.5057255029678345, "problem_reward_std_0": 0.4124282896518707, "problem_reward_std_1": 0.3915742039680481, "reward_mean_0": 0.3569924831390381, "reward_mean_1": 0.5057255029678345, "reward_std_0": 0.29860803484916687, "reward_std_1": 0.1031903624534607, "rewards/compute_reward/mean": 0.4313589632511139, "rewards/compute_reward/std": 0.4020012468099594, "step": 39, "updated_reward_mean_0": 0.6131981611251831, "updated_reward_mean_1": 0.3631473481655121, "updated_reward_std_0": 0.17164501547813416, "updated_reward_std_1": 0.0939832478761673 }, { "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": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.5, "completions/mean_length": 1805.9102172851562, "completions/mean_terminated_length": 1440.7766723632812, "completions/min_length": 708.0, "completions/min_terminated_length": 708.0, "epoch": 0.16, "format_reward_mean_0": 0.4166666865348816, "format_reward_mean_1": 0.9505208730697632, "format_reward_std_0": 0.49822381138801575, "format_reward_std_1": 0.21714918315410614, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4375, "grad_norm": 0.931061327457428, "kl": 9.186976548488345e+28, "learning_rate": 8.439999999999999e-07, "loss": 0.0009, "num_tokens": 80675774.0, "problem_reward_mean_0": 0.13359789550304413, "problem_reward_mean_1": 0.323203980922699, "problem_reward_std_0": 0.29371392726898193, "problem_reward_std_1": 0.33973267674446106, "reward_mean_0": 0.13359789550304413, "reward_mean_1": 0.323203980922699, "reward_std_0": 0.2616574168205261, "reward_std_1": 0.12076856940984726, "rewards/compute_reward/mean": 0.22840092331171036, "rewards/compute_reward/std": 0.3167233020067215, "step": 40, "updated_reward_mean_0": 0.43468913435935974, "updated_reward_mean_1": 0.2685164511203766, "updated_reward_std_0": 0.2645796835422516, "updated_reward_std_1": 0.10998629778623581 }, { "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": 0.5260416666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.5, "completions/mean_length": 1666.8372802734375, "completions/mean_terminated_length": 1334.1519165039062, "completions/min_length": 576.0, "completions/min_terminated_length": 576.0, "epoch": 0.164, "format_reward_mean_0": 0.6458333730697632, "format_reward_mean_1": 0.9401041865348816, "format_reward_std_0": 0.48332110047340393, "format_reward_std_1": 0.2376033216714859, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 2.079678773880005, "kl": 2.077538483707376e+30, "learning_rate": 8.399999999999999e-07, "loss": 0.0007, "num_tokens": 82737086.0, "problem_reward_mean_0": 0.25282546877861023, "problem_reward_mean_1": 0.40985724329948425, "problem_reward_std_0": 0.40988409519195557, "problem_reward_std_1": 0.43239107728004456, "reward_mean_0": 0.25282546877861023, "reward_mean_1": 0.40985724329948425, "reward_std_0": 0.2136237472295761, "reward_std_1": 0.13810434937477112, "rewards/compute_reward/mean": 0.33134135603904724, "rewards/compute_reward/std": 0.4211375415325165, "step": 41, "updated_reward_mean_0": 0.576238214969635, "updated_reward_mean_1": 0.2396756410598755, "updated_reward_std_0": 0.2082596719264984, "updated_reward_std_1": 0.10669559240341187 }, { "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": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.5, "completions/mean_length": 1770.9791870117188, "completions/mean_terminated_length": 1358.4357299804688, "completions/min_length": 520.5, "completions/min_terminated_length": 520.5, "epoch": 0.168, "format_reward_mean_0": 0.625, "format_reward_mean_1": 0.875, "format_reward_std_0": 0.48924607038497925, "format_reward_std_1": 0.3311503827571869, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 0.7902534008026123, "kl": 3.460017913204313e+33, "learning_rate": 8.359999999999999e-07, "loss": 0.001, "num_tokens": 84815188.0, "problem_reward_mean_0": 0.28166669607162476, "problem_reward_mean_1": 0.31830471754074097, "problem_reward_std_0": 0.37137657403945923, "problem_reward_std_1": 0.41014862060546875, "reward_mean_0": 0.28166669607162476, "reward_mean_1": 0.31830471754074097, "reward_std_0": 0.25428327918052673, "reward_std_1": 0.19833168387413025, "rewards/compute_reward/mean": 0.29998569190502167, "rewards/compute_reward/std": 0.390762597322464, "step": 42, "updated_reward_mean_0": 0.5573943853378296, "updated_reward_mean_1": 0.28445059061050415, "updated_reward_std_0": 0.12949439883232117, "updated_reward_std_1": 0.17983011901378632 }, { "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": 0.7317708333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.5, "completions/mean_length": 1846.875, "completions/mean_terminated_length": 1318.6077270507812, "completions/min_length": 617.0, "completions/min_terminated_length": 617.0, "epoch": 0.172, "format_reward_mean_0": 0.25, "format_reward_mean_1": 0.8723958730697632, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.33408382534980774, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 1.1578903198242188, "kl": -1.164477926872037e+32, "learning_rate": 8.319999999999999e-07, "loss": 0.001, "num_tokens": 86915549.0, "problem_reward_mean_0": 0.07111378759145737, "problem_reward_mean_1": 0.35877108573913574, "problem_reward_std_0": 0.19041231274604797, "problem_reward_std_1": 0.32522696256637573, "reward_mean_0": 0.07111378759145737, "reward_mean_1": 0.35877108573913574, "reward_std_0": 0.16143730282783508, "reward_std_1": 0.18203359842300415, "rewards/compute_reward/mean": 0.21494243666529655, "rewards/compute_reward/std": 0.25781963020563126, "step": 43, "updated_reward_mean_0": 0.5325745344161987, "updated_reward_mean_1": 0.33793777227401733, "updated_reward_std_0": 0.13640311360359192, "updated_reward_std_1": 0.18203362822532654 }, { "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": 0.7994791666666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.5, "completions/mean_length": 1920.4219360351562, "completions/mean_terminated_length": 1447.1948852539062, "completions/min_length": 726.0, "completions/min_terminated_length": 726.0, "epoch": 0.176, "format_reward_mean_0": 0.1875, "format_reward_mean_1": 0.7317708730697632, "format_reward_std_0": 0.3944427967071533, "format_reward_std_1": 0.443615585565567, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 0.9102089405059814, "kl": 3.461531240803365e+35, "learning_rate": 8.28e-07, "loss": 0.0011, "num_tokens": 88981545.0, "problem_reward_mean_0": 0.025833334773778915, "problem_reward_mean_1": 0.19697165489196777, "problem_reward_std_0": 0.10574624687433243, "problem_reward_std_1": 0.2757859528064728, "reward_mean_0": 0.025833334773778915, "reward_mean_1": 0.19697165489196777, "reward_std_0": 0.07306770980358124, "reward_std_1": 0.1314355432987213, "rewards/compute_reward/mean": 0.11140249390155077, "rewards/compute_reward/std": 0.1907660886645317, "step": 44, "updated_reward_mean_0": 0.376339852809906, "updated_reward_mean_1": 0.19697165489196777, "updated_reward_std_0": 0.1645078808069229, "updated_reward_std_1": 0.1314355581998825 }, { "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": 0.3671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.5, "completions/mean_length": 1436.3607177734375, "completions/mean_terminated_length": 1110.3486328125, "completions/min_length": 443.5, "completions/min_terminated_length": 443.5, "epoch": 0.18, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.9557291865348816, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.20596453547477722, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.5625, "grad_norm": 1.4784208536148071, "kl": -1.697698161128511e+32, "learning_rate": 8.24e-07, "loss": 0.0007, "num_tokens": 90791294.0, "problem_reward_mean_0": 0.4286525249481201, "problem_reward_mean_1": 0.4845340847969055, "problem_reward_std_0": 0.47943222522735596, "problem_reward_std_1": 0.45884326100349426, "reward_mean_0": 0.4286525249481201, "reward_mean_1": 0.4845340847969055, "reward_std_0": 0.13781920075416565, "reward_std_1": 0.08362507820129395, "rewards/compute_reward/mean": 0.4565932899713516, "rewards/compute_reward/std": 0.4691377431154251, "step": 45, "updated_reward_mean_0": 0.5839461088180542, "updated_reward_mean_1": 0.12378039956092834, "updated_reward_std_0": 0.128187894821167, "updated_reward_std_1": 0.04510257765650749 }, { "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": 0.28515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.5, "completions/mean_length": 1286.3398742675781, "completions/mean_terminated_length": 1027.1424560546875, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.184, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.14301246404647827, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.5, "grad_norm": 1.3099095821380615, "kl": 1.2890598048392273e+33, "learning_rate": 8.199999999999999e-07, "loss": 0.0007, "num_tokens": 92488411.0, "problem_reward_mean_0": 0.4597793519496918, "problem_reward_mean_1": 0.5676552653312683, "problem_reward_std_0": 0.4663116931915283, "problem_reward_std_1": 0.4320697784423828, "reward_mean_0": 0.4597793519496918, "reward_mean_1": 0.5676552653312683, "reward_std_0": 0.20392009615898132, "reward_std_1": 0.11393110454082489, "rewards/compute_reward/mean": 0.51371730864048, "rewards/compute_reward/std": 0.4491907060146332, "step": 46, "updated_reward_mean_0": 0.7019529342651367, "updated_reward_mean_1": 0.22318407893180847, "updated_reward_std_0": 0.1288362443447113, "updated_reward_std_1": 0.09491443634033203 }, { "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": 0.3932291666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1460.1732177734375, "completions/mean_terminated_length": 1125.1326904296875, "completions/min_length": 370.5, "completions/min_terminated_length": 370.5, "epoch": 0.188, "format_reward_mean_0": 0.5416666865348816, "format_reward_mean_1": 0.984375, "format_reward_std_0": 0.503533661365509, "format_reward_std_1": 0.12418138235807419, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.375, "grad_norm": 1.3549622297286987, "kl": -2.091158619133026e+35, "learning_rate": 8.159999999999999e-07, "loss": 0.0009, "num_tokens": 94371977.0, "problem_reward_mean_0": 0.30149635672569275, "problem_reward_mean_1": 0.5270750522613525, "problem_reward_std_0": 0.39472678303718567, "problem_reward_std_1": 0.37106749415397644, "reward_mean_0": 0.30149635672569275, "reward_mean_1": 0.5270750522613525, "reward_std_0": 0.2790029048919678, "reward_std_1": 0.12066902965307236, "rewards/compute_reward/mean": 0.41428570449352264, "rewards/compute_reward/std": 0.38289710879325867, "step": 47, "updated_reward_mean_0": 0.6696000099182129, "updated_reward_mean_1": 0.37082505226135254, "updated_reward_std_0": 0.2428307980298996, "updated_reward_std_1": 0.11423971503973007 }, { "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": 0.26041666666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1278.8059997558594, "completions/mean_terminated_length": 1037.4898376464844, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.192, "format_reward_mean_0": 0.6041666865348816, "format_reward_mean_1": 0.9583333730697632, "format_reward_std_0": 0.49420401453971863, "format_reward_std_1": 0.2000870257616043, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.7467588782310486, "kl": 1313317745096.7686, "learning_rate": 8.12e-07, "loss": 0.0008, "num_tokens": 96024967.0, "problem_reward_mean_0": 0.3661075234413147, "problem_reward_mean_1": 0.3876059949398041, "problem_reward_std_0": 0.44995033740997314, "problem_reward_std_1": 0.4195745289325714, "reward_mean_0": 0.3661075234413147, "reward_mean_1": 0.3876059949398041, "reward_std_0": 0.24140623211860657, "reward_std_1": 0.12012609094381332, "rewards/compute_reward/mean": 0.3768567442893982, "rewards/compute_reward/std": 0.4347624182701111, "step": 48, "updated_reward_mean_0": 0.5351055264472961, "updated_reward_mean_1": 0.11627660691738129, "updated_reward_std_0": 0.24029135704040527, "updated_reward_std_1": 0.0704243928194046 }, { "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": 0.5338541666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1720.475341796875, "completions/mean_terminated_length": 1400.9077758789062, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.196, "format_reward_mean_0": 0.4166666865348816, "format_reward_mean_1": 0.9505208730697632, "format_reward_std_0": 0.49822381138801575, "format_reward_std_1": 0.21714918315410614, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3125, "grad_norm": 2.1229441165924072, "kl": -5.3417250244057414e+32, "learning_rate": 8.08e-07, "loss": 0.0011, "num_tokens": 98068604.0, "problem_reward_mean_0": 0.09919752925634384, "problem_reward_mean_1": 0.21097031235694885, "problem_reward_std_0": 0.2144337147474289, "problem_reward_std_1": 0.28133800625801086, "reward_mean_0": 0.09919752925634384, "reward_mean_1": 0.21097031235694885, "reward_std_0": 0.19539368152618408, "reward_std_1": 0.1534106284379959, "rewards/compute_reward/mean": 0.15508391335606575, "rewards/compute_reward/std": 0.24788586050271988, "step": 49, "updated_reward_mean_0": 0.45206791162490845, "updated_reward_mean_1": 0.19013695418834686, "updated_reward_std_0": 0.08968322724103928, "updated_reward_std_1": 0.1534106433391571 }, { "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": 0.3190104166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1526.4323120117188, "completions/mean_terminated_length": 1350.4307861328125, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.2, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.1016615703701973, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 0.7596748471260071, "kl": 9.915844709420791e+32, "learning_rate": 8.04e-07, "loss": 0.0009, "num_tokens": 100002939.0, "problem_reward_mean_0": 0.32921457290649414, "problem_reward_mean_1": 0.5120910406112671, "problem_reward_std_0": 0.3531428873538971, "problem_reward_std_1": 0.33090081810951233, "reward_mean_0": 0.32921457290649414, "reward_mean_1": 0.5120911002159119, "reward_std_0": 0.22802993655204773, "reward_std_1": 0.1293172687292099, "rewards/compute_reward/mean": 0.4206528067588806, "rewards/compute_reward/std": 0.3420218229293823, "step": 50, "updated_reward_mean_0": 0.6597024202346802, "updated_reward_mean_1": 0.43315786123275757, "updated_reward_std_0": 0.13297486305236816, "updated_reward_std_1": 0.11218631267547607 }, { "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": 0.46744791666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1551.7643432617188, "completions/mean_terminated_length": 1075.7029724121094, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.204, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.14301247894763947, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 1.5726449489593506, "kl": 1.973072844477534e+35, "learning_rate": 8e-07, "loss": 0.0014, "num_tokens": 102001670.0, "problem_reward_mean_0": 0.2440476268529892, "problem_reward_mean_1": 0.4022727906703949, "problem_reward_std_0": 0.4083492159843445, "problem_reward_std_1": 0.39565378427505493, "reward_mean_0": 0.2440476268529892, "reward_mean_1": 0.4022727906703949, "reward_std_0": 0.10321750491857529, "reward_std_1": 0.12444710731506348, "rewards/compute_reward/mean": 0.32316020876169205, "rewards/compute_reward/std": 0.4020014852285385, "step": 51, "updated_reward_mean_0": 0.5532190799713135, "updated_reward_mean_1": 0.1939394474029541, "updated_reward_std_0": 0.15184760093688965, "updated_reward_std_1": 0.12444710731506348 }, { "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": 0.5677083333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1768.5091552734375, "completions/mean_terminated_length": 1438.5211181640625, "completions/min_length": 705.5, "completions/min_terminated_length": 705.5, "epoch": 0.208, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9479166865348816, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.22248496115207672, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 1.778757929801941, "kl": 9.190079353195503e+25, "learning_rate": 7.96e-07, "loss": 0.001, "num_tokens": 104083563.0, "problem_reward_mean_0": 0.16869214177131653, "problem_reward_mean_1": 0.33662471175193787, "problem_reward_std_0": 0.2756715416908264, "problem_reward_std_1": 0.36116233468055725, "reward_mean_0": 0.16869214177131653, "reward_mean_1": 0.33662471175193787, "reward_std_0": 0.1767314374446869, "reward_std_1": 0.17280423641204834, "rewards/compute_reward/mean": 0.2526584193110466, "rewards/compute_reward/std": 0.31841690838336945, "step": 52, "updated_reward_mean_0": 0.5633680820465088, "updated_reward_mean_1": 0.30016636848449707, "updated_reward_std_0": 0.2367755025625229, "updated_reward_std_1": 0.1631602942943573 }, { "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": 0.49088541666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.5, "completions/mean_length": 1697.2513427734375, "completions/mean_terminated_length": 1338.2392578125, "completions/min_length": 490.5, "completions/min_terminated_length": 490.5, "epoch": 0.212, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.14301247894763947, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4375, "grad_norm": 1.2287437915802002, "kl": 4.8583582524150936e+35, "learning_rate": 7.92e-07, "loss": 0.0009, "num_tokens": 106214631.0, "problem_reward_mean_0": 0.2154720276594162, "problem_reward_mean_1": 0.2523699998855591, "problem_reward_std_0": 0.288810670375824, "problem_reward_std_1": 0.3266875147819519, "reward_mean_0": 0.2154720276594162, "reward_mean_1": 0.25237002968788147, "reward_std_0": 0.1623694747686386, "reward_std_1": 0.14496967196464539, "rewards/compute_reward/mean": 0.23392101377248764, "rewards/compute_reward/std": 0.30774909257888794, "step": 53, "updated_reward_mean_0": 0.4166874587535858, "updated_reward_mean_1": 0.25237002968788147, "updated_reward_std_0": 0.13933101296424866, "updated_reward_std_1": 0.14496967196464539 }, { "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": 0.5091145833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.5, "completions/mean_length": 1689.7513427734375, "completions/mean_terminated_length": 1283.1093139648438, "completions/min_length": 605.0, "completions/min_terminated_length": 605.0, "epoch": 0.216, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9765625, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.15148578584194183, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.125, "grad_norm": 2.0095176696777344, "kl": 5.1814851482703176e+35, "learning_rate": 7.88e-07, "loss": 0.0008, "num_tokens": 108367709.0, "problem_reward_mean_0": 0.26565808057785034, "problem_reward_mean_1": 0.36167535185813904, "problem_reward_std_0": 0.38632965087890625, "problem_reward_std_1": 0.3689184784889221, "reward_mean_0": 0.26565808057785034, "reward_mean_1": 0.36167535185813904, "reward_std_0": 0.1604686975479126, "reward_std_1": 0.16235165297985077, "rewards/compute_reward/mean": 0.3136667013168335, "rewards/compute_reward/std": 0.3776240646839142, "step": 54, "updated_reward_mean_0": 0.5876736640930176, "updated_reward_mean_1": 0.22718462347984314, "updated_reward_std_0": 0.14212831854820251, "updated_reward_std_1": 0.13436387479305267 }, { "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": 0.3658854166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.5, "completions/mean_length": 1595.8515625, "completions/mean_terminated_length": 1350.8760375976562, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.22, "format_reward_mean_0": 0.8333333730697632, "format_reward_mean_1": 0.9401041865348816, "format_reward_std_0": 0.3766217827796936, "format_reward_std_1": 0.2376033067703247, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 0.8218652606010437, "kl": 1.1324343524471604e+33, "learning_rate": 7.84e-07, "loss": 0.0008, "num_tokens": 110482518.0, "problem_reward_mean_0": 0.5147420763969421, "problem_reward_mean_1": 0.35361504554748535, "problem_reward_std_0": 0.4227263331413269, "problem_reward_std_1": 0.4089951515197754, "reward_mean_0": 0.5147420763969421, "reward_mean_1": 0.35361504554748535, "reward_std_0": 0.20971159636974335, "reward_std_1": 0.14845725893974304, "rewards/compute_reward/mean": 0.43417856097221375, "rewards/compute_reward/std": 0.41586072742938995, "step": 55, "updated_reward_mean_0": 0.8121603727340698, "updated_reward_mean_1": 0.2953150272369385, "updated_reward_std_0": 0.1945478618144989, "updated_reward_std_1": 0.11904865503311157 }, { "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": 0.6471354166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1798.6758422851562, "completions/mean_terminated_length": 1375.3202514648438, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.224, "format_reward_mean_0": 0.7083333730697632, "format_reward_mean_1": 0.9583333730697632, "format_reward_std_0": 0.4593396484851837, "format_reward_std_1": 0.2000870257616043, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2083333432674408, "grad_norm": 1.0690768957138062, "kl": -9.634149706385531e+31, "learning_rate": 7.799999999999999e-07, "loss": 0.0008, "num_tokens": 112612216.0, "problem_reward_mean_0": 0.27820512652397156, "problem_reward_mean_1": 0.47232574224472046, "problem_reward_std_0": 0.3984052538871765, "problem_reward_std_1": 0.4029066264629364, "reward_mean_0": 0.27820512652397156, "reward_mean_1": 0.47232574224472046, "reward_std_0": 0.15364299714565277, "reward_std_1": 0.15229734778404236, "rewards/compute_reward/mean": 0.3752654194831848, "rewards/compute_reward/std": 0.40065594017505646, "step": 56, "updated_reward_mean_0": 0.6466346383094788, "updated_reward_mean_1": 0.2906250059604645, "updated_reward_std_0": 0.13404768705368042, "updated_reward_std_1": 0.1361488699913025 }, { "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": 0.5859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1791.5846557617188, "completions/mean_terminated_length": 1459.2077026367188, "completions/min_length": 699.0, "completions/min_terminated_length": 699.0, "epoch": 0.228, "format_reward_mean_0": 0.6875, "format_reward_mean_1": 0.984375, "format_reward_std_0": 0.4684174358844757, "format_reward_std_1": 0.12418138235807419, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1875, "grad_norm": 0.6320978999137878, "kl": -1.206645620218379e+35, "learning_rate": 7.76e-07, "loss": 0.001, "num_tokens": 114741395.0, "problem_reward_mean_0": 0.3138354420661926, "problem_reward_mean_1": 0.4223197400569916, "problem_reward_std_0": 0.37596607208251953, "problem_reward_std_1": 0.3882826864719391, "reward_mean_0": 0.313835471868515, "reward_mean_1": 0.4223197400569916, "reward_std_0": 0.2643830180168152, "reward_std_1": 0.1709332913160324, "rewards/compute_reward/mean": 0.3680775910615921, "rewards/compute_reward/std": 0.3821243643760681, "step": 57, "updated_reward_mean_0": 0.6580812931060791, "updated_reward_mean_1": 0.30398643016815186, "updated_reward_std_0": 0.23494552075862885, "updated_reward_std_1": 0.15339726209640503 }, { "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": 0.35026041666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1510.23828125, "completions/mean_terminated_length": 1284.0902709960938, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.232, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.05103104189038277, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.1026179790496826, "kl": -1.7346320544679175e+35, "learning_rate": 7.72e-07, "loss": 0.001, "num_tokens": 116570982.0, "problem_reward_mean_0": 0.33552631735801697, "problem_reward_mean_1": 0.4439232051372528, "problem_reward_std_0": 0.4277532398700714, "problem_reward_std_1": 0.45214277505874634, "reward_mean_0": 0.33552631735801697, "reward_mean_1": 0.4439232051372528, "reward_std_0": 0.24367183446884155, "reward_std_1": 0.18202440440654755, "rewards/compute_reward/mean": 0.3897247612476349, "rewards/compute_reward/std": 0.4399479925632477, "step": 58, "updated_reward_mean_0": 0.6595394611358643, "updated_reward_mean_1": 0.22067509591579437, "updated_reward_std_0": 0.10038220882415771, "updated_reward_std_1": 0.1471182107925415 }, { "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": 0.5638020833333334, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1674.8619995117188, "completions/mean_terminated_length": 1203.9779052734375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.236, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.11351143568754196, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.5, "grad_norm": 1.5195142030715942, "kl": Infinity, "learning_rate": 7.68e-07, "loss": 0.0009, "num_tokens": 118531623.0, "problem_reward_mean_0": 0.21282681822776794, "problem_reward_mean_1": 0.24859580397605896, "problem_reward_std_0": 0.40372973680496216, "problem_reward_std_1": 0.407775342464447, "reward_mean_0": 0.21282681822776794, "reward_mean_1": 0.24859580397605896, "reward_std_0": 0.10972578823566437, "reward_std_1": 0.13224518299102783, "rewards/compute_reward/mean": 0.23071129620075226, "rewards/compute_reward/std": 0.4057525396347046, "step": 59, "updated_reward_mean_0": 0.481004923582077, "updated_reward_mean_1": 0.06890829652547836, "updated_reward_std_0": 0.22570618987083435, "updated_reward_std_1": 0.11523552238941193 }, { "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": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.5, "completions/mean_length": 1919.5117797851562, "completions/mean_terminated_length": 1506.6595458984375, "completions/min_length": 937.5, "completions/min_terminated_length": 937.5, "epoch": 0.24, "format_reward_mean_0": 0.5416666865348816, "format_reward_mean_1": 0.9322916865348816, "format_reward_std_0": 0.503533661365509, "format_reward_std_1": 0.25157251954078674, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 1.5447924137115479, "kl": 6.758104514169101e+35, "learning_rate": 7.64e-07, "loss": 0.001, "num_tokens": 120684469.0, "problem_reward_mean_0": 0.1138952374458313, "problem_reward_mean_1": 0.38038742542266846, "problem_reward_std_0": 0.22676359117031097, "problem_reward_std_1": 0.3759339153766632, "reward_mean_0": 0.1138952374458313, "reward_mean_1": 0.38038739562034607, "reward_std_0": 0.2052919864654541, "reward_std_1": 0.2516074776649475, "rewards/compute_reward/mean": 0.24714133143424988, "rewards/compute_reward/std": 0.3013487383723259, "step": 60, "updated_reward_mean_0": 0.6989147067070007, "updated_reward_mean_1": 0.36997076869010925, "updated_reward_std_0": 0.160629540681839, "updated_reward_std_1": 0.24047158658504486 }, { "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": 0.8580729166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 1705.5, "completions/mean_length": 1949.609375, "completions/mean_terminated_length": 1186.2938232421875, "completions/min_length": 485.0, "completions/min_terminated_length": 485.0, "epoch": 0.244, "format_reward_mean_0": 0.5833333730697632, "format_reward_mean_1": 0.9479166865348816, "format_reward_std_0": 0.49822381138801575, "format_reward_std_1": 0.22248497605323792, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.0416666679084301, "grad_norm": 1.6712886095046997, "kl": Infinity, "learning_rate": 7.599999999999999e-07, "loss": 0.0011, "num_tokens": 122954618.0, "problem_reward_mean_0": 0.10592889040708542, "problem_reward_mean_1": 0.3691001832485199, "problem_reward_std_0": 0.1933557242155075, "problem_reward_std_1": 0.3072221279144287, "reward_mean_0": 0.10592889040708542, "reward_mean_1": 0.3691001832485199, "reward_std_0": 0.10121267288923264, "reward_std_1": 0.21408221125602722, "rewards/compute_reward/mean": 0.23751453682780266, "rewards/compute_reward/std": 0.2502889037132263, "step": 61, "updated_reward_mean_0": 0.6347588300704956, "updated_reward_mean_1": 0.3691001832485199, "updated_reward_std_0": 0.16504479944705963, "updated_reward_std_1": 0.21408221125602722 }, { "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": 0.8333333333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1960.9492797851562, "completions/mean_terminated_length": 1538.8802490234375, "completions/min_length": 813.0, "completions/min_terminated_length": 813.0, "epoch": 0.248, "format_reward_mean_0": 0.5625, "format_reward_mean_1": 0.9375, "format_reward_std_0": 0.5013279914855957, "format_reward_std_1": 0.24237726628780365, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 1.4107134342193604, "kl": -5.907569894468129e+35, "learning_rate": 7.559999999999999e-07, "loss": 0.0011, "num_tokens": 125139728.0, "problem_reward_mean_0": 0.11712827533483505, "problem_reward_mean_1": 0.281910240650177, "problem_reward_std_0": 0.18519656360149384, "problem_reward_std_1": 0.25681498646736145, "reward_mean_0": 0.11712827533483505, "reward_mean_1": 0.281910240650177, "reward_std_0": 0.14425019919872284, "reward_std_1": 0.17541107535362244, "rewards/compute_reward/mean": 0.19951924309134483, "rewards/compute_reward/std": 0.22100578248500824, "step": 62, "updated_reward_mean_0": 0.5131581425666809, "updated_reward_mean_1": 0.281910240650177, "updated_reward_std_0": 0.2257710099220276, "updated_reward_std_1": 0.17541107535362244 }, { "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": 0.6197916666666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.5, "completions/mean_length": 1810.2461547851562, "completions/mean_terminated_length": 1431.6486206054688, "completions/min_length": 615.0, "completions/min_terminated_length": 615.0, "epoch": 0.252, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.14301246404647827, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.25, "grad_norm": 1.1740312576293945, "kl": -9.014272629698407e+30, "learning_rate": 7.52e-07, "loss": 0.0008, "num_tokens": 127367848.0, "problem_reward_mean_0": 0.24267815053462982, "problem_reward_mean_1": 0.36510419845581055, "problem_reward_std_0": 0.32973265647888184, "problem_reward_std_1": 0.39547276496887207, "reward_mean_0": 0.24267813563346863, "reward_mean_1": 0.36510419845581055, "reward_std_0": 0.21026691794395447, "reward_std_1": 0.17577961087226868, "rewards/compute_reward/mean": 0.3038911744952202, "rewards/compute_reward/std": 0.36260271072387695, "step": 63, "updated_reward_mean_0": 0.5400060415267944, "updated_reward_mean_1": 0.31432294845581055, "updated_reward_std_0": 0.06954167038202286, "updated_reward_std_1": 0.15280884504318237 }, { "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": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1969.0, "completions/mean_length": 1865.5846557617188, "completions/mean_terminated_length": 1600.4058227539062, "completions/min_length": 790.0, "completions/min_terminated_length": 790.0, "epoch": 0.256, "format_reward_mean_0": 0.6875, "format_reward_mean_1": 0.984375, "format_reward_std_0": 0.4684174358844757, "format_reward_std_1": 0.12418138980865479, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 3.125887632369995, "kl": 9.84553703681025e+34, "learning_rate": 7.48e-07, "loss": 0.001, "num_tokens": 129518239.0, "problem_reward_mean_0": 0.1421455442905426, "problem_reward_mean_1": 0.2631451487541199, "problem_reward_std_0": 0.30235692858695984, "problem_reward_std_1": 0.38105764985084534, "reward_mean_0": 0.1421455442905426, "reward_mean_1": 0.2631451487541199, "reward_std_0": 0.24992744624614716, "reward_std_1": 0.1862521767616272, "rewards/compute_reward/mean": 0.20264533162117004, "rewards/compute_reward/std": 0.3417072892189026, "step": 64, "updated_reward_mean_0": 0.5482755899429321, "updated_reward_mean_1": 0.22787567973136902, "updated_reward_std_0": 0.2425118386745453, "updated_reward_std_1": 0.16834688186645508 }, { "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": 0.9453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1018.5, "completions/mean_length": 2026.9505615234375, "completions/mean_terminated_length": 831.547607421875, "completions/min_length": 1519.0, "completions/min_terminated_length": 495.0, "epoch": 0.26, "format_reward_mean_0": 0.2708333432674408, "format_reward_mean_1": 0.7890625, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.4085063934326172, "frac_reward_zero_std_0": 0.6666666865348816, "frac_reward_zero_std_1": 0.25, "grad_norm": 3.7183735370635986, "kl": 1.5657337180602633e+33, "learning_rate": 7.44e-07, "loss": 0.0009, "num_tokens": 131708369.0, "problem_reward_mean_0": 0.0491071455180645, "problem_reward_mean_1": 0.2386377900838852, "problem_reward_std_0": 0.20478230714797974, "problem_reward_std_1": 0.34240785241127014, "reward_mean_0": 0.0491071455180645, "reward_mean_1": 0.2386377900838852, "reward_std_0": 0.09391222894191742, "reward_std_1": 0.12182196974754333, "rewards/compute_reward/mean": 0.14387246780097485, "rewards/compute_reward/std": 0.27359506487846375, "step": 65, "updated_reward_mean_0": 0.3976934850215912, "updated_reward_mean_1": 0.1969711035490036, "updated_reward_std_0": 0.2225004881620407, "updated_reward_std_1": 0.12182198464870453 }, { "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": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1758.5, "completions/mean_length": 1942.6250610351562, "completions/mean_terminated_length": 1206.9649047851562, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.264, "format_reward_mean_0": 0.5208333730697632, "format_reward_mean_1": 0.890625, "format_reward_std_0": 0.5048523545265198, "format_reward_std_1": 0.31251633167266846, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 2.194742441177368, "kl": -1.2030920271794968e+31, "learning_rate": 7.4e-07, "loss": 0.0008, "num_tokens": 133864176.0, "problem_reward_mean_0": 0.24286508560180664, "problem_reward_mean_1": 0.5335766077041626, "problem_reward_std_0": 0.3908400237560272, "problem_reward_std_1": 0.43979790806770325, "reward_mean_0": 0.24286508560180664, "reward_mean_1": 0.5335766077041626, "reward_std_0": 0.30628812313079834, "reward_std_1": 0.19463734328746796, "rewards/compute_reward/mean": 0.3882208466529846, "rewards/compute_reward/std": 0.41531895101070404, "step": 66, "updated_reward_mean_0": 0.703204333782196, "updated_reward_mean_1": 0.36454248428344727, "updated_reward_std_0": 0.1157386526465416, "updated_reward_std_1": 0.1588101089000702 }, { "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": 0.8294270833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.5, "completions/mean_length": 1937.3841552734375, "completions/mean_terminated_length": 1405.2667236328125, "completions/min_length": 617.5, "completions/min_terminated_length": 617.5, "epoch": 0.268, "format_reward_mean_0": 0.4375, "format_reward_mean_1": 0.8854166865348816, "format_reward_std_0": 0.5013279914855957, "format_reward_std_1": 0.31893399357795715, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.0833333358168602, "grad_norm": 1.3249965906143188, "kl": 3.1966065577043296e+27, "learning_rate": 7.359999999999999e-07, "loss": 0.0009, "num_tokens": 136042394.0, "problem_reward_mean_0": 0.328125, "problem_reward_mean_1": 0.5272141695022583, "problem_reward_std_0": 0.46458700299263, "problem_reward_std_1": 0.4162200391292572, "reward_mean_0": 0.328125, "reward_mean_1": 0.5272141695022583, "reward_std_0": 0.10311973094940186, "reward_std_1": 0.28483912348747253, "rewards/compute_reward/mean": 0.42766958475112915, "rewards/compute_reward/std": 0.4404035359621048, "step": 67, "updated_reward_mean_0": 0.8695575594902039, "updated_reward_mean_1": 0.2597290575504303, "updated_reward_std_0": 0.14507055282592773, "updated_reward_std_1": 0.21469378471374512 }, { "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": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1970.49609375, "completions/mean_terminated_length": 1634.6458129882812, "completions/min_length": 529.0, "completions/min_terminated_length": 529.0, "epoch": 0.272, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.8203125, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.38442784547805786, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1666666716337204, "grad_norm": 0.7107228636741638, "kl": -5.27117672524355e+35, "learning_rate": 7.319999999999999e-07, "loss": 0.0008, "num_tokens": 138455435.0, "problem_reward_mean_0": 0.2881265878677368, "problem_reward_mean_1": 0.3350094258785248, "problem_reward_std_0": 0.3647603988647461, "problem_reward_std_1": 0.32032907009124756, "reward_mean_0": 0.2881265878677368, "reward_mean_1": 0.3350094258785248, "reward_std_0": 0.11436723172664642, "reward_std_1": 0.16676941514015198, "rewards/compute_reward/mean": 0.3115680068731308, "rewards/compute_reward/std": 0.3425447344779968, "step": 68, "updated_reward_mean_0": 0.7092627286911011, "updated_reward_mean_1": 0.316036194562912, "updated_reward_std_0": 0.1467166543006897, "updated_reward_std_1": 0.1572992354631424 }, { "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": 0.22916666666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.5, "completions/mean_length": 1319.6224365234375, "completions/mean_terminated_length": 1108.8424682617188, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.276, "format_reward_mean_0": 1.0, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.6041666865348816, "grad_norm": 0.20323900878429413, "kl": 7.177997053883648e+23, "learning_rate": 7.28e-07, "loss": 0.0007, "num_tokens": 140234943.0, "problem_reward_mean_0": 0.6811342239379883, "problem_reward_mean_1": 0.7094241976737976, "problem_reward_std_0": 0.38112780451774597, "problem_reward_std_1": 0.3671373426914215, "reward_mean_0": 0.6811342239379883, "reward_mean_1": 0.7094241976737976, "reward_std_0": 0.25903549790382385, "reward_std_1": 0.09830693900585175, "rewards/compute_reward/mean": 0.6952792108058929, "rewards/compute_reward/std": 0.37413257360458374, "step": 69, "updated_reward_mean_0": 0.8159722089767456, "updated_reward_mean_1": 0.24703416228294373, "updated_reward_std_0": 0.1063167154788971, "updated_reward_std_1": 0.04992803931236267 }, { "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": 0.7122395833333334, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.5, "completions/mean_length": 1864.0859985351562, "completions/mean_terminated_length": 1398.2393798828125, "completions/min_length": 640.5, "completions/min_terminated_length": 640.5, "epoch": 0.28, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1666666716337204, "grad_norm": 1.3438600301742554, "kl": -1.1661258721559868e+34, "learning_rate": 7.24e-07, "loss": 0.001, "num_tokens": 142484951.0, "problem_reward_mean_0": 0.4291289448738098, "problem_reward_mean_1": 0.6346383094787598, "problem_reward_std_0": 0.3843865990638733, "problem_reward_std_1": 0.3865700364112854, "reward_mean_0": 0.4291289448738098, "reward_mean_1": 0.6346383094787598, "reward_std_0": 0.33200737833976746, "reward_std_1": 0.24253062903881073, "rewards/compute_reward/mean": 0.5318835824728012, "rewards/compute_reward/std": 0.38547831773757935, "step": 70, "updated_reward_mean_0": 0.8830389976501465, "updated_reward_mean_1": 0.5193756222724915, "updated_reward_std_0": 0.13830962777137756, "updated_reward_std_1": 0.1862405240535736 }, { "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": 0.7369791666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.5, "completions/mean_length": 1854.3099365234375, "completions/mean_terminated_length": 1396.5444946289062, "completions/min_length": 822.5, "completions/min_terminated_length": 822.5, "epoch": 0.284, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.0720745101571083, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 0.3951137363910675, "kl": -3.005792569067561e+28, "learning_rate": 7.2e-07, "loss": 0.0011, "num_tokens": 144611395.0, "problem_reward_mean_0": 0.2975694537162781, "problem_reward_mean_1": 0.37161463499069214, "problem_reward_std_0": 0.32771849632263184, "problem_reward_std_1": 0.3688775300979614, "reward_mean_0": 0.2975694537162781, "reward_mean_1": 0.37161463499069214, "reward_std_0": 0.20526444911956787, "reward_std_1": 0.18205910921096802, "rewards/compute_reward/mean": 0.3345920294523239, "rewards/compute_reward/std": 0.34829799830913544, "step": 71, "updated_reward_mean_0": 0.6072916984558105, "updated_reward_mean_1": 0.30683594942092896, "updated_reward_std_0": 0.21665966510772705, "updated_reward_std_1": 0.15999051928520203 }, { "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": 0.7434895833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1897.5508422851562, "completions/mean_terminated_length": 1507.7086791992188, "completions/min_length": 851.0, "completions/min_terminated_length": 851.0, "epoch": 0.288, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.19399647414684296, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 2.036374568939209, "kl": -1.6464602522495544e+29, "learning_rate": 7.159999999999999e-07, "loss": 0.0009, "num_tokens": 146769522.0, "problem_reward_mean_0": 0.25851190090179443, "problem_reward_mean_1": 0.41569942235946655, "problem_reward_std_0": 0.41693586111068726, "problem_reward_std_1": 0.45548683404922485, "reward_mean_0": 0.25851190090179443, "reward_mean_1": 0.41569942235946655, "reward_std_0": 0.35737037658691406, "reward_std_1": 0.2830284833908081, "rewards/compute_reward/mean": 0.3371056467294693, "rewards/compute_reward/std": 0.43621133267879486, "step": 72, "updated_reward_mean_0": 0.8068453073501587, "updated_reward_mean_1": 0.27507442235946655, "updated_reward_std_0": 0.2272893190383911, "updated_reward_std_1": 0.20798365771770477 }, { "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": 0.7864583333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1937.1510620117188, "completions/mean_terminated_length": 1746.3446044921875, "completions/min_length": 1256.0, "completions/min_terminated_length": 1256.0, "epoch": 0.292, "format_reward_mean_0": 0.5416666865348816, "format_reward_mean_1": 0.96875, "format_reward_std_0": 0.503533661365509, "format_reward_std_1": 0.17421962320804596, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.0833333358168602, "grad_norm": 2.0035996437072754, "kl": 1.4423046830942474e+32, "learning_rate": 7.119999999999999e-07, "loss": 0.001, "num_tokens": 148962143.0, "problem_reward_mean_0": 0.2513706088066101, "problem_reward_mean_1": 0.4641845226287842, "problem_reward_std_0": 0.39327099919319153, "problem_reward_std_1": 0.4110954701900482, "reward_mean_0": 0.2513706088066101, "reward_mean_1": 0.4641845226287842, "reward_std_0": 0.280097097158432, "reward_std_1": 0.2299395054578781, "rewards/compute_reward/mean": 0.35777756571769714, "rewards/compute_reward/std": 0.40218324959278107, "step": 73, "updated_reward_mean_0": 0.700334906578064, "updated_reward_mean_1": 0.3302001655101776, "updated_reward_std_0": 0.09973889589309692, "updated_reward_std_1": 0.19329723715782166 }, { "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": 0.51953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.5, "completions/mean_length": 1758.5924682617188, "completions/mean_terminated_length": 1469.8856201171875, "completions/min_length": 584.0, "completions/min_terminated_length": 584.0, "epoch": 0.296, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.125, "grad_norm": 1.6148369312286377, "kl": -5.215853809188435e+29, "learning_rate": 7.079999999999999e-07, "loss": 0.0012, "num_tokens": 151178777.0, "problem_reward_mean_0": 0.20592007040977478, "problem_reward_mean_1": 0.37552815675735474, "problem_reward_std_0": 0.30689144134521484, "problem_reward_std_1": 0.39559388160705566, "reward_mean_0": 0.20592007040977478, "reward_mean_1": 0.37552815675735474, "reward_std_0": 0.2588387429714203, "reward_std_1": 0.2267875373363495, "rewards/compute_reward/mean": 0.29072409123182297, "rewards/compute_reward/std": 0.35124266147613525, "step": 74, "updated_reward_mean_0": 0.6808068752288818, "updated_reward_mean_1": 0.31042397022247314, "updated_reward_std_0": 0.23277340829372406, "updated_reward_std_1": 0.2048693746328354 }, { "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": 0.35416666666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.5, "completions/mean_length": 1538.1354370117188, "completions/mean_terminated_length": 1293.7713623046875, "completions/min_length": 463.0, "completions/min_terminated_length": 463.0, "epoch": 0.3, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9427083730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.23270228505134583, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 1.2659318447113037, "kl": 3.692323036775549e+34, "learning_rate": 7.04e-07, "loss": 0.001, "num_tokens": 153151761.0, "problem_reward_mean_0": 0.462738037109375, "problem_reward_mean_1": 0.3422487676143646, "problem_reward_std_0": 0.43496155738830566, "problem_reward_std_1": 0.43171581625938416, "reward_mean_0": 0.462738037109375, "reward_mean_1": 0.3422487676143646, "reward_std_0": 0.2531580626964569, "reward_std_1": 0.19648918509483337, "rewards/compute_reward/mean": 0.4024934023618698, "rewards/compute_reward/std": 0.4333386719226837, "step": 75, "updated_reward_mean_0": 0.6603688597679138, "updated_reward_mean_1": 0.12175153195858002, "updated_reward_std_0": 0.1113782450556755, "updated_reward_std_1": 0.1153714656829834 }, { "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": 0.6575520833333334, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.5, "completions/mean_length": 1870.4896240234375, "completions/mean_terminated_length": 1575.8932495117188, "completions/min_length": 912.5, "completions/min_terminated_length": 912.5, "epoch": 0.304, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.9921875, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.08815725892782211, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.02083333395421505, "grad_norm": 1.4400324821472168, "kl": -2.9560563874755713e+31, "learning_rate": 7e-07, "loss": 0.0012, "num_tokens": 155389320.0, "problem_reward_mean_0": 0.21731960773468018, "problem_reward_mean_1": 0.47712141275405884, "problem_reward_std_0": 0.23660384118556976, "problem_reward_std_1": 0.33292993903160095, "reward_mean_0": 0.21731960773468018, "reward_mean_1": 0.47712141275405884, "reward_std_0": 0.12066736072301865, "reward_std_1": 0.24428141117095947, "rewards/compute_reward/mean": 0.3472204953432083, "rewards/compute_reward/std": 0.28476687520742416, "step": 76, "updated_reward_mean_0": 0.7702394723892212, "updated_reward_mean_1": 0.4771214425563812, "updated_reward_std_0": 0.18419519066810608, "updated_reward_std_1": 0.24428144097328186 }, { "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": 0.7135416666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.5, "completions/mean_length": 1859.8190307617188, "completions/mean_terminated_length": 1407.7504272460938, "completions/min_length": 651.5, "completions/min_terminated_length": 651.5, "epoch": 0.308, "format_reward_mean_0": 0.6666666865348816, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.47639307379722595, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.02083333395421505, "grad_norm": 1.3512004613876343, "kl": -5.97649941188434e+28, "learning_rate": 6.959999999999999e-07, "loss": 0.0011, "num_tokens": 157495377.0, "problem_reward_mean_0": 0.19381439685821533, "problem_reward_mean_1": 0.46716439723968506, "problem_reward_std_0": 0.29746773838996887, "problem_reward_std_1": 0.360067754983902, "reward_mean_0": 0.19381439685821533, "reward_mean_1": 0.46716439723968506, "reward_std_0": 0.2624613046646118, "reward_std_1": 0.2426001876592636, "rewards/compute_reward/mean": 0.3304893970489502, "rewards/compute_reward/std": 0.32876773178577423, "step": 77, "updated_reward_mean_0": 0.7733675241470337, "updated_reward_mean_1": 0.42810189723968506, "updated_reward_std_0": 0.2179100215435028, "updated_reward_std_1": 0.23523449897766113 }, { "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": 0.6627604166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.5, "completions/mean_length": 1827.936279296875, "completions/mean_terminated_length": 1444.3016357421875, "completions/min_length": 498.0, "completions/min_terminated_length": 498.0, "epoch": 0.312, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9921875, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.08815725892782211, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 1.1700859069824219, "kl": -3.7374565342532026e+32, "learning_rate": 6.919999999999999e-07, "loss": 0.001, "num_tokens": 159661591.0, "problem_reward_mean_0": 0.33468109369277954, "problem_reward_mean_1": 0.5979245901107788, "problem_reward_std_0": 0.36744576692581177, "problem_reward_std_1": 0.37545791268348694, "reward_mean_0": 0.33468109369277954, "reward_mean_1": 0.5979245901107788, "reward_std_0": 0.2529590129852295, "reward_std_1": 0.1872342824935913, "rewards/compute_reward/mean": 0.466302827000618, "rewards/compute_reward/std": 0.37145185470581055, "step": 78, "updated_reward_mean_0": 0.8158520460128784, "updated_reward_mean_1": 0.4781329333782196, "updated_reward_std_0": 0.1597723364830017, "updated_reward_std_1": 0.17250290513038635 }, { "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": 0.6432291666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1778.4010620117188, "completions/mean_terminated_length": 1198.2750244140625, "completions/min_length": 435.0, "completions/min_terminated_length": 435.0, "epoch": 0.316, "format_reward_mean_0": 0.4583333432674408, "format_reward_mean_1": 0.96875, "format_reward_std_0": 0.503533661365509, "format_reward_std_1": 0.17421963810920715, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 1.6409627199172974, "kl": 2.976916865559741e+35, "learning_rate": 6.879999999999999e-07, "loss": 0.0011, "num_tokens": 161774745.0, "problem_reward_mean_0": 0.2889610528945923, "problem_reward_mean_1": 0.6071344017982483, "problem_reward_std_0": 0.39619180560112, "problem_reward_std_1": 0.35125598311424255, "reward_mean_0": 0.2889610528945923, "reward_mean_1": 0.6071344017982483, "reward_std_0": 0.2901078462600708, "reward_std_1": 0.229107066988945, "rewards/compute_reward/mean": 0.4480477273464203, "rewards/compute_reward/std": 0.3737238645553589, "step": 79, "updated_reward_mean_0": 0.8480452299118042, "updated_reward_mean_1": 0.4391656816005707, "updated_reward_std_0": 0.13468977808952332, "updated_reward_std_1": 0.18823817372322083 }, { "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": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.5, "completions/mean_length": 1814.1016235351562, "completions/mean_terminated_length": 1420.97412109375, "completions/min_length": 561.0, "completions/min_terminated_length": 561.0, "epoch": 0.32, "format_reward_mean_0": 0.7708333730697632, "format_reward_mean_1": 0.9375, "format_reward_std_0": 0.4247443974018097, "format_reward_std_1": 0.24237726628780365, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 1.305829644203186, "kl": 8.185229485321758e+33, "learning_rate": 6.84e-07, "loss": 0.0009, "num_tokens": 163869583.0, "problem_reward_mean_0": 0.2526041865348816, "problem_reward_mean_1": 0.5088820457458496, "problem_reward_std_0": 0.39632585644721985, "problem_reward_std_1": 0.42743730545043945, "reward_mean_0": 0.2526041865348816, "reward_mean_1": 0.5088821053504944, "reward_std_0": 0.33719751238822937, "reward_std_1": 0.31223705410957336, "rewards/compute_reward/mean": 0.3807431012392044, "rewards/compute_reward/std": 0.41188161075115204, "step": 80, "updated_reward_mean_0": 0.8921130895614624, "updated_reward_mean_1": 0.3412853479385376, "updated_reward_std_0": 0.1443384438753128, "updated_reward_std_1": 0.2722409963607788 }, { "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": 0.8020833333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 1831.5, "completions/mean_length": 1950.069091796875, "completions/mean_terminated_length": 1585.2117919921875, "completions/min_length": 1076.0, "completions/min_terminated_length": 1076.0, "epoch": 0.324, "format_reward_mean_0": 0.5833333730697632, "format_reward_mean_1": 0.9635416865348816, "format_reward_std_0": 0.49822381138801575, "format_reward_std_1": 0.1876721829175949, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.125, "grad_norm": 0.708548903465271, "kl": 4.256359160680077e+27, "learning_rate": 6.800000000000001e-07, "loss": 0.0012, "num_tokens": 166078073.0, "problem_reward_mean_0": 0.18017999827861786, "problem_reward_mean_1": 0.4898608326911926, "problem_reward_std_0": 0.2660451829433441, "problem_reward_std_1": 0.4119383692741394, "reward_mean_0": 0.18017999827861786, "reward_mean_1": 0.489860862493515, "reward_std_0": 0.16628842055797577, "reward_std_1": 0.1834920346736908, "rewards/compute_reward/mean": 0.33502040058374405, "rewards/compute_reward/std": 0.33899179100990295, "step": 81, "updated_reward_mean_0": 0.6492717266082764, "updated_reward_mean_1": 0.46902748942375183, "updated_reward_std_0": 0.14933684468269348, "updated_reward_std_1": 0.1834920346736908 }, { "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": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.5, "completions/mean_length": 1841.8880615234375, "completions/mean_terminated_length": 1569.424072265625, "completions/min_length": 927.5, "completions/min_terminated_length": 927.5, "epoch": 0.328, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.875, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.3311503827571869, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.1875, "grad_norm": 1.0815684795379639, "kl": -5.607810198407704e+21, "learning_rate": 6.76e-07, "loss": 0.0009, "num_tokens": 168200600.0, "problem_reward_mean_0": 0.44897469878196716, "problem_reward_mean_1": 0.6243945956230164, "problem_reward_std_0": 0.3633441925048828, "problem_reward_std_1": 0.43782347440719604, "reward_mean_0": 0.44897469878196716, "reward_mean_1": 0.6243945956230164, "reward_std_0": 0.1117510199546814, "reward_std_1": 0.1713825911283493, "rewards/compute_reward/mean": 0.5366846472024918, "rewards/compute_reward/std": 0.40058383345603943, "step": 82, "updated_reward_mean_0": 0.8133814334869385, "updated_reward_mean_1": 0.40754759311676025, "updated_reward_std_0": 0.06232956051826477, "updated_reward_std_1": 0.13653717935085297 }, { "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": 0.7018229166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1836.58984375, "completions/mean_terminated_length": 1352.0855102539062, "completions/min_length": 562.5, "completions/min_terminated_length": 562.5, "epoch": 0.332, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.0720745176076889, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.25, "grad_norm": 2.744021415710449, "kl": -5.923094611877324e+35, "learning_rate": 6.72e-07, "loss": 0.0009, "num_tokens": 170392609.0, "problem_reward_mean_0": 0.3906646966934204, "problem_reward_mean_1": 0.5491792559623718, "problem_reward_std_0": 0.4129965305328369, "problem_reward_std_1": 0.45052263140678406, "reward_mean_0": 0.3906647562980652, "reward_mean_1": 0.5491793155670166, "reward_std_0": 0.16917097568511963, "reward_std_1": 0.12281760573387146, "rewards/compute_reward/mean": 0.4699219763278961, "rewards/compute_reward/std": 0.4317595958709717, "step": 83, "updated_reward_mean_0": 0.6515768766403198, "updated_reward_mean_1": 0.30225691199302673, "updated_reward_std_0": 0.06864359974861145, "updated_reward_std_1": 0.08176423609256744 }, { "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": 0.6458333333333334, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.5, "completions/mean_length": 1836.6055297851562, "completions/mean_terminated_length": 1490.3363647460938, "completions/min_length": 863.0, "completions/min_terminated_length": 863.0, "epoch": 0.336, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.1016615703701973, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 0.35614779591560364, "kl": 2.951913161720796e+35, "learning_rate": 6.68e-07, "loss": 0.0009, "num_tokens": 172499367.0, "problem_reward_mean_0": 0.4602007567882538, "problem_reward_mean_1": 0.6443415880203247, "problem_reward_std_0": 0.3964860737323761, "problem_reward_std_1": 0.34995508193969727, "reward_mean_0": 0.4602007567882538, "reward_mean_1": 0.6443415880203247, "reward_std_0": 0.1666417419910431, "reward_std_1": 0.1406029313802719, "rewards/compute_reward/mean": 0.5522711724042892, "rewards/compute_reward/std": 0.3732205778360367, "step": 84, "updated_reward_mean_0": 0.8340627551078796, "updated_reward_mean_1": 0.3552790880203247, "updated_reward_std_0": 0.14906421303749084, "updated_reward_std_1": 0.1344357281923294 }, { "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": 0.8723958333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.5, "completions/mean_length": 2005.3698120117188, "completions/mean_terminated_length": 1737.0315551757812, "completions/min_length": 1198.0, "completions/min_terminated_length": 1198.0, "epoch": 0.34, "format_reward_mean_0": 0.6875, "format_reward_mean_1": 0.9348958730697632, "format_reward_std_0": 0.4684174358844757, "format_reward_std_1": 0.24703145027160645, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0625, "grad_norm": 1.3220572471618652, "kl": 1.3835484380038259e+26, "learning_rate": 6.64e-07, "loss": 0.0013, "num_tokens": 174743422.0, "problem_reward_mean_0": 0.3068230450153351, "problem_reward_mean_1": 0.5549226999282837, "problem_reward_std_0": 0.33524805307388306, "problem_reward_std_1": 0.35392311215400696, "reward_mean_0": 0.3068230450153351, "reward_mean_1": 0.5549227595329285, "reward_std_0": 0.19333159923553467, "reward_std_1": 0.21789270639419556, "rewards/compute_reward/mean": 0.4308728724718094, "rewards/compute_reward/std": 0.3445855528116226, "step": 85, "updated_reward_mean_0": 0.7718362212181091, "updated_reward_mean_1": 0.4507560729980469, "updated_reward_std_0": 0.17312990128993988, "updated_reward_std_1": 0.17931687831878662 }, { "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": 0.5390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.5, "completions/mean_length": 1762.8672485351562, "completions/mean_terminated_length": 1432.2804565429688, "completions/min_length": 496.5, "completions/min_terminated_length": 496.5, "epoch": 0.344, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.96875, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.17421965301036835, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3541666865348816, "grad_norm": 1.25790536403656, "kl": -1.0070763853760436e+31, "learning_rate": 6.6e-07, "loss": 0.001, "num_tokens": 176919732.0, "problem_reward_mean_0": 0.5134742856025696, "problem_reward_mean_1": 0.5232462882995605, "problem_reward_std_0": 0.36464664340019226, "problem_reward_std_1": 0.4457317292690277, "reward_mean_0": 0.5134742856025696, "reward_mean_1": 0.5232462882995605, "reward_std_0": 0.218515083193779, "reward_std_1": 0.13795581459999084, "rewards/compute_reward/mean": 0.5183602869510651, "rewards/compute_reward/std": 0.4051891714334488, "step": 86, "updated_reward_mean_0": 0.709330677986145, "updated_reward_mean_1": 0.28730881214141846, "updated_reward_std_0": 0.016613416373729706, "updated_reward_std_1": 0.07895666360855103 }, { "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": 0.8372395833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1968.7943115234375, "completions/mean_terminated_length": 1674.8504638671875, "completions/min_length": 1160.0, "completions/min_terminated_length": 1160.0, "epoch": 0.348, "format_reward_mean_0": 0.75, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.19399647414684296, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1875, "grad_norm": 0.8450292944908142, "kl": -5.79168196273572e+35, "learning_rate": 6.56e-07, "loss": 0.0012, "num_tokens": 179167731.0, "problem_reward_mean_0": 0.19591347873210907, "problem_reward_mean_1": 0.5170688033103943, "problem_reward_std_0": 0.23079556226730347, "problem_reward_std_1": 0.3967122435569763, "reward_mean_0": 0.19591347873210907, "reward_mean_1": 0.5170687437057495, "reward_std_0": 0.19176971912384033, "reward_std_1": 0.20506712794303894, "rewards/compute_reward/mean": 0.3564911410212517, "rewards/compute_reward/std": 0.3137539103627205, "step": 87, "updated_reward_mean_0": 0.7286286354064941, "updated_reward_mean_1": 0.5170687437057495, "updated_reward_std_0": 0.16613782942295074, "updated_reward_std_1": 0.20506712794303894 }, { "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": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.5, "completions/mean_length": 1970.4688110351562, "completions/mean_terminated_length": 1669.317626953125, "completions/min_length": 688.0, "completions/min_terminated_length": 688.0, "epoch": 0.352, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9661458730697632, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.1810896396636963, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.0625, "grad_norm": 1.3763843774795532, "kl": 1.4133149482817699e+28, "learning_rate": 6.52e-07, "loss": 0.0012, "num_tokens": 181569109.0, "problem_reward_mean_0": 0.24178659915924072, "problem_reward_mean_1": 0.2939561903476715, "problem_reward_std_0": 0.3097529113292694, "problem_reward_std_1": 0.28487610816955566, "reward_mean_0": 0.24178659915924072, "reward_mean_1": 0.2939562201499939, "reward_std_0": 0.20969483256340027, "reward_std_1": 0.19198068976402283, "rewards/compute_reward/mean": 0.2678713947534561, "rewards/compute_reward/std": 0.29731450974941254, "step": 88, "updated_reward_mean_0": 0.5463153719902039, "updated_reward_mean_1": 0.27879077196121216, "updated_reward_std_0": 0.09982340782880783, "updated_reward_std_1": 0.17441622912883759 }, { "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": 0.7447916666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1882.5, "completions/mean_length": 1895.1836547851562, "completions/mean_terminated_length": 1455.82373046875, "completions/min_length": 730.5, "completions/min_terminated_length": 730.5, "epoch": 0.356, "format_reward_mean_0": 0.75, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.0720745176076889, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2083333432674408, "grad_norm": 0.805962860584259, "kl": -1.2334710310481977e+34, "learning_rate": 6.48e-07, "loss": 0.001, "num_tokens": 183665625.0, "problem_reward_mean_0": 0.3387255072593689, "problem_reward_mean_1": 0.48942869901657104, "problem_reward_std_0": 0.43142175674438477, "problem_reward_std_1": 0.4364200830459595, "reward_mean_0": 0.3387255072593689, "reward_mean_1": 0.48942869901657104, "reward_std_0": 0.28965646028518677, "reward_std_1": 0.2527877688407898, "rewards/compute_reward/mean": 0.4140770733356476, "rewards/compute_reward/std": 0.4339209198951721, "step": 89, "updated_reward_mean_0": 0.7584850788116455, "updated_reward_mean_1": 0.3128661811351776, "updated_reward_std_0": 0.1711348295211792, "updated_reward_std_1": 0.1688731610774994 }, { "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": 0.6940104166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1887.686279296875, "completions/mean_terminated_length": 1520.033935546875, "completions/min_length": 674.5, "completions/min_terminated_length": 674.5, "epoch": 0.36, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9583333730697632, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.20008701086044312, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0, "grad_norm": 1.4325562715530396, "kl": -1.0481657019278797e+35, "learning_rate": 6.44e-07, "loss": 0.001, "num_tokens": 186003375.0, "problem_reward_mean_0": 0.3659135699272156, "problem_reward_mean_1": 0.5173202753067017, "problem_reward_std_0": 0.3233485519886017, "problem_reward_std_1": 0.3829236924648285, "reward_mean_0": 0.3659135699272156, "reward_mean_1": 0.5173202753067017, "reward_std_0": 0.23805835843086243, "reward_std_1": 0.3138420283794403, "rewards/compute_reward/mean": 0.4416169077157974, "rewards/compute_reward/std": 0.3531361073255539, "step": 90, "updated_reward_mean_0": 0.8419277667999268, "updated_reward_mean_1": 0.48346611857414246, "updated_reward_std_0": 0.0663045272231102, "updated_reward_std_1": 0.2968323230743408 }, { "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": 0.546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1751.4323120117188, "completions/mean_terminated_length": 1397.1405639648438, "completions/min_length": 614.5, "completions/min_terminated_length": 614.5, "epoch": 0.364, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 0.6216562986373901, "kl": -4.77132476722722e+34, "learning_rate": 6.4e-07, "loss": 0.0007, "num_tokens": 188252547.0, "problem_reward_mean_0": 0.5359995365142822, "problem_reward_mean_1": 0.5690120458602905, "problem_reward_std_0": 0.40461623668670654, "problem_reward_std_1": 0.41529810428619385, "reward_mean_0": 0.5359995365142822, "reward_mean_1": 0.5690120458602905, "reward_std_0": 0.24918588995933533, "reward_std_1": 0.22144848108291626, "rewards/compute_reward/mean": 0.5525057315826416, "rewards/compute_reward/std": 0.4099571406841278, "step": 91, "updated_reward_mean_0": 0.8682911396026611, "updated_reward_mean_1": 0.3262486159801483, "updated_reward_std_0": 0.18074676394462585, "updated_reward_std_1": 0.15500573813915253 }, { "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": 0.39713541666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.5, "completions/mean_length": 1618.4479370117188, "completions/mean_terminated_length": 1344.8380126953125, "completions/min_length": 235.5, "completions/min_terminated_length": 235.5, "epoch": 0.368, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9713541865348816, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.16702663898468018, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 0.6107704043388367, "kl": 1.2800453142187708e+33, "learning_rate": 6.36e-07, "loss": 0.0008, "num_tokens": 190328247.0, "problem_reward_mean_0": 0.5966435670852661, "problem_reward_mean_1": 0.7505292296409607, "problem_reward_std_0": 0.4000832736492157, "problem_reward_std_1": 0.39956778287887573, "reward_mean_0": 0.5966435670852661, "reward_mean_1": 0.7505292296409607, "reward_std_0": 0.17630629241466522, "reward_std_1": 0.17432259023189545, "rewards/compute_reward/mean": 0.673586368560791, "rewards/compute_reward/std": 0.3998255133628845, "step": 92, "updated_reward_mean_0": 0.9016203880310059, "updated_reward_mean_1": 0.3459925055503845, "updated_reward_std_0": 0.05251838266849518, "updated_reward_std_1": 0.09676190465688705 }, { "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": 0.52734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1756.053466796875, "completions/mean_terminated_length": 1445.6663208007812, "completions/min_length": 700.0, "completions/min_terminated_length": 700.0, "epoch": 0.372, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 0.623790442943573, "kl": 9.018399369873265e+19, "learning_rate": 6.319999999999999e-07, "loss": 0.0009, "num_tokens": 192475308.0, "problem_reward_mean_0": 0.6101974248886108, "problem_reward_mean_1": 0.608484148979187, "problem_reward_std_0": 0.4012915790081024, "problem_reward_std_1": 0.43317022919654846, "reward_mean_0": 0.6101974248886108, "reward_mean_1": 0.608484148979187, "reward_std_0": 0.21858260035514832, "reward_std_1": 0.1742812842130661, "rewards/compute_reward/mean": 0.6093407273292542, "rewards/compute_reward/std": 0.41723087430000305, "step": 93, "updated_reward_mean_0": 0.878061056137085, "updated_reward_mean_1": 0.32202577590942383, "updated_reward_std_0": 0.13211825489997864, "updated_reward_std_1": 0.11079919338226318 }, { "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": 0.26041666666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1556.7955932617188, "completions/mean_terminated_length": 1378.1712646484375, "completions/min_length": 464.5, "completions/min_terminated_length": 464.5, "epoch": 0.376, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.14301246404647827, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 1.953001618385315, "kl": -3.0072052638626348e+35, "learning_rate": 6.28e-07, "loss": 0.0009, "num_tokens": 194489912.0, "problem_reward_mean_0": 0.43408817052841187, "problem_reward_mean_1": 0.7855180501937866, "problem_reward_std_0": 0.3983567953109741, "problem_reward_std_1": 0.3234117031097412, "reward_mean_0": 0.43408817052841187, "reward_mean_1": 0.7855180501937866, "reward_std_0": 0.23248764872550964, "reward_std_1": 0.17355656623840332, "rewards/compute_reward/mean": 0.6098030656576157, "rewards/compute_reward/std": 0.36088423430919647, "step": 94, "updated_reward_mean_0": 0.9040900468826294, "updated_reward_mean_1": 0.5355180501937866, "updated_reward_std_0": 0.026900719851255417, "updated_reward_std_1": 0.13804534077644348 }, { "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": 0.515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1698.6171875, "completions/mean_terminated_length": 1310.332763671875, "completions/min_length": 519.0, "completions/min_terminated_length": 519.0, "epoch": 0.38, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9505208730697632, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.21714921295642853, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0833333358168602, "grad_norm": 1.1218184232711792, "kl": -4.117378879261009e+25, "learning_rate": 6.24e-07, "loss": 0.001, "num_tokens": 196606283.0, "problem_reward_mean_0": 0.32923582196235657, "problem_reward_mean_1": 0.6596024632453918, "problem_reward_std_0": 0.34897732734680176, "problem_reward_std_1": 0.35419079661369324, "reward_mean_0": 0.32923582196235657, "reward_mean_1": 0.6596024036407471, "reward_std_0": 0.23600511252880096, "reward_std_1": 0.256372332572937, "rewards/compute_reward/mean": 0.4944191426038742, "rewards/compute_reward/std": 0.3515840470790863, "step": 95, "updated_reward_mean_0": 0.880002498626709, "updated_reward_mean_1": 0.5478530526161194, "updated_reward_std_0": 0.11521536111831665, "updated_reward_std_1": 0.23350703716278076 }, { "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": 0.19661458333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1419.79296875, "completions/mean_terminated_length": 1278.2119750976562, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.384, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.11351143568754196, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2083333432674408, "grad_norm": 1.44371497631073, "kl": -9.62558087190369e+34, "learning_rate": 6.2e-07, "loss": 0.0008, "num_tokens": 198432365.0, "problem_reward_mean_0": 0.5490448474884033, "problem_reward_mean_1": 0.7306036353111267, "problem_reward_std_0": 0.3999405801296234, "problem_reward_std_1": 0.3597347140312195, "reward_mean_0": 0.5490447878837585, "reward_mean_1": 0.7306036353111267, "reward_std_0": 0.32167431712150574, "reward_std_1": 0.21829204261302948, "rewards/compute_reward/mean": 0.6398242115974426, "rewards/compute_reward/std": 0.37983764708042145, "step": 96, "updated_reward_mean_0": 0.9213603734970093, "updated_reward_mean_1": 0.4011765122413635, "updated_reward_std_0": 0.03640090674161911, "updated_reward_std_1": 0.13702470064163208 }, { "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": 0.2838541666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.5, "completions/mean_length": 1500.076904296875, "completions/mean_terminated_length": 1278.2977905273438, "completions/min_length": 521.5, "completions/min_terminated_length": 521.5, "epoch": 0.388, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.96875, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.17421963810920715, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0, "grad_norm": 1.662169098854065, "kl": -1.8029261797936302e+33, "learning_rate": 6.16e-07, "loss": 0.0011, "num_tokens": 200641851.0, "problem_reward_mean_0": 0.312666118144989, "problem_reward_mean_1": 0.6379632353782654, "problem_reward_std_0": 0.21851298213005066, "problem_reward_std_1": 0.29623857140541077, "reward_mean_0": 0.3126660883426666, "reward_mean_1": 0.6379632353782654, "reward_std_0": 0.1135893166065216, "reward_std_1": 0.16529694199562073, "rewards/compute_reward/mean": 0.475314661860466, "rewards/compute_reward/std": 0.2573757842183113, "step": 97, "updated_reward_mean_0": 0.8181812167167664, "updated_reward_mean_1": 0.6379632353782654, "updated_reward_std_0": 0.10758104920387268, "updated_reward_std_1": 0.16529695689678192 }, { "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": 0.32421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.5, "completions/mean_length": 1472.5755615234375, "completions/mean_terminated_length": 1189.9312133789062, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.392, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.953125, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.21164682507514954, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2291666716337204, "grad_norm": 1.7016583681106567, "kl": -3.938440794725022e+35, "learning_rate": 6.119999999999999e-07, "loss": 0.0009, "num_tokens": 202556234.0, "problem_reward_mean_0": 0.4038194417953491, "problem_reward_mean_1": 0.6648437976837158, "problem_reward_std_0": 0.4071709215641022, "problem_reward_std_1": 0.4051482677459717, "reward_mean_0": 0.4038194417953491, "reward_mean_1": 0.6648437976837158, "reward_std_0": 0.19601166248321533, "reward_std_1": 0.2586866617202759, "rewards/compute_reward/mean": 0.5343315899372101, "rewards/compute_reward/std": 0.40615957975387573, "step": 98, "updated_reward_mean_0": 0.9019964933395386, "updated_reward_mean_1": 0.39401042461395264, "updated_reward_std_0": 0.1270817369222641, "updated_reward_std_1": 0.21107590198516846 }, { "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": 0.3072916666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.5, "completions/mean_length": 1404.5221557617188, "completions/mean_terminated_length": 1119.5081787109375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.396, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.921875, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.26871830224990845, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 1.0905659198760986, "kl": -9.865364030781412e+34, "learning_rate": 6.079999999999999e-07, "loss": 0.0006, "num_tokens": 204360624.0, "problem_reward_mean_0": 0.6054559946060181, "problem_reward_mean_1": 0.6299232840538025, "problem_reward_std_0": 0.4381123185157776, "problem_reward_std_1": 0.44654712080955505, "reward_mean_0": 0.6054559946060181, "reward_mean_1": 0.6299232840538025, "reward_std_0": 0.13030777871608734, "reward_std_1": 0.12063834071159363, "rewards/compute_reward/mean": 0.6176896393299103, "rewards/compute_reward/std": 0.4423297196626663, "step": 99, "updated_reward_mean_0": 0.7260264158248901, "updated_reward_mean_1": 0.17216861248016357, "updated_reward_std_0": 0.024188369512557983, "updated_reward_std_1": 0.04165288433432579 }, { "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": 0.17708333333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.5, "completions/mean_length": 1192.3789672851562, "completions/mean_terminated_length": 1006.8509521484375, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.4, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.1016615629196167, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.3932225704193115, "kl": 1.9384574938530338e+35, "learning_rate": 6.04e-07, "loss": 0.0009, "num_tokens": 206050670.0, "problem_reward_mean_0": 0.6125440001487732, "problem_reward_mean_1": 0.8340513110160828, "problem_reward_std_0": 0.39187556505203247, "problem_reward_std_1": 0.27684497833251953, "reward_mean_0": 0.6125440001487732, "reward_mean_1": 0.8340513110160828, "reward_std_0": 0.17830735445022583, "reward_std_1": 0.1689213216304779, "rewards/compute_reward/mean": 0.723297655582428, "rewards/compute_reward/std": 0.3343602567911148, "step": 100, "updated_reward_mean_0": 0.9735106229782104, "updated_reward_mean_1": 0.3966381549835205, "updated_reward_std_0": 0.019737549126148224, "updated_reward_std_1": 0.12737694382667542 }, { "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": 0.5078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1720.6185302734375, "completions/mean_terminated_length": 1448.3722534179688, "completions/min_length": 721.0, "completions/min_terminated_length": 721.0, "epoch": 0.404, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.11351142823696136, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 1.0572221279144287, "kl": -4.1264932070086904e+34, "learning_rate": 6e-07, "loss": 0.0012, "num_tokens": 208114674.0, "problem_reward_mean_0": 0.2698822617530823, "problem_reward_mean_1": 0.688877284526825, "problem_reward_std_0": 0.33193308115005493, "problem_reward_std_1": 0.42273348569869995, "reward_mean_0": 0.2698822617530823, "reward_mean_1": 0.688877284526825, "reward_std_0": 0.1660836935043335, "reward_std_1": 0.15320947766304016, "rewards/compute_reward/mean": 0.4793797582387924, "rewards/compute_reward/std": 0.37733326852321625, "step": 101, "updated_reward_mean_0": 0.7985342741012573, "updated_reward_mean_1": 0.6107522249221802, "updated_reward_std_0": 0.030256425961852074, "updated_reward_std_1": 0.138478085398674 }, { "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": 0.28125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1370.4075927734375, "completions/mean_terminated_length": 1107.301513671875, "completions/min_length": 137.5, "completions/min_terminated_length": 137.5, "epoch": 0.408, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9583333730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.20008701086044312, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.4158227741718292, "kl": -8.382674820835851e+35, "learning_rate": 5.96e-07, "loss": 0.0009, "num_tokens": 210016457.0, "problem_reward_mean_0": 0.6963143348693848, "problem_reward_mean_1": 0.8320579528808594, "problem_reward_std_0": 0.24772444367408752, "problem_reward_std_1": 0.2775608003139496, "reward_mean_0": 0.6963143348693848, "reward_mean_1": 0.8320579528808594, "reward_std_0": 0.14265727996826172, "reward_std_1": 0.16499197483062744, "rewards/compute_reward/mean": 0.7641861438751221, "rewards/compute_reward/std": 0.26264261454343796, "step": 102, "updated_reward_mean_0": 0.939609706401825, "updated_reward_mean_1": 0.573546826839447, "updated_reward_std_0": 0.0185447558760643, "updated_reward_std_1": 0.14108537137508392 }, { "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": 0.27473958333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1507.658935546875, "completions/mean_terminated_length": 1344.0994262695312, "completions/min_length": 518.5, "completions/min_terminated_length": 518.5, "epoch": 0.412, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 1.9574143886566162, "kl": 9.761360518033365e+35, "learning_rate": 5.919999999999999e-07, "loss": 0.001, "num_tokens": 211957646.0, "problem_reward_mean_0": 0.46945494413375854, "problem_reward_mean_1": 0.7225944399833679, "problem_reward_std_0": 0.39149045944213867, "problem_reward_std_1": 0.3231395184993744, "reward_mean_0": 0.46945494413375854, "reward_mean_1": 0.7225944399833679, "reward_std_0": 0.17262551188468933, "reward_std_1": 0.15322858095169067, "rewards/compute_reward/mean": 0.596024677157402, "rewards/compute_reward/std": 0.35731498897075653, "step": 103, "updated_reward_mean_0": 0.8596380949020386, "updated_reward_mean_1": 0.5032657384872437, "updated_reward_std_0": 0.0472087562084198, "updated_reward_std_1": 0.09499363601207733 }, { "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": 0.23046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1489.18359375, "completions/mean_terminated_length": 1325.945068359375, "completions/min_length": 437.5, "completions/min_terminated_length": 437.5, "epoch": 0.416, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.15946701169013977, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 0.7089760303497314, "kl": 1.7438872818208562e+31, "learning_rate": 5.879999999999999e-07, "loss": 0.001, "num_tokens": 214061458.0, "problem_reward_mean_0": 0.4485360383987427, "problem_reward_mean_1": 0.7549217939376831, "problem_reward_std_0": 0.34163662791252136, "problem_reward_std_1": 0.3320671021938324, "reward_mean_0": 0.4485360383987427, "reward_mean_1": 0.7549216747283936, "reward_std_0": 0.10421093553304672, "reward_std_1": 0.12317123264074326, "rewards/compute_reward/mean": 0.6017288863658905, "rewards/compute_reward/std": 0.3368518501520157, "step": 104, "updated_reward_mean_0": 0.8929147124290466, "updated_reward_mean_1": 0.575162947177887, "updated_reward_std_0": 0.12351435422897339, "updated_reward_std_1": 0.11878278851509094 }, { "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": 0.15234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1336.5625610351562, "completions/mean_terminated_length": 1210.8805541992188, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.42, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9921875, "format_reward_std_0": 0.0, "format_reward_std_1": 0.08815725892782211, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4375, "grad_norm": 0.9203000664710999, "kl": 2.253802919380644e+30, "learning_rate": 5.839999999999999e-07, "loss": 0.0013, "num_tokens": 216001318.0, "problem_reward_mean_0": 0.27778610587120056, "problem_reward_mean_1": 0.5119227170944214, "problem_reward_std_0": 0.3257428705692291, "problem_reward_std_1": 0.45884910225868225, "reward_mean_0": 0.27778610587120056, "reward_mean_1": 0.5119227170944214, "reward_std_0": 0.1356445848941803, "reward_std_1": 0.1630730777978897, "rewards/compute_reward/mean": 0.394854411482811, "rewards/compute_reward/std": 0.3922959566116333, "step": 105, "updated_reward_mean_0": 0.6829289197921753, "updated_reward_mean_1": 0.3985704183578491, "updated_reward_std_0": 0.03562727943062782, "updated_reward_std_1": 0.13616347312927246 }, { "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": 0.3502604166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1533.1029052734375, "completions/mean_terminated_length": 1271.0296630859375, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.424, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.3529976606369019, "kl": 1.71273680960292e+32, "learning_rate": 5.8e-07, "loss": 0.001, "num_tokens": 217971919.0, "problem_reward_mean_0": 0.49775707721710205, "problem_reward_mean_1": 0.5626989603042603, "problem_reward_std_0": 0.45901358127593994, "problem_reward_std_1": 0.4520709812641144, "reward_mean_0": 0.4977571368217468, "reward_mean_1": 0.5626989603042603, "reward_std_0": 0.3344268500804901, "reward_std_1": 0.17851287126541138, "rewards/compute_reward/mean": 0.5302279889583588, "rewards/compute_reward/std": 0.45554229617118835, "step": 106, "updated_reward_mean_0": 0.8079408407211304, "updated_reward_mean_1": 0.22780311107635498, "updated_reward_std_0": 0.1782611608505249, "updated_reward_std_1": 0.11472389101982117 }, { "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": 0.4700520833333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.5, "completions/mean_length": 1653.2044677734375, "completions/mean_terminated_length": 1296.3284912109375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.428, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.1016615703701973, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 1.4467134475708008, "kl": -4.2334211797292415e+22, "learning_rate": 5.76e-07, "loss": 0.0014, "num_tokens": 220177998.0, "problem_reward_mean_0": 0.19223737716674805, "problem_reward_mean_1": 0.3030022084712982, "problem_reward_std_0": 0.31271564960479736, "problem_reward_std_1": 0.32255005836486816, "reward_mean_0": 0.19223737716674805, "reward_mean_1": 0.3030022084712982, "reward_std_0": 0.1454014629125595, "reward_std_1": 0.10090981423854828, "rewards/compute_reward/mean": 0.24761979281902313, "rewards/compute_reward/std": 0.31763285398483276, "step": 107, "updated_reward_mean_0": 0.4695506691932678, "updated_reward_mean_1": 0.2587313652038574, "updated_reward_std_0": 0.13215532898902893, "updated_reward_std_1": 0.09363623708486557 }, { "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": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.5, "completions/mean_length": 1844.6836547851562, "completions/mean_terminated_length": 1531.722900390625, "completions/min_length": 696.5, "completions/min_terminated_length": 696.5, "epoch": 0.432, "format_reward_mean_0": 0.7291666865348816, "format_reward_mean_1": 0.9817708730697632, "format_reward_std_0": 0.4490928649902344, "format_reward_std_1": 0.13395369052886963, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.9493458867073059, "kl": 2.20672615652856e+34, "learning_rate": 5.719999999999999e-07, "loss": 0.0013, "num_tokens": 222312626.0, "problem_reward_mean_0": 0.29273056983947754, "problem_reward_mean_1": 0.6084877252578735, "problem_reward_std_0": 0.35340961813926697, "problem_reward_std_1": 0.4363510310649872, "reward_mean_0": 0.29273051023483276, "reward_mean_1": 0.6084877252578735, "reward_std_0": 0.23090426623821259, "reward_std_1": 0.11462394893169403, "rewards/compute_reward/mean": 0.45060910284519196, "rewards/compute_reward/std": 0.3948802947998047, "step": 108, "updated_reward_mean_0": 0.7295202016830444, "updated_reward_mean_1": 0.47697728872299194, "updated_reward_std_0": 0.15075837075710297, "updated_reward_std_1": 0.10102143883705139 }, { "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": 0.7955729166666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.5, "completions/mean_length": 1948.3294677734375, "completions/mean_terminated_length": 1547.3983154296875, "completions/min_length": 775.0, "completions/min_terminated_length": 775.0, "epoch": 0.436, "format_reward_mean_0": 0.4791666865348816, "format_reward_mean_1": 0.90625, "format_reward_std_0": 0.5048523545265198, "format_reward_std_1": 0.2918608784675598, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1875, "grad_norm": 1.6280370950698853, "kl": -1.4377271826138594e+23, "learning_rate": 5.679999999999999e-07, "loss": 0.0014, "num_tokens": 224502882.0, "problem_reward_mean_0": 0.15127316117286682, "problem_reward_mean_1": 0.4144260883331299, "problem_reward_std_0": 0.2568710446357727, "problem_reward_std_1": 0.4110967814922333, "reward_mean_0": 0.15127316117286682, "reward_mean_1": 0.4144260585308075, "reward_std_0": 0.12943512201309204, "reward_std_1": 0.1456976681947708, "rewards/compute_reward/mean": 0.28284962475299835, "rewards/compute_reward/std": 0.3339838832616806, "step": 109, "updated_reward_mean_0": 0.5569930076599121, "updated_reward_mean_1": 0.4144260585308075, "updated_reward_std_0": 0.16884170472621918, "updated_reward_std_1": 0.145697683095932 }, { "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": 0.7291666666666666, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1932.5690307617188, "completions/mean_terminated_length": 1642.1556396484375, "completions/min_length": 959.0, "completions/min_terminated_length": 959.0, "epoch": 0.44, "format_reward_mean_0": 0.75, "format_reward_mean_1": 0.9348958730697632, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.24703145027160645, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0625, "grad_norm": 1.334069013595581, "kl": 7.233157984909811e+34, "learning_rate": 5.639999999999999e-07, "loss": 0.0013, "num_tokens": 226781382.0, "problem_reward_mean_0": 0.22597861289978027, "problem_reward_mean_1": 0.5695818662643433, "problem_reward_std_0": 0.1752566546201706, "problem_reward_std_1": 0.3696480989456177, "reward_mean_0": 0.22597861289978027, "reward_mean_1": 0.5695818662643433, "reward_std_0": 0.12309014797210693, "reward_std_1": 0.17143985629081726, "rewards/compute_reward/mean": 0.3977802097797394, "rewards/compute_reward/std": 0.27245237678289413, "step": 110, "updated_reward_mean_0": 0.7479392886161804, "updated_reward_mean_1": 0.5695818662643433, "updated_reward_std_0": 0.1522241234779358, "updated_reward_std_1": 0.17143985629081726 }, { "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": 0.375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1601.16015625, "completions/mean_terminated_length": 1343.7962646484375, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.444, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9817708730697632, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.13395369052886963, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.5416666865348816, "grad_norm": 0.6394209861755371, "kl": -2.9434627822196634e+28, "learning_rate": 5.6e-07, "loss": 0.0009, "num_tokens": 228717642.0, "problem_reward_mean_0": 0.5796167850494385, "problem_reward_mean_1": 0.7310277819633484, "problem_reward_std_0": 0.4340033531188965, "problem_reward_std_1": 0.40277233719825745, "reward_mean_0": 0.5796167850494385, "reward_mean_1": 0.7310277819633484, "reward_std_0": 0.10334613174200058, "reward_std_1": 0.09113983809947968, "rewards/compute_reward/mean": 0.655322253704071, "rewards/compute_reward/std": 0.41838783025741577, "step": 111, "updated_reward_mean_0": 0.8026682734489441, "updated_reward_mean_1": 0.26947468519210815, "updated_reward_std_0": 0.051948439329862595, "updated_reward_std_1": 0.052974335849285126 }, { "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": 0.44921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.5, "completions/mean_length": 1594.2591552734375, "completions/mean_terminated_length": 1223.2914428710938, "completions/min_length": 175.5, "completions/min_terminated_length": 175.5, "epoch": 0.448, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9635416865348816, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.18767216801643372, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 1.8409779071807861, "kl": 5.261245167090968e+27, "learning_rate": 5.560000000000001e-07, "loss": 0.0009, "num_tokens": 230653102.0, "problem_reward_mean_0": 0.448722243309021, "problem_reward_mean_1": 0.7952827215194702, "problem_reward_std_0": 0.43413081765174866, "problem_reward_std_1": 0.3065131902694702, "reward_mean_0": 0.448722243309021, "reward_mean_1": 0.7952827215194702, "reward_std_0": 0.1974605917930603, "reward_std_1": 0.1676422655582428, "rewards/compute_reward/mean": 0.6220024526119232, "rewards/compute_reward/std": 0.37032201886177063, "step": 112, "updated_reward_mean_0": 0.9254268407821655, "updated_reward_mean_1": 0.47204047441482544, "updated_reward_std_0": 0.044528089463710785, "updated_reward_std_1": 0.15041500329971313 }, { "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": 0.4296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1726.963623046875, "completions/mean_terminated_length": 1481.10107421875, "completions/min_length": 674.0, "completions/min_terminated_length": 674.0, "epoch": 0.452, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.19399645924568176, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.125, "grad_norm": 0.760766327381134, "kl": -7.384599976627595e+34, "learning_rate": 5.520000000000001e-07, "loss": 0.0012, "num_tokens": 232902042.0, "problem_reward_mean_0": 0.4155661463737488, "problem_reward_mean_1": 0.8044408559799194, "problem_reward_std_0": 0.34254711866378784, "problem_reward_std_1": 0.2809666395187378, "reward_mean_0": 0.4155661463737488, "reward_mean_1": 0.8044408559799194, "reward_std_0": 0.19049564003944397, "reward_std_1": 0.19595468044281006, "rewards/compute_reward/mean": 0.6100034564733505, "rewards/compute_reward/std": 0.311756893992424, "step": 113, "updated_reward_mean_0": 0.9562026858329773, "updated_reward_mean_1": 0.6651179194450378, "updated_reward_std_0": 0.04271625727415085, "updated_reward_std_1": 0.17754045128822327 }, { "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": 0.3776041666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.5, "completions/mean_length": 1600.3255615234375, "completions/mean_terminated_length": 1340.7520751953125, "completions/min_length": 479.0, "completions/min_terminated_length": 479.0, "epoch": 0.456, "format_reward_mean_0": 0.625, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.48924607038497925, "format_reward_std_1": 0.15946698188781738, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.8481889963150024, "kl": -1.0534804075187963e+24, "learning_rate": 5.48e-07, "loss": 0.0013, "num_tokens": 234835682.0, "problem_reward_mean_0": 0.20441019535064697, "problem_reward_mean_1": 0.6336960792541504, "problem_reward_std_0": 0.3212360739707947, "problem_reward_std_1": 0.40711140632629395, "reward_mean_0": 0.20441019535064697, "reward_mean_1": 0.6336960792541504, "reward_std_0": 0.2063441276550293, "reward_std_1": 0.18638041615486145, "rewards/compute_reward/mean": 0.4190531373023987, "rewards/compute_reward/std": 0.3641737252473831, "step": 114, "updated_reward_mean_0": 0.8225446939468384, "updated_reward_mean_1": 0.5379929542541504, "updated_reward_std_0": 0.13778382539749146, "updated_reward_std_1": 0.17375671863555908 }, { "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": 0.6901041666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1838.0065307617188, "completions/mean_terminated_length": 1383.9761352539062, "completions/min_length": 506.0, "completions/min_terminated_length": 506.0, "epoch": 0.46, "format_reward_mean_0": 0.75, "format_reward_mean_1": 0.9114583730697632, "format_reward_std_0": 0.4375949800014496, "format_reward_std_1": 0.2844516634941101, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0625, "grad_norm": 0.700484573841095, "kl": 7.3834051168569445e+34, "learning_rate": 5.44e-07, "loss": 0.001, "num_tokens": 237107689.0, "problem_reward_mean_0": 0.26310625672340393, "problem_reward_mean_1": 0.4755783677101135, "problem_reward_std_0": 0.3358604609966278, "problem_reward_std_1": 0.35077083110809326, "reward_mean_0": 0.26310622692108154, "reward_mean_1": 0.4755783677101135, "reward_std_0": 0.24087560176849365, "reward_std_1": 0.21044576168060303, "rewards/compute_reward/mean": 0.36934229731559753, "rewards/compute_reward/std": 0.34331564605236053, "step": 115, "updated_reward_mean_0": 0.7134528160095215, "updated_reward_mean_1": 0.36750543117523193, "updated_reward_std_0": 0.17070260643959045, "updated_reward_std_1": 0.16346287727355957 }, { "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": 0.4114583333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.5, "completions/mean_length": 1653.408935546875, "completions/mean_terminated_length": 1377.6685791015625, "completions/min_length": 531.5, "completions/min_terminated_length": 531.5, "epoch": 0.464, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9296875, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.25600647926330566, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.0625, "grad_norm": 0.3549410104751587, "kl": -1.8052329397325684e+31, "learning_rate": 5.4e-07, "loss": 0.0009, "num_tokens": 239301419.0, "problem_reward_mean_0": 0.4086044430732727, "problem_reward_mean_1": 0.6918016672134399, "problem_reward_std_0": 0.41246750950813293, "problem_reward_std_1": 0.3508613407611847, "reward_mean_0": 0.4086044430732727, "reward_mean_1": 0.6918016672134399, "reward_std_0": 0.1974153220653534, "reward_std_1": 0.19527791440486908, "rewards/compute_reward/mean": 0.5502030104398727, "rewards/compute_reward/std": 0.3816644251346588, "step": 116, "updated_reward_mean_0": 0.8641723394393921, "updated_reward_mean_1": 0.4339838922023773, "updated_reward_std_0": 0.11821456998586655, "updated_reward_std_1": 0.1310281753540039 }, { "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": 0.2760416666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.5, "completions/mean_length": 1520.1172485351562, "completions/mean_terminated_length": 1317.893310546875, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.468, "format_reward_mean_0": 1.0, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 0.6306811571121216, "kl": -5.047892831075133e+34, "learning_rate": 5.36e-07, "loss": 0.0009, "num_tokens": 241462760.0, "problem_reward_mean_0": 0.49106916785240173, "problem_reward_mean_1": 0.9182124137878418, "problem_reward_std_0": 0.37121227383613586, "problem_reward_std_1": 0.21508526802062988, "reward_mean_0": 0.49106916785240173, "reward_mean_1": 0.9182124137878418, "reward_std_0": 0.16401797533035278, "reward_std_1": 0.12436497211456299, "rewards/compute_reward/mean": 0.7046407908201218, "rewards/compute_reward/std": 0.2931487560272217, "step": 117, "updated_reward_mean_0": 1.0, "updated_reward_mean_1": 0.6322927474975586, "updated_reward_std_0": 0.0, "updated_reward_std_1": 0.0811978429555893 }, { "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": 0.41015624999999994, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1662.2994995117188, "completions/mean_terminated_length": 1416.6953735351562, "completions/min_length": 498.5, "completions/min_terminated_length": 498.5, "epoch": 0.472, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.984375, "format_reward_std_0": 0.0, "format_reward_std_1": 0.12418138980865479, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.5625, "grad_norm": 0.2853580713272095, "kl": 7.106331863163145e+25, "learning_rate": 5.32e-07, "loss": 0.001, "num_tokens": 243490136.0, "problem_reward_mean_0": 0.43963679671287537, "problem_reward_mean_1": 0.7323884963989258, "problem_reward_std_0": 0.4062879681587219, "problem_reward_std_1": 0.38794785737991333, "reward_mean_0": 0.43963679671287537, "reward_mean_1": 0.7323884963989258, "reward_std_0": 0.26638612151145935, "reward_std_1": 0.12467338889837265, "rewards/compute_reward/mean": 0.5860126465559006, "rewards/compute_reward/std": 0.3971179127693176, "step": 118, "updated_reward_mean_0": 0.8576388359069824, "updated_reward_mean_1": 0.4587339758872986, "updated_reward_std_0": 0.055180005729198456, "updated_reward_std_1": 0.08493881672620773 }, { "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": 0.5859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1735.248779296875, "completions/mean_terminated_length": 1286.7594604492188, "completions/min_length": 434.5, "completions/min_terminated_length": 434.5, "epoch": 0.476, "format_reward_mean_0": 0.8333333730697632, "format_reward_mean_1": 0.9479166865348816, "format_reward_std_0": 0.3766217827796936, "format_reward_std_1": 0.22248496115207672, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2291666716337204, "grad_norm": 1.502332091331482, "kl": -5.855511393064051e+31, "learning_rate": 5.28e-07, "loss": 0.001, "num_tokens": 245729390.0, "problem_reward_mean_0": 0.3197817802429199, "problem_reward_mean_1": 0.37144801020622253, "problem_reward_std_0": 0.3671751916408539, "problem_reward_std_1": 0.3833369016647339, "reward_mean_0": 0.3197817802429199, "reward_mean_1": 0.37144801020622253, "reward_std_0": 0.13913121819496155, "reward_std_1": 0.09710101783275604, "rewards/compute_reward/mean": 0.34561489522457123, "rewards/compute_reward/std": 0.3752560168504715, "step": 119, "updated_reward_mean_0": 0.6479067802429199, "updated_reward_mean_1": 0.3342991769313812, "updated_reward_std_0": 0.20727473497390747, "updated_reward_std_1": 0.08852499723434448 }, { "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": 0.26041666666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.5, "completions/mean_length": 1471.967529296875, "completions/mean_terminated_length": 1266.8869018554688, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.48, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 0.8150061368942261, "kl": -6.490371073216021e+33, "learning_rate": 5.24e-07, "loss": 0.0008, "num_tokens": 247643700.0, "problem_reward_mean_0": 0.6445574760437012, "problem_reward_mean_1": 0.6582525968551636, "problem_reward_std_0": 0.33509454131126404, "problem_reward_std_1": 0.3484085500240326, "reward_mean_0": 0.6445574760437012, "reward_mean_1": 0.6582525968551636, "reward_std_0": 0.13122057914733887, "reward_std_1": 0.12009565532207489, "rewards/compute_reward/mean": 0.6514050364494324, "rewards/compute_reward/std": 0.3417515456676483, "step": 120, "updated_reward_mean_0": 0.7880975604057312, "updated_reward_mean_1": 0.24158594012260437, "updated_reward_std_0": 0.05405968427658081, "updated_reward_std_1": 0.08067431300878525 }, { "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": 0.35677083333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.5, "completions/mean_length": 1349.9583740234375, "completions/mean_terminated_length": 967.29150390625, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.484, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.8307291865348816, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.37548020482063293, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 1.01509428024292, "kl": 4.6442209320174385e+34, "learning_rate": 5.2e-07, "loss": 0.0007, "num_tokens": 249571600.0, "problem_reward_mean_0": 0.5593501329421997, "problem_reward_mean_1": 0.6643652319908142, "problem_reward_std_0": 0.4379761815071106, "problem_reward_std_1": 0.4504590630531311, "reward_mean_0": 0.5593501329421997, "reward_mean_1": 0.6643652319908142, "reward_std_0": 0.06549635529518127, "reward_std_1": 0.19841808080673218, "rewards/compute_reward/mean": 0.6118576526641846, "rewards/compute_reward/std": 0.44421759247779846, "step": 121, "updated_reward_mean_0": 0.8617424368858337, "updated_reward_mean_1": 0.21174024045467377, "updated_reward_std_0": 0.05955884978175163, "updated_reward_std_1": 0.1363115906715393 }, { "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": 0.44010416666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1681.3958740234375, "completions/mean_terminated_length": 1398.3632202148438, "completions/min_length": 615.0, "completions/min_terminated_length": 615.0, "epoch": 0.488, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9114583730697632, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.2844516634941101, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.125, "grad_norm": 1.145526647567749, "kl": 2.7051300249951927e+30, "learning_rate": 5.16e-07, "loss": 0.0011, "num_tokens": 251692802.0, "problem_reward_mean_0": 0.3363882303237915, "problem_reward_mean_1": 0.6195026636123657, "problem_reward_std_0": 0.30298393964767456, "problem_reward_std_1": 0.37748685479164124, "reward_mean_0": 0.3363882303237915, "reward_mean_1": 0.6195026636123657, "reward_std_0": 0.1676650047302246, "reward_std_1": 0.20722854137420654, "rewards/compute_reward/mean": 0.4779454469680786, "rewards/compute_reward/std": 0.3402353972196579, "step": 122, "updated_reward_mean_0": 0.8206211924552917, "updated_reward_mean_1": 0.6033568382263184, "updated_reward_std_0": 0.11838778108358383, "updated_reward_std_1": 0.1984778344631195 }, { "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": 0.20703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1423.9244995117188, "completions/mean_terminated_length": 1257.3297119140625, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.492, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.8932291865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.30922460556030273, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.786781370639801, "kl": -8.867556944465404e+22, "learning_rate": 5.12e-07, "loss": 0.001, "num_tokens": 253646046.0, "problem_reward_mean_0": 0.4732028841972351, "problem_reward_mean_1": 0.6684221029281616, "problem_reward_std_0": 0.37080028653144836, "problem_reward_std_1": 0.4096193015575409, "reward_mean_0": 0.4732028841972351, "reward_mean_1": 0.6684221029281616, "reward_std_0": 0.1750127077102661, "reward_std_1": 0.17125974595546722, "rewards/compute_reward/mean": 0.5708124488592148, "rewards/compute_reward/std": 0.39020979404449463, "step": 123, "updated_reward_mean_0": 0.8830906748771667, "updated_reward_mean_1": 0.43665120005607605, "updated_reward_std_0": 0.04664095491170883, "updated_reward_std_1": 0.12154966592788696 }, { "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": 0.125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1158.8112182617188, "completions/mean_terminated_length": 1031.7842712402344, "completions/min_length": 165.5, "completions/min_terminated_length": 165.5, "epoch": 0.496, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.1016615778207779, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.4375, "grad_norm": 0.756936252117157, "kl": -3.694874897706506e+29, "learning_rate": 5.079999999999999e-07, "loss": 0.001, "num_tokens": 255301027.0, "problem_reward_mean_0": 0.5619725584983826, "problem_reward_mean_1": 0.6731367111206055, "problem_reward_std_0": 0.35829389095306396, "problem_reward_std_1": 0.41455158591270447, "reward_mean_0": 0.5619725584983826, "reward_mean_1": 0.6731367111206055, "reward_std_0": 0.20615176856517792, "reward_std_1": 0.1732945293188095, "rewards/compute_reward/mean": 0.617554634809494, "rewards/compute_reward/std": 0.3864227533340454, "step": 124, "updated_reward_mean_0": 0.8637582659721375, "updated_reward_mean_1": 0.38928258419036865, "updated_reward_std_0": 0.019287919625639915, "updated_reward_std_1": 0.126216322183609 }, { "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": 0.140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1258.7969360351562, "completions/mean_terminated_length": 1128.8088073730469, "completions/min_length": 375.5, "completions/min_terminated_length": 375.5, "epoch": 0.5, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9505208730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.21714921295642853, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 0.7874290347099304, "kl": 3.784607460833037e+35, "learning_rate": 5.04e-07, "loss": 0.0009, "num_tokens": 257133935.0, "problem_reward_mean_0": 0.6854739189147949, "problem_reward_mean_1": 0.8179911971092224, "problem_reward_std_0": 0.2943665683269501, "problem_reward_std_1": 0.29932767152786255, "reward_mean_0": 0.6854739189147949, "reward_mean_1": 0.8179911971092224, "reward_std_0": 0.11973455548286438, "reward_std_1": 0.17870251834392548, "rewards/compute_reward/mean": 0.7517325580120087, "rewards/compute_reward/std": 0.2968471199274063, "step": 125, "updated_reward_mean_0": 0.984150767326355, "updated_reward_mean_1": 0.4764113426208496, "updated_reward_std_0": 0.02147727459669113, "updated_reward_std_1": 0.14917021989822388 }, { "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": 0.14453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.5, "completions/mean_length": 1299.1224365234375, "completions/mean_terminated_length": 1173.602294921875, "completions/min_length": 260.5, "completions/min_terminated_length": 260.5, "epoch": 0.504, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.375, "grad_norm": 0.5011930465698242, "kl": 7.355753883001576e+34, "learning_rate": 5e-07, "loss": 0.0008, "num_tokens": 258841355.0, "problem_reward_mean_0": 0.6006567478179932, "problem_reward_mean_1": 0.8760662078857422, "problem_reward_std_0": 0.3798164427280426, "problem_reward_std_1": 0.28873705863952637, "reward_mean_0": 0.6006567478179932, "reward_mean_1": 0.8760662078857422, "reward_std_0": 0.1255209743976593, "reward_std_1": 0.13362464308738708, "rewards/compute_reward/mean": 0.7383614480495453, "rewards/compute_reward/std": 0.3342767506837845, "step": 126, "updated_reward_mean_0": 0.96263587474823, "updated_reward_mean_1": 0.5205974578857422, "updated_reward_std_0": 0.07006970047950745, "updated_reward_std_1": 0.09364422410726547 }, { "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": 0.17838541666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.5, "completions/mean_length": 1321.8021240234375, "completions/mean_terminated_length": 1165.0025024414062, "completions/min_length": 295.5, "completions/min_terminated_length": 295.5, "epoch": 0.508, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9505208730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.21714918315410614, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3541666865348816, "grad_norm": 0.6785868406295776, "kl": 1.6992124020139912e+25, "learning_rate": 4.96e-07, "loss": 0.0008, "num_tokens": 260790733.0, "problem_reward_mean_0": 0.6115133166313171, "problem_reward_mean_1": 0.6100690364837646, "problem_reward_std_0": 0.4115138053894043, "problem_reward_std_1": 0.45934244990348816, "reward_mean_0": 0.6115133762359619, "reward_mean_1": 0.6100690364837646, "reward_std_0": 0.16640329360961914, "reward_std_1": 0.14064884185791016, "rewards/compute_reward/mean": 0.6107911765575409, "rewards/compute_reward/std": 0.43542812764644623, "step": 127, "updated_reward_mean_0": 0.7561561465263367, "updated_reward_mean_1": 0.20478200912475586, "updated_reward_std_0": 0.06664437055587769, "updated_reward_std_1": 0.051290612667798996 }, { "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": 0.23177083333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.5, "completions/mean_length": 1406.6979370117188, "completions/mean_terminated_length": 1213.1873779296875, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.512, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9244791865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.2645746171474457, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 0.7034213542938232, "kl": 7.501113206386754e+33, "learning_rate": 4.92e-07, "loss": 0.0013, "num_tokens": 262758495.0, "problem_reward_mean_0": 0.4283220171928406, "problem_reward_mean_1": 0.6546205878257751, "problem_reward_std_0": 0.3630545735359192, "problem_reward_std_1": 0.34303760528564453, "reward_mean_0": 0.4283220171928406, "reward_mean_1": 0.6546205878257751, "reward_std_0": 0.19003622233867645, "reward_std_1": 0.10383886098861694, "rewards/compute_reward/mean": 0.5414712876081467, "rewards/compute_reward/std": 0.35304608941078186, "step": 128, "updated_reward_mean_0": 0.7984954118728638, "updated_reward_mean_1": 0.5833312273025513, "updated_reward_std_0": 0.1088639423251152, "updated_reward_std_1": 0.08960666507482529 }, { "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": 0.3463541666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1575.5833740234375, "completions/mean_terminated_length": 1357.4707641601562, "completions/min_length": 469.5, "completions/min_terminated_length": 469.5, "epoch": 0.516, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9479166865348816, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.22248496115207672, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 1.0167416334152222, "kl": -9.870350123723989e+34, "learning_rate": 4.879999999999999e-07, "loss": 0.0011, "num_tokens": 264724275.0, "problem_reward_mean_0": 0.4385913014411926, "problem_reward_mean_1": 0.6934021711349487, "problem_reward_std_0": 0.3854098916053772, "problem_reward_std_1": 0.34241020679473877, "reward_mean_0": 0.4385913014411926, "reward_mean_1": 0.6934021711349487, "reward_std_0": 0.23316437005996704, "reward_std_1": 0.14835625886917114, "rewards/compute_reward/mean": 0.5659966915845871, "rewards/compute_reward/std": 0.3639100342988968, "step": 129, "updated_reward_mean_0": 0.8734701871871948, "updated_reward_mean_1": 0.6030003428459167, "updated_reward_std_0": 0.12021619081497192, "updated_reward_std_1": 0.13874387741088867 }, { "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": 0.1171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.5, "completions/mean_length": 1226.4922485351562, "completions/mean_terminated_length": 1120.1719970703125, "completions/min_length": 278.5, "completions/min_terminated_length": 278.5, "epoch": 0.52, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.11351142078638077, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.5208333730697632, "grad_norm": 0.3823786675930023, "kl": -3.76351378201092e+32, "learning_rate": 4.839999999999999e-07, "loss": 0.0009, "num_tokens": 266408149.0, "problem_reward_mean_0": 0.4799107015132904, "problem_reward_mean_1": 0.6828497648239136, "problem_reward_std_0": 0.4445419907569885, "problem_reward_std_1": 0.41486412286758423, "reward_mean_0": 0.4799107015132904, "reward_mean_1": 0.6828497648239136, "reward_std_0": 0.28598666191101074, "reward_std_1": 0.13724148273468018, "rewards/compute_reward/mean": 0.5813802033662796, "rewards/compute_reward/std": 0.4297030419111252, "step": 130, "updated_reward_mean_0": 0.8035714626312256, "updated_reward_mean_1": 0.3151041865348816, "updated_reward_std_0": 0.07576143741607666, "updated_reward_std_1": 0.09435179829597473 }, { "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": 0.23697916666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1539.5899047851562, "completions/mean_terminated_length": 1387.6234741210938, "completions/min_length": 653.0, "completions/min_terminated_length": 653.0, "epoch": 0.524, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.19399645924568176, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.0416666679084301, "grad_norm": 2.0509226322174072, "kl": -2.848664744950818e+33, "learning_rate": 4.8e-07, "loss": 0.0012, "num_tokens": 268381901.0, "problem_reward_mean_0": 0.3185078501701355, "problem_reward_mean_1": 0.661020815372467, "problem_reward_std_0": 0.23197227716445923, "problem_reward_std_1": 0.34183502197265625, "reward_mean_0": 0.3185078501701355, "reward_mean_1": 0.661020815372467, "reward_std_0": 0.12523798644542694, "reward_std_1": 0.1726803034543991, "rewards/compute_reward/mean": 0.48976433277130127, "rewards/compute_reward/std": 0.28690364956855774, "step": 131, "updated_reward_mean_0": 0.8435437083244324, "updated_reward_mean_1": 0.644527792930603, "updated_reward_std_0": 0.09083735942840576, "updated_reward_std_1": 0.16908621788024902 }, { "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": 0.3177083333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1529.463623046875, "completions/mean_terminated_length": 1277.4077758789062, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.528, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.15946698188781738, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3125, "grad_norm": 0.7544950842857361, "kl": -1.378363145749898e+33, "learning_rate": 4.76e-07, "loss": 0.0008, "num_tokens": 270400265.0, "problem_reward_mean_0": 0.4229166507720947, "problem_reward_mean_1": 0.4896484315395355, "problem_reward_std_0": 0.4015576243400574, "problem_reward_std_1": 0.4212180972099304, "reward_mean_0": 0.4229166507720947, "reward_mean_1": 0.4896484315395355, "reward_std_0": 0.2503393888473511, "reward_std_1": 0.14832672476768494, "rewards/compute_reward/mean": 0.4562825411558151, "rewards/compute_reward/std": 0.4113878458738327, "step": 132, "updated_reward_mean_0": 0.625, "updated_reward_mean_1": 0.22662761807441711, "updated_reward_std_0": 0.10902484506368637, "updated_reward_std_1": 0.10352512449026108 }, { "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": 0.12630208333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.5, "completions/mean_length": 1148.72265625, "completions/mean_terminated_length": 1018.9928588867188, "completions/min_length": 164.5, "completions/min_terminated_length": 164.5, "epoch": 0.532, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.375, "grad_norm": 2.050503730773926, "kl": -4.52238086223278e+22, "learning_rate": 4.7199999999999994e-07, "loss": 0.0007, "num_tokens": 272207222.0, "problem_reward_mean_0": 0.5325901508331299, "problem_reward_mean_1": 0.6818804740905762, "problem_reward_std_0": 0.4669823944568634, "problem_reward_std_1": 0.4221253991127014, "reward_mean_0": 0.5325901508331299, "reward_mean_1": 0.6818804740905762, "reward_std_0": 0.2774626612663269, "reward_std_1": 0.22443680465221405, "rewards/compute_reward/mean": 0.607235312461853, "rewards/compute_reward/std": 0.4445538818836212, "step": 133, "updated_reward_mean_0": 0.9364035129547119, "updated_reward_mean_1": 0.27563047409057617, "updated_reward_std_0": 0.15574640035629272, "updated_reward_std_1": 0.16305828094482422 }, { "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": 0.3177083333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 1969.0, "completions/mean_length": 1496.9583740234375, "completions/mean_terminated_length": 1239.8906860351562, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.536, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.9765625, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.15148578584194183, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.375, "grad_norm": 1.2994359731674194, "kl": -1.8389384707310847e+34, "learning_rate": 4.68e-07, "loss": 0.0009, "num_tokens": 274156667.0, "problem_reward_mean_0": 0.49483737349510193, "problem_reward_mean_1": 0.7238320112228394, "problem_reward_std_0": 0.44702085852622986, "problem_reward_std_1": 0.3746803402900696, "reward_mean_0": 0.49483737349510193, "reward_mean_1": 0.7238320112228394, "reward_std_0": 0.3257061243057251, "reward_std_1": 0.15424224734306335, "rewards/compute_reward/mean": 0.6093346625566483, "rewards/compute_reward/std": 0.41085056960582733, "step": 134, "updated_reward_mean_0": 0.8539474010467529, "updated_reward_mean_1": 0.3216066360473633, "updated_reward_std_0": 0.05108107253909111, "updated_reward_std_1": 0.1196829304099083 }, { "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": 0.22786458333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.5, "completions/mean_length": 1478.002685546875, "completions/mean_terminated_length": 1311.9828491210938, "completions/min_length": 366.5, "completions/min_terminated_length": 366.5, "epoch": 0.54, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4375, "grad_norm": 0.3133186399936676, "kl": 1.5111508269139153e+30, "learning_rate": 4.64e-07, "loss": 0.0007, "num_tokens": 276191255.0, "problem_reward_mean_0": 0.7501543760299683, "problem_reward_mean_1": 0.905160129070282, "problem_reward_std_0": 0.3677656352519989, "problem_reward_std_1": 0.23292405903339386, "reward_mean_0": 0.7501543760299683, "reward_mean_1": 0.905160129070282, "reward_std_0": 0.21276089549064636, "reward_std_1": 0.1220514178276062, "rewards/compute_reward/mean": 0.8276572227478027, "rewards/compute_reward/std": 0.3003448471426964, "step": 135, "updated_reward_mean_0": 0.979205310344696, "updated_reward_mean_1": 0.34097224473953247, "updated_reward_std_0": 0.058816444128751755, "updated_reward_std_1": 0.0788278728723526 }, { "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": 0.3776041666666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.5, "completions/mean_length": 1634.7864990234375, "completions/mean_terminated_length": 1383.4221801757812, "completions/min_length": 729.0, "completions/min_terminated_length": 729.0, "epoch": 0.544, "format_reward_mean_0": 0.7916666865348816, "format_reward_mean_1": 0.9427083730697632, "format_reward_std_0": 0.41041406989097595, "format_reward_std_1": 0.23270227015018463, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.25, "grad_norm": 1.381812334060669, "kl": 5.956718340486202e+35, "learning_rate": 4.6e-07, "loss": 0.0012, "num_tokens": 278327667.0, "problem_reward_mean_0": 0.2962833642959595, "problem_reward_mean_1": 0.40121567249298096, "problem_reward_std_0": 0.3576852083206177, "problem_reward_std_1": 0.3666583299636841, "reward_mean_0": 0.29628342390060425, "reward_mean_1": 0.40121567249298096, "reward_std_0": 0.21566295623779297, "reward_std_1": 0.1811845898628235, "rewards/compute_reward/mean": 0.3487495183944702, "rewards/compute_reward/std": 0.3621717542409897, "step": 136, "updated_reward_mean_0": 0.5671131610870361, "updated_reward_mean_1": 0.29704901576042175, "updated_reward_std_0": 0.08966273069381714, "updated_reward_std_1": 0.14488866925239563 }, { "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": 0.24479166666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.5, "completions/mean_length": 1511.91015625, "completions/mean_terminated_length": 1382.2269287109375, "completions/min_length": 690.0, "completions/min_terminated_length": 690.0, "epoch": 0.548, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.0720745176076889, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3125, "grad_norm": 0.5390975475311279, "kl": 2.772253266203841e+34, "learning_rate": 4.56e-07, "loss": 0.0013, "num_tokens": 280220732.0, "problem_reward_mean_0": 0.4372023940086365, "problem_reward_mean_1": 0.7896019816398621, "problem_reward_std_0": 0.3571062386035919, "problem_reward_std_1": 0.35261455178260803, "reward_mean_0": 0.4372023940086365, "reward_mean_1": 0.7896019816398621, "reward_std_0": 0.2960062623023987, "reward_std_1": 0.1913420855998993, "rewards/compute_reward/mean": 0.6134021729230881, "rewards/compute_reward/std": 0.3548603802919388, "step": 137, "updated_reward_mean_0": 0.941964328289032, "updated_reward_mean_1": 0.5965216159820557, "updated_reward_std_0": 0.14405637979507446, "updated_reward_std_1": 0.1250441074371338 }, { "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": 0.2864583333333333, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1485.8919677734375, "completions/mean_terminated_length": 1272.5224609375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.552, "format_reward_mean_0": 0.8333333730697632, "format_reward_mean_1": 0.8984375, "format_reward_std_0": 0.3766217827796936, "format_reward_std_1": 0.3024662137031555, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 1.4223382472991943, "kl": 2.406236055921564e+31, "learning_rate": 4.5199999999999997e-07, "loss": 0.001, "num_tokens": 282110581.0, "problem_reward_mean_0": 0.48184525966644287, "problem_reward_mean_1": 0.7062263488769531, "problem_reward_std_0": 0.39938032627105713, "problem_reward_std_1": 0.3976229727268219, "reward_mean_0": 0.48184525966644287, "reward_mean_1": 0.7062263488769531, "reward_std_0": 0.24612578749656677, "reward_std_1": 0.19253391027450562, "rewards/compute_reward/mean": 0.594035804271698, "rewards/compute_reward/std": 0.3985016644001007, "step": 138, "updated_reward_mean_0": 0.9269841909408569, "updated_reward_mean_1": 0.43539297580718994, "updated_reward_std_0": 0.12089534103870392, "updated_reward_std_1": 0.15342716872692108 }, { "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": 0.23046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.5, "completions/mean_length": 1391.9596557617188, "completions/mean_terminated_length": 1202.5043334960938, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.556, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.19399645924568176, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.375, "grad_norm": 0.880232036113739, "kl": 9.919973802855046e+34, "learning_rate": 4.48e-07, "loss": 0.0012, "num_tokens": 283965559.0, "problem_reward_mean_0": 0.27905961871147156, "problem_reward_mean_1": 0.6380784511566162, "problem_reward_std_0": 0.3532927930355072, "problem_reward_std_1": 0.4138459861278534, "reward_mean_0": 0.27905961871147156, "reward_mean_1": 0.6380784511566162, "reward_std_0": 0.22151832282543182, "reward_std_1": 0.17097651958465576, "rewards/compute_reward/mean": 0.4585690349340439, "rewards/compute_reward/std": 0.3835693895816803, "step": 139, "updated_reward_mean_0": 0.7758399844169617, "updated_reward_mean_1": 0.494849294424057, "updated_reward_std_0": 0.08730538934469223, "updated_reward_std_1": 0.16361084580421448 }, { "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": 0.16145833333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1350.328125, "completions/mean_terminated_length": 1218.9100952148438, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.56, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9817708730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.13395369052886963, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 1.1878442764282227, "kl": 6.517848849465921e+20, "learning_rate": 4.44e-07, "loss": 0.001, "num_tokens": 285902024.0, "problem_reward_mean_0": 0.374627947807312, "problem_reward_mean_1": 0.7538101077079773, "problem_reward_std_0": 0.3719237744808197, "problem_reward_std_1": 0.3418193757534027, "reward_mean_0": 0.374627947807312, "reward_mean_1": 0.7538101673126221, "reward_std_0": 0.11379903554916382, "reward_std_1": 0.12451478838920593, "rewards/compute_reward/mean": 0.5642190277576447, "rewards/compute_reward/std": 0.3568715751171112, "step": 140, "updated_reward_mean_0": 0.8327133059501648, "updated_reward_mean_1": 0.5308935046195984, "updated_reward_std_0": 0.05623961240053177, "updated_reward_std_1": 0.10683713853359222 }, { "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": 0.20052083333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.5, "completions/mean_length": 1514.6185302734375, "completions/mean_terminated_length": 1386.8685302734375, "completions/min_length": 540.0, "completions/min_terminated_length": 540.0, "epoch": 0.564, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.19399645924568176, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2708333432674408, "grad_norm": 0.5141999125480652, "kl": -Infinity, "learning_rate": 4.3999999999999997e-07, "loss": 0.0011, "num_tokens": 287925975.0, "problem_reward_mean_0": 0.39118146896362305, "problem_reward_mean_1": 0.7983565330505371, "problem_reward_std_0": 0.38106563687324524, "problem_reward_std_1": 0.3068094849586487, "reward_mean_0": 0.39118146896362305, "reward_mean_1": 0.7983565330505371, "reward_std_0": 0.20498153567314148, "reward_std_1": 0.17083458602428436, "rewards/compute_reward/mean": 0.5947690010070801, "rewards/compute_reward/std": 0.34393754601478577, "step": 141, "updated_reward_mean_0": 0.9242343306541443, "updated_reward_mean_1": 0.6447107791900635, "updated_reward_std_0": 0.07401987910270691, "updated_reward_std_1": 0.14418095350265503 }, { "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": 0.3515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1655.8932495117188, "completions/mean_terminated_length": 1443.2581787109375, "completions/min_length": 699.0, "completions/min_terminated_length": 699.0, "epoch": 0.568, "format_reward_mean_0": 0.8125, "format_reward_mean_1": 0.8072916865348816, "format_reward_std_0": 0.3944427967071533, "format_reward_std_1": 0.3949405252933502, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1666666716337204, "grad_norm": 1.6649343967437744, "kl": -7.380543494631391e+31, "learning_rate": 4.36e-07, "loss": 0.0008, "num_tokens": 290132935.0, "problem_reward_mean_0": 0.37850260734558105, "problem_reward_mean_1": 0.5540118217468262, "problem_reward_std_0": 0.4051676094532013, "problem_reward_std_1": 0.4160943627357483, "reward_mean_0": 0.37850260734558105, "reward_mean_1": 0.5540118217468262, "reward_std_0": 0.1628473848104477, "reward_std_1": 0.25139501690864563, "rewards/compute_reward/mean": 0.4662572145462036, "rewards/compute_reward/std": 0.410631000995636, "step": 142, "updated_reward_mean_0": 0.8010234832763672, "updated_reward_mean_1": 0.3592911660671234, "updated_reward_std_0": 0.1717694252729416, "updated_reward_std_1": 0.17564350366592407 }, { "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": 0.1171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1954.5, "completions/mean_length": 1345.40625, "completions/mean_terminated_length": 1252.4825439453125, "completions/min_length": 653.5, "completions/min_terminated_length": 653.5, "epoch": 0.572, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0720745176076889, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1458333432674408, "grad_norm": 2.0634281635284424, "kl": -8.047212310907288e+30, "learning_rate": 4.3199999999999995e-07, "loss": 0.0016, "num_tokens": 292115613.0, "problem_reward_mean_0": 0.4380517899990082, "problem_reward_mean_1": 0.7506627440452576, "problem_reward_std_0": 0.25926917791366577, "problem_reward_std_1": 0.3573533296585083, "reward_mean_0": 0.43805181980133057, "reward_mean_1": 0.7506627440452576, "reward_std_0": 0.10772191733121872, "reward_std_1": 0.08287813514471054, "rewards/compute_reward/mean": 0.5943572670221329, "rewards/compute_reward/std": 0.30831125378608704, "step": 143, "updated_reward_mean_0": 0.8270286917686462, "updated_reward_mean_1": 0.7506627440452576, "updated_reward_std_0": 0.018343374133110046, "updated_reward_std_1": 0.08287814259529114 }, { "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": 0.03385416666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 1907.5, "completions/mean_length": 1123.9557495117188, "completions/mean_terminated_length": 1091.7223510742188, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.576, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1666666716337204, "grad_norm": 1.4378365278244019, "kl": -4.90541344269651e+34, "learning_rate": 4.2799999999999997e-07, "loss": 0.0012, "num_tokens": 293895354.0, "problem_reward_mean_0": 0.5320332050323486, "problem_reward_mean_1": 0.8304901123046875, "problem_reward_std_0": 0.29679644107818604, "problem_reward_std_1": 0.26506322622299194, "reward_mean_0": 0.5320332050323486, "reward_mean_1": 0.8304901123046875, "reward_std_0": 0.25358936190605164, "reward_std_1": 0.16859325766563416, "rewards/compute_reward/mean": 0.6812616586685181, "rewards/compute_reward/std": 0.280929833650589, "step": 144, "updated_reward_mean_0": 0.9715369939804077, "updated_reward_mean_1": 0.683002233505249, "updated_reward_std_0": 0.05060317739844322, "updated_reward_std_1": 0.12273178994655609 }, { "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": 0.078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.5, "completions/mean_length": 1192.5026245117188, "completions/mean_terminated_length": 1120.49951171875, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.58, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.96875, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.17421962320804596, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2291666716337204, "grad_norm": 0.8423560857772827, "kl": -4.622428753770106e+35, "learning_rate": 4.24e-07, "loss": 0.001, "num_tokens": 295692070.0, "problem_reward_mean_0": 0.5450748801231384, "problem_reward_mean_1": 0.679495096206665, "problem_reward_std_0": 0.3662242889404297, "problem_reward_std_1": 0.40628519654273987, "reward_mean_0": 0.5450748801231384, "reward_mean_1": 0.679495096206665, "reward_std_0": 0.16708998382091522, "reward_std_1": 0.13472560048103333, "rewards/compute_reward/mean": 0.6122849881649017, "rewards/compute_reward/std": 0.3862547129392624, "step": 145, "updated_reward_mean_0": 0.8104519248008728, "updated_reward_mean_1": 0.40382546186447144, "updated_reward_std_0": 0.055936168879270554, "updated_reward_std_1": 0.09060697257518768 }, { "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": 0.15234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1438.2005615234375, "completions/mean_terminated_length": 1331.69921875, "completions/min_length": 578.5, "completions/min_terminated_length": 578.5, "epoch": 0.584, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0833333358168602, "grad_norm": 1.1473488807678223, "kl": -5.0389379125647015e+29, "learning_rate": 4.1999999999999995e-07, "loss": 0.0011, "num_tokens": 297722176.0, "problem_reward_mean_0": 0.30353671312332153, "problem_reward_mean_1": 0.6662158966064453, "problem_reward_std_0": 0.348347544670105, "problem_reward_std_1": 0.3295122981071472, "reward_mean_0": 0.30353671312332153, "reward_mean_1": 0.6662158966064453, "reward_std_0": 0.2619614601135254, "reward_std_1": 0.1815815418958664, "rewards/compute_reward/mean": 0.4848763048648834, "rewards/compute_reward/std": 0.3389298915863037, "step": 146, "updated_reward_mean_0": 0.8564633131027222, "updated_reward_mean_1": 0.5744190812110901, "updated_reward_std_0": 0.07693927735090256, "updated_reward_std_1": 0.15841850638389587 }, { "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": 0.16145833333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1399.1927490234375, "completions/mean_terminated_length": 1279.5173950195312, "completions/min_length": 423.5, "completions/min_terminated_length": 423.5, "epoch": 0.588, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.984375, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.12418138235807419, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 1.0303088426589966, "kl": 2.6884559289746997e+35, "learning_rate": 4.1599999999999997e-07, "loss": 0.002, "num_tokens": 299579272.0, "problem_reward_mean_0": 0.4564593732357025, "problem_reward_mean_1": 0.7530735731124878, "problem_reward_std_0": 0.4416990280151367, "problem_reward_std_1": 0.3708077073097229, "reward_mean_0": 0.4564593732357025, "reward_mean_1": 0.7530735731124878, "reward_std_0": 0.15424057841300964, "reward_std_1": 0.05724193900823593, "rewards/compute_reward/mean": 0.6047664433717728, "rewards/compute_reward/std": 0.4062533527612686, "step": 147, "updated_reward_mean_0": 0.8058878779411316, "updated_reward_mean_1": 0.3832818269729614, "updated_reward_std_0": 0.09022840857505798, "updated_reward_std_1": 0.0475979745388031 }, { "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": 0.2109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.5, "completions/mean_length": 1388.4284057617188, "completions/mean_terminated_length": 1217.172119140625, "completions/min_length": 384.5, "completions/min_terminated_length": 384.5, "epoch": 0.592, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9635416865348816, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.18767215311527252, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2083333432674408, "grad_norm": 0.4873388409614563, "kl": -1.2563891284604495e+32, "learning_rate": 4.12e-07, "loss": 0.0009, "num_tokens": 301431492.0, "problem_reward_mean_0": 0.6063098311424255, "problem_reward_mean_1": 0.8106030225753784, "problem_reward_std_0": 0.4002746343612671, "problem_reward_std_1": 0.28863927721977234, "reward_mean_0": 0.6063098311424255, "reward_mean_1": 0.8106030225753784, "reward_std_0": 0.19492948055267334, "reward_std_1": 0.2007373422384262, "rewards/compute_reward/mean": 0.7084563970565796, "rewards/compute_reward/std": 0.3444569557905197, "step": 148, "updated_reward_mean_0": 0.9583333730697632, "updated_reward_mean_1": 0.4746062755584717, "updated_reward_std_0": 0.04577406868338585, "updated_reward_std_1": 0.1251227855682373 }, { "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": 0.1328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1375.6588745117188, "completions/mean_terminated_length": 1274.7808227539062, "completions/min_length": 572.0, "completions/min_terminated_length": 572.0, "epoch": 0.596, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 1.7272558212280273, "kl": -6.625587138620083e+32, "learning_rate": 4.0799999999999995e-07, "loss": 0.0009, "num_tokens": 303404913.0, "problem_reward_mean_0": 0.5909336805343628, "problem_reward_mean_1": 0.9025294184684753, "problem_reward_std_0": 0.39890918135643005, "problem_reward_std_1": 0.2327379286289215, "reward_mean_0": 0.5909336805343628, "reward_mean_1": 0.9025294184684753, "reward_std_0": 0.328091025352478, "reward_std_1": 0.1257166713476181, "rewards/compute_reward/mean": 0.7467315196990967, "rewards/compute_reward/std": 0.3158235549926758, "step": 149, "updated_reward_mean_0": 0.9772377014160156, "updated_reward_mean_1": 0.5483627319335938, "updated_reward_std_0": 0.010294496081769466, "updated_reward_std_1": 0.08205340057611465 }, { "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": 0.17447916666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.5, "completions/mean_length": 1447.69140625, "completions/mean_terminated_length": 1324.858154296875, "completions/min_length": 753.0, "completions/min_terminated_length": 753.0, "epoch": 0.6, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9765625, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.15148577094078064, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2291666716337204, "grad_norm": 3.829416513442993, "kl": 254109561709462.22, "learning_rate": 4.04e-07, "loss": 0.0012, "num_tokens": 305377195.0, "problem_reward_mean_0": 0.43682894110679626, "problem_reward_mean_1": 0.8254278898239136, "problem_reward_std_0": 0.354164719581604, "problem_reward_std_1": 0.2716018855571747, "reward_mean_0": 0.43682894110679626, "reward_mean_1": 0.8254278898239136, "reward_std_0": 0.2044822871685028, "reward_std_1": 0.1588263362646103, "rewards/compute_reward/mean": 0.6311283856630325, "rewards/compute_reward/std": 0.31288328766822815, "step": 150, "updated_reward_mean_0": 0.9551091194152832, "updated_reward_mean_1": 0.7264695167541504, "updated_reward_std_0": 0.03352333977818489, "updated_reward_std_1": 0.14409494400024414 }, { "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": 0.20703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1528.1497802734375, "completions/mean_terminated_length": 1396.4998779296875, "completions/min_length": 512.0, "completions/min_terminated_length": 512.0, "epoch": 0.604, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9661458730697632, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.1810896396636963, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 1.2043017148971558, "kl": -3.8536643499024114e+32, "learning_rate": 4e-07, "loss": 0.0011, "num_tokens": 307337379.0, "problem_reward_mean_0": 0.42624199390411377, "problem_reward_mean_1": 0.7040514945983887, "problem_reward_std_0": 0.38129615783691406, "problem_reward_std_1": 0.3820696175098419, "reward_mean_0": 0.42624199390411377, "reward_mean_1": 0.7040514945983887, "reward_std_0": 0.12723739445209503, "reward_std_1": 0.1657784879207611, "rewards/compute_reward/mean": 0.5651467442512512, "rewards/compute_reward/std": 0.381682887673378, "step": 151, "updated_reward_mean_0": 0.886591911315918, "updated_reward_mean_1": 0.5399889945983887, "updated_reward_std_0": 0.1605849266052246, "updated_reward_std_1": 0.15841279923915863 }, { "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": 0.19791666666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1534.1002807617188, "completions/mean_terminated_length": 1415.293212890625, "completions/min_length": 693.5, "completions/min_terminated_length": 693.5, "epoch": 0.608, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.1016615629196167, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 0.6190908551216125, "kl": 4.31296617190879e+30, "learning_rate": 3.96e-07, "loss": 0.0011, "num_tokens": 309284591.0, "problem_reward_mean_0": 0.48721933364868164, "problem_reward_mean_1": 0.9132817983627319, "problem_reward_std_0": 0.29656219482421875, "problem_reward_std_1": 0.19790887832641602, "reward_mean_0": 0.48721933364868164, "reward_mean_1": 0.9132817983627319, "reward_std_0": 0.18208977580070496, "reward_std_1": 0.10573910176753998, "rewards/compute_reward/mean": 0.7002505362033844, "rewards/compute_reward/std": 0.24723555147647858, "step": 152, "updated_reward_mean_0": 0.9713694453239441, "updated_reward_mean_1": 0.7700526118278503, "updated_reward_std_0": 0.07506769895553589, "updated_reward_std_1": 0.0983734130859375 }, { "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": 0.08333333333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.5, "completions/mean_length": 1190.9193115234375, "completions/mean_terminated_length": 1114.275634765625, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.612, "format_reward_mean_0": 1.0, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.7916666865348816, "grad_norm": 0.01299639604985714, "kl": -1.8661953451323517e+21, "learning_rate": 3.92e-07, "loss": 0.0009, "num_tokens": 310771971.0, "problem_reward_mean_0": 0.5802083611488342, "problem_reward_mean_1": 0.7826822996139526, "problem_reward_std_0": 0.42757225036621094, "problem_reward_std_1": 0.4061732888221741, "reward_mean_0": 0.5802083611488342, "reward_mean_1": 0.7826822996139526, "reward_std_0": 0.20432528853416443, "reward_std_1": 0.06175501272082329, "rewards/compute_reward/mean": 0.681445300579071, "rewards/compute_reward/std": 0.4168727546930313, "step": 153, "updated_reward_mean_0": 0.8333333730697632, "updated_reward_mean_1": 0.37252604961395264, "updated_reward_std_0": 0.0, "updated_reward_std_1": 0.0433407723903656 }, { "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": 0.19401041666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.5, "completions/mean_length": 1456.1015625, "completions/mean_terminated_length": 1318.4886474609375, "completions/min_length": 562.0, "completions/min_terminated_length": 562.0, "epoch": 0.616, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9427083730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.23270225524902344, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 0.6203358769416809, "kl": 4.543380626605168e+33, "learning_rate": 3.88e-07, "loss": 0.0013, "num_tokens": 312832363.0, "problem_reward_mean_0": 0.43810173869132996, "problem_reward_mean_1": 0.7269057035446167, "problem_reward_std_0": 0.33510449528694153, "problem_reward_std_1": 0.41824862360954285, "reward_mean_0": 0.43810173869132996, "reward_mean_1": 0.7269057035446167, "reward_std_0": 0.15178316831588745, "reward_std_1": 0.12092380225658417, "rewards/compute_reward/mean": 0.5825036913156509, "rewards/compute_reward/std": 0.3766765743494034, "step": 154, "updated_reward_mean_0": 0.8245443105697632, "updated_reward_mean_1": 0.7269057035446167, "updated_reward_std_0": 0.018775688484311104, "updated_reward_std_1": 0.12092380225658417 }, { "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": 0.16796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1334.0859985351562, "completions/mean_terminated_length": 1189.9989013671875, "completions/min_length": 347.5, "completions/min_terminated_length": 347.5, "epoch": 0.62, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.15946698188781738, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 2.05790114402771, "kl": 1.9437309821899936e+31, "learning_rate": 3.84e-07, "loss": 0.001, "num_tokens": 314608041.0, "problem_reward_mean_0": 0.5250000357627869, "problem_reward_mean_1": 0.7166088819503784, "problem_reward_std_0": 0.38179272413253784, "problem_reward_std_1": 0.3572044372558594, "reward_mean_0": 0.5250000357627869, "reward_mean_1": 0.7166088819503784, "reward_std_0": 0.15027418732643127, "reward_std_1": 0.16225233674049377, "rewards/compute_reward/mean": 0.6208044290542603, "rewards/compute_reward/std": 0.3694985657930374, "step": 155, "updated_reward_mean_0": 0.859259307384491, "updated_reward_mean_1": 0.45879632234573364, "updated_reward_std_0": 0.048501960933208466, "updated_reward_std_1": 0.10894672572612762 }, { "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": 0.17578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.5, "completions/mean_length": 1276.8203735351562, "completions/mean_terminated_length": 1111.4175415039062, "completions/min_length": 295.5, "completions/min_terminated_length": 295.5, "epoch": 0.624, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 1.0131584405899048, "kl": 1.1110319284130486e+33, "learning_rate": 3.7999999999999996e-07, "loss": 0.0008, "num_tokens": 316288631.0, "problem_reward_mean_0": 0.6688385009765625, "problem_reward_mean_1": 0.789291262626648, "problem_reward_std_0": 0.4268232583999634, "problem_reward_std_1": 0.3556593358516693, "reward_mean_0": 0.6688385009765625, "reward_mean_1": 0.789291262626648, "reward_std_0": 0.19541780650615692, "reward_std_1": 0.185441255569458, "rewards/compute_reward/mean": 0.7290648519992828, "rewards/compute_reward/std": 0.39124128222465515, "step": 156, "updated_reward_mean_0": 0.9628378748893738, "updated_reward_mean_1": 0.32574960589408875, "updated_reward_std_0": 0.020677311345934868, "updated_reward_std_1": 0.14124707877635956 }, { "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": 0.3111979166666667, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1429.6640625, "completions/mean_terminated_length": 1148.1495361328125, "completions/min_length": 298.5, "completions/min_terminated_length": 298.5, "epoch": 0.628, "format_reward_mean_0": 0.8125, "format_reward_mean_1": 0.8125, "format_reward_std_0": 0.3944427967071533, "format_reward_std_1": 0.3908216059207916, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.1875, "grad_norm": 0.6160537600517273, "kl": -Infinity, "learning_rate": 3.76e-07, "loss": 0.0009, "num_tokens": 318143490.0, "problem_reward_mean_0": 0.3386724889278412, "problem_reward_mean_1": 0.6085162162780762, "problem_reward_std_0": 0.3987845778465271, "problem_reward_std_1": 0.4192735552787781, "reward_mean_0": 0.3386724889278412, "reward_mean_1": 0.6085162162780762, "reward_std_0": 0.15398070216178894, "reward_std_1": 0.2161800116300583, "rewards/compute_reward/mean": 0.4735943526029587, "rewards/compute_reward/std": 0.4090290069580078, "step": 157, "updated_reward_mean_0": 0.8495838642120361, "updated_reward_mean_1": 0.44862040877342224, "updated_reward_std_0": 0.13363581895828247, "updated_reward_std_1": 0.17761710286140442 }, { "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": 0.19270833333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.5, "completions/mean_length": 1406.944091796875, "completions/mean_terminated_length": 1252.5896606445312, "completions/min_length": 565.5, "completions/min_terminated_length": 565.5, "epoch": 0.632, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9088541865348816, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.28819188475608826, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.125, "grad_norm": 1.0256520509719849, "kl": 1.8118952294513178e+33, "learning_rate": 3.72e-07, "loss": 0.0011, "num_tokens": 320101814.0, "problem_reward_mean_0": 0.4048035740852356, "problem_reward_mean_1": 0.6994962692260742, "problem_reward_std_0": 0.31998610496520996, "problem_reward_std_1": 0.37096846103668213, "reward_mean_0": 0.4048035144805908, "reward_mean_1": 0.6994962692260742, "reward_std_0": 0.18904821574687958, "reward_std_1": 0.16801679134368896, "rewards/compute_reward/mean": 0.5521499067544937, "rewards/compute_reward/std": 0.34547725319862366, "step": 158, "updated_reward_mean_0": 0.8903180956840515, "updated_reward_mean_1": 0.6044442057609558, "updated_reward_std_0": 0.08229231834411621, "updated_reward_std_1": 0.14909400045871735 }, { "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": 0.19661458333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.5, "completions/mean_length": 1418.2435302734375, "completions/mean_terminated_length": 1263.779296875, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.636, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.117498755455017, "kl": 7.292322583577344e+35, "learning_rate": 3.6799999999999996e-07, "loss": 0.0011, "num_tokens": 321941805.0, "problem_reward_mean_0": 0.3070436716079712, "problem_reward_mean_1": 0.5724361538887024, "problem_reward_std_0": 0.37695959210395813, "problem_reward_std_1": 0.43972721695899963, "reward_mean_0": 0.3070436716079712, "reward_mean_1": 0.5724360942840576, "reward_std_0": 0.19059275090694427, "reward_std_1": 0.1836763620376587, "rewards/compute_reward/mean": 0.4397398978471756, "rewards/compute_reward/std": 0.4083433896303177, "step": 159, "updated_reward_mean_0": 0.786210298538208, "updated_reward_mean_1": 0.4535125195980072, "updated_reward_std_0": 0.19130802154541016, "updated_reward_std_1": 0.14585904777050018 }, { "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": 0.18619791666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1389.807373046875, "completions/mean_terminated_length": 1243.0034790039062, "completions/min_length": 324.5, "completions/min_terminated_length": 324.5, "epoch": 0.64, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 1.237623929977417, "kl": -144120013712.8026, "learning_rate": 3.64e-07, "loss": 0.0007, "num_tokens": 323770811.0, "problem_reward_mean_0": 0.6753697395324707, "problem_reward_mean_1": 0.9198458194732666, "problem_reward_std_0": 0.417127788066864, "problem_reward_std_1": 0.22284764051437378, "reward_mean_0": 0.6753697395324707, "reward_mean_1": 0.9198458194732666, "reward_std_0": 0.26574963331222534, "reward_std_1": 0.13679981231689453, "rewards/compute_reward/mean": 0.7976077795028687, "rewards/compute_reward/std": 0.3199877142906189, "step": 160, "updated_reward_mean_0": 0.9924242496490479, "updated_reward_mean_1": 0.3910815715789795, "updated_reward_std_0": 0.016197651624679565, "updated_reward_std_1": 0.050023313611745834 }, { "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": 0.17447916666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1336.6497802734375, "completions/mean_terminated_length": 1181.3291015625, "completions/min_length": 176.5, "completions/min_terminated_length": 176.5, "epoch": 0.644, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9140625, "format_reward_std_0": 0.0, "format_reward_std_1": 0.28063738346099854, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 1.0548734664916992, "kl": 6.661046141490336e+34, "learning_rate": 3.6e-07, "loss": 0.0012, "num_tokens": 325727250.0, "problem_reward_mean_0": 0.40667232871055603, "problem_reward_mean_1": 0.665886640548706, "problem_reward_std_0": 0.324449747800827, "problem_reward_std_1": 0.39856159687042236, "reward_mean_0": 0.40667232871055603, "reward_mean_1": 0.665886640548706, "reward_std_0": 0.1434021294116974, "reward_std_1": 0.1014556735754013, "rewards/compute_reward/mean": 0.536279484629631, "rewards/compute_reward/std": 0.3615056872367859, "step": 161, "updated_reward_mean_0": 0.7775846719741821, "updated_reward_mean_1": 0.5009560585021973, "updated_reward_std_0": 0.03541749715805054, "updated_reward_std_1": 0.09654520452022552 }, { "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": 0.22786458333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.5, "completions/mean_length": 1506.6563110351562, "completions/mean_terminated_length": 1349.51904296875, "completions/min_length": 638.5, "completions/min_terminated_length": 638.5, "epoch": 0.648, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2083333432674408, "grad_norm": 1.0371465682983398, "kl": -2.309316071584413e+33, "learning_rate": 3.5599999999999996e-07, "loss": 0.0012, "num_tokens": 327698577.0, "problem_reward_mean_0": 0.24091151356697083, "problem_reward_mean_1": 0.5563050508499146, "problem_reward_std_0": 0.3367159366607666, "problem_reward_std_1": 0.4104936122894287, "reward_mean_0": 0.24091152846813202, "reward_mean_1": 0.5563050508499146, "reward_std_0": 0.08313427865505219, "reward_std_1": 0.19390080869197845, "rewards/compute_reward/mean": 0.3986082747578621, "rewards/compute_reward/std": 0.37360477447509766, "step": 162, "updated_reward_mean_0": 0.7779802083969116, "updated_reward_mean_1": 0.41568002104759216, "updated_reward_std_0": 0.09725698083639145, "updated_reward_std_1": 0.1791694164276123 }, { "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": 0.1796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1309.4544677734375, "completions/mean_terminated_length": 1147.6045532226562, "completions/min_length": 256.5, "completions/min_terminated_length": 256.5, "epoch": 0.652, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9921875, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.0881572738289833, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.125, "grad_norm": 0.6892989277839661, "kl": 2.494688550669984e+30, "learning_rate": 3.52e-07, "loss": 0.0011, "num_tokens": 329623065.0, "problem_reward_mean_0": 0.3177378475666046, "problem_reward_mean_1": 0.7301591634750366, "problem_reward_std_0": 0.35328209400177, "problem_reward_std_1": 0.33403703570365906, "reward_mean_0": 0.3177378475666046, "reward_mean_1": 0.7301592230796814, "reward_std_0": 0.0997675284743309, "reward_std_1": 0.12828406691551208, "rewards/compute_reward/mean": 0.5239484757184982, "rewards/compute_reward/std": 0.34365956485271454, "step": 163, "updated_reward_mean_0": 0.8210126757621765, "updated_reward_mean_1": 0.5687007904052734, "updated_reward_std_0": 0.19948813319206238, "updated_reward_std_1": 0.1167975664138794 }, { "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": 0.06380208333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.5, "completions/mean_length": 1283.2005615234375, "completions/mean_terminated_length": 1233.1166381835938, "completions/min_length": 383.5, "completions/min_terminated_length": 383.5, "epoch": 0.656, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 0.8372791409492493, "kl": 1.2151135253521038e+34, "learning_rate": 3.4799999999999994e-07, "loss": 0.0007, "num_tokens": 331286405.0, "problem_reward_mean_0": 0.6804654002189636, "problem_reward_mean_1": 0.8920454978942871, "problem_reward_std_0": 0.3745321035385132, "problem_reward_std_1": 0.2855236530303955, "reward_mean_0": 0.6804654002189636, "reward_mean_1": 0.8920454978942871, "reward_std_0": 0.3074204921722412, "reward_std_1": 0.18725916743278503, "rewards/compute_reward/mean": 0.7862554490566254, "rewards/compute_reward/std": 0.33002789318561554, "step": 164, "updated_reward_mean_0": 0.9981061220169067, "updated_reward_mean_1": 0.4658301770687103, "updated_reward_std_0": 0.005356866400688887, "updated_reward_std_1": 0.0901133120059967 }, { "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": 0.09765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.5, "completions/mean_length": 1168.3529052734375, "completions/mean_terminated_length": 1072.0276489257812, "completions/min_length": 270.5, "completions/min_terminated_length": 270.5, "epoch": 0.66, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9817708730697632, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.13395369052886963, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.3497333824634552, "kl": 2.5590971220959008e+29, "learning_rate": 3.4399999999999996e-07, "loss": 0.0011, "num_tokens": 333142293.0, "problem_reward_mean_0": 0.43549299240112305, "problem_reward_mean_1": 0.8561701774597168, "problem_reward_std_0": 0.32748305797576904, "problem_reward_std_1": 0.2860587537288666, "reward_mean_0": 0.43549299240112305, "reward_mean_1": 0.8561701774597168, "reward_std_0": 0.2037455439567566, "reward_std_1": 0.14103981852531433, "rewards/compute_reward/mean": 0.6458315849304199, "rewards/compute_reward/std": 0.3067709058523178, "step": 165, "updated_reward_mean_0": 0.9618055820465088, "updated_reward_mean_1": 0.7222893238067627, "updated_reward_std_0": 0.08869580924510956, "updated_reward_std_1": 0.09351211786270142 }, { "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": 0.24739583333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1425.6211547851562, "completions/mean_terminated_length": 1225.597412109375, "completions/min_length": 308.5, "completions/min_terminated_length": 308.5, "epoch": 0.664, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.875, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.3311503827571869, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.6666666865348816, "grad_norm": 0.6368861198425293, "kl": -5.915702436918144e+30, "learning_rate": 3.4000000000000003e-07, "loss": 0.0008, "num_tokens": 335028188.0, "problem_reward_mean_0": 0.4485863149166107, "problem_reward_mean_1": 0.5836402773857117, "problem_reward_std_0": 0.4696843922138214, "problem_reward_std_1": 0.4692504405975342, "reward_mean_0": 0.4485863149166107, "reward_mean_1": 0.5836402773857117, "reward_std_0": 0.2110671103000641, "reward_std_1": 0.08847910165786743, "rewards/compute_reward/mean": 0.5161132961511612, "rewards/compute_reward/std": 0.4694673866033554, "step": 166, "updated_reward_mean_0": 0.6636161208152771, "updated_reward_mean_1": 0.20082776248455048, "updated_reward_std_0": 0.012519977055490017, "updated_reward_std_1": 0.056738048791885376 }, { "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": 0.18619791666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1464.9154052734375, "completions/mean_terminated_length": 1332.1090087890625, "completions/min_length": 363.5, "completions/min_terminated_length": 363.5, "epoch": 0.668, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9921875, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.08815725892782211, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.0625, "grad_norm": 1.4895868301391602, "kl": -1.6935388460659674e+28, "learning_rate": 3.36e-07, "loss": 0.0013, "num_tokens": 337084837.0, "problem_reward_mean_0": 0.1759137511253357, "problem_reward_mean_1": 0.6112732887268066, "problem_reward_std_0": 0.18314190208911896, "problem_reward_std_1": 0.3870725929737091, "reward_mean_0": 0.1759137511253357, "reward_mean_1": 0.6112733483314514, "reward_std_0": 0.10109062492847443, "reward_std_1": 0.20035293698310852, "rewards/compute_reward/mean": 0.39359351992607117, "rewards/compute_reward/std": 0.28510725498199463, "step": 167, "updated_reward_mean_0": 0.7939876317977905, "updated_reward_mean_1": 0.6112732887268066, "updated_reward_std_0": 0.14658042788505554, "updated_reward_std_1": 0.20035293698310852 }, { "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": 0.140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1318.783935546875, "completions/mean_terminated_length": 1203.5101928710938, "completions/min_length": 455.5, "completions/min_terminated_length": 455.5, "epoch": 0.672, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.15946701169013977, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.1875, "grad_norm": 0.6989606618881226, "kl": 1.6551946069711803e+33, "learning_rate": 3.32e-07, "loss": 0.0009, "num_tokens": 338812367.0, "problem_reward_mean_0": 0.5092728734016418, "problem_reward_mean_1": 0.8002154231071472, "problem_reward_std_0": 0.3863068222999573, "problem_reward_std_1": 0.3222363591194153, "reward_mean_0": 0.5092728734016418, "reward_mean_1": 0.8002154231071472, "reward_std_0": 0.18097275495529175, "reward_std_1": 0.2244255244731903, "rewards/compute_reward/mean": 0.6547441482543945, "rewards/compute_reward/std": 0.3542715907096863, "step": 168, "updated_reward_mean_0": 0.9694927930831909, "updated_reward_mean_1": 0.49292370676994324, "updated_reward_std_0": 0.0654030442237854, "updated_reward_std_1": 0.1660308539867401 }, { "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": 0.10546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1254.9114990234375, "completions/mean_terminated_length": 1158.3486938476562, "completions/min_length": 331.5, "completions/min_terminated_length": 331.5, "epoch": 0.676, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.0720745101571083, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.375, "grad_norm": 0.4247985780239105, "kl": -5.674238153598002e+27, "learning_rate": 3.28e-07, "loss": 0.0008, "num_tokens": 340656024.0, "problem_reward_mean_0": 0.4520934224128723, "problem_reward_mean_1": 0.9291579127311707, "problem_reward_std_0": 0.38401976227760315, "problem_reward_std_1": 0.2144925743341446, "reward_mean_0": 0.4520934224128723, "reward_mean_1": 0.9291579127311707, "reward_std_0": 0.24956148862838745, "reward_std_1": 0.09750201553106308, "rewards/compute_reward/mean": 0.6906256526708603, "rewards/compute_reward/std": 0.29925617575645447, "step": 169, "updated_reward_mean_0": 0.9895833730697632, "updated_reward_mean_1": 0.6417350172996521, "updated_reward_std_0": 0.02946278266608715, "updated_reward_std_1": 0.08853331208229065 }, { "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": 0.11067708333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1343.2396240234375, "completions/mean_terminated_length": 1256.2522583007812, "completions/min_length": 622.5, "completions/min_terminated_length": 622.5, "epoch": 0.68, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.9765625, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.15148580074310303, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.2778096199035645, "kl": 3.408008213680249e+34, "learning_rate": 3.24e-07, "loss": 0.0008, "num_tokens": 342558316.0, "problem_reward_mean_0": 0.5902559757232666, "problem_reward_mean_1": 0.8546392321586609, "problem_reward_std_0": 0.3579597473144531, "problem_reward_std_1": 0.30286261439323425, "reward_mean_0": 0.5902559757232666, "reward_mean_1": 0.8546392321586609, "reward_std_0": 0.20364223420619965, "reward_std_1": 0.16889360547065735, "rewards/compute_reward/mean": 0.7224475741386414, "rewards/compute_reward/std": 0.3304111957550049, "step": 170, "updated_reward_mean_0": 0.9813432693481445, "updated_reward_mean_1": 0.640001118183136, "updated_reward_std_0": 0.01842428371310234, "updated_reward_std_1": 0.10330542176961899 }, { "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": 0.041666666666666685, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1265.5899047851562, "completions/mean_terminated_length": 1231.9104614257812, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.684, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.14301247894763947, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4375, "grad_norm": 0.6484044790267944, "kl": 3.946145607436866e+35, "learning_rate": 3.2e-07, "loss": 0.0011, "num_tokens": 344412041.0, "problem_reward_mean_0": 0.33056116104125977, "problem_reward_mean_1": 0.8650892972946167, "problem_reward_std_0": 0.28310683369636536, "problem_reward_std_1": 0.2716405689716339, "reward_mean_0": 0.33056116104125977, "reward_mean_1": 0.8650891780853271, "reward_std_0": 0.20374846458435059, "reward_std_1": 0.10173477977514267, "rewards/compute_reward/mean": 0.5978251993656158, "rewards/compute_reward/std": 0.27737370133399963, "step": 171, "updated_reward_mean_0": 0.9583075046539307, "updated_reward_mean_1": 0.8234225511550903, "updated_reward_std_0": 0.04686264321208, "updated_reward_std_1": 0.10173477977514267 }, { "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": 0.16015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1330.0573120117188, "completions/mean_terminated_length": 1192.4346923828125, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.688, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.8619791865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.3453718423843384, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3958333432674408, "grad_norm": 0.9328606724739075, "kl": 2.944717056583287e+33, "learning_rate": 3.1599999999999997e-07, "loss": 0.0006, "num_tokens": 346321040.0, "problem_reward_mean_0": 0.6576704382896423, "problem_reward_mean_1": 0.7402090430259705, "problem_reward_std_0": 0.4003784656524658, "problem_reward_std_1": 0.40172556042671204, "reward_mean_0": 0.6576704382896423, "reward_mean_1": 0.7402090430259705, "reward_std_0": 0.2704579830169678, "reward_std_1": 0.17345790565013885, "rewards/compute_reward/mean": 0.6989397406578064, "rewards/compute_reward/std": 0.40105199813842773, "step": 172, "updated_reward_mean_0": 0.9803165793418884, "updated_reward_mean_1": 0.3724888563156128, "updated_reward_std_0": 0.025025859475135803, "updated_reward_std_1": 0.09145953506231308 }, { "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": 0.14453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.5, "completions/mean_length": 1336.1484985351562, "completions/mean_terminated_length": 1215.826904296875, "completions/min_length": 390.5, "completions/min_terminated_length": 390.5, "epoch": 0.692, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9322916865348816, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.25157251954078674, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.2083333432674408, "grad_norm": 1.1590218544006348, "kl": -7.355032729991901e+35, "learning_rate": 3.12e-07, "loss": 0.0013, "num_tokens": 348352769.0, "problem_reward_mean_0": 0.32641106843948364, "problem_reward_mean_1": 0.7906882166862488, "problem_reward_std_0": 0.26831474900245667, "problem_reward_std_1": 0.3067144453525543, "reward_mean_0": 0.32641109824180603, "reward_mean_1": 0.7906882166862488, "reward_std_0": 0.15196740627288818, "reward_std_1": 0.16858872771263123, "rewards/compute_reward/mean": 0.558549627661705, "rewards/compute_reward/std": 0.2875145971775055, "step": 173, "updated_reward_mean_0": 0.9176158905029297, "updated_reward_mean_1": 0.7765512466430664, "updated_reward_std_0": 0.06103629991412163, "updated_reward_std_1": 0.16462934017181396 }, { "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": 0.09635416666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.5, "completions/mean_length": 1197.7591552734375, "completions/mean_terminated_length": 1112.133544921875, "completions/min_length": 597.0, "completions/min_terminated_length": 597.0, "epoch": 0.696, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.984375, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.12418139725923538, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.5208333730697632, "grad_norm": 0.3767165541648865, "kl": 3.0658439618290556e+35, "learning_rate": 3.08e-07, "loss": 0.0011, "num_tokens": 349950303.0, "problem_reward_mean_0": 0.4776986241340637, "problem_reward_mean_1": 0.8179736137390137, "problem_reward_std_0": 0.36725184321403503, "problem_reward_std_1": 0.338344931602478, "reward_mean_0": 0.4776986241340637, "reward_mean_1": 0.8179736137390137, "reward_std_0": 0.2195589244365692, "reward_std_1": 0.058479007333517075, "rewards/compute_reward/mean": 0.6478361040353775, "rewards/compute_reward/std": 0.35279838740825653, "step": 174, "updated_reward_mean_0": 0.8638493418693542, "updated_reward_mean_1": 0.5757861137390137, "updated_reward_std_0": 0.03738704323768616, "updated_reward_std_1": 0.03638192266225815 }, { "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": 0.13802083333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1240.3060302734375, "completions/mean_terminated_length": 1111.8981018066406, "completions/min_length": 276.5, "completions/min_terminated_length": 276.5, "epoch": 0.7, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9557291865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.2059645652770996, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.4061529636383057, "kl": -1.915560485231398e+32, "learning_rate": 3.0399999999999997e-07, "loss": 0.0013, "num_tokens": 351804379.0, "problem_reward_mean_0": 0.3407866358757019, "problem_reward_mean_1": 0.623661994934082, "problem_reward_std_0": 0.3278515338897705, "problem_reward_std_1": 0.42496299743652344, "reward_mean_0": 0.3407866656780243, "reward_mean_1": 0.623661994934082, "reward_std_0": 0.1522875726222992, "reward_std_1": 0.13383737206459045, "rewards/compute_reward/mean": 0.4822242707014084, "rewards/compute_reward/std": 0.3764072507619858, "step": 175, "updated_reward_mean_0": 0.725853443145752, "updated_reward_mean_1": 0.5064744353294373, "updated_reward_std_0": 0.028129786252975464, "updated_reward_std_1": 0.11174027621746063 }, { "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": 0.028645833333333315, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.5, "completions/mean_length": 1120.1966552734375, "completions/mean_terminated_length": 1093.09326171875, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.704, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.125, "grad_norm": 1.0765483379364014, "kl": -7.481179541004432e+19, "learning_rate": 3e-07, "loss": 0.001, "num_tokens": 353531426.0, "problem_reward_mean_0": 0.5604933500289917, "problem_reward_mean_1": 0.9027724266052246, "problem_reward_std_0": 0.3365018367767334, "problem_reward_std_1": 0.22336654365062714, "reward_mean_0": 0.5604933500289917, "reward_mean_1": 0.9027724266052246, "reward_std_0": 0.2024308145046234, "reward_std_1": 0.15641403198242188, "rewards/compute_reward/mean": 0.7316328883171082, "rewards/compute_reward/std": 0.27993418276309967, "step": 176, "updated_reward_mean_0": 0.995039701461792, "updated_reward_mean_1": 0.6387099623680115, "updated_reward_std_0": 0.01402989961206913, "updated_reward_std_1": 0.08851252496242523 }, { "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": 0.014322916666666685, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.5, "completions/mean_length": 986.8020935058594, "completions/mean_terminated_length": 971.9827575683594, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.708, "format_reward_mean_0": 1.0, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 2.857191801071167, "kl": -2.5618042049086366e+27, "learning_rate": 2.9599999999999995e-07, "loss": 0.0011, "num_tokens": 355278226.0, "problem_reward_mean_0": 0.5061588287353516, "problem_reward_mean_1": 0.8372690081596375, "problem_reward_std_0": 0.3232247531414032, "problem_reward_std_1": 0.28431424498558044, "reward_mean_0": 0.5061588287353516, "reward_mean_1": 0.8372690081596375, "reward_std_0": 0.2232469767332077, "reward_std_1": 0.11387088894844055, "rewards/compute_reward/mean": 0.6717139184474945, "rewards/compute_reward/std": 0.3037694841623306, "step": 177, "updated_reward_mean_0": 0.9482938051223755, "updated_reward_mean_1": 0.6504201292991638, "updated_reward_std_0": 0.026048801839351654, "updated_reward_std_1": 0.07646796107292175 }, { "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": 0.06640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1268.8594360351562, "completions/mean_terminated_length": 1214.279296875, "completions/min_length": 548.0, "completions/min_terminated_length": 548.0, "epoch": 0.712, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.0720745176076889, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.375, "grad_norm": 9.882614135742188, "kl": -7.247545268992617e+33, "learning_rate": 2.9199999999999997e-07, "loss": 0.0009, "num_tokens": 357018233.0, "problem_reward_mean_0": 0.5102183222770691, "problem_reward_mean_1": 0.9412760734558105, "problem_reward_std_0": 0.4033553898334503, "problem_reward_std_1": 0.18089164793491364, "reward_mean_0": 0.5102183222770691, "reward_mean_1": 0.9412760734558105, "reward_std_0": 0.11022254079580307, "reward_std_1": 0.0953003466129303, "rewards/compute_reward/mean": 0.7257471978664398, "rewards/compute_reward/std": 0.2921235039830208, "step": 178, "updated_reward_mean_0": 0.9992560148239136, "updated_reward_mean_1": 0.5779948234558105, "updated_reward_std_0": 0.002104482613503933, "updated_reward_std_1": 0.075362429022789 }, { "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": 0.10416666666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.5, "completions/mean_length": 1233.9961547851562, "completions/mean_terminated_length": 1141.6721801757812, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.716, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9609375, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.19399645924568176, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.6458333730697632, "grad_norm": 1.4158504009246826, "kl": 9.459961023529187e+28, "learning_rate": 2.88e-07, "loss": 0.0008, "num_tokens": 358787096.0, "problem_reward_mean_0": 0.5961805582046509, "problem_reward_mean_1": 0.7048766613006592, "problem_reward_std_0": 0.40600571036338806, "problem_reward_std_1": 0.42444318532943726, "reward_mean_0": 0.5961805582046509, "reward_mean_1": 0.7048766613006592, "reward_std_0": 0.09338458627462387, "reward_std_1": 0.07816743850708008, "rewards/compute_reward/mean": 0.650528609752655, "rewards/compute_reward/std": 0.41522446274757385, "step": 179, "updated_reward_mean_0": 0.7994431257247925, "updated_reward_mean_1": 0.2725850045681, "updated_reward_std_0": 0.03522856533527374, "updated_reward_std_1": 0.06343605369329453 }, { "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": 0.18489583333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.5, "completions/mean_length": 1370.8021240234375, "completions/mean_terminated_length": 1220.6782836914062, "completions/min_length": 556.0, "completions/min_terminated_length": 556.0, "epoch": 0.72, "format_reward_mean_0": 0.9583333730697632, "format_reward_mean_1": 0.9739583730697632, "format_reward_std_0": 0.20194093883037567, "format_reward_std_1": 0.15946698188781738, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4166666865348816, "grad_norm": 0.5202940106391907, "kl": 4.648052200724521e+31, "learning_rate": 2.8399999999999995e-07, "loss": 0.0008, "num_tokens": 360565474.0, "problem_reward_mean_0": 0.5751869678497314, "problem_reward_mean_1": 0.8781384229660034, "problem_reward_std_0": 0.4244050979614258, "problem_reward_std_1": 0.2938694953918457, "reward_mean_0": 0.5751869678497314, "reward_mean_1": 0.8781384229660034, "reward_std_0": 0.19222784042358398, "reward_std_1": 0.18445704877376556, "rewards/compute_reward/mean": 0.726662665605545, "rewards/compute_reward/std": 0.35913731157779694, "step": 180, "updated_reward_mean_0": 0.9883813858032227, "updated_reward_mean_1": 0.45105504989624023, "updated_reward_std_0": 0.032862335443496704, "updated_reward_std_1": 0.10115613043308258 }, { "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": 0.1796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1437.4649047851562, "completions/mean_terminated_length": 1305.8464965820312, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.724, "format_reward_mean_0": 0.8125, "format_reward_mean_1": 0.9713541865348816, "format_reward_std_0": 0.3944427967071533, "format_reward_std_1": 0.16702665388584137, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3541666865348816, "grad_norm": 0.7345628142356873, "kl": -2.840989860823995e+31, "learning_rate": 2.8e-07, "loss": 0.0011, "num_tokens": 362553235.0, "problem_reward_mean_0": 0.4129680395126343, "problem_reward_mean_1": 0.7851080894470215, "problem_reward_std_0": 0.37406009435653687, "problem_reward_std_1": 0.36264756321907043, "reward_mean_0": 0.4129680395126343, "reward_mean_1": 0.7851080894470215, "reward_std_0": 0.20360004901885986, "reward_std_1": 0.18389318883419037, "rewards/compute_reward/mean": 0.5990380644798279, "rewards/compute_reward/std": 0.36835381388664246, "step": 181, "updated_reward_mean_0": 0.9581449031829834, "updated_reward_mean_1": 0.7095872759819031, "updated_reward_std_0": 0.013088878244161606, "updated_reward_std_1": 0.16179610788822174 }, { "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": 0.10026041666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.5, "completions/mean_length": 1275.4166870117188, "completions/mean_terminated_length": 1192.58349609375, "completions/min_length": 418.5, "completions/min_terminated_length": 418.5, "epoch": 0.728, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.1016615629196167, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.4375, "grad_norm": 2.126969575881958, "kl": 3.6188412004891047e+22, "learning_rate": 2.7600000000000004e-07, "loss": 0.0011, "num_tokens": 364282128.0, "problem_reward_mean_0": 0.5549361109733582, "problem_reward_mean_1": 0.887279212474823, "problem_reward_std_0": 0.4399021863937378, "problem_reward_std_1": 0.261767715215683, "reward_mean_0": 0.5549361109733582, "reward_mean_1": 0.887279212474823, "reward_std_0": 0.20986686646938324, "reward_std_1": 0.09700610488653183, "rewards/compute_reward/mean": 0.7211076617240906, "rewards/compute_reward/std": 0.3508349359035492, "step": 182, "updated_reward_mean_0": 0.9761905074119568, "updated_reward_mean_1": 0.44511908292770386, "updated_reward_std_0": 0.02545345202088356, "updated_reward_std_1": 0.08634698390960693 }, { "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": 0.07682291666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.5, "completions/mean_length": 1282.4453125, "completions/mean_terminated_length": 1217.8801879882812, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.732, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.11351143568754196, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.0833333358168602, "grad_norm": 1.6132224798202515, "kl": -2.8615321455887914e+35, "learning_rate": 2.72e-07, "loss": 0.0011, "num_tokens": 366258368.0, "problem_reward_mean_0": 0.5251488089561462, "problem_reward_mean_1": 0.716369092464447, "problem_reward_std_0": 0.31873440742492676, "problem_reward_std_1": 0.30370286107063293, "reward_mean_0": 0.5251488089561462, "reward_mean_1": 0.716369092464447, "reward_std_0": 0.19679586589336395, "reward_std_1": 0.17774860560894012, "rewards/compute_reward/mean": 0.6207589507102966, "rewards/compute_reward/std": 0.31121861934661865, "step": 183, "updated_reward_mean_0": 0.8663691282272339, "updated_reward_mean_1": 0.6062500476837158, "updated_reward_std_0": 0.09297367185354233, "updated_reward_std_1": 0.15527042746543884 }, { "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": 0.11067708333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.5, "completions/mean_length": 1297.8138427734375, "completions/mean_terminated_length": 1206.1802978515625, "completions/min_length": 263.5, "completions/min_terminated_length": 263.5, "epoch": 0.736, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.11351142823696136, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.25, "grad_norm": 0.7947391867637634, "kl": 1.6089660108735526e+20, "learning_rate": 2.68e-07, "loss": 0.0011, "num_tokens": 368105458.0, "problem_reward_mean_0": 0.40956124663352966, "problem_reward_mean_1": 0.7334945201873779, "problem_reward_std_0": 0.41102033853530884, "problem_reward_std_1": 0.4195856750011444, "reward_mean_0": 0.40956124663352966, "reward_mean_1": 0.7334944605827332, "reward_std_0": 0.21359769999980927, "reward_std_1": 0.163508802652359, "rewards/compute_reward/mean": 0.5715278834104538, "rewards/compute_reward/std": 0.41530299186706543, "step": 184, "updated_reward_mean_0": 0.8445833921432495, "updated_reward_mean_1": 0.5230406522750854, "updated_reward_std_0": 0.0063773468136787415, "updated_reward_std_1": 0.12218593806028366 }, { "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": 0.22265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.5, "completions/mean_length": 1524.8450927734375, "completions/mean_terminated_length": 1373.4669189453125, "completions/min_length": 570.0, "completions/min_terminated_length": 570.0, "epoch": 0.74, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9192708730697632, "format_reward_std_0": 0.0, "format_reward_std_1": 0.2727741301059723, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.25, "grad_norm": 1.0170122385025024, "kl": -1.9282996471717725e+22, "learning_rate": 2.64e-07, "loss": 0.0009, "num_tokens": 370202243.0, "problem_reward_mean_0": 0.4696166515350342, "problem_reward_mean_1": 0.802888810634613, "problem_reward_std_0": 0.35570377111434937, "problem_reward_std_1": 0.3290998637676239, "reward_mean_0": 0.4696166515350342, "reward_mean_1": 0.802888810634613, "reward_std_0": 0.22258228063583374, "reward_std_1": 0.144419327378273, "rewards/compute_reward/mean": 0.6362527310848236, "rewards/compute_reward/std": 0.3424018323421478, "step": 185, "updated_reward_mean_0": 0.9579812288284302, "updated_reward_mean_1": 0.6414304971694946, "updated_reward_std_0": 0.029092034325003624, "updated_reward_std_1": 0.12968793511390686 }, { "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": 0.13020833333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1324.7044677734375, "completions/mean_terminated_length": 1214.220458984375, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.744, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 1.0, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.0, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3125, "grad_norm": 1.3036712408065796, "kl": 1.7331666252797866e+28, "learning_rate": 2.6e-07, "loss": 0.0011, "num_tokens": 372140103.0, "problem_reward_mean_0": 0.3778724670410156, "problem_reward_mean_1": 0.6921402215957642, "problem_reward_std_0": 0.3708103597164154, "problem_reward_std_1": 0.39094915986061096, "reward_mean_0": 0.3778724670410156, "reward_mean_1": 0.6921402215957642, "reward_std_0": 0.2723449468612671, "reward_std_1": 0.14365322887897491, "rewards/compute_reward/mean": 0.5350063145160675, "rewards/compute_reward/std": 0.3808797597885132, "step": 186, "updated_reward_mean_0": 0.8752209544181824, "updated_reward_mean_1": 0.5776002407073975, "updated_reward_std_0": 0.10504507273435593, "updated_reward_std_1": 0.1317960023880005 }, { "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": 0.203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.5, "completions/mean_length": 1419.9557495117188, "completions/mean_terminated_length": 1257.2675170898438, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.748, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9895833730697632, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.1016615703701973, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3541666865348816, "grad_norm": 1.381913423538208, "kl": -4.619544306164372e+31, "learning_rate": 2.56e-07, "loss": 0.0012, "num_tokens": 374026315.0, "problem_reward_mean_0": 0.4878842234611511, "problem_reward_mean_1": 0.7926443815231323, "problem_reward_std_0": 0.38575559854507446, "problem_reward_std_1": 0.29191863536834717, "reward_mean_0": 0.4878842234611511, "reward_mean_1": 0.7926443815231323, "reward_std_0": 0.3003426194190979, "reward_std_1": 0.16481223702430725, "rewards/compute_reward/mean": 0.6402642577886581, "rewards/compute_reward/std": 0.3388371169567108, "step": 187, "updated_reward_mean_0": 0.9516288042068481, "updated_reward_mean_1": 0.5960298776626587, "updated_reward_std_0": 0.04820685088634491, "updated_reward_std_1": 0.14053787291049957 }, { "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": 0.19140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.5, "completions/mean_length": 1432.0104370117188, "completions/mean_terminated_length": 1280.35546875, "completions/min_length": 506.5, "completions/min_terminated_length": 506.5, "epoch": 0.752, "format_reward_mean_0": 0.8958333730697632, "format_reward_mean_1": 0.9791666865348816, "format_reward_std_0": 0.3087092638015747, "format_reward_std_1": 0.14301246404647827, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1041666716337204, "grad_norm": 1.4094663858413696, "kl": 1.251919226837634e+35, "learning_rate": 2.52e-07, "loss": 0.0013, "num_tokens": 376019658.0, "problem_reward_mean_0": 0.2578703761100769, "problem_reward_mean_1": 0.7834824919700623, "problem_reward_std_0": 0.17209331691265106, "problem_reward_std_1": 0.2955227494239807, "reward_mean_0": 0.2578703761100769, "reward_mean_1": 0.7834824919700623, "reward_std_0": 0.11433462798595428, "reward_std_1": 0.1422155797481537, "rewards/compute_reward/mean": 0.5206764191389084, "rewards/compute_reward/std": 0.2338080331683159, "step": 188, "updated_reward_mean_0": 0.9153218269348145, "updated_reward_mean_1": 0.783482551574707, "updated_reward_std_0": 0.15789946913719177, "updated_reward_std_1": 0.1422155797481537 }, { "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": 0.265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1354.9010620117188, "completions/mean_terminated_length": 1104.4096069335938, "completions/min_length": 153.5, "completions/min_terminated_length": 153.5, "epoch": 0.756, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9010416865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.29899561405181885, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 0.5308528542518616, "kl": -4.867778304876401e+33, "learning_rate": 2.48e-07, "loss": 0.0006, "num_tokens": 377825907.0, "problem_reward_mean_0": 0.6254850625991821, "problem_reward_mean_1": 0.7262285947799683, "problem_reward_std_0": 0.42630326747894287, "problem_reward_std_1": 0.3910098671913147, "reward_mean_0": 0.6254850625991821, "reward_mean_1": 0.7262285947799683, "reward_std_0": 0.16331720352172852, "reward_std_1": 0.11733686923980713, "rewards/compute_reward/mean": 0.6758567690849304, "rewards/compute_reward/std": 0.4086565375328064, "step": 189, "updated_reward_mean_0": 0.846451997756958, "updated_reward_mean_1": 0.23949456214904785, "updated_reward_std_0": 0.049936018884181976, "updated_reward_std_1": 0.048116933554410934 }, { "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": 0.28385416666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1573.0872802734375, "completions/mean_terminated_length": 1414.7013549804688, "completions/min_length": 654.0, "completions/min_terminated_length": 654.0, "epoch": 0.76, "format_reward_mean_0": 0.7083333730697632, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.4593396484851837, "format_reward_std_1": 0.11351142823696136, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 1.00436270236969, "kl": -6.866440751236242e+29, "learning_rate": 2.4399999999999996e-07, "loss": 0.001, "num_tokens": 379823886.0, "problem_reward_mean_0": 0.5063603520393372, "problem_reward_mean_1": 0.6377334594726562, "problem_reward_std_0": 0.44574496150016785, "problem_reward_std_1": 0.43192803859710693, "reward_mean_0": 0.5063603520393372, "reward_mean_1": 0.6377334594726562, "reward_std_0": 0.31191709637641907, "reward_std_1": 0.17916327714920044, "rewards/compute_reward/mean": 0.5720469057559967, "rewards/compute_reward/std": 0.4388365149497986, "step": 190, "updated_reward_mean_0": 0.8948385715484619, "updated_reward_mean_1": 0.40681490302085876, "updated_reward_std_0": 0.13995371758937836, "updated_reward_std_1": 0.13010330498218536 }, { "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": 0.17317708333333331, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.5, "completions/mean_length": 1406.1641235351562, "completions/mean_terminated_length": 1281.3705444335938, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.764, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.953125, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.21164682507514954, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.7263819575309753, "kl": -9.563015277830758e+32, "learning_rate": 2.4e-07, "loss": 0.001, "num_tokens": 381826217.0, "problem_reward_mean_0": 0.39862892031669617, "problem_reward_mean_1": 0.6560249328613281, "problem_reward_std_0": 0.3452571928501129, "problem_reward_std_1": 0.3819993734359741, "reward_mean_0": 0.39862892031669617, "reward_mean_1": 0.6560249328613281, "reward_std_0": 0.16021648049354553, "reward_std_1": 0.1430257111787796, "rewards/compute_reward/mean": 0.5273268967866898, "rewards/compute_reward/std": 0.3636282682418823, "step": 191, "updated_reward_mean_0": 0.8553368449211121, "updated_reward_mean_1": 0.48935824632644653, "updated_reward_std_0": 0.11023262143135071, "updated_reward_std_1": 0.1430256962776184 }, { "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": 0.171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.5, "completions/mean_length": 1396.8997802734375, "completions/mean_terminated_length": 1263.8052368164062, "completions/min_length": 356.5, "completions/min_terminated_length": 356.5, "epoch": 0.768, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.953125, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.21164685487747192, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.5208333730697632, "grad_norm": 0.7457505464553833, "kl": 7.061550313060245e+35, "learning_rate": 2.3599999999999997e-07, "loss": 0.0011, "num_tokens": 383858348.0, "problem_reward_mean_0": 0.32500892877578735, "problem_reward_mean_1": 0.62871915102005, "problem_reward_std_0": 0.3632108271121979, "problem_reward_std_1": 0.43824976682662964, "reward_mean_0": 0.32500892877578735, "reward_mean_1": 0.6287190914154053, "reward_std_0": 0.06270094215869904, "reward_std_1": 0.07625553756952286, "rewards/compute_reward/mean": 0.4768640249967575, "rewards/compute_reward/std": 0.40073029696941376, "step": 192, "updated_reward_mean_0": 0.682109534740448, "updated_reward_mean_1": 0.46436724066734314, "updated_reward_std_0": 0.050225429236888885, "updated_reward_std_1": 0.07044120132923126 }, { "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": 0.15885416666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1442.4258422851562, "completions/mean_terminated_length": 1331.2101440429688, "completions/min_length": 584.5, "completions/min_terminated_length": 584.5, "epoch": 0.772, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9947916865348816, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.07207450270652771, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.125, "grad_norm": 0.7875493168830872, "kl": -2.887426367186523e+30, "learning_rate": 2.32e-07, "loss": 0.001, "num_tokens": 385895757.0, "problem_reward_mean_0": 0.6571556329727173, "problem_reward_mean_1": 0.8048035502433777, "problem_reward_std_0": 0.3178899884223938, "problem_reward_std_1": 0.27377641201019287, "reward_mean_0": 0.6571556329727173, "reward_mean_1": 0.8048035502433777, "reward_std_0": 0.24971023201942444, "reward_std_1": 0.17952758073806763, "rewards/compute_reward/mean": 0.7309795618057251, "rewards/compute_reward/std": 0.29583320021629333, "step": 193, "updated_reward_mean_0": 0.963921070098877, "updated_reward_mean_1": 0.5593411326408386, "updated_reward_std_0": 0.022950515151023865, "updated_reward_std_1": 0.13034167885780334 }, { "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": 0.25, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1432.9232177734375, "completions/mean_terminated_length": 1230.424072265625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.776, "format_reward_mean_0": 0.8541666865348816, "format_reward_mean_1": 0.7578125, "format_reward_std_0": 0.3566739559173584, "format_reward_std_1": 0.42896607518196106, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.1666666716337204, "grad_norm": 1.2944223880767822, "kl": -3.3240654962244606e+31, "learning_rate": 2.28e-07, "loss": 0.0017, "num_tokens": 388062750.0, "problem_reward_mean_0": 0.4117136597633362, "problem_reward_mean_1": 0.46424752473831177, "problem_reward_std_0": 0.36560121178627014, "problem_reward_std_1": 0.4147360324859619, "reward_mean_0": 0.4117136597633362, "reward_mean_1": 0.464247465133667, "reward_std_0": 0.1553313136100769, "reward_std_1": 0.154381662607193, "rewards/compute_reward/mean": 0.4379805624485016, "rewards/compute_reward/std": 0.3901686370372772, "step": 194, "updated_reward_mean_0": 0.7660870552062988, "updated_reward_mean_1": 0.4281485676765442, "updated_reward_std_0": 0.12367391586303711, "updated_reward_std_1": 0.1344684660434723 }, { "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": 0.15234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.5, "completions/mean_length": 1328.8021240234375, "completions/mean_terminated_length": 1203.4198608398438, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.78, "format_reward_mean_0": 0.875, "format_reward_mean_1": 0.9713541865348816, "format_reward_std_0": 0.33421868085861206, "format_reward_std_1": 0.16702665388584137, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4791666865348816, "grad_norm": 2.0142502784729004, "kl": 1.850654211468877e+34, "learning_rate": 2.24e-07, "loss": 0.0011, "num_tokens": 389753205.0, "problem_reward_mean_0": 0.4913254976272583, "problem_reward_mean_1": 0.844022274017334, "problem_reward_std_0": 0.363505482673645, "problem_reward_std_1": 0.3045122027397156, "reward_mean_0": 0.4913254976272583, "reward_mean_1": 0.844022274017334, "reward_std_0": 0.3178296983242035, "reward_std_1": 0.13575206696987152, "rewards/compute_reward/mean": 0.6676738858222961, "rewards/compute_reward/std": 0.3340088576078415, "step": 195, "updated_reward_mean_0": 0.9369918704032898, "updated_reward_mean_1": 0.6265742778778076, "updated_reward_std_0": 0.07260189950466156, "updated_reward_std_1": 0.10515015572309494 }, { "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": 0.13541666666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.5, "completions/mean_length": 1297.533935546875, "completions/mean_terminated_length": 1180.7265014648438, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.784, "format_reward_mean_0": 0.9166666865348816, "format_reward_mean_1": 0.953125, "format_reward_std_0": 0.2793101966381073, "format_reward_std_1": 0.21164685487747192, "frac_reward_zero_std_0": 0.1666666716337204, "frac_reward_zero_std_1": 0.3333333432674408, "grad_norm": 0.605797290802002, "kl": -3.0829262600908705e+33, "learning_rate": 2.1999999999999998e-07, "loss": 0.0009, "num_tokens": 391667040.0, "problem_reward_mean_0": 0.5451697111129761, "problem_reward_mean_1": 0.7276796102523804, "problem_reward_std_0": 0.40584349632263184, "problem_reward_std_1": 0.3632848560810089, "reward_mean_0": 0.5451697111129761, "reward_mean_1": 0.7276796102523804, "reward_std_0": 0.16431266069412231, "reward_std_1": 0.1268884539604187, "rewards/compute_reward/mean": 0.6364246010780334, "rewards/compute_reward/std": 0.3845641762018204, "step": 196, "updated_reward_mean_0": 0.8318789005279541, "updated_reward_mean_1": 0.41968685388565063, "updated_reward_std_0": 0.029797527939081192, "updated_reward_std_1": 0.11414013803005219 }, { "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": 0.01953125, "completions/max_length": 2038.0, "completions/max_terminated_length": 2035.5, "completions/mean_length": 1210.662841796875, "completions/mean_terminated_length": 1193.6893920898438, "completions/min_length": 337.5, "completions/min_terminated_length": 337.5, "epoch": 0.788, "format_reward_mean_0": 1.0, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.0, "format_reward_std_1": 0.11351142078638077, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.4583333432674408, "grad_norm": 0.2924601137638092, "kl": -7.199985969583521e+35, "learning_rate": 2.1599999999999998e-07, "loss": 0.0009, "num_tokens": 393446977.0, "problem_reward_mean_0": 0.4513698220252991, "problem_reward_mean_1": 0.8406403064727783, "problem_reward_std_0": 0.42463433742523193, "problem_reward_std_1": 0.2785504162311554, "reward_mean_0": 0.4513698220252991, "reward_mean_1": 0.8406403064727783, "reward_std_0": 0.2023860663175583, "reward_std_1": 0.11858195066452026, "rewards/compute_reward/mean": 0.6460050493478775, "rewards/compute_reward/std": 0.35159240663051605, "step": 197, "updated_reward_mean_0": 0.941620945930481, "updated_reward_mean_1": 0.5190256834030151, "updated_reward_std_0": 0.02364530973136425, "updated_reward_std_1": 0.09610031545162201 }, { "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": 0.05208333333333337, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1138.4609375, "completions/mean_terminated_length": 1087.09765625, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.792, "format_reward_mean_0": 0.9375, "format_reward_mean_1": 0.9869791865348816, "format_reward_std_0": 0.24462303519248962, "format_reward_std_1": 0.11351144313812256, "frac_reward_zero_std_0": 0.3333333432674408, "frac_reward_zero_std_1": 0.6875, "grad_norm": 0.4689013361930847, "kl": -4.5972949398658976e+32, "learning_rate": 2.12e-07, "loss": 0.0008, "num_tokens": 395314142.0, "problem_reward_mean_0": 0.46462053060531616, "problem_reward_mean_1": 0.840737521648407, "problem_reward_std_0": 0.365509569644928, "problem_reward_std_1": 0.33058962225914, "reward_mean_0": 0.46462056040763855, "reward_mean_1": 0.840737521648407, "reward_std_0": 0.18528637290000916, "reward_std_1": 0.08216734975576401, "rewards/compute_reward/mean": 0.6526790112257004, "rewards/compute_reward/std": 0.348049595952034, "step": 198, "updated_reward_mean_0": 0.932198703289032, "updated_reward_mean_1": 0.574288547039032, "updated_reward_std_0": 0.0745026022195816, "updated_reward_std_1": 0.06990798562765121 }, { "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": 0.17447916666666663, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1433.1875610351562, "completions/mean_terminated_length": 1288.08349609375, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.796, "format_reward_mean_0": 0.8125, "format_reward_mean_1": 0.9765625, "format_reward_std_0": 0.3944427967071533, "format_reward_std_1": 0.15148578584194183, "frac_reward_zero_std_0": 0.0, "frac_reward_zero_std_1": 0.2916666865348816, "grad_norm": 0.7068716883659363, "kl": -3.0590778570968136e+29, "learning_rate": 2.0799999999999998e-07, "loss": 0.001, "num_tokens": 397330098.0, "problem_reward_mean_0": 0.4098748564720154, "problem_reward_mean_1": 0.7635288238525391, "problem_reward_std_0": 0.3946942985057831, "problem_reward_std_1": 0.3067418932914734, "reward_mean_0": 0.4098748564720154, "reward_mean_1": 0.7635288238525391, "reward_std_0": 0.19979169964790344, "reward_std_1": 0.10052882879972458, "rewards/compute_reward/mean": 0.586701825261116, "rewards/compute_reward/std": 0.35071809589862823, "step": 199, "updated_reward_mean_0": 0.8756868243217468, "updated_reward_mean_1": 0.6202996373176575, "updated_reward_std_0": 0.14398854970932007, "updated_reward_std_1": 0.0931631326675415 }, { "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": 0.08463541666666669, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1224.84375, "completions/mean_terminated_length": 1148.5762939453125, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.8, "format_reward_mean_0": 0.9791666865348816, "format_reward_mean_1": 0.9973958730697632, "format_reward_std_0": 0.14433756470680237, "format_reward_std_1": 0.05103103071451187, "frac_reward_zero_std_0": 0.5, "frac_reward_zero_std_1": 0.7083333730697632, "grad_norm": 0.35364776849746704, "kl": 1.4076893705361088e+35, "learning_rate": 2.0399999999999997e-07, "loss": 0.0006, "num_tokens": 399119100.0, "problem_reward_mean_0": 0.6866320371627808, "problem_reward_mean_1": 0.9180822372436523, "problem_reward_std_0": 0.39433059096336365, "problem_reward_std_1": 0.18218779563903809, "reward_mean_0": 0.6866320371627808, "reward_mean_1": 0.9180822372436523, "reward_std_0": 0.110328309237957, "reward_std_1": 0.04240766167640686, "rewards/compute_reward/mean": 0.8023571074008942, "rewards/compute_reward/std": 0.28825919330120087, "step": 200, "updated_reward_mean_0": 0.9466146230697632, "updated_reward_mean_1": 0.3418511152267456, "updated_reward_std_0": 0.007232969626784325, "updated_reward_std_1": 0.023000992834568024 } ], "logging_steps": 1, "max_steps": 250, "num_input_tokens_seen": 399119100, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }