diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,38443 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 14.951603498542275, + "eval_steps": 500, + "global_step": 1600, + "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": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3276.0, + "completions/mean_length": 583.7980346679688, + "completions/mean_terminated_length": 536.12109375, + "completions/min_length": 18.0, + "completions/min_terminated_length": 18.0, + "epoch": 0.009329446064139942, + "grad_norm": 0.16279126703739166, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 560435.0, + "reward": 0.5234375, + "reward_std": 0.260808527469635, + "rewards/verify_math_reward/mean": 0.5234375, + "rewards/verify_math_reward/std": 0.49972933530807495, + "step": 1 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3958.0, + "completions/mean_length": 590.8482666015625, + "completions/mean_terminated_length": 547.2813720703125, + "completions/min_length": 3.0, + "completions/min_terminated_length": 3.0, + "epoch": 0.018658892128279883, + "grad_norm": 0.12443200498819351, + "learning_rate": 1e-06, + "loss": 0.0046, + "num_tokens": 1132035.0, + "reward": 0.4854910969734192, + "reward_std": 0.21601703763008118, + "rewards/verify_math_reward/mean": 0.4854910671710968, + "rewards/verify_math_reward/std": 0.5000686049461365, + "step": 2 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3586.0, + "completions/mean_length": 606.2734375, + "completions/mean_terminated_length": 550.8809814453125, + "completions/min_length": 50.0, + "completions/min_terminated_length": 50.0, + "epoch": 0.027988338192419825, + "grad_norm": 0.12766121327877045, + "learning_rate": 1e-06, + "loss": 0.0184, + "num_tokens": 1716456.0, + "reward": 0.515625, + "reward_std": 0.2055736631155014, + "rewards/verify_math_reward/mean": 0.515625, + "rewards/verify_math_reward/std": 0.5000349283218384, + "step": 3 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2674.0, + "completions/mean_length": 553.9866333007812, + "completions/mean_terminated_length": 518.0473022460938, + "completions/min_length": 14.0, + "completions/min_terminated_length": 14.0, + "epoch": 0.037317784256559766, + "grad_norm": 0.1797182410955429, + "learning_rate": 1e-06, + "loss": 0.0142, + "num_tokens": 2255396.0, + "reward": 0.53125, + "reward_std": 0.2643796503543854, + "rewards/verify_math_reward/mean": 0.53125, + "rewards/verify_math_reward/std": 0.4993011951446533, + "step": 4 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3977.0, + "completions/mean_length": 602.3192138671875, + "completions/mean_terminated_length": 554.8936767578125, + "completions/min_length": 19.0, + "completions/min_terminated_length": 19.0, + "epoch": 0.04664723032069971, + "grad_norm": 0.14559607207775116, + "learning_rate": 1e-06, + "loss": -0.0004, + "num_tokens": 2829490.0, + "reward": 0.5145089626312256, + "reward_std": 0.24784532189369202, + "rewards/verify_math_reward/mean": 0.5145089030265808, + "rewards/verify_math_reward/std": 0.5000685453414917, + "step": 5 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3390.0, + "completions/mean_length": 620.560302734375, + "completions/mean_terminated_length": 561.3870849609375, + "completions/min_length": 12.0, + "completions/min_terminated_length": 12.0, + "epoch": 0.05597667638483965, + "grad_norm": 0.134445458650589, + "learning_rate": 1e-06, + "loss": 0.0071, + "num_tokens": 3414496.0, + "reward": 0.5223214626312256, + "reward_std": 0.2517502009868622, + "rewards/verify_math_reward/mean": 0.5223214030265808, + "rewards/verify_math_reward/std": 0.49978047609329224, + "step": 6 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2842.0, + "completions/mean_length": 557.359375, + "completions/mean_terminated_length": 517.4198608398438, + "completions/min_length": 59.0, + "completions/min_terminated_length": 59.0, + "epoch": 0.0653061224489796, + "grad_norm": 0.14972200989723206, + "learning_rate": 1e-06, + "loss": -0.0008, + "num_tokens": 3971386.0, + "reward": 0.5412946939468384, + "reward_std": 0.23314937949180603, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 7 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4069.0, + "completions/mean_length": 646.013427734375, + "completions/mean_terminated_length": 587.2735595703125, + "completions/min_length": 25.0, + "completions/min_terminated_length": 25.0, + "epoch": 0.07463556851311953, + "grad_norm": 0.12313701957464218, + "learning_rate": 1e-06, + "loss": 0.0025, + "num_tokens": 4563166.0, + "reward": 0.4955357313156128, + "reward_std": 0.23596841096878052, + "rewards/verify_math_reward/mean": 0.4955357015132904, + "rewards/verify_math_reward/std": 0.500259280204773, + "step": 8 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3397.0, + "completions/mean_length": 648.646240234375, + "completions/mean_terminated_length": 569.939453125, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "epoch": 0.08396501457725948, + "grad_norm": 0.13047420978546143, + "learning_rate": 1e-06, + "loss": 0.0073, + "num_tokens": 5151409.0, + "reward": 0.5245535969734192, + "reward_std": 0.226617693901062, + "rewards/verify_math_reward/mean": 0.5245535969734192, + "rewards/verify_math_reward/std": 0.4996756911277771, + "step": 9 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2405.0, + "completions/mean_length": 623.4832763671875, + "completions/mean_terminated_length": 552.292724609375, + "completions/min_length": 27.0, + "completions/min_terminated_length": 27.0, + "epoch": 0.09329446064139942, + "grad_norm": 0.12889985740184784, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 5731562.0, + "reward": 0.5066964626312256, + "reward_std": 0.23826003074645996, + "rewards/verify_math_reward/mean": 0.5066964030265808, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 10 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 666.0926513671875, + "completions/mean_terminated_length": 587.7842407226562, + "completions/min_length": 33.0, + "completions/min_terminated_length": 33.0, + "epoch": 0.10262390670553936, + "grad_norm": 0.12700113654136658, + "learning_rate": 1e-06, + "loss": 0.0002, + "num_tokens": 6330805.0, + "reward": 0.5011160969734192, + "reward_std": 0.19437934458255768, + "rewards/verify_math_reward/mean": 0.5011160969734192, + "rewards/verify_math_reward/std": 0.5002779960632324, + "step": 11 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2313.0, + "completions/mean_length": 632.5022583007812, + "completions/mean_terminated_length": 520.7764892578125, + "completions/min_length": 18.0, + "completions/min_terminated_length": 18.0, + "epoch": 0.1119533527696793, + "grad_norm": 0.12584556639194489, + "learning_rate": 1e-06, + "loss": -0.0068, + "num_tokens": 6885319.0, + "reward": 0.5502232313156128, + "reward_std": 0.19475941359996796, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 12 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3132.0, + "completions/mean_length": 616.6116333007812, + "completions/mean_terminated_length": 573.364990234375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 0.12128279883381925, + "grad_norm": 0.12127769738435745, + "learning_rate": 1e-06, + "loss": -0.0073, + "num_tokens": 7480739.0, + "reward": 0.5602678656578064, + "reward_std": 0.221996009349823, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317117214203, + "step": 13 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3769.0, + "completions/mean_length": 562.4442138671875, + "completions/mean_terminated_length": 526.5907592773438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 0.1306122448979592, + "grad_norm": 0.12756912410259247, + "learning_rate": 1e-06, + "loss": 0.0102, + "num_tokens": 8043937.0, + "reward": 0.5491071939468384, + "reward_std": 0.2145100086927414, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 14 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3477.0, + "completions/mean_length": 636.7254638671875, + "completions/mean_terminated_length": 577.8275146484375, + "completions/min_length": 23.0, + "completions/min_terminated_length": 23.0, + "epoch": 0.13994169096209913, + "grad_norm": 0.12179768085479736, + "learning_rate": 1e-06, + "loss": 0.013, + "num_tokens": 8646467.0, + "reward": 0.5535714626312256, + "reward_std": 0.19050543010234833, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973994791507721, + "step": 15 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 615.0123291015625, + "completions/mean_terminated_length": 555.74462890625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 0.14927113702623906, + "grad_norm": 0.11943355947732925, + "learning_rate": 1e-06, + "loss": -0.0037, + "num_tokens": 9227542.0, + "reward": 0.5658482313156128, + "reward_std": 0.19520196318626404, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 16 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2099.0, + "completions/mean_length": 580.3136596679688, + "completions/mean_terminated_length": 556.6123657226562, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 0.158600583090379, + "grad_norm": 0.12313390523195267, + "learning_rate": 1e-06, + "loss": 0.022, + "num_tokens": 9808551.0, + "reward": 0.551339328289032, + "reward_std": 0.20590998232364655, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 17 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3671.0, + "completions/mean_length": 655.5592041015625, + "completions/mean_terminated_length": 581.0227661132812, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "epoch": 0.16793002915451896, + "grad_norm": 0.130902960896492, + "learning_rate": 1e-06, + "loss": 0.0051, + "num_tokens": 10399340.0, + "reward": 0.5111607313156128, + "reward_std": 0.24374611675739288, + "rewards/verify_math_reward/mean": 0.5111607313156128, + "rewards/verify_math_reward/std": 0.5001546144485474, + "step": 18 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3546.0, + "completions/mean_length": 642.9342041015625, + "completions/mean_terminated_length": 551.9599609375, + "completions/min_length": 4.0, + "completions/min_terminated_length": 4.0, + "epoch": 0.1772594752186589, + "grad_norm": 0.1357765793800354, + "learning_rate": 1e-06, + "loss": -0.0144, + "num_tokens": 10975657.0, + "reward": 0.6194196939468384, + "reward_std": 0.2508493959903717, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 19 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.004464285714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3992.0, + "completions/mean_length": 550.7623291015625, + "completions/mean_terminated_length": 534.8643798828125, + "completions/min_length": 11.0, + "completions/min_terminated_length": 11.0, + "epoch": 0.18658892128279883, + "grad_norm": 0.12462402135133743, + "learning_rate": 1e-06, + "loss": 0.0274, + "num_tokens": 11543924.0, + "reward": 0.6350446939468384, + "reward_std": 0.21155014634132385, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 20 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4090.0, + "completions/mean_length": 595.3136596679688, + "completions/mean_terminated_length": 535.7105712890625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 0.19591836734693877, + "grad_norm": 0.1224556565284729, + "learning_rate": 1e-06, + "loss": 0.0134, + "num_tokens": 12098077.0, + "reward": 0.590401828289032, + "reward_std": 0.19666734337806702, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 21 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3172.0, + "completions/mean_length": 631.0580444335938, + "completions/mean_terminated_length": 603.7750244140625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.20524781341107873, + "grad_norm": 0.11804339289665222, + "learning_rate": 1e-06, + "loss": 0.0182, + "num_tokens": 12730265.0, + "reward": 0.5256696939468384, + "reward_std": 0.1963321417570114, + "rewards/verify_math_reward/mean": 0.5256696343421936, + "rewards/verify_math_reward/std": 0.4996195137500763, + "step": 22 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3255.0, + "completions/mean_length": 602.216552734375, + "completions/mean_terminated_length": 554.7896118164062, + "completions/min_length": 11.0, + "completions/min_terminated_length": 11.0, + "epoch": 0.21457725947521866, + "grad_norm": 0.12200116366147995, + "learning_rate": 1e-06, + "loss": 0.0015, + "num_tokens": 13305035.0, + "reward": 0.6316964626312256, + "reward_std": 0.19313082098960876, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 23 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2276.0, + "completions/mean_length": 615.6116333007812, + "completions/mean_terminated_length": 580.297607421875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 0.2239067055393586, + "grad_norm": 0.1176367923617363, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 13904055.0, + "reward": 0.5267857313156128, + "reward_std": 0.2082500159740448, + "rewards/verify_math_reward/mean": 0.5267857313156128, + "rewards/verify_math_reward/std": 0.4995608627796173, + "step": 24 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3909.0, + "completions/mean_length": 663.9553833007812, + "completions/mean_terminated_length": 601.5545043945312, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "epoch": 0.23323615160349853, + "grad_norm": 0.11231141537427902, + "learning_rate": 1e-06, + "loss": 0.0037, + "num_tokens": 14518655.0, + "reward": 0.5345982313156128, + "reward_std": 0.17476637661457062, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 25 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2841.0, + "completions/mean_length": 632.375, + "completions/mean_terminated_length": 565.387939453125, + "completions/min_length": 38.0, + "completions/min_terminated_length": 38.0, + "epoch": 0.2425655976676385, + "grad_norm": 0.12935595214366913, + "learning_rate": 1e-06, + "loss": 0.0131, + "num_tokens": 15103103.0, + "reward": 0.5022321939468384, + "reward_std": 0.2314223051071167, + "rewards/verify_math_reward/mean": 0.5022321343421936, + "rewards/verify_math_reward/std": 0.5002742409706116, + "step": 26 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3503.0, + "completions/mean_length": 594.5424194335938, + "completions/mean_terminated_length": 551.021484375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 0.2518950437317784, + "grad_norm": 0.1343274563550949, + "learning_rate": 1e-06, + "loss": 0.0008, + "num_tokens": 15678877.0, + "reward": 0.5814732313156128, + "reward_std": 0.21113713085651398, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 27 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2757.0, + "completions/mean_length": 592.0256958007812, + "completions/mean_terminated_length": 552.4774169921875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.2612244897959184, + "grad_norm": 0.12321079522371292, + "learning_rate": 1e-06, + "loss": 0.0121, + "num_tokens": 16258364.0, + "reward": 0.5479910969734192, + "reward_std": 0.20388120412826538, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 28 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3989.0, + "completions/mean_length": 614.122802734375, + "completions/mean_terminated_length": 562.8607177734375, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 0.2705539358600583, + "grad_norm": 0.12643910944461823, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 16844354.0, + "reward": 0.512276828289032, + "reward_std": 0.20929309725761414, + "rewards/verify_math_reward/mean": 0.5122767686843872, + "rewards/verify_math_reward/std": 0.500128448009491, + "step": 29 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3742.0, + "completions/mean_length": 569.3917846679688, + "completions/mean_terminated_length": 537.6204833984375, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 0.27988338192419826, + "grad_norm": 0.1229817271232605, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 17417089.0, + "reward": 0.5189732313156128, + "reward_std": 0.20490150153636932, + "rewards/verify_math_reward/mean": 0.5189732313156128, + "rewards/verify_math_reward/std": 0.49991893768310547, + "step": 30 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4007.0, + "completions/mean_length": 663.896240234375, + "completions/mean_terminated_length": 613.366943359375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 0.2892128279883382, + "grad_norm": 0.12167727202177048, + "learning_rate": 1e-06, + "loss": 0.0107, + "num_tokens": 18043572.0, + "reward": 0.5323660969734192, + "reward_std": 0.24134893715381622, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 31 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2534.0, + "completions/mean_length": 590.7745971679688, + "completions/mean_terminated_length": 551.2122192382812, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 0.29854227405247813, + "grad_norm": 0.1275821179151535, + "learning_rate": 1e-06, + "loss": 0.0074, + "num_tokens": 18622786.0, + "reward": 0.574776828289032, + "reward_std": 0.19047291576862335, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 32 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2392.0, + "completions/mean_length": 558.90625, + "completions/mean_terminated_length": 510.89141845703125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 0.30787172011661806, + "grad_norm": 0.14067736268043518, + "learning_rate": 1e-06, + "loss": 0.0011, + "num_tokens": 19160950.0, + "reward": 0.5412946939468384, + "reward_std": 0.20786459743976593, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 33 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3821.0, + "completions/mean_length": 602.005615234375, + "completions/mean_terminated_length": 550.5651245117188, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.317201166180758, + "grad_norm": 0.11986638605594635, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 19730491.0, + "reward": 0.5379464626312256, + "reward_std": 0.20639601349830627, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 34 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2325.0, + "completions/mean_length": 610.3638916015625, + "completions/mean_terminated_length": 555.0363159179688, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 0.32653061224489793, + "grad_norm": 0.12306813150644302, + "learning_rate": 1e-06, + "loss": -0.0012, + "num_tokens": 20303689.0, + "reward": 0.5848214626312256, + "reward_std": 0.21800485253334045, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 35 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2622.0, + "completions/mean_length": 609.1373291015625, + "completions/mean_terminated_length": 577.72412109375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 0.3358600583090379, + "grad_norm": 0.11342012882232666, + "learning_rate": 1e-06, + "loss": 0.0058, + "num_tokens": 20909532.0, + "reward": 0.543526828289032, + "reward_std": 0.1877259612083435, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 36 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3879.0, + "completions/mean_length": 583.880615234375, + "completions/mean_terminated_length": 536.2047729492188, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 0.34518950437317786, + "grad_norm": 0.13498900830745697, + "learning_rate": 1e-06, + "loss": 0.0142, + "num_tokens": 21460265.0, + "reward": 0.5524553656578064, + "reward_std": 0.2318705916404724, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 37 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2287.0, + "completions/mean_length": 638.3236694335938, + "completions/mean_terminated_length": 559.3812255859375, + "completions/min_length": 38.0, + "completions/min_terminated_length": 38.0, + "epoch": 0.3545189504373178, + "grad_norm": 0.12872715294361115, + "learning_rate": 1e-06, + "loss": -0.0033, + "num_tokens": 22039619.0, + "reward": 0.5479910969734192, + "reward_std": 0.21260139346122742, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 38 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2261.0, + "completions/mean_length": 559.3125, + "completions/mean_terminated_length": 535.4696655273438, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 0.3638483965014577, + "grad_norm": 0.13344311714172363, + "learning_rate": 1e-06, + "loss": 0.0075, + "num_tokens": 22612051.0, + "reward": 0.5390625, + "reward_std": 0.2111714631319046, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 39 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3793.0, + "completions/mean_length": 604.6428833007812, + "completions/mean_terminated_length": 541.1636352539062, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 0.37317784256559766, + "grad_norm": 0.1348295360803604, + "learning_rate": 1e-06, + "loss": 0.0006, + "num_tokens": 23190627.0, + "reward": 0.5390625, + "reward_std": 0.200280100107193, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 40 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2643.0, + "completions/mean_length": 653.5535888671875, + "completions/mean_terminated_length": 606.8235473632812, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 0.3825072886297376, + "grad_norm": 0.12652353942394257, + "learning_rate": 1e-06, + "loss": 0.0113, + "num_tokens": 23828843.0, + "reward": 0.5089285969734192, + "reward_std": 0.24242384731769562, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 41 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 612.0424194335938, + "completions/mean_terminated_length": 556.7415161132812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 0.39183673469387753, + "grad_norm": 0.10542068630456924, + "learning_rate": 1e-06, + "loss": 0.0141, + "num_tokens": 24411089.0, + "reward": 0.5212053656578064, + "reward_std": 0.17005804181098938, + "rewards/verify_math_reward/mean": 0.5212053656578064, + "rewards/verify_math_reward/std": 0.49982914328575134, + "step": 42 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0022321428571429047, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3583.0, + "completions/mean_length": 569.7756958007812, + "completions/mean_terminated_length": 561.8870239257812, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 0.40116618075801747, + "grad_norm": 0.1200394555926323, + "learning_rate": 1e-06, + "loss": 0.0196, + "num_tokens": 24999552.0, + "reward": 0.5524553656578064, + "reward_std": 0.21748380362987518, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 43 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2246.0, + "completions/mean_length": 562.7902221679688, + "completions/mean_terminated_length": 494.45733642578125, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 0.41049562682215746, + "grad_norm": 0.1317327469587326, + "learning_rate": 1e-06, + "loss": 0.0007, + "num_tokens": 25520372.0, + "reward": 0.6049107313156128, + "reward_std": 0.207753986120224, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 44 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3878.0, + "completions/mean_length": 669.0803833007812, + "completions/mean_terminated_length": 578.7949829101562, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 0.4198250728862974, + "grad_norm": 0.1366340070962906, + "learning_rate": 1e-06, + "loss": 0.0144, + "num_tokens": 26119052.0, + "reward": 0.551339328289032, + "reward_std": 0.2380324900150299, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 45 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3879.0, + "completions/mean_length": 686.3627319335938, + "completions/mean_terminated_length": 624.3693237304688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 0.4291545189504373, + "grad_norm": 0.11587977409362793, + "learning_rate": 1e-06, + "loss": 0.0118, + "num_tokens": 26756945.0, + "reward": 0.543526828289032, + "reward_std": 0.21872051060199738, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 46 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3860.0, + "completions/mean_length": 649.734375, + "completions/mean_terminated_length": 587.074951171875, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 0.43848396501457726, + "grad_norm": 0.1326758861541748, + "learning_rate": 1e-06, + "loss": 0.0117, + "num_tokens": 27358467.0, + "reward": 0.535714328289032, + "reward_std": 0.2493869662284851, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 47 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3103.0, + "completions/mean_length": 626.5011596679688, + "completions/mean_terminated_length": 547.288818359375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 0.4478134110787172, + "grad_norm": 0.11908731609582901, + "learning_rate": 1e-06, + "loss": 0.0142, + "num_tokens": 27932540.0, + "reward": 0.5234375, + "reward_std": 0.1784840077161789, + "rewards/verify_math_reward/mean": 0.5234375, + "rewards/verify_math_reward/std": 0.49972933530807495, + "step": 48 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3629.0, + "completions/mean_length": 613.0089721679688, + "completions/mean_terminated_length": 573.697509765625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 0.45714285714285713, + "grad_norm": 0.14169055223464966, + "learning_rate": 1e-06, + "loss": 0.0217, + "num_tokens": 28523852.0, + "reward": 0.5011160969734192, + "reward_std": 0.21553102135658264, + "rewards/verify_math_reward/mean": 0.5011160969734192, + "rewards/verify_math_reward/std": 0.5002779960632324, + "step": 49 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4037.0, + "completions/mean_length": 663.5145263671875, + "completions/mean_terminated_length": 544.6062622070312, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 0.46647230320699706, + "grad_norm": 0.1248873695731163, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 29098017.0, + "reward": 0.5479910969734192, + "reward_std": 0.19670717418193817, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 50 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3823.0, + "completions/mean_length": 674.2455444335938, + "completions/mean_terminated_length": 584.0962524414062, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 0.47580174927113705, + "grad_norm": 0.13538333773612976, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 29689141.0, + "reward": 0.5881696939468384, + "reward_std": 0.2436712384223938, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 51 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3670.0, + "completions/mean_length": 587.8917846679688, + "completions/mean_terminated_length": 540.2703857421875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 0.485131195335277, + "grad_norm": 0.12987874448299408, + "learning_rate": 1e-06, + "loss": 0.0109, + "num_tokens": 30246428.0, + "reward": 0.6328125, + "reward_std": 0.20177756249904633, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 52 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2864.0, + "completions/mean_length": 602.796875, + "completions/mean_terminated_length": 543.3212280273438, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 0.4944606413994169, + "grad_norm": 0.12448890507221222, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 30810846.0, + "reward": 0.5524553656578064, + "reward_std": 0.19840146601200104, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 53 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3736.0, + "completions/mean_length": 533.4129638671875, + "completions/mean_terminated_length": 497.2649230957031, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 0.5037900874635568, + "grad_norm": 0.13337939977645874, + "learning_rate": 1e-06, + "loss": 0.0189, + "num_tokens": 31341760.0, + "reward": 0.5725446939468384, + "reward_std": 0.18340355157852173, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 54 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2066.0, + "completions/mean_length": 530.4910888671875, + "completions/mean_terminated_length": 510.4826354980469, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 0.5131195335276968, + "grad_norm": 0.14338673651218414, + "learning_rate": 1e-06, + "loss": 0.0199, + "num_tokens": 31890992.0, + "reward": 0.6127232313156128, + "reward_std": 0.19637449085712433, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 55 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3688.0, + "completions/mean_length": 608.536865234375, + "completions/mean_terminated_length": 557.1925048828125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 0.5224489795918368, + "grad_norm": 0.12986308336257935, + "learning_rate": 1e-06, + "loss": 0.0114, + "num_tokens": 32474401.0, + "reward": 0.4966517984867096, + "reward_std": 0.2038070410490036, + "rewards/verify_math_reward/mean": 0.4966517984867096, + "rewards/verify_math_reward/std": 0.5002680420875549, + "step": 56 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2142.0, + "completions/mean_length": 584.9475708007812, + "completions/mean_terminated_length": 529.216552734375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.5317784256559767, + "grad_norm": 0.12951917946338654, + "learning_rate": 1e-06, + "loss": 0.0186, + "num_tokens": 33027922.0, + "reward": 0.5848214626312256, + "reward_std": 0.18798990547657013, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 57 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3472.0, + "completions/mean_length": 604.9832763671875, + "completions/mean_terminated_length": 553.5866088867188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 0.5411078717201167, + "grad_norm": 0.12152307480573654, + "learning_rate": 1e-06, + "loss": 0.0098, + "num_tokens": 33606603.0, + "reward": 0.5412946939468384, + "reward_std": 0.1991407573223114, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 58 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3860.0, + "completions/mean_length": 585.0402221679688, + "completions/mean_terminated_length": 525.26220703125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 0.5504373177842565, + "grad_norm": 0.1423441469669342, + "learning_rate": 1e-06, + "loss": -0.007, + "num_tokens": 34163503.0, + "reward": 0.5803571939468384, + "reward_std": 0.19084423780441284, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 59 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1863.0, + "completions/mean_length": 567.099365234375, + "completions/mean_terminated_length": 527.269775390625, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 0.5597667638483965, + "grad_norm": 0.12532348930835724, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 34715672.0, + "reward": 0.5915178656578064, + "reward_std": 0.17904217541217804, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 60 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2935.0, + "completions/mean_length": 659.638427734375, + "completions/mean_terminated_length": 597.1590576171875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.5690962099125364, + "grad_norm": 0.11917824298143387, + "learning_rate": 1e-06, + "loss": 0.0149, + "num_tokens": 35325196.0, + "reward": 0.5167410969734192, + "reward_std": 0.21068432927131653, + "rewards/verify_math_reward/mean": 0.5167410969734192, + "rewards/verify_math_reward/std": 0.4999987483024597, + "step": 61 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3161.0, + "completions/mean_length": 649.9676513671875, + "completions/mean_terminated_length": 599.2332763671875, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 0.5784256559766764, + "grad_norm": 0.12504246830940247, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 35938287.0, + "reward": 0.5457589626312256, + "reward_std": 0.2385701686143875, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 62 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2267.0, + "completions/mean_length": 597.2902221679688, + "completions/mean_terminated_length": 545.7802734375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 0.5877551020408164, + "grad_norm": 0.13810665905475616, + "learning_rate": 1e-06, + "loss": 0.0033, + "num_tokens": 36498867.0, + "reward": 0.6026785969734192, + "reward_std": 0.22822391986846924, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 63 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3502.0, + "completions/mean_length": 611.9921875, + "completions/mean_terminated_length": 560.69873046875, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 0.5970845481049563, + "grad_norm": 0.1360047310590744, + "learning_rate": 1e-06, + "loss": 0.0114, + "num_tokens": 37083732.0, + "reward": 0.6160714626312256, + "reward_std": 0.2054981142282486, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 64 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3711.0, + "completions/mean_length": 617.9386596679688, + "completions/mean_terminated_length": 582.6482543945312, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 0.6064139941690962, + "grad_norm": 0.126007542014122, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 37698221.0, + "reward": 0.5379464626312256, + "reward_std": 0.20400065183639526, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 65 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1989.0, + "completions/mean_length": 596.2142944335938, + "completions/mean_terminated_length": 528.52783203125, + "completions/min_length": 59.0, + "completions/min_terminated_length": 59.0, + "epoch": 0.6157434402332361, + "grad_norm": 0.12345066666603088, + "learning_rate": 1e-06, + "loss": 0.0018, + "num_tokens": 38261333.0, + "reward": 0.5915178656578064, + "reward_std": 0.17427602410316467, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 66 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 708.4420166015625, + "completions/mean_terminated_length": 627.1405639648438, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.6250728862973761, + "grad_norm": 0.12230608612298965, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 38905753.0, + "reward": 0.5, + "reward_std": 0.2242519110441208, + "rewards/verify_math_reward/mean": 0.5, + "rewards/verify_math_reward/std": 0.5002792477607727, + "step": 67 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4021.0, + "completions/mean_length": 632.646240234375, + "completions/mean_terminated_length": 565.6643676757812, + "completions/min_length": 3.0, + "completions/min_terminated_length": 3.0, + "epoch": 0.634402332361516, + "grad_norm": 0.13022969663143158, + "learning_rate": 1e-06, + "loss": -0.0006, + "num_tokens": 39486596.0, + "reward": 0.5758928656578064, + "reward_std": 0.18257826566696167, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 68 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3286.0, + "completions/mean_length": 641.9866333007812, + "completions/mean_terminated_length": 603.0022583007812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.643731778425656, + "grad_norm": 0.12673784792423248, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 40111224.0, + "reward": 0.5569196939468384, + "reward_std": 0.23615315556526184, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 69 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 652.265625, + "completions/mean_terminated_length": 605.5181274414062, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 0.6530612244897959, + "grad_norm": 0.13908004760742188, + "learning_rate": 1e-06, + "loss": 0.0019, + "num_tokens": 40736542.0, + "reward": 0.551339328289032, + "reward_std": 0.26088592410087585, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 70 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 612.810302734375, + "completions/mean_terminated_length": 569.5164184570312, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 0.6623906705539359, + "grad_norm": 0.13196811079978943, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 41340148.0, + "reward": 0.5424107313156128, + "reward_std": 0.2030581682920456, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 71 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2886.0, + "completions/mean_length": 633.4420166015625, + "completions/mean_terminated_length": 574.4880981445312, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 0.6717201166180758, + "grad_norm": 0.2375737726688385, + "learning_rate": 1e-06, + "loss": 0.0013, + "num_tokens": 41943712.0, + "reward": 0.5245535969734192, + "reward_std": 0.22184516489505768, + "rewards/verify_math_reward/mean": 0.5245535969734192, + "rewards/verify_math_reward/std": 0.4996756613254547, + "step": 72 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3683.0, + "completions/mean_length": 668.9464721679688, + "completions/mean_terminated_length": 578.6575317382812, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 0.6810495626822157, + "grad_norm": 0.125754714012146, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 42545656.0, + "reward": 0.5401785969734192, + "reward_std": 0.208957239985466, + "rewards/verify_math_reward/mean": 0.5401785969734192, + "rewards/verify_math_reward/std": 0.49866142868995667, + "step": 73 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3217.0, + "completions/mean_length": 711.8594360351562, + "completions/mean_terminated_length": 610.72412109375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 0.6903790087463557, + "grad_norm": 0.12093528360128403, + "learning_rate": 1e-06, + "loss": 0.008, + "num_tokens": 43166690.0, + "reward": 0.5569196939468384, + "reward_std": 0.2078709453344345, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 74 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3360.0, + "completions/mean_length": 659.029052734375, + "completions/mean_terminated_length": 596.5386352539062, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 0.6997084548104956, + "grad_norm": 0.11361652612686157, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 43776708.0, + "reward": 0.551339328289032, + "reward_std": 0.1917535364627838, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 75 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2998.0, + "completions/mean_length": 711.169677734375, + "completions/mean_terminated_length": 621.9931640625, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.7090379008746356, + "grad_norm": 0.13189241290092468, + "learning_rate": 1e-06, + "loss": 0.0128, + "num_tokens": 44412948.0, + "reward": 0.53125, + "reward_std": 0.25039586424827576, + "rewards/verify_math_reward/mean": 0.53125, + "rewards/verify_math_reward/std": 0.4993011951446533, + "step": 76 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2839.0, + "completions/mean_length": 649.984375, + "completions/mean_terminated_length": 571.3081665039062, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.7183673469387755, + "grad_norm": 0.14138217270374298, + "learning_rate": 1e-06, + "loss": 0.0179, + "num_tokens": 45005038.0, + "reward": 0.546875, + "reward_std": 0.23957975208759308, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 77 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 648.763427734375, + "completions/mean_terminated_length": 570.059326171875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 0.7276967930029155, + "grad_norm": 0.1287572979927063, + "learning_rate": 1e-06, + "loss": 0.006, + "num_tokens": 45590322.0, + "reward": 0.574776828289032, + "reward_std": 0.20313303172588348, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 78 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3246.0, + "completions/mean_length": 588.5413208007812, + "completions/mean_terminated_length": 548.9537353515625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.7370262390670554, + "grad_norm": 0.139412060379982, + "learning_rate": 1e-06, + "loss": 0.0011, + "num_tokens": 46165783.0, + "reward": 0.5859375, + "reward_std": 0.22034701704978943, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 79 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 575.8939819335938, + "completions/mean_terminated_length": 544.1813354492188, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 0.7463556851311953, + "grad_norm": 0.14733393490314484, + "learning_rate": 1e-06, + "loss": 0.0116, + "num_tokens": 46739064.0, + "reward": 0.6037946939468384, + "reward_std": 0.2509681284427643, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 80 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2903.0, + "completions/mean_length": 666.5736694335938, + "completions/mean_terminated_length": 600.2479858398438, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 0.7556851311953353, + "grad_norm": 0.12558312714099884, + "learning_rate": 1e-06, + "loss": 0.0132, + "num_tokens": 47362498.0, + "reward": 0.5703125, + "reward_std": 0.22872063517570496, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 81 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4041.0, + "completions/mean_length": 667.9017944335938, + "completions/mean_terminated_length": 573.5504150390625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.7650145772594752, + "grad_norm": 0.12746962904930115, + "learning_rate": 1e-06, + "loss": -0.0071, + "num_tokens": 47956802.0, + "reward": 0.566964328289032, + "reward_std": 0.19772042334079742, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 82 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 609.3973388671875, + "completions/mean_terminated_length": 566.06103515625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 0.7743440233236152, + "grad_norm": 0.12530255317687988, + "learning_rate": 1e-06, + "loss": -0.0014, + "num_tokens": 48542982.0, + "reward": 0.6305803656578064, + "reward_std": 0.2122662216424942, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 83 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3931.0, + "completions/mean_length": 689.1395263671875, + "completions/mean_terminated_length": 599.3826293945312, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.7836734693877551, + "grad_norm": 0.13452297449111938, + "learning_rate": 1e-06, + "loss": 0.0059, + "num_tokens": 49161859.0, + "reward": 0.5, + "reward_std": 0.22192074358463287, + "rewards/verify_math_reward/mean": 0.5, + "rewards/verify_math_reward/std": 0.5002792477607727, + "step": 84 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2056.0, + "completions/mean_length": 623.0245971679688, + "completions/mean_terminated_length": 547.7833251953125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.793002915451895, + "grad_norm": 0.12535934150218964, + "learning_rate": 1e-06, + "loss": 0.0068, + "num_tokens": 49731993.0, + "reward": 0.574776828289032, + "reward_std": 0.18697144091129303, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 85 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2100.0, + "completions/mean_length": 600.349365234375, + "completions/mean_terminated_length": 568.8569946289062, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 0.8023323615160349, + "grad_norm": 0.12047363072633743, + "learning_rate": 1e-06, + "loss": -0.0128, + "num_tokens": 50324002.0, + "reward": 0.606026828289032, + "reward_std": 0.18708907067775726, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 86 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3408.0, + "completions/mean_length": 573.5725708007812, + "completions/mean_terminated_length": 537.8319702148438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 0.8116618075801749, + "grad_norm": 0.13725604116916656, + "learning_rate": 1e-06, + "loss": 0.0138, + "num_tokens": 50881723.0, + "reward": 0.6127232313156128, + "reward_std": 0.201027974486351, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 87 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2509.0, + "completions/mean_length": 607.1395263671875, + "completions/mean_terminated_length": 547.7378540039062, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 0.8209912536443149, + "grad_norm": 0.1279614120721817, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 51444352.0, + "reward": 0.6171875, + "reward_std": 0.18298916518688202, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 88 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2580.0, + "completions/mean_length": 672.2355346679688, + "completions/mean_terminated_length": 598.0604248046875, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 0.8303206997084548, + "grad_norm": 0.1233532503247261, + "learning_rate": 1e-06, + "loss": -0.0085, + "num_tokens": 52070699.0, + "reward": 0.5580357313156128, + "reward_std": 0.1935872733592987, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 89 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2521.0, + "completions/mean_length": 639.6908569335938, + "completions/mean_terminated_length": 572.8452758789062, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 0.8396501457725948, + "grad_norm": 0.12155285477638245, + "learning_rate": 1e-06, + "loss": 0.0034, + "num_tokens": 52661782.0, + "reward": 0.5535714626312256, + "reward_std": 0.17690393328666687, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973995089530945, + "step": 90 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3007.0, + "completions/mean_length": 644.9464721679688, + "completions/mean_terminated_length": 578.2025146484375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.8489795918367347, + "grad_norm": 0.12493814527988434, + "learning_rate": 1e-06, + "loss": 0.0022, + "num_tokens": 53257982.0, + "reward": 0.6116071939468384, + "reward_std": 0.18588514626026154, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 91 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 651.6585083007812, + "completions/mean_terminated_length": 596.9863891601562, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 0.8583090379008746, + "grad_norm": 0.11729107052087784, + "learning_rate": 1e-06, + "loss": 0.0186, + "num_tokens": 53871604.0, + "reward": 0.6149553656578064, + "reward_std": 0.1893460899591446, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 92 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3905.0, + "completions/mean_length": 634.122802734375, + "completions/mean_terminated_length": 563.150390625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 0.8676384839650145, + "grad_norm": 0.14197830855846405, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 54463586.0, + "reward": 0.5814732313156128, + "reward_std": 0.25926899909973145, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 93 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3285.0, + "completions/mean_length": 618.9520263671875, + "completions/mean_terminated_length": 579.7077026367188, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 0.8769679300291545, + "grad_norm": 0.12423322349786758, + "learning_rate": 1e-06, + "loss": -0.0008, + "num_tokens": 55053959.0, + "reward": 0.5859375, + "reward_std": 0.19234876334667206, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 94 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3804.0, + "completions/mean_length": 676.2890625, + "completions/mean_terminated_length": 602.2017822265625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 0.8862973760932945, + "grad_norm": 0.11907894164323807, + "learning_rate": 1e-06, + "loss": 0.0036, + "num_tokens": 55669586.0, + "reward": 0.5535714626312256, + "reward_std": 0.19294928014278412, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973995089530945, + "step": 95 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3194.0, + "completions/mean_length": 643.1004638671875, + "completions/mean_terminated_length": 592.2650146484375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 0.8956268221574344, + "grad_norm": 0.12358750402927399, + "learning_rate": 1e-06, + "loss": 0.0149, + "num_tokens": 56276012.0, + "reward": 0.5915178656578064, + "reward_std": 0.21958816051483154, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 96 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3991.0, + "completions/mean_length": 662.0636596679688, + "completions/mean_terminated_length": 603.5971069335938, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 0.9049562682215744, + "grad_norm": 0.11952099204063416, + "learning_rate": 1e-06, + "loss": 0.0138, + "num_tokens": 56899077.0, + "reward": 0.5714285969734192, + "reward_std": 0.2015172392129898, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 97 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2761.0, + "completions/mean_length": 702.7935791015625, + "completions/mean_terminated_length": 637.1683349609375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.9142857142857143, + "grad_norm": 0.11333039402961731, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 57545076.0, + "reward": 0.5133928656578064, + "reward_std": 0.19054116308689117, + "rewards/verify_math_reward/mean": 0.5133928656578064, + "rewards/verify_math_reward/std": 0.500099778175354, + "step": 98 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3370.0, + "completions/mean_length": 671.5201416015625, + "completions/mean_terminated_length": 601.3143920898438, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 0.9236151603498542, + "grad_norm": 0.1330297440290451, + "learning_rate": 1e-06, + "loss": 0.0103, + "num_tokens": 58172246.0, + "reward": 0.5145089626312256, + "reward_std": 0.22120507061481476, + "rewards/verify_math_reward/mean": 0.5145089030265808, + "rewards/verify_math_reward/std": 0.5000685453414917, + "step": 99 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 693.1641235351562, + "completions/mean_terminated_length": 603.51318359375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.9329446064139941, + "grad_norm": 0.1167035773396492, + "learning_rate": 1e-06, + "loss": -0.0026, + "num_tokens": 58789377.0, + "reward": 0.5703125, + "reward_std": 0.17581875622272491, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 100 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3859.0, + "completions/mean_length": 664.2801513671875, + "completions/mean_terminated_length": 589.9327392578125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 0.9422740524781341, + "grad_norm": 0.12442203611135483, + "learning_rate": 1e-06, + "loss": -0.0085, + "num_tokens": 59392268.0, + "reward": 0.5691964626312256, + "reward_std": 0.1979147344827652, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 101 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3652.0, + "completions/mean_length": 638.84375, + "completions/mean_terminated_length": 595.8734741210938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 0.9516034985422741, + "grad_norm": 0.1264164000749588, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 60019528.0, + "reward": 0.5491071939468384, + "reward_std": 0.2003892958164215, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 102 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 734.8482666015625, + "completions/mean_terminated_length": 646.2955932617188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 0.960932944606414, + "grad_norm": 0.11680614203214645, + "learning_rate": 1e-06, + "loss": 0.0154, + "num_tokens": 60671208.0, + "reward": 0.546875, + "reward_std": 0.18077746033668518, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 103 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3937.0, + "completions/mean_length": 614.7957763671875, + "completions/mean_terminated_length": 575.5045166015625, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "epoch": 0.970262390670554, + "grad_norm": 0.139450803399086, + "learning_rate": 1e-06, + "loss": 0.0195, + "num_tokens": 61274049.0, + "reward": 0.59375, + "reward_std": 0.24026376008987427, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 591.904052734375, + "completions/mean_terminated_length": 536.283447265625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.9795918367346939, + "grad_norm": 0.1420077532529831, + "learning_rate": 1e-06, + "loss": -0.0005, + "num_tokens": 61834947.0, + "reward": 0.546875, + "reward_std": 0.2228643298149109, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3632.0, + "completions/mean_length": 633.6864013671875, + "completions/mean_terminated_length": 574.7366943359375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 0.9889212827988338, + "grad_norm": 0.1315501481294632, + "learning_rate": 1e-06, + "loss": -0.0082, + "num_tokens": 62423482.0, + "reward": 0.5736607313156128, + "reward_std": 0.18070074915885925, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 106 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.019886363636363646, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3818.0, + "completions/mean_length": 602.875, + "completions/mean_terminated_length": 532.0, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 0.9982507288629737, + "grad_norm": 0.13619235157966614, + "learning_rate": 1e-06, + "loss": 0.0101, + "num_tokens": 63009874.0, + "reward": 0.551339328289032, + "reward_std": 0.22102029621601105, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2732.0, + "completions/mean_length": 620.9989013671875, + "completions/mean_terminated_length": 569.8380126953125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.00932944606414, + "grad_norm": 0.13184566795825958, + "learning_rate": 1e-06, + "loss": 0.0037, + "num_tokens": 63603465.0, + "reward": 0.5714285969734192, + "reward_std": 0.22684520483016968, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 652.6551513671875, + "completions/mean_terminated_length": 570.0148315429688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 1.01865889212828, + "grad_norm": 0.13099420070648193, + "learning_rate": 1e-06, + "loss": 0.0235, + "num_tokens": 64198676.0, + "reward": 0.5636160969734192, + "reward_std": 0.21658366918563843, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 109 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3688.0, + "completions/mean_length": 675.78125, + "completions/mean_terminated_length": 601.6829833984375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 1.0279883381924197, + "grad_norm": 0.12204349040985107, + "learning_rate": 1e-06, + "loss": -0.0039, + "num_tokens": 64819352.0, + "reward": 0.5290178656578064, + "reward_std": 0.19715909659862518, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943602085113525, + "step": 110 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3496.0, + "completions/mean_length": 687.6842041015625, + "completions/mean_terminated_length": 573.6804809570312, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "epoch": 1.0373177842565597, + "grad_norm": 0.11746125668287277, + "learning_rate": 1e-06, + "loss": -0.0001, + "num_tokens": 65421109.0, + "reward": 0.5066964626312256, + "reward_std": 0.17341159284114838, + "rewards/verify_math_reward/mean": 0.5066964030265808, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4081.0, + "completions/mean_length": 616.5045166015625, + "completions/mean_terminated_length": 561.2744140625, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 1.0466472303206997, + "grad_norm": 0.12566648423671722, + "learning_rate": 1e-06, + "loss": -0.0085, + "num_tokens": 66010729.0, + "reward": 0.6361607313156128, + "reward_std": 0.19186089932918549, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2760.0, + "completions/mean_length": 630.2623291015625, + "completions/mean_terminated_length": 567.2488403320312, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 1.0559766763848397, + "grad_norm": 0.11994180083274841, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 66600540.0, + "reward": 0.5424107313156128, + "reward_std": 0.17904217541217804, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2407.0, + "completions/mean_length": 659.8504638671875, + "completions/mean_terminated_length": 577.3828735351562, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 1.0653061224489795, + "grad_norm": 0.1303776204586029, + "learning_rate": 1e-06, + "loss": -0.0064, + "num_tokens": 67217198.0, + "reward": 0.5066964626312256, + "reward_std": 0.21001102030277252, + "rewards/verify_math_reward/mean": 0.5066964030265808, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 114 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3434.0, + "completions/mean_length": 661.591552734375, + "completions/mean_terminated_length": 563.014892578125, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 1.0746355685131195, + "grad_norm": 0.12177181243896484, + "learning_rate": 1e-06, + "loss": 0.0079, + "num_tokens": 67795496.0, + "reward": 0.5558035969734192, + "reward_std": 0.18798990547657013, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715372920036316, + "step": 115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3010.0, + "completions/mean_length": 641.4921875, + "completions/mean_terminated_length": 586.6587524414062, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 1.0839650145772595, + "grad_norm": 0.13219384849071503, + "learning_rate": 1e-06, + "loss": 0.0078, + "num_tokens": 68400769.0, + "reward": 0.6127232313156128, + "reward_std": 0.22323454916477203, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 116 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3392.0, + "completions/mean_length": 622.1629638671875, + "completions/mean_terminated_length": 563.0170288085938, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 1.0932944606413995, + "grad_norm": 0.13247352838516235, + "learning_rate": 1e-06, + "loss": 0.0075, + "num_tokens": 68981907.0, + "reward": 0.5680803656578064, + "reward_std": 0.22901484370231628, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 117 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2248.0, + "completions/mean_length": 650.8873291015625, + "completions/mean_terminated_length": 592.23046875, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "epoch": 1.1026239067055394, + "grad_norm": 0.11174468696117401, + "learning_rate": 1e-06, + "loss": 0.0116, + "num_tokens": 69578014.0, + "reward": 0.6261160969734192, + "reward_std": 0.17223836481571198, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2884.0, + "completions/mean_length": 620.458740234375, + "completions/mean_terminated_length": 537.0457153320312, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 1.1119533527696792, + "grad_norm": 0.12506945431232452, + "learning_rate": 1e-06, + "loss": 0.0044, + "num_tokens": 70121553.0, + "reward": 0.6595982313156128, + "reward_std": 0.16901816427707672, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 119 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3575.0, + "completions/mean_length": 668.9174194335938, + "completions/mean_terminated_length": 598.6583251953125, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 1.1212827988338192, + "grad_norm": 0.13134749233722687, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 70743815.0, + "reward": 0.5881696939468384, + "reward_std": 0.23278099298477173, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 120 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3602.0, + "completions/mean_length": 678.8147583007812, + "completions/mean_terminated_length": 596.8023071289062, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 1.1306122448979592, + "grad_norm": 0.13624440133571625, + "learning_rate": 1e-06, + "loss": -0.0165, + "num_tokens": 71363217.0, + "reward": 0.5345982313156128, + "reward_std": 0.22443917393684387, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 121 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2014.0, + "completions/mean_length": 645.9609375, + "completions/mean_terminated_length": 563.1599731445312, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 1.1399416909620992, + "grad_norm": 0.12669922411441803, + "learning_rate": 1e-06, + "loss": -0.0141, + "num_tokens": 71934550.0, + "reward": 0.6517857313156128, + "reward_std": 0.18670859932899475, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 122 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3769.0, + "completions/mean_length": 704.208740234375, + "completions/mean_terminated_length": 610.8566284179688, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.149271137026239, + "grad_norm": 0.12893548607826233, + "learning_rate": 1e-06, + "loss": 0.0176, + "num_tokens": 72557305.0, + "reward": 0.5770089626312256, + "reward_std": 0.23244328796863556, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 123 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2902.0, + "completions/mean_length": 642.6171875, + "completions/mean_terminated_length": 563.7728271484375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "epoch": 1.158600583090379, + "grad_norm": 0.1255681812763214, + "learning_rate": 1e-06, + "loss": 0.005, + "num_tokens": 73140682.0, + "reward": 0.6104910969734192, + "reward_std": 0.17577557265758514, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 124 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3902.0, + "completions/mean_length": 713.0971069335938, + "completions/mean_terminated_length": 623.9713745117188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 1.167930029154519, + "grad_norm": 0.10720223933458328, + "learning_rate": 1e-06, + "loss": 0.0077, + "num_tokens": 73775753.0, + "reward": 0.5066964626312256, + "reward_std": 0.16247648000717163, + "rewards/verify_math_reward/mean": 0.5066964030265808, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 125 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 633.5279541015625, + "completions/mean_terminated_length": 578.5680541992188, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.177259475218659, + "grad_norm": 0.12771426141262054, + "learning_rate": 1e-06, + "loss": -0.0002, + "num_tokens": 74374530.0, + "reward": 0.6205357313156128, + "reward_std": 0.1678474396467209, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 126 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3944.0, + "completions/mean_length": 665.1261596679688, + "completions/mean_terminated_length": 582.78515625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.186588921282799, + "grad_norm": 0.12777753174304962, + "learning_rate": 1e-06, + "loss": 0.005, + "num_tokens": 74967067.0, + "reward": 0.6082589626312256, + "reward_std": 0.17107835412025452, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 127 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1992.0, + "completions/mean_length": 646.2042846679688, + "completions/mean_terminated_length": 579.484619140625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 1.1959183673469387, + "grad_norm": 0.13895373046398163, + "learning_rate": 1e-06, + "loss": 0.0112, + "num_tokens": 75568354.0, + "reward": 0.5569196939468384, + "reward_std": 0.23165053129196167, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2431.0, + "completions/mean_length": 651.9832763671875, + "completions/mean_terminated_length": 589.36474609375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 1.2052478134110787, + "grad_norm": 0.13100416958332062, + "learning_rate": 1e-06, + "loss": 0.0138, + "num_tokens": 76180083.0, + "reward": 0.5502232313156128, + "reward_std": 0.2245786488056183, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 129 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4023.0, + "completions/mean_length": 755.8381958007812, + "completions/mean_terminated_length": 652.0587158203125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 1.2145772594752187, + "grad_norm": 0.11624865978956223, + "learning_rate": 1e-06, + "loss": -0.0078, + "num_tokens": 76839866.0, + "reward": 0.5245535969734192, + "reward_std": 0.2003878951072693, + "rewards/verify_math_reward/mean": 0.5245535969734192, + "rewards/verify_math_reward/std": 0.4996756613254547, + "step": 130 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3677.0, + "completions/mean_length": 581.8192138671875, + "completions/mean_terminated_length": 534.1154174804688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.2239067055393587, + "grad_norm": 0.11987384408712387, + "learning_rate": 1e-06, + "loss": 0.0089, + "num_tokens": 77397176.0, + "reward": 0.625, + "reward_std": 0.174800306558609, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 131 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3192.0, + "completions/mean_length": 628.9085083007812, + "completions/mean_terminated_length": 545.6983032226562, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 1.2332361516034984, + "grad_norm": 0.1218874454498291, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 77960158.0, + "reward": 0.590401828289032, + "reward_std": 0.16709604859352112, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 132 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2199.0, + "completions/mean_length": 640.3795166015625, + "completions/mean_terminated_length": 569.5353393554688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.2425655976676384, + "grad_norm": 0.12592343986034393, + "learning_rate": 1e-06, + "loss": 0.0114, + "num_tokens": 78551082.0, + "reward": 0.53125, + "reward_std": 0.20174364745616913, + "rewards/verify_math_reward/mean": 0.53125, + "rewards/verify_math_reward/std": 0.4993011951446533, + "step": 133 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3893.0, + "completions/mean_length": 589.3203125, + "completions/mean_terminated_length": 565.6798095703125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.2518950437317784, + "grad_norm": 0.1288214921951294, + "learning_rate": 1e-06, + "loss": 0.0035, + "num_tokens": 79135801.0, + "reward": 0.629464328289032, + "reward_std": 0.21053165197372437, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 134 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2917.0, + "completions/mean_length": 628.3549194335938, + "completions/mean_terminated_length": 549.1849365234375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.2612244897959184, + "grad_norm": 0.14329035580158234, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 79700903.0, + "reward": 0.6584821939468384, + "reward_std": 0.22620469331741333, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 135 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3282.0, + "completions/mean_length": 614.4408569335938, + "completions/mean_terminated_length": 559.1780395507812, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 1.2705539358600584, + "grad_norm": 0.12524747848510742, + "learning_rate": 1e-06, + "loss": 0.0084, + "num_tokens": 80285698.0, + "reward": 0.6004464626312256, + "reward_std": 0.19212302565574646, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 136 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2938.0, + "completions/mean_length": 677.765625, + "completions/mean_terminated_length": 583.6857299804688, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 1.2798833819241984, + "grad_norm": 0.12303480505943298, + "learning_rate": 1e-06, + "loss": -0.0103, + "num_tokens": 80880896.0, + "reward": 0.5792410969734192, + "reward_std": 0.18772847950458527, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 137 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 667.2846069335938, + "completions/mean_terminated_length": 589.00341796875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 1.2892128279883381, + "grad_norm": 0.11458269506692886, + "learning_rate": 1e-06, + "loss": -0.0034, + "num_tokens": 81489175.0, + "reward": 0.5558035969734192, + "reward_std": 0.17348577082157135, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715372920036316, + "step": 138 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3072.0, + "completions/mean_length": 659.4542846679688, + "completions/mean_terminated_length": 600.9432983398438, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 1.2985422740524781, + "grad_norm": 0.14296361804008484, + "learning_rate": 1e-06, + "loss": 0.0144, + "num_tokens": 82114390.0, + "reward": 0.5066964626312256, + "reward_std": 0.22808054089546204, + "rewards/verify_math_reward/mean": 0.5066964030265808, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 139 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2636.0, + "completions/mean_length": 691.6451416015625, + "completions/mean_terminated_length": 617.8905029296875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 1.3078717201166181, + "grad_norm": 0.11705105751752853, + "learning_rate": 1e-06, + "loss": 0.0053, + "num_tokens": 82749336.0, + "reward": 0.4933035969734192, + "reward_std": 0.1844961941242218, + "rewards/verify_math_reward/mean": 0.4933035671710968, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 140 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 689.5513916015625, + "completions/mean_terminated_length": 607.7965698242188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.3172011661807579, + "grad_norm": 0.12601278722286224, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 83375086.0, + "reward": 0.5089285969734192, + "reward_std": 0.2058332860469818, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 141 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3122.0, + "completions/mean_length": 697.591552734375, + "completions/mean_terminated_length": 604.0573120117188, + "completions/min_length": 52.0, + "completions/min_terminated_length": 52.0, + "epoch": 1.3265306122448979, + "grad_norm": 0.12080052495002747, + "learning_rate": 1e-06, + "loss": 0.0115, + "num_tokens": 83995232.0, + "reward": 0.543526828289032, + "reward_std": 0.18517090380191803, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 142 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 699.6986694335938, + "completions/mean_terminated_length": 598.2000122070312, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 1.3358600583090379, + "grad_norm": 0.13864567875862122, + "learning_rate": 1e-06, + "loss": 0.0198, + "num_tokens": 84607234.0, + "reward": 0.5212053656578064, + "reward_std": 0.2281247079372406, + "rewards/verify_math_reward/mean": 0.5212053656578064, + "rewards/verify_math_reward/std": 0.49982914328575134, + "step": 143 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2969.0, + "completions/mean_length": 718.8314819335938, + "completions/mean_terminated_length": 633.8226318359375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.3451895043731779, + "grad_norm": 0.1279512494802475, + "learning_rate": 1e-06, + "loss": 0.0016, + "num_tokens": 85257411.0, + "reward": 0.5401785969734192, + "reward_std": 0.198696106672287, + "rewards/verify_math_reward/mean": 0.5401785969734192, + "rewards/verify_math_reward/std": 0.49866142868995667, + "step": 144 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 664.677490234375, + "completions/mean_terminated_length": 574.2760620117188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.3545189504373178, + "grad_norm": 0.14231976866722107, + "learning_rate": 1e-06, + "loss": -0.0036, + "num_tokens": 85845410.0, + "reward": 0.609375, + "reward_std": 0.21752910315990448, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 145 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3667.0, + "completions/mean_length": 714.4017944335938, + "completions/mean_terminated_length": 621.3302612304688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.3638483965014578, + "grad_norm": 0.12819808721542358, + "learning_rate": 1e-06, + "loss": -0.0091, + "num_tokens": 86490594.0, + "reward": 0.5636160969734192, + "reward_std": 0.21673564612865448, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 146 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 747.0558471679688, + "completions/mean_terminated_length": 606.867431640625, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 1.3731778425655976, + "grad_norm": 0.13006004691123962, + "learning_rate": 1e-06, + "loss": 0.0244, + "num_tokens": 87100108.0, + "reward": 0.5267857313156128, + "reward_std": 0.196895569562912, + "rewards/verify_math_reward/mean": 0.5267857313156128, + "rewards/verify_math_reward/std": 0.4995608627796173, + "step": 147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3927.0, + "completions/mean_length": 730.810302734375, + "completions/mean_terminated_length": 630.2413940429688, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 1.3825072886297376, + "grad_norm": 0.133043572306633, + "learning_rate": 1e-06, + "loss": 0.0112, + "num_tokens": 87731866.0, + "reward": 0.5915178656578064, + "reward_std": 0.2325078845024109, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 148 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3835.0, + "completions/mean_length": 629.466552734375, + "completions/mean_terminated_length": 558.398681640625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 1.3918367346938776, + "grad_norm": 0.10918683558702469, + "learning_rate": 1e-06, + "loss": -0.0012, + "num_tokens": 88306764.0, + "reward": 0.6015625, + "reward_std": 0.15059703588485718, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 149 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1983.0, + "completions/mean_length": 653.9252319335938, + "completions/mean_terminated_length": 579.3534545898438, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 1.4011661807580174, + "grad_norm": 0.12218791991472244, + "learning_rate": 1e-06, + "loss": -0.0124, + "num_tokens": 88911113.0, + "reward": 0.6261160969734192, + "reward_std": 0.1806679666042328, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 150 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3154.0, + "completions/mean_length": 712.6027221679688, + "completions/mean_terminated_length": 599.4324951171875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.4104956268221573, + "grad_norm": 0.12704800069332123, + "learning_rate": 1e-06, + "loss": 0.0132, + "num_tokens": 89528021.0, + "reward": 0.5446428656578064, + "reward_std": 0.19745828211307526, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.49828118085861206, + "step": 151 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3326.0, + "completions/mean_length": 631.6897583007812, + "completions/mean_terminated_length": 568.7022705078125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.4198250728862973, + "grad_norm": 0.12975318729877472, + "learning_rate": 1e-06, + "loss": 0.0073, + "num_tokens": 90112959.0, + "reward": 0.6417410969734192, + "reward_std": 0.20463047921657562, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2905.0, + "completions/mean_length": 638.380615234375, + "completions/mean_terminated_length": 567.4954833984375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 1.4291545189504373, + "grad_norm": 0.12570594251155853, + "learning_rate": 1e-06, + "loss": 0.0101, + "num_tokens": 90700444.0, + "reward": 0.5803571939468384, + "reward_std": 0.18257547914981842, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3382.0, + "completions/mean_length": 630.8504638671875, + "completions/mean_terminated_length": 575.8480834960938, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 1.4384839650145773, + "grad_norm": 0.11297712475061417, + "learning_rate": 1e-06, + "loss": 0.0059, + "num_tokens": 91290870.0, + "reward": 0.6227678656578064, + "reward_std": 0.18347588181495667, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644601345062, + "step": 154 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3250.0, + "completions/mean_length": 595.1138916015625, + "completions/mean_terminated_length": 539.5442504882812, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 1.4478134110787173, + "grad_norm": 0.12884287536144257, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 91851076.0, + "reward": 0.629464328289032, + "reward_std": 0.2043701708316803, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3097.0, + "completions/mean_length": 618.2433471679688, + "completions/mean_terminated_length": 559.0307006835938, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 1.457142857142857, + "grad_norm": 0.1330742985010147, + "learning_rate": 1e-06, + "loss": -0.0053, + "num_tokens": 92432886.0, + "reward": 0.5993303656578064, + "reward_std": 0.1982824206352234, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 156 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2748.0, + "completions/mean_length": 706.7098388671875, + "completions/mean_terminated_length": 625.3668212890625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.466472303206997, + "grad_norm": 0.12443883717060089, + "learning_rate": 1e-06, + "loss": -0.0002, + "num_tokens": 93071586.0, + "reward": 0.559151828289032, + "reward_std": 0.18761970102787018, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3262.0, + "completions/mean_length": 676.2824096679688, + "completions/mean_terminated_length": 598.2066040039062, + "completions/min_length": 202.0, + "completions/min_terminated_length": 202.0, + "epoch": 1.475801749271137, + "grad_norm": 0.12442443519830704, + "learning_rate": 1e-06, + "loss": -0.0001, + "num_tokens": 93686471.0, + "reward": 0.5580357313156128, + "reward_std": 0.18829302489757538, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 158 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2785.0, + "completions/mean_length": 687.3047485351562, + "completions/mean_terminated_length": 601.5022583007812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 1.485131195335277, + "grad_norm": 0.12545685470104218, + "learning_rate": 1e-06, + "loss": 0.0019, + "num_tokens": 94290728.0, + "reward": 0.606026828289032, + "reward_std": 0.18678458034992218, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 159 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 688.3013916015625, + "completions/mean_terminated_length": 606.5165405273438, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 1.4944606413994168, + "grad_norm": 0.12918633222579956, + "learning_rate": 1e-06, + "loss": -0.0027, + "num_tokens": 94911534.0, + "reward": 0.5926339626312256, + "reward_std": 0.211582213640213, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 160 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4007.0, + "completions/mean_length": 674.7667846679688, + "completions/mean_terminated_length": 608.5995483398438, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 1.5037900874635568, + "grad_norm": 0.1253700703382492, + "learning_rate": 1e-06, + "loss": 0.0148, + "num_tokens": 95528885.0, + "reward": 0.6004464626312256, + "reward_std": 0.20711347460746765, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 161 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4010.0, + "completions/mean_length": 735.2533569335938, + "completions/mean_terminated_length": 622.8408203125, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 1.5131195335276968, + "grad_norm": 0.12896056473255157, + "learning_rate": 1e-06, + "loss": 0.0068, + "num_tokens": 96164328.0, + "reward": 0.546875, + "reward_std": 0.19381371140480042, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 162 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2627.0, + "completions/mean_length": 666.7678833007812, + "completions/mean_terminated_length": 592.4743041992188, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 1.5224489795918368, + "grad_norm": 0.13119478523731232, + "learning_rate": 1e-06, + "loss": 0.0047, + "num_tokens": 96785544.0, + "reward": 0.5256696939468384, + "reward_std": 0.2169952541589737, + "rewards/verify_math_reward/mean": 0.5256696343421936, + "rewards/verify_math_reward/std": 0.4996195137500763, + "step": 163 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3694.0, + "completions/mean_length": 639.7980346679688, + "completions/mean_terminated_length": 552.7997436523438, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.5317784256559768, + "grad_norm": 0.12215036153793335, + "learning_rate": 1e-06, + "loss": -0.0031, + "num_tokens": 97362179.0, + "reward": 0.5658482313156128, + "reward_std": 0.17577417194843292, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 164 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3940.0, + "completions/mean_length": 729.3917846679688, + "completions/mean_terminated_length": 624.7905883789062, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 1.5411078717201168, + "grad_norm": 0.12878678739070892, + "learning_rate": 1e-06, + "loss": 0.0081, + "num_tokens": 97997010.0, + "reward": 0.551339328289032, + "reward_std": 0.21469798684120178, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2869.0, + "completions/mean_length": 629.8772583007812, + "completions/mean_terminated_length": 546.6903076171875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.5504373177842565, + "grad_norm": 0.12062986195087433, + "learning_rate": 1e-06, + "loss": 0.0082, + "num_tokens": 98568924.0, + "reward": 0.5691964626312256, + "reward_std": 0.1649598926305771, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 166 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 670.068115234375, + "completions/mean_terminated_length": 603.8099975585938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.5597667638483965, + "grad_norm": 0.1336459517478943, + "learning_rate": 1e-06, + "loss": 0.0079, + "num_tokens": 99182577.0, + "reward": 0.5870535969734192, + "reward_std": 0.24498917162418365, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 167 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2902.0, + "completions/mean_length": 708.825927734375, + "completions/mean_terminated_length": 615.6008911132812, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.5690962099125363, + "grad_norm": 0.10686108469963074, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 99803725.0, + "reward": 0.5714285969734192, + "reward_std": 0.17739246785640717, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514806270599365, + "step": 168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3401.0, + "completions/mean_length": 664.8270263671875, + "completions/mean_terminated_length": 582.4788208007812, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 1.5784256559766763, + "grad_norm": 0.13156671822071075, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 100401194.0, + "reward": 0.5792410969734192, + "reward_std": 0.22101962566375732, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 169 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3616.0, + "completions/mean_length": 720.364990234375, + "completions/mean_terminated_length": 607.4544067382812, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 1.5877551020408163, + "grad_norm": 0.12857548892498016, + "learning_rate": 1e-06, + "loss": 0.0009, + "num_tokens": 101017025.0, + "reward": 0.590401828289032, + "reward_std": 0.20026984810829163, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 170 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2680.0, + "completions/mean_length": 676.6194458007812, + "completions/mean_terminated_length": 594.5542602539062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.5970845481049563, + "grad_norm": 0.1252916157245636, + "learning_rate": 1e-06, + "loss": 0.0008, + "num_tokens": 101628348.0, + "reward": 0.5703125, + "reward_std": 0.18362995982170105, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 171 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2716.0, + "completions/mean_length": 682.0424194335938, + "completions/mean_terminated_length": 627.8526000976562, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 1.6064139941690962, + "grad_norm": 0.12653912603855133, + "learning_rate": 1e-06, + "loss": 0.0212, + "num_tokens": 102273866.0, + "reward": 0.6183035969734192, + "reward_std": 0.21647420525550842, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 172 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3430.0, + "completions/mean_length": 669.6239013671875, + "completions/mean_terminated_length": 571.27783203125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.6157434402332362, + "grad_norm": 0.12435685098171234, + "learning_rate": 1e-06, + "loss": -0.0051, + "num_tokens": 102851897.0, + "reward": 0.6171875, + "reward_std": 0.176642507314682, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 173 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3958.0, + "completions/mean_length": 651.1842041015625, + "completions/mean_terminated_length": 564.4725341796875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.6250728862973762, + "grad_norm": 0.13036301732063293, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 103435558.0, + "reward": 0.5848214626312256, + "reward_std": 0.1837814897298813, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 174 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3609.0, + "completions/mean_length": 665.9453125, + "completions/mean_terminated_length": 611.5, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 1.634402332361516, + "grad_norm": 0.12362460047006607, + "learning_rate": 1e-06, + "loss": 0.0183, + "num_tokens": 104083269.0, + "reward": 0.5792410969734192, + "reward_std": 0.20080046355724335, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 175 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 699.818115234375, + "completions/mean_terminated_length": 610.342529296875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 1.643731778425656, + "grad_norm": 0.126058429479599, + "learning_rate": 1e-06, + "loss": 0.0012, + "num_tokens": 104706402.0, + "reward": 0.574776828289032, + "reward_std": 0.2046714425086975, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 176 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2311.0, + "completions/mean_length": 702.239990234375, + "completions/mean_terminated_length": 588.72314453125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 1.6530612244897958, + "grad_norm": 0.1367228776216507, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 105307441.0, + "reward": 0.5457589626312256, + "reward_std": 0.21527138352394104, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 177 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3785.0, + "completions/mean_length": 677.0580444335938, + "completions/mean_terminated_length": 574.8827514648438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 1.6623906705539357, + "grad_norm": 0.13413365185260773, + "learning_rate": 1e-06, + "loss": 0.0233, + "num_tokens": 105891045.0, + "reward": 0.625, + "reward_std": 0.19622154533863068, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 178 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3353.0, + "completions/mean_length": 635.6685791015625, + "completions/mean_terminated_length": 576.7525634765625, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 1.6717201166180757, + "grad_norm": 0.12873566150665283, + "learning_rate": 1e-06, + "loss": 0.0031, + "num_tokens": 106481348.0, + "reward": 0.6026785969734192, + "reward_std": 0.182315856218338, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 179 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3820.0, + "completions/mean_length": 642.4163208007812, + "completions/mean_terminated_length": 563.5673217773438, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.6810495626822157, + "grad_norm": 0.12520509958267212, + "learning_rate": 1e-06, + "loss": -0.0019, + "num_tokens": 107061937.0, + "reward": 0.6082589626312256, + "reward_std": 0.17938552796840668, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 180 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3912.0, + "completions/mean_length": 707.4498291015625, + "completions/mean_terminated_length": 590.0634765625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.6903790087463557, + "grad_norm": 0.12314064055681229, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 107655780.0, + "reward": 0.6205357313156128, + "reward_std": 0.17964836955070496, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 181 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3921.0, + "completions/mean_length": 773.8717041015625, + "completions/mean_terminated_length": 654.8126831054688, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 1.6997084548104957, + "grad_norm": 0.11460225284099579, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 108326697.0, + "reward": 0.5033482313156128, + "reward_std": 0.20655006170272827, + "rewards/verify_math_reward/mean": 0.5033482313156128, + "rewards/verify_math_reward/std": 0.5002680420875549, + "step": 182 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3508.0, + "completions/mean_length": 568.2210083007812, + "completions/mean_terminated_length": 536.439208984375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 1.7090379008746357, + "grad_norm": 0.15026937425136566, + "learning_rate": 1e-06, + "loss": 0.0064, + "num_tokens": 108898831.0, + "reward": 0.6439732313156128, + "reward_std": 0.2181578129529953, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 183 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3214.0, + "completions/mean_length": 674.3426513671875, + "completions/mean_terminated_length": 604.1947631835938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 1.7183673469387755, + "grad_norm": 0.11818968504667282, + "learning_rate": 1e-06, + "loss": 0.0194, + "num_tokens": 109522162.0, + "reward": 0.566964328289032, + "reward_std": 0.18693752586841583, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 184 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3660.0, + "completions/mean_length": 703.763427734375, + "completions/mean_terminated_length": 606.397216796875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 1.7276967930029155, + "grad_norm": 0.12568622827529907, + "learning_rate": 1e-06, + "loss": 0.0105, + "num_tokens": 110136118.0, + "reward": 0.559151828289032, + "reward_std": 0.19580180943012238, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 185 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3221.0, + "completions/mean_length": 674.255615234375, + "completions/mean_terminated_length": 600.124267578125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.7370262390670554, + "grad_norm": 0.1335996836423874, + "learning_rate": 1e-06, + "loss": 0.0032, + "num_tokens": 110754275.0, + "reward": 0.5290178656578064, + "reward_std": 0.2127886563539505, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943602085113525, + "step": 186 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3063.0, + "completions/mean_length": 618.6942138671875, + "completions/mean_terminated_length": 555.470458984375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 1.7463556851311952, + "grad_norm": 0.12500326335430145, + "learning_rate": 1e-06, + "loss": 0.019, + "num_tokens": 111330833.0, + "reward": 0.6383928656578064, + "reward_std": 0.18979327380657196, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 187 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4048.0, + "completions/mean_length": 766.2076416015625, + "completions/mean_terminated_length": 630.8501586914062, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 1.7556851311953352, + "grad_norm": 0.130299910902977, + "learning_rate": 1e-06, + "loss": -0.0104, + "num_tokens": 111965011.0, + "reward": 0.5703125, + "reward_std": 0.19926205277442932, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 188 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3126.0, + "completions/mean_length": 680.6015625, + "completions/mean_terminated_length": 594.6304321289062, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.7650145772594752, + "grad_norm": 0.13165517151355743, + "learning_rate": 1e-06, + "loss": 0.0071, + "num_tokens": 112581406.0, + "reward": 0.6004464626312256, + "reward_std": 0.20080116391181946, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 189 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3959.0, + "completions/mean_length": 742.8147583007812, + "completions/mean_terminated_length": 638.630615234375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 1.7743440233236152, + "grad_norm": 0.1261264979839325, + "learning_rate": 1e-06, + "loss": -0.0072, + "num_tokens": 113221160.0, + "reward": 0.590401828289032, + "reward_std": 0.20489898324012756, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 190 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3226.0, + "completions/mean_length": 634.4866333007812, + "completions/mean_terminated_length": 567.5403442382812, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 1.7836734693877552, + "grad_norm": 0.1222400888800621, + "learning_rate": 1e-06, + "loss": -0.0081, + "num_tokens": 113810380.0, + "reward": 0.6361607313156128, + "reward_std": 0.15537773072719574, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 191 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3573.0, + "completions/mean_length": 660.5859375, + "completions/mean_terminated_length": 590.1560668945312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.7930029154518952, + "grad_norm": 0.12068247050046921, + "learning_rate": 1e-06, + "loss": 0.0157, + "num_tokens": 114412185.0, + "reward": 0.6116071939468384, + "reward_std": 0.20117954909801483, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 192 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3504.0, + "completions/mean_length": 751.6295166015625, + "completions/mean_terminated_length": 639.7647094726562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.802332361516035, + "grad_norm": 0.11500409245491028, + "learning_rate": 1e-06, + "loss": -0.0104, + "num_tokens": 115059357.0, + "reward": 0.5613839626312256, + "reward_std": 0.16717413067817688, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 193 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3045.0, + "completions/mean_length": 645.4129638671875, + "completions/mean_terminated_length": 570.6567993164062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 1.811661807580175, + "grad_norm": 0.12167953699827194, + "learning_rate": 1e-06, + "loss": -0.0072, + "num_tokens": 115645279.0, + "reward": 0.6049107313156128, + "reward_std": 0.17164427042007446, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 194 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3930.0, + "completions/mean_length": 658.0100708007812, + "completions/mean_terminated_length": 583.5267944335938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 1.820991253644315, + "grad_norm": 0.13267289102077484, + "learning_rate": 1e-06, + "loss": 0.0144, + "num_tokens": 116241416.0, + "reward": 0.6026785969734192, + "reward_std": 0.18809755146503448, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 695.2678833007812, + "completions/mean_terminated_length": 621.591796875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 1.8303206997084547, + "grad_norm": 0.12682850658893585, + "learning_rate": 1e-06, + "loss": 0.0023, + "num_tokens": 116869424.0, + "reward": 0.640625, + "reward_std": 0.17852042615413666, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3840.0, + "completions/mean_length": 757.6295166015625, + "completions/mean_terminated_length": 669.677001953125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 1.8396501457725947, + "grad_norm": 0.13293851912021637, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 117554860.0, + "reward": 0.4542410969734192, + "reward_std": 0.2379986047744751, + "rewards/verify_math_reward/mean": 0.4542410671710968, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 197 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3875.0, + "completions/mean_length": 712.5145263671875, + "completions/mean_terminated_length": 599.3413696289062, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.8489795918367347, + "grad_norm": 0.13029628992080688, + "learning_rate": 1e-06, + "loss": 0.0055, + "num_tokens": 118170025.0, + "reward": 0.5245535969734192, + "reward_std": 0.19133026897907257, + "rewards/verify_math_reward/mean": 0.5245535969734192, + "rewards/verify_math_reward/std": 0.4996756911277771, + "step": 198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3970.0, + "completions/mean_length": 711.6428833007812, + "completions/mean_terminated_length": 582.2294311523438, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 1.8583090379008746, + "grad_norm": 0.12420004606246948, + "learning_rate": 1e-06, + "loss": -0.003, + "num_tokens": 118753057.0, + "reward": 0.5736607313156128, + "reward_std": 0.18742355704307556, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 199 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4067.0, + "completions/mean_length": 716.904052734375, + "completions/mean_terminated_length": 607.9009399414062, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 1.8676384839650146, + "grad_norm": 0.12810364365577698, + "learning_rate": 1e-06, + "loss": -0.009, + "num_tokens": 119372107.0, + "reward": 0.6283482313156128, + "reward_std": 0.1928398311138153, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 200 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 668.75, + "completions/mean_terminated_length": 582.4805297851562, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 1.8769679300291546, + "grad_norm": 0.14725738763809204, + "learning_rate": 1e-06, + "loss": -0.0109, + "num_tokens": 119961811.0, + "reward": 0.5892857313156128, + "reward_std": 0.23773685097694397, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 201 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4029.0, + "completions/mean_length": 741.6160888671875, + "completions/mean_terminated_length": 661.11083984375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.8862973760932946, + "grad_norm": 0.12467392534017563, + "learning_rate": 1e-06, + "loss": 0.0189, + "num_tokens": 120620363.0, + "reward": 0.5658482313156128, + "reward_std": 0.21312315762043, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 202 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 698.6283569335938, + "completions/mean_terminated_length": 609.1214599609375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.8956268221574344, + "grad_norm": 0.12918609380722046, + "learning_rate": 1e-06, + "loss": 0.0125, + "num_tokens": 121243918.0, + "reward": 0.546875, + "reward_std": 0.2172248810529709, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4060.0, + "completions/mean_length": 678.474365234375, + "completions/mean_terminated_length": 620.2872314453125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.9049562682215744, + "grad_norm": 0.12341257184743881, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 121887343.0, + "reward": 0.5970982313156128, + "reward_std": 0.1837066262960434, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 204 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3210.0, + "completions/mean_length": 719.779052734375, + "completions/mean_terminated_length": 622.87255859375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.9142857142857141, + "grad_norm": 0.12145639955997467, + "learning_rate": 1e-06, + "loss": -0.0055, + "num_tokens": 122515681.0, + "reward": 0.5424107313156128, + "reward_std": 0.18502570688724518, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 205 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3777.0, + "completions/mean_length": 753.404052734375, + "completions/mean_terminated_length": 673.1817016601562, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.9236151603498541, + "grad_norm": 0.12220670282840729, + "learning_rate": 1e-06, + "loss": -0.0007, + "num_tokens": 123184339.0, + "reward": 0.5658482313156128, + "reward_std": 0.20143873989582062, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 206 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4022.0, + "completions/mean_length": 680.841552734375, + "completions/mean_terminated_length": 614.7918090820312, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 1.9329446064139941, + "grad_norm": 0.14044316112995148, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 123818189.0, + "reward": 0.6194196939468384, + "reward_std": 0.18122069537639618, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 207 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3026.0, + "completions/mean_length": 759.3694458007812, + "completions/mean_terminated_length": 643.78173828125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.9422740524781341, + "grad_norm": 0.13641150295734406, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 124466696.0, + "reward": 0.5379464626312256, + "reward_std": 0.21744990348815918, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4062.0, + "completions/mean_length": 624.958740234375, + "completions/mean_terminated_length": 533.5109252929688, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 1.951603498542274, + "grad_norm": 0.14100514352321625, + "learning_rate": 1e-06, + "loss": -0.0111, + "num_tokens": 125017675.0, + "reward": 0.6540178656578064, + "reward_std": 0.20139819383621216, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 659.8125, + "completions/mean_terminated_length": 561.184814453125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.960932944606414, + "grad_norm": 0.13975530862808228, + "learning_rate": 1e-06, + "loss": -0.0073, + "num_tokens": 125589091.0, + "reward": 0.6618303656578064, + "reward_std": 0.19460716843605042, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 210 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3929.0, + "completions/mean_length": 671.4553833007812, + "completions/mean_terminated_length": 597.2633666992188, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 1.970262390670554, + "grad_norm": 0.13062146306037903, + "learning_rate": 1e-06, + "loss": -0.0045, + "num_tokens": 126202811.0, + "reward": 0.606026828289032, + "reward_std": 0.21560658514499664, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2173.0, + "completions/mean_length": 647.5435791015625, + "completions/mean_terminated_length": 568.8116455078125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.9795918367346939, + "grad_norm": 0.1245775818824768, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 126785626.0, + "reward": 0.6037946939468384, + "reward_std": 0.178518608212471, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 212 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3920.0, + "completions/mean_length": 589.3080444335938, + "completions/mean_terminated_length": 565.6674194335938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.9889212827988338, + "grad_norm": 0.11517234146595001, + "learning_rate": 1e-06, + "loss": 0.0182, + "num_tokens": 127381822.0, + "reward": 0.613839328289032, + "reward_std": 0.17085151374340057, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.02840909090909094, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4095.0, + "completions/mean_length": 719.5767211914062, + "completions/mean_terminated_length": 620.8508911132812, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 1.9982507288629736, + "grad_norm": 0.11962586641311646, + "learning_rate": 1e-06, + "loss": 0.0143, + "num_tokens": 127996977.0, + "reward": 0.6272321939468384, + "reward_std": 0.17303043603897095, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 214 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 635.7020263671875, + "completions/mean_terminated_length": 592.6926879882812, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 2.00932944606414, + "grad_norm": 0.1271495372056961, + "learning_rate": 1e-06, + "loss": -0.0045, + "num_tokens": 128609998.0, + "reward": 0.6037946939468384, + "reward_std": 0.1884024739265442, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 742.349365234375, + "completions/mean_terminated_length": 606.0220336914062, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 2.01865889212828, + "grad_norm": 0.13169968128204346, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 129221463.0, + "reward": 0.5959821939468384, + "reward_std": 0.18701167404651642, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 216 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3781.0, + "completions/mean_length": 677.7533569335938, + "completions/mean_terminated_length": 587.6964721679688, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 2.02798833819242, + "grad_norm": 0.13871188461780548, + "learning_rate": 1e-06, + "loss": 0.0143, + "num_tokens": 129828218.0, + "reward": 0.5680803656578064, + "reward_std": 0.18175272643566132, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 217 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3714.0, + "completions/mean_length": 753.8281860351562, + "completions/mean_terminated_length": 661.8417358398438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.03731778425656, + "grad_norm": 0.11747800558805466, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 130496656.0, + "reward": 0.5569196939468384, + "reward_std": 0.18555065989494324, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.4970270097255707, + "step": 218 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4003.0, + "completions/mean_length": 625.396240234375, + "completions/mean_terminated_length": 538.0354614257812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.0466472303206995, + "grad_norm": 0.15057526528835297, + "learning_rate": 1e-06, + "loss": -0.0103, + "num_tokens": 131054243.0, + "reward": 0.6674107313156128, + "reward_std": 0.21027062833309174, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 727.2299194335938, + "completions/mean_terminated_length": 642.4324951171875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 2.0559766763848395, + "grad_norm": 0.12068066000938416, + "learning_rate": 1e-06, + "loss": 0.0101, + "num_tokens": 131710265.0, + "reward": 0.5680803656578064, + "reward_std": 0.1876097172498703, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 220 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3359.0, + "completions/mean_length": 681.5301513671875, + "completions/mean_terminated_length": 607.556396484375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 2.0653061224489795, + "grad_norm": 0.11410924047231674, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 132331004.0, + "reward": 0.6305803656578064, + "reward_std": 0.15172751247882843, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 221 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 734.872802734375, + "completions/mean_terminated_length": 622.447509765625, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 2.0746355685131195, + "grad_norm": 0.12705783545970917, + "learning_rate": 1e-06, + "loss": -0.0088, + "num_tokens": 132962314.0, + "reward": 0.6026785969734192, + "reward_std": 0.19133026897907257, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 222 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3392.0, + "completions/mean_length": 668.7511596679688, + "completions/mean_terminated_length": 574.4231567382812, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 2.0839650145772595, + "grad_norm": 0.13876819610595703, + "learning_rate": 1e-06, + "loss": 0.0223, + "num_tokens": 133564331.0, + "reward": 0.6495535969734192, + "reward_std": 0.2012985646724701, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 728.5189819335938, + "completions/mean_terminated_length": 603.7974853515625, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 2.0932944606413995, + "grad_norm": 0.11487426608800888, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 134171716.0, + "reward": 0.652901828289032, + "reward_std": 0.15736696124076843, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 224 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3972.0, + "completions/mean_length": 748.2667846679688, + "completions/mean_terminated_length": 652.1779174804688, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 2.1026239067055394, + "grad_norm": 0.12201671302318573, + "learning_rate": 1e-06, + "loss": 0.0081, + "num_tokens": 134838171.0, + "reward": 0.5379464626312256, + "reward_std": 0.19918353855609894, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 225 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 764.7120971679688, + "completions/mean_terminated_length": 653.2848510742188, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 2.1119533527696794, + "grad_norm": 0.12113168090581894, + "learning_rate": 1e-06, + "loss": 0.0116, + "num_tokens": 135496417.0, + "reward": 0.6004464626312256, + "reward_std": 0.19569344818592072, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 226 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2793.0, + "completions/mean_length": 681.3046875, + "completions/mean_terminated_length": 575.2094116210938, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 2.1212827988338194, + "grad_norm": 0.12550842761993408, + "learning_rate": 1e-06, + "loss": -0.0108, + "num_tokens": 136079026.0, + "reward": 0.6495535969734192, + "reward_std": 0.18287856876850128, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 227 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3315.0, + "completions/mean_length": 721.482177734375, + "completions/mean_terminated_length": 600.545654296875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 2.130612244897959, + "grad_norm": 0.1321675032377243, + "learning_rate": 1e-06, + "loss": 0.0066, + "num_tokens": 136686906.0, + "reward": 0.6261160969734192, + "reward_std": 0.18547257781028748, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 228 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3592.0, + "completions/mean_length": 777.9989013671875, + "completions/mean_terminated_length": 674.907958984375, + "completions/min_length": 205.0, + "completions/min_terminated_length": 205.0, + "epoch": 2.139941690962099, + "grad_norm": 0.11579850316047668, + "learning_rate": 1e-06, + "loss": 0.0031, + "num_tokens": 137368337.0, + "reward": 0.5680803656578064, + "reward_std": 0.1889663189649582, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 229 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3466.0, + "completions/mean_length": 707.9989013671875, + "completions/mean_terminated_length": 586.5791625976562, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 2.149271137026239, + "grad_norm": 0.150935098528862, + "learning_rate": 1e-06, + "loss": 0.0037, + "num_tokens": 137961320.0, + "reward": 0.629464328289032, + "reward_std": 0.19437752664089203, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 230 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3770.0, + "completions/mean_length": 685.2366333007812, + "completions/mean_terminated_length": 591.3623657226562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 2.158600583090379, + "grad_norm": 0.1222238540649414, + "learning_rate": 1e-06, + "loss": -0.0043, + "num_tokens": 138566724.0, + "reward": 0.5837053656578064, + "reward_std": 0.16589422523975372, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 231 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 802.4810791015625, + "completions/mean_terminated_length": 632.3931884765625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 2.167930029154519, + "grad_norm": 0.12585654854774475, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 139195899.0, + "reward": 0.5625, + "reward_std": 0.17926861345767975, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 232 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3567.0, + "completions/mean_length": 746.1797485351562, + "completions/mean_terminated_length": 665.7839965820312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.177259475218659, + "grad_norm": 0.11441010981798172, + "learning_rate": 1e-06, + "loss": -0.0108, + "num_tokens": 139869724.0, + "reward": 0.5580357313156128, + "reward_std": 0.17773906886577606, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 233 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3891.0, + "completions/mean_length": 657.271240234375, + "completions/mean_terminated_length": 582.7719116210938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.186588921282799, + "grad_norm": 0.13363678753376007, + "learning_rate": 1e-06, + "loss": 0.0159, + "num_tokens": 140488079.0, + "reward": 0.5558035969734192, + "reward_std": 0.20894725620746613, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715372920036316, + "step": 234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2298.0, + "completions/mean_length": 610.5346069335938, + "completions/mean_terminated_length": 551.1907348632812, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.195918367346939, + "grad_norm": 0.11896467953920364, + "learning_rate": 1e-06, + "loss": -0.0025, + "num_tokens": 141060974.0, + "reward": 0.6305803656578064, + "reward_std": 0.16822539269924164, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2972.0, + "completions/mean_length": 651.185302734375, + "completions/mean_terminated_length": 572.5364990234375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 2.205247813411079, + "grad_norm": 0.13944171369075775, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 141651948.0, + "reward": 0.6194196939468384, + "reward_std": 0.18187259137630463, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3066.0, + "completions/mean_length": 720.4542846679688, + "completions/mean_terminated_length": 607.5466918945312, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 2.2145772594752184, + "grad_norm": 0.1344020813703537, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 142262099.0, + "reward": 0.59375, + "reward_std": 0.2035774141550064, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 237 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2830.0, + "completions/mean_length": 689.2645263671875, + "completions/mean_terminated_length": 587.4540405273438, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 2.2239067055393584, + "grad_norm": 0.1321393847465515, + "learning_rate": 1e-06, + "loss": -0.0051, + "num_tokens": 142854920.0, + "reward": 0.5491071939468384, + "reward_std": 0.19820503890514374, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 238 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3189.0, + "completions/mean_length": 670.794677734375, + "completions/mean_terminated_length": 580.554443359375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.2332361516034984, + "grad_norm": 0.13054966926574707, + "learning_rate": 1e-06, + "loss": 0.0046, + "num_tokens": 143452936.0, + "reward": 0.6361607313156128, + "reward_std": 0.18362672626972198, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 239 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3671.0, + "completions/mean_length": 719.732177734375, + "completions/mean_terminated_length": 582.4854736328125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 2.2425655976676384, + "grad_norm": 0.1302865445613861, + "learning_rate": 1e-06, + "loss": -0.0003, + "num_tokens": 144029600.0, + "reward": 0.684151828289032, + "reward_std": 0.17281359434127808, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 240 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3573.0, + "completions/mean_length": 699.5892944335938, + "completions/mean_terminated_length": 598.0873413085938, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 2.2518950437317784, + "grad_norm": 0.12489066272974014, + "learning_rate": 1e-06, + "loss": -0.0171, + "num_tokens": 144638416.0, + "reward": 0.6037946939468384, + "reward_std": 0.17345185577869415, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 241 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2900.0, + "completions/mean_length": 677.6328125, + "completions/mean_terminated_length": 587.57275390625, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 2.2612244897959184, + "grad_norm": 0.1454547494649887, + "learning_rate": 1e-06, + "loss": -0.0081, + "num_tokens": 145231335.0, + "reward": 0.6573660969734192, + "reward_std": 0.19444352388381958, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 242 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3579.0, + "completions/mean_length": 681.349365234375, + "completions/mean_terminated_length": 575.2554931640625, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 2.2705539358600584, + "grad_norm": 0.12814916670322418, + "learning_rate": 1e-06, + "loss": -0.0106, + "num_tokens": 145816888.0, + "reward": 0.5848214626312256, + "reward_std": 0.18712298572063446, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 243 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4016.0, + "completions/mean_length": 666.9386596679688, + "completions/mean_terminated_length": 552.2410278320312, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 2.2798833819241984, + "grad_norm": 0.11985107511281967, + "learning_rate": 1e-06, + "loss": 0.0085, + "num_tokens": 146372745.0, + "reward": 0.629464328289032, + "reward_std": 0.12354449182748795, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 244 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2443.0, + "completions/mean_length": 716.0960083007812, + "completions/mean_terminated_length": 586.8528442382812, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.2892128279883384, + "grad_norm": 0.1258755773305893, + "learning_rate": 1e-06, + "loss": -0.0053, + "num_tokens": 146962455.0, + "reward": 0.5959821939468384, + "reward_std": 0.1714908927679062, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 245 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2302.0, + "completions/mean_length": 727.6707763671875, + "completions/mean_terminated_length": 602.9178466796875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.298542274052478, + "grad_norm": 0.13353858888149261, + "learning_rate": 1e-06, + "loss": -0.0051, + "num_tokens": 147576368.0, + "reward": 0.551339328289032, + "reward_std": 0.2144358605146408, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 246 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3315.0, + "completions/mean_length": 699.911865234375, + "completions/mean_terminated_length": 614.4267578125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 2.307871720116618, + "grad_norm": 0.11353510618209839, + "learning_rate": 1e-06, + "loss": 0.0191, + "num_tokens": 148209065.0, + "reward": 0.5770089626312256, + "reward_std": 0.1661524474620819, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 247 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4091.0, + "completions/mean_length": 701.9542846679688, + "completions/mean_terminated_length": 563.98486328125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 2.317201166180758, + "grad_norm": 0.1472627818584442, + "learning_rate": 1e-06, + "loss": 0.0039, + "num_tokens": 148785656.0, + "reward": 0.5959821939468384, + "reward_std": 0.21891415119171143, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 248 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3494.0, + "completions/mean_length": 752.6317138671875, + "completions/mean_terminated_length": 628.80322265625, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 2.326530612244898, + "grad_norm": 0.1291126012802124, + "learning_rate": 1e-06, + "loss": -0.0087, + "num_tokens": 149419598.0, + "reward": 0.5915178656578064, + "reward_std": 0.17051492631435394, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 249 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3729.0, + "completions/mean_length": 707.0078735351562, + "completions/mean_terminated_length": 589.6062622070312, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 2.335860058309038, + "grad_norm": 0.1351911574602127, + "learning_rate": 1e-06, + "loss": 0.0214, + "num_tokens": 150013261.0, + "reward": 0.6584821939468384, + "reward_std": 0.1923915296792984, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 250 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2664.0, + "completions/mean_length": 664.7020263671875, + "completions/mean_terminated_length": 586.3618774414062, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 2.345189504373178, + "grad_norm": 0.13551536202430725, + "learning_rate": 1e-06, + "loss": 0.0191, + "num_tokens": 150618754.0, + "reward": 0.546875, + "reward_std": 0.19723185896873474, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 251 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3959.0, + "completions/mean_length": 669.0379638671875, + "completions/mean_terminated_length": 554.4105834960938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.354518950437318, + "grad_norm": 0.14088645577430725, + "learning_rate": 1e-06, + "loss": -0.0077, + "num_tokens": 151184756.0, + "reward": 0.6104910969734192, + "reward_std": 0.20218941569328308, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 252 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2801.0, + "completions/mean_length": 703.1004638671875, + "completions/mean_terminated_length": 593.652099609375, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 2.363848396501458, + "grad_norm": 0.11449974030256271, + "learning_rate": 1e-06, + "loss": -0.0032, + "num_tokens": 151785182.0, + "reward": 0.6428571939468384, + "reward_std": 0.15492378175258636, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 253 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3804.0, + "completions/mean_length": 785.0580444335938, + "completions/mean_terminated_length": 662.4305419921875, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 2.373177842565598, + "grad_norm": 0.12557293474674225, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 152439978.0, + "reward": 0.5457589626312256, + "reward_std": 0.21376506984233856, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 254 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3737.0, + "completions/mean_length": 650.6015625, + "completions/mean_terminated_length": 575.9578247070312, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 2.3825072886297374, + "grad_norm": 0.13001284003257751, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 153030125.0, + "reward": 0.6651785969734192, + "reward_std": 0.18930330872535706, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 255 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 757.9408569335938, + "completions/mean_terminated_length": 622.2473754882812, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 2.3918367346938774, + "grad_norm": 0.13443338871002197, + "learning_rate": 1e-06, + "loss": -0.0028, + "num_tokens": 153659576.0, + "reward": 0.5892857313156128, + "reward_std": 0.1672501266002655, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 256 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3411.0, + "completions/mean_length": 726.8426513671875, + "completions/mean_terminated_length": 614.1487426757812, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 2.4011661807580174, + "grad_norm": 0.13200843334197998, + "learning_rate": 1e-06, + "loss": -0.0071, + "num_tokens": 154280555.0, + "reward": 0.59375, + "reward_std": 0.20801475644111633, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 257 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2817.0, + "completions/mean_length": 624.950927734375, + "completions/mean_terminated_length": 577.8325805664062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 2.4104956268221573, + "grad_norm": 0.1377565860748291, + "learning_rate": 1e-06, + "loss": 0.0149, + "num_tokens": 154876263.0, + "reward": 0.6428571939468384, + "reward_std": 0.19505153596401215, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 258 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3386.0, + "completions/mean_length": 687.318115234375, + "completions/mean_terminated_length": 569.2344360351562, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 2.4198250728862973, + "grad_norm": 0.13844400644302368, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 155460756.0, + "reward": 0.6584821939468384, + "reward_std": 0.18911676108837128, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3983.0, + "completions/mean_length": 739.1652221679688, + "completions/mean_terminated_length": 622.8776245117188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 2.4291545189504373, + "grad_norm": 0.14645332098007202, + "learning_rate": 1e-06, + "loss": 0.0106, + "num_tokens": 156091800.0, + "reward": 0.5580357313156128, + "reward_std": 0.22571474313735962, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689781665802, + "step": 260 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2604.0, + "completions/mean_length": 726.5781860351562, + "completions/mean_terminated_length": 609.8544921875, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 2.4384839650145773, + "grad_norm": 0.13708554208278656, + "learning_rate": 1e-06, + "loss": -0.0192, + "num_tokens": 156715214.0, + "reward": 0.6127232313156128, + "reward_std": 0.1992606520652771, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 261 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2073.0, + "completions/mean_length": 740.450927734375, + "completions/mean_terminated_length": 583.6495361328125, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 2.4478134110787173, + "grad_norm": 0.13642618060112, + "learning_rate": 1e-06, + "loss": 0.0046, + "num_tokens": 157303658.0, + "reward": 0.5837053656578064, + "reward_std": 0.19475801289081573, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 262 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 725.1752319335938, + "completions/mean_terminated_length": 579.9825439453125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 2.4571428571428573, + "grad_norm": 0.13021457195281982, + "learning_rate": 1e-06, + "loss": -0.0077, + "num_tokens": 157886671.0, + "reward": 0.6316964626312256, + "reward_std": 0.18246811628341675, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 263 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2715.0, + "completions/mean_length": 723.2846069335938, + "completions/mean_terminated_length": 594.3163452148438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.466472303206997, + "grad_norm": 0.1289193332195282, + "learning_rate": 1e-06, + "loss": -0.0137, + "num_tokens": 158493798.0, + "reward": 0.598214328289032, + "reward_std": 0.19361935555934906, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 779.4542846679688, + "completions/mean_terminated_length": 616.3453979492188, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 2.4758017492711373, + "grad_norm": 0.11994383484125137, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 159108901.0, + "reward": 0.5959821939468384, + "reward_std": 0.16236519813537598, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 265 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3940.0, + "completions/mean_length": 684.7142944335938, + "completions/mean_terminated_length": 578.7249755859375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 2.485131195335277, + "grad_norm": 0.13918307423591614, + "learning_rate": 1e-06, + "loss": 0.015, + "num_tokens": 159695157.0, + "reward": 0.6696428656578064, + "reward_std": 0.19024720788002014, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 266 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 727.8438110351562, + "completions/mean_terminated_length": 627.1862182617188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 2.494460641399417, + "grad_norm": 0.13502292335033417, + "learning_rate": 1e-06, + "loss": -0.0048, + "num_tokens": 160327017.0, + "reward": 0.6082589626312256, + "reward_std": 0.1848343014717102, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.4884119927883148, + "step": 267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 814.6574096679688, + "completions/mean_terminated_length": 669.329833984375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 2.503790087463557, + "grad_norm": 0.11430999636650085, + "learning_rate": 1e-06, + "loss": -0.0025, + "num_tokens": 161000278.0, + "reward": 0.5959821939468384, + "reward_std": 0.1821650117635727, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 268 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3885.0, + "completions/mean_length": 738.9799194335938, + "completions/mean_terminated_length": 610.61181640625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 2.513119533527697, + "grad_norm": 0.12878777086734772, + "learning_rate": 1e-06, + "loss": -0.0188, + "num_tokens": 161612076.0, + "reward": 0.6395089626312256, + "reward_std": 0.19692835211753845, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 269 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2845.0, + "completions/mean_length": 705.771240234375, + "completions/mean_terminated_length": 616.4525146484375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 2.522448979591837, + "grad_norm": 0.11359921842813492, + "learning_rate": 1e-06, + "loss": 0.0023, + "num_tokens": 162244015.0, + "reward": 0.5714285969734192, + "reward_std": 0.1498495638370514, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514806270599365, + "step": 270 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2981.0, + "completions/mean_length": 798.5592041015625, + "completions/mean_terminated_length": 668.4976806640625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 2.5317784256559768, + "grad_norm": 0.12128040194511414, + "learning_rate": 1e-06, + "loss": 0.0121, + "num_tokens": 162931260.0, + "reward": 0.5479910969734192, + "reward_std": 0.1702548712491989, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2581.0, + "completions/mean_length": 704.677490234375, + "completions/mean_terminated_length": 595.2799682617188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 2.5411078717201168, + "grad_norm": 0.13675227761268616, + "learning_rate": 1e-06, + "loss": 0.0066, + "num_tokens": 163543051.0, + "reward": 0.6116071939468384, + "reward_std": 0.17412516474723816, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 272 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2391.0, + "completions/mean_length": 736.5111694335938, + "completions/mean_terminated_length": 612.0856323242188, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 2.5504373177842563, + "grad_norm": 0.1361248940229416, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 164154029.0, + "reward": 0.613839328289032, + "reward_std": 0.21515193581581116, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 882.2366333007812, + "completions/mean_terminated_length": 704.3251342773438, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 2.5597667638483967, + "grad_norm": 0.12098653614521027, + "learning_rate": 1e-06, + "loss": -0.0127, + "num_tokens": 164836809.0, + "reward": 0.5546875, + "reward_std": 0.20298220217227936, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 274 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3235.0, + "completions/mean_length": 737.6295166015625, + "completions/mean_terminated_length": 637.2643432617188, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.5690962099125363, + "grad_norm": 0.15183310210704803, + "learning_rate": 1e-06, + "loss": 0.0126, + "num_tokens": 165478309.0, + "reward": 0.566964328289032, + "reward_std": 0.23291975259780884, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 275 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2284.0, + "completions/mean_length": 723.8047485351562, + "completions/mean_terminated_length": 619.0299072265625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 2.5784256559766763, + "grad_norm": 0.13333465158939362, + "learning_rate": 1e-06, + "loss": 0.0059, + "num_tokens": 166113518.0, + "reward": 0.6316964626312256, + "reward_std": 0.1810370534658432, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 276 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1856.0, + "completions/mean_length": 727.8839721679688, + "completions/mean_terminated_length": 590.9686279296875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 2.5877551020408163, + "grad_norm": 0.1330011785030365, + "learning_rate": 1e-06, + "loss": -0.0001, + "num_tokens": 166709278.0, + "reward": 0.6071428656578064, + "reward_std": 0.18495012819766998, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 277 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4088.0, + "completions/mean_length": 763.1517944335938, + "completions/mean_terminated_length": 647.6951293945312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.5970845481049563, + "grad_norm": 0.1292235106229782, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 167362774.0, + "reward": 0.5602678656578064, + "reward_std": 0.2015914022922516, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 278 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3500.0, + "completions/mean_length": 740.6495971679688, + "completions/mean_terminated_length": 583.8574829101562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.6064139941690962, + "grad_norm": 0.13222847878932953, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 167949580.0, + "reward": 0.6383928656578064, + "reward_std": 0.1944110095500946, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 279 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3005.0, + "completions/mean_length": 707.5413208007812, + "completions/mean_terminated_length": 594.2018432617188, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 2.6157434402332362, + "grad_norm": 0.1384594887495041, + "learning_rate": 1e-06, + "loss": -0.0113, + "num_tokens": 168549137.0, + "reward": 0.59375, + "reward_std": 0.185399129986763, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 280 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3861.0, + "completions/mean_length": 646.1796875, + "completions/mean_terminated_length": 567.4166259765625, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 2.6250728862973762, + "grad_norm": 0.13485532999038696, + "learning_rate": 1e-06, + "loss": 0.0129, + "num_tokens": 169127114.0, + "reward": 0.6830357313156128, + "reward_std": 0.17964836955070496, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 281 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3154.0, + "completions/mean_length": 752.1864013671875, + "completions/mean_terminated_length": 608.1571655273438, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 2.6344023323615158, + "grad_norm": 0.13395307958126068, + "learning_rate": 1e-06, + "loss": 0.0014, + "num_tokens": 169734185.0, + "reward": 0.676339328289032, + "reward_std": 0.20538821816444397, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 282 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3916.0, + "completions/mean_length": 713.1328735351562, + "completions/mean_terminated_length": 579.7018432617188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 2.643731778425656, + "grad_norm": 0.138084277510643, + "learning_rate": 1e-06, + "loss": -0.0026, + "num_tokens": 170314400.0, + "reward": 0.6316964626312256, + "reward_std": 0.198629692196846, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3249.0, + "completions/mean_length": 681.8214721679688, + "completions/mean_terminated_length": 575.7422485351562, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 2.6530612244897958, + "grad_norm": 0.13710369169712067, + "learning_rate": 1e-06, + "loss": -0.0013, + "num_tokens": 170909976.0, + "reward": 0.6897321939468384, + "reward_std": 0.20421819388866425, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 284 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 738.904052734375, + "completions/mean_terminated_length": 614.567138671875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.6623906705539357, + "grad_norm": 0.12634806334972382, + "learning_rate": 1e-06, + "loss": 0.0028, + "num_tokens": 171530530.0, + "reward": 0.59375, + "reward_std": 0.172130286693573, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 869.8359985351562, + "completions/mean_terminated_length": 703.2265625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 2.6717201166180757, + "grad_norm": 0.12923769652843475, + "learning_rate": 1e-06, + "loss": -0.0099, + "num_tokens": 172218639.0, + "reward": 0.5100446939468384, + "reward_std": 0.2254137545824051, + "rewards/verify_math_reward/mean": 0.5100446343421936, + "rewards/verify_math_reward/std": 0.5001782774925232, + "step": 286 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2561.0, + "completions/mean_length": 794.9464721679688, + "completions/mean_terminated_length": 652.759033203125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 2.6810495626822157, + "grad_norm": 0.12405488640069962, + "learning_rate": 1e-06, + "loss": -0.0091, + "num_tokens": 172873975.0, + "reward": 0.5636160969734192, + "reward_std": 0.1823907196521759, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621349573135376, + "step": 287 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3543.0, + "completions/mean_length": 753.411865234375, + "completions/mean_terminated_length": 641.606689453125, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 2.6903790087463557, + "grad_norm": 0.1366748958826065, + "learning_rate": 1e-06, + "loss": -0.0158, + "num_tokens": 173515096.0, + "reward": 0.5870535969734192, + "reward_std": 0.2054981291294098, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 288 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3486.0, + "completions/mean_length": 786.0535888671875, + "completions/mean_terminated_length": 639.459228515625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.6997084548104957, + "grad_norm": 0.12155113369226456, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 174148040.0, + "reward": 0.546875, + "reward_std": 0.16935263574123383, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 289 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3424.0, + "completions/mean_length": 716.2031860351562, + "completions/mean_terminated_length": 607.1774291992188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 2.7090379008746357, + "grad_norm": 0.13820956647396088, + "learning_rate": 1e-06, + "loss": 0.0021, + "num_tokens": 174762518.0, + "reward": 0.6361607313156128, + "reward_std": 0.21245235204696655, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 290 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2889.0, + "completions/mean_length": 677.9330444335938, + "completions/mean_terminated_length": 623.6780395507812, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 2.7183673469387752, + "grad_norm": 0.13216546177864075, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 175402898.0, + "reward": 0.6283482313156128, + "reward_std": 0.17915913462638855, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 291 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3723.0, + "completions/mean_length": 825.4654541015625, + "completions/mean_terminated_length": 648.4717407226562, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 2.7276967930029157, + "grad_norm": 0.14216077327728271, + "learning_rate": 1e-06, + "loss": -0.0208, + "num_tokens": 176056467.0, + "reward": 0.5379464626312256, + "reward_std": 0.19384941458702087, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 292 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2566.0, + "completions/mean_length": 862.8303833007812, + "completions/mean_terminated_length": 659.5587158203125, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.7370262390670552, + "grad_norm": 0.1314835399389267, + "learning_rate": 1e-06, + "loss": -0.0108, + "num_tokens": 176703435.0, + "reward": 0.578125, + "reward_std": 0.2044457346200943, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 293 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4069.0, + "completions/mean_length": 780.060302734375, + "completions/mean_terminated_length": 625.1098022460938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.746355685131195, + "grad_norm": 0.14325650036334991, + "learning_rate": 1e-06, + "loss": 0.0059, + "num_tokens": 177339361.0, + "reward": 0.5569196939468384, + "reward_std": 0.19219790399074554, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2970.0, + "completions/mean_length": 710.513427734375, + "completions/mean_terminated_length": 613.3410034179688, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 2.755685131195335, + "grad_norm": 0.11966270953416824, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 177959501.0, + "reward": 0.6428571939468384, + "reward_std": 0.1828799694776535, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3709.0, + "completions/mean_length": 752.7064819335938, + "completions/mean_terminated_length": 652.7919311523438, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 2.765014577259475, + "grad_norm": 0.1251486837863922, + "learning_rate": 1e-06, + "loss": 0.0038, + "num_tokens": 178616094.0, + "reward": 0.6417410969734192, + "reward_std": 0.17198985815048218, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 296 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3888.0, + "completions/mean_length": 713.4553833007812, + "completions/mean_terminated_length": 620.3577880859375, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 2.774344023323615, + "grad_norm": 0.13381893932819366, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 179245334.0, + "reward": 0.5625, + "reward_std": 0.1802888810634613, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 297 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 786.1361694335938, + "completions/mean_terminated_length": 659.5712280273438, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 2.783673469387755, + "grad_norm": 0.12793391942977905, + "learning_rate": 1e-06, + "loss": 0.0126, + "num_tokens": 179906816.0, + "reward": 0.574776828289032, + "reward_std": 0.17284680902957916, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 298 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3119.0, + "completions/mean_length": 745.4777221679688, + "completions/mean_terminated_length": 625.401123046875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.793002915451895, + "grad_norm": 0.12774276733398438, + "learning_rate": 1e-06, + "loss": 0.0023, + "num_tokens": 180535620.0, + "reward": 0.5892857313156128, + "reward_std": 0.18306542932987213, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 299 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2319.0, + "completions/mean_length": 800.6116333007812, + "completions/mean_terminated_length": 662.6651000976562, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 2.8023323615160347, + "grad_norm": 0.1357034295797348, + "learning_rate": 1e-06, + "loss": 0.0108, + "num_tokens": 181189472.0, + "reward": 0.6049107313156128, + "reward_std": 0.20654937624931335, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 300 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3837.0, + "completions/mean_length": 751.1585083007812, + "completions/mean_terminated_length": 598.9428100585938, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 2.811661807580175, + "grad_norm": 0.13493631780147552, + "learning_rate": 1e-06, + "loss": -0.0055, + "num_tokens": 181783414.0, + "reward": 0.5680803656578064, + "reward_std": 0.18926052749156952, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 301 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2910.0, + "completions/mean_length": 755.9699096679688, + "completions/mean_terminated_length": 640.264404296875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.8209912536443147, + "grad_norm": 0.12465789914131165, + "learning_rate": 1e-06, + "loss": -0.0059, + "num_tokens": 182421571.0, + "reward": 0.578125, + "reward_std": 0.18336711823940277, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4045.0, + "completions/mean_length": 845.216552734375, + "completions/mean_terminated_length": 697.28125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.8303206997084547, + "grad_norm": 0.11650644987821579, + "learning_rate": 1e-06, + "loss": 0.0007, + "num_tokens": 183109005.0, + "reward": 0.5703125, + "reward_std": 0.20267769694328308, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 303 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2570.0, + "completions/mean_length": 723.9285888671875, + "completions/mean_terminated_length": 607.1131591796875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 2.8396501457725947, + "grad_norm": 0.13706143200397491, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 183725789.0, + "reward": 0.6506696939468384, + "reward_std": 0.20681221783161163, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 304 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 762.0781860351562, + "completions/mean_terminated_length": 626.5528564453125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.8489795918367347, + "grad_norm": 0.12602370977401733, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 184364515.0, + "reward": 0.598214328289032, + "reward_std": 0.18149100244045258, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 305 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3020.0, + "completions/mean_length": 733.9598388671875, + "completions/mean_terminated_length": 601.350341796875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 2.8583090379008746, + "grad_norm": 0.15349425375461578, + "learning_rate": 1e-06, + "loss": -0.0039, + "num_tokens": 184966015.0, + "reward": 0.6160714626312256, + "reward_std": 0.22172611951828003, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 306 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2517.0, + "completions/mean_length": 731.9453735351562, + "completions/mean_terminated_length": 615.4075927734375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 2.8676384839650146, + "grad_norm": 0.12386087328195572, + "learning_rate": 1e-06, + "loss": 0.0014, + "num_tokens": 185582518.0, + "reward": 0.5970982313156128, + "reward_std": 0.1680399477481842, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.4907552897930145, + "step": 307 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2568.0, + "completions/mean_length": 706.6920166015625, + "completions/mean_terminated_length": 593.3240966796875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 2.8769679300291546, + "grad_norm": 0.16650456190109253, + "learning_rate": 1e-06, + "loss": -0.0219, + "num_tokens": 186183738.0, + "reward": 0.6316964626312256, + "reward_std": 0.17479778826236725, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 308 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3092.0, + "completions/mean_length": 726.3214721679688, + "completions/mean_terminated_length": 621.6248779296875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 2.8862973760932946, + "grad_norm": 0.130840465426445, + "learning_rate": 1e-06, + "loss": 0.0138, + "num_tokens": 186811066.0, + "reward": 0.5703125, + "reward_std": 0.19952942430973053, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 309 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2262.0, + "completions/mean_length": 759.6004638671875, + "completions/mean_terminated_length": 675.6177978515625, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 2.8956268221574346, + "grad_norm": 0.12991474568843842, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 187486892.0, + "reward": 0.5613839626312256, + "reward_std": 0.20181076228618622, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 310 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3804.0, + "completions/mean_length": 706.1964721679688, + "completions/mean_terminated_length": 600.8745727539062, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.904956268221574, + "grad_norm": 0.132023423910141, + "learning_rate": 1e-06, + "loss": -0.0003, + "num_tokens": 188108236.0, + "reward": 0.6741071939468384, + "reward_std": 0.17179398238658905, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2984.0, + "completions/mean_length": 751.4475708007812, + "completions/mean_terminated_length": 619.52783203125, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 2.914285714285714, + "grad_norm": 0.11564072221517563, + "learning_rate": 1e-06, + "loss": -0.0132, + "num_tokens": 188732613.0, + "reward": 0.6227678656578064, + "reward_std": 0.15973344445228577, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644601345062, + "step": 312 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3927.0, + "completions/mean_length": 833.536865234375, + "completions/mean_terminated_length": 665.0528564453125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.923615160349854, + "grad_norm": 0.12622685730457306, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 189392894.0, + "reward": 0.574776828289032, + "reward_std": 0.18216319382190704, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 313 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 803.0558471679688, + "completions/mean_terminated_length": 673.1716918945312, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 2.932944606413994, + "grad_norm": 0.13950712978839874, + "learning_rate": 1e-06, + "loss": 0.0135, + "num_tokens": 190062640.0, + "reward": 0.546875, + "reward_std": 0.21984636783599854, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 314 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3873.0, + "completions/mean_length": 762.4207763671875, + "completions/mean_terminated_length": 634.948974609375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.942274052478134, + "grad_norm": 0.1219010129570961, + "learning_rate": 1e-06, + "loss": 0.0027, + "num_tokens": 190705313.0, + "reward": 0.551339328289032, + "reward_std": 0.16506867110729218, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 315 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3457.0, + "completions/mean_length": 728.9397583007812, + "completions/mean_terminated_length": 624.3245239257812, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 2.951603498542274, + "grad_norm": 0.12655006349086761, + "learning_rate": 1e-06, + "loss": 0.0094, + "num_tokens": 191342923.0, + "reward": 0.6540178656578064, + "reward_std": 0.18121999502182007, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 316 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3419.0, + "completions/mean_length": 805.3951416015625, + "completions/mean_terminated_length": 675.6032104492188, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 2.960932944606414, + "grad_norm": 0.13242557644844055, + "learning_rate": 1e-06, + "loss": 0.0036, + "num_tokens": 192012605.0, + "reward": 0.5792410969734192, + "reward_std": 0.21222274005413055, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 317 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3360.0, + "completions/mean_length": 706.591552734375, + "completions/mean_terminated_length": 601.2819213867188, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 2.970262390670554, + "grad_norm": 0.12549835443496704, + "learning_rate": 1e-06, + "loss": -0.002, + "num_tokens": 192624271.0, + "reward": 0.6383928656578064, + "reward_std": 0.17325684428215027, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3669.0, + "completions/mean_length": 780.0949096679688, + "completions/mean_terminated_length": 653.2989501953125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 2.979591836734694, + "grad_norm": 0.12499081343412399, + "learning_rate": 1e-06, + "loss": 0.0011, + "num_tokens": 193277852.0, + "reward": 0.5959821939468384, + "reward_std": 0.17833498120307922, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 319 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3491.0, + "completions/mean_length": 766.114990234375, + "completions/mean_terminated_length": 650.760986328125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 2.9889212827988336, + "grad_norm": 0.12210965901613235, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 193938115.0, + "reward": 0.6506696939468384, + "reward_std": 0.18914924561977386, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 320 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2620.0, + "completions/mean_length": 752.946044921875, + "completions/mean_terminated_length": 645.1055908203125, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 2.9982507288629736, + "grad_norm": 0.12024641036987305, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 194592890.0, + "reward": 0.629464328289032, + "reward_std": 0.1607832908630371, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 321 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3246.0, + "completions/mean_length": 975.0569458007812, + "completions/mean_terminated_length": 726.8855590820312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 3.00932944606414, + "grad_norm": 0.12204635888338089, + "learning_rate": 1e-06, + "loss": -0.023, + "num_tokens": 195286493.0, + "reward": 0.5212053656578064, + "reward_std": 0.1892682909965515, + "rewards/verify_math_reward/mean": 0.5212053656578064, + "rewards/verify_math_reward/std": 0.49982914328575134, + "step": 322 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3803.0, + "completions/mean_length": 844.849365234375, + "completions/mean_terminated_length": 696.8973388671875, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 3.01865889212828, + "grad_norm": 0.1216764971613884, + "learning_rate": 1e-06, + "loss": 0.0064, + "num_tokens": 195966230.0, + "reward": 0.621651828289032, + "reward_std": 0.1701001226902008, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3690.0, + "completions/mean_length": 815.982177734375, + "completions/mean_terminated_length": 638.4752807617188, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.02798833819242, + "grad_norm": 0.13047727942466736, + "learning_rate": 1e-06, + "loss": 0.0147, + "num_tokens": 196598598.0, + "reward": 0.5736607313156128, + "reward_std": 0.18979185819625854, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3157.0, + "completions/mean_length": 739.0881958007812, + "completions/mean_terminated_length": 638.7666625976562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 3.03731778425656, + "grad_norm": 0.12554390728473663, + "learning_rate": 1e-06, + "loss": 0.0021, + "num_tokens": 197244917.0, + "reward": 0.6127232313156128, + "reward_std": 0.18123069405555725, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 325 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2938.0, + "completions/mean_length": 745.0792846679688, + "completions/mean_terminated_length": 604.8081665039062, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 3.0466472303206995, + "grad_norm": 0.1269174963235855, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 197845316.0, + "reward": 0.6540178656578064, + "reward_std": 0.1638319492340088, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 326 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3319.0, + "completions/mean_length": 800.5100708007812, + "completions/mean_terminated_length": 662.559326171875, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 3.0559766763848395, + "grad_norm": 0.13453570008277893, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 198507269.0, + "reward": 0.5959821939468384, + "reward_std": 0.19384831190109253, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 327 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 790.5201416015625, + "completions/mean_terminated_length": 623.8897705078125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.0653061224489795, + "grad_norm": 0.13485023379325867, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 199132199.0, + "reward": 0.5993303656578064, + "reward_std": 0.17788033187389374, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3760.0, + "completions/mean_length": 741.2969360351562, + "completions/mean_terminated_length": 608.976806640625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.0746355685131195, + "grad_norm": 0.12389012426137924, + "learning_rate": 1e-06, + "loss": -0.0048, + "num_tokens": 199747945.0, + "reward": 0.6339285969734192, + "reward_std": 0.14342734217643738, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199838399887085, + "step": 329 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3197.0, + "completions/mean_length": 829.8928833007812, + "completions/mean_terminated_length": 669.2645874023438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.0839650145772595, + "grad_norm": 0.13717308640480042, + "learning_rate": 1e-06, + "loss": 0.0025, + "num_tokens": 200415233.0, + "reward": 0.5323660969734192, + "reward_std": 0.17844556272029877, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 330 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4031.0, + "completions/mean_length": 737.4219360351562, + "completions/mean_terminated_length": 625.0818481445312, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 3.0932944606413995, + "grad_norm": 0.13934089243412018, + "learning_rate": 1e-06, + "loss": 0.0058, + "num_tokens": 201043955.0, + "reward": 0.637276828289032, + "reward_std": 0.20286203920841217, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3564.0, + "completions/mean_length": 708.052490234375, + "completions/mean_terminated_length": 622.7722778320312, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 3.1026239067055394, + "grad_norm": 0.11321843415498734, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 201675474.0, + "reward": 0.6328125, + "reward_std": 0.1518772393465042, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 332 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3975.0, + "completions/mean_length": 708.044677734375, + "completions/mean_terminated_length": 586.6265869140625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 3.1119533527696794, + "grad_norm": 0.11942648887634277, + "learning_rate": 1e-06, + "loss": 0.0064, + "num_tokens": 202278234.0, + "reward": 0.6551339626312256, + "reward_std": 0.13970929384231567, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 333 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3900.0, + "completions/mean_length": 791.6574096679688, + "completions/mean_terminated_length": 677.188232421875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 3.1212827988338194, + "grad_norm": 0.11885570734739304, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 202965511.0, + "reward": 0.598214328289032, + "reward_std": 0.1627795696258545, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3705.0, + "completions/mean_length": 707.5413208007812, + "completions/mean_terminated_length": 569.799072265625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 3.130612244897959, + "grad_norm": 0.18239238858222961, + "learning_rate": 1e-06, + "loss": 0.0097, + "num_tokens": 203537652.0, + "reward": 0.6718750596046448, + "reward_std": 0.18246419727802277, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3574.0, + "completions/mean_length": 770.1641235351562, + "completions/mean_terminated_length": 666.8297119140625, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 3.139941690962099, + "grad_norm": 0.13014322519302368, + "learning_rate": 1e-06, + "loss": 0.0044, + "num_tokens": 204207463.0, + "reward": 0.5680803656578064, + "reward_std": 0.19701138138771057, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 336 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2921.0, + "completions/mean_length": 731.4676513671875, + "completions/mean_terminated_length": 598.7598266601562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 3.149271137026239, + "grad_norm": 0.13176432251930237, + "learning_rate": 1e-06, + "loss": -0.0086, + "num_tokens": 204810898.0, + "reward": 0.6808035969734192, + "reward_std": 0.16777075827121735, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 337 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3937.0, + "completions/mean_length": 957.404052734375, + "completions/mean_terminated_length": 724.0791625976562, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "epoch": 3.158600583090379, + "grad_norm": 0.125205859541893, + "learning_rate": 1e-06, + "loss": -0.0144, + "num_tokens": 205529364.0, + "reward": 0.486607164144516, + "reward_std": 0.17325684428215027, + "rewards/verify_math_reward/mean": 0.4866071343421936, + "rewards/verify_math_reward/std": 0.500099778175354, + "step": 338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 805.255615234375, + "completions/mean_terminated_length": 655.5017700195312, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 3.167930029154519, + "grad_norm": 0.1224595233798027, + "learning_rate": 1e-06, + "loss": 0.0038, + "num_tokens": 206183657.0, + "reward": 0.6026785969734192, + "reward_std": 0.151692196726799, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 339 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3550.0, + "completions/mean_length": 797.7813110351562, + "completions/mean_terminated_length": 635.57373046875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 3.177259475218659, + "grad_norm": 0.13148732483386993, + "learning_rate": 1e-06, + "loss": -0.0173, + "num_tokens": 206811333.0, + "reward": 0.6551339626312256, + "reward_std": 0.18419378995895386, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 340 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3885.0, + "completions/mean_length": 792.4699096679688, + "completions/mean_terminated_length": 642.1342163085938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.186588921282799, + "grad_norm": 0.13803736865520477, + "learning_rate": 1e-06, + "loss": -0.0067, + "num_tokens": 207444314.0, + "reward": 0.6004464626312256, + "reward_std": 0.1982075572013855, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 341 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4011.0, + "completions/mean_length": 840.5279541015625, + "completions/mean_terminated_length": 700.3038330078125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 3.195918367346939, + "grad_norm": 0.12066961079835892, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 208139347.0, + "reward": 0.582589328289032, + "reward_std": 0.18193748593330383, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 757.6875610351562, + "completions/mean_terminated_length": 638.0485229492188, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.205247813411079, + "grad_norm": 0.131647989153862, + "learning_rate": 1e-06, + "loss": -0.0089, + "num_tokens": 208777363.0, + "reward": 0.6428571939468384, + "reward_std": 0.19358982145786285, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 343 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2681.0, + "completions/mean_length": 792.4944458007812, + "completions/mean_terminated_length": 662.1937255859375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 3.2145772594752184, + "grad_norm": 0.12827526032924652, + "learning_rate": 1e-06, + "loss": -0.0065, + "num_tokens": 209436734.0, + "reward": 0.5736607313156128, + "reward_std": 0.1675097495317459, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 344 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3706.0, + "completions/mean_length": 836.8683471679688, + "completions/mean_terminated_length": 684.5723876953125, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 3.2239067055393584, + "grad_norm": 0.12524019181728363, + "learning_rate": 1e-06, + "loss": -0.004, + "num_tokens": 210117512.0, + "reward": 0.5625, + "reward_std": 0.17171843349933624, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 345 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2176.0, + "completions/mean_length": 752.7210083007812, + "completions/mean_terminated_length": 620.8515014648438, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.2332361516034984, + "grad_norm": 0.125301331281662, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 210736638.0, + "reward": 0.6674107313156128, + "reward_std": 0.1475936770439148, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 346 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3931.0, + "completions/mean_length": 782.5111694335938, + "completions/mean_terminated_length": 639.7881469726562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.2425655976676384, + "grad_norm": 0.13713519275188446, + "learning_rate": 1e-06, + "loss": 0.0009, + "num_tokens": 211377448.0, + "reward": 0.606026828289032, + "reward_std": 0.1965574324131012, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 347 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 724.2344360351562, + "completions/mean_terminated_length": 607.4295654296875, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 3.2518950437317784, + "grad_norm": 0.13097605109214783, + "learning_rate": 1e-06, + "loss": -0.0064, + "num_tokens": 211994682.0, + "reward": 0.6116071939468384, + "reward_std": 0.18738895654678345, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 348 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3868.0, + "completions/mean_length": 972.9163208007812, + "completions/mean_terminated_length": 748.771484375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 3.2612244897959184, + "grad_norm": 0.1196514368057251, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 212715503.0, + "reward": 0.504464328289032, + "reward_std": 0.18821631371974945, + "rewards/verify_math_reward/mean": 0.5044642686843872, + "rewards/verify_math_reward/std": 0.5002593398094177, + "step": 349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4051.0, + "completions/mean_length": 785.9297485351562, + "completions/mean_terminated_length": 647.3685913085938, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 3.2705539358600584, + "grad_norm": 0.12302932888269424, + "learning_rate": 1e-06, + "loss": -0.0072, + "num_tokens": 213364704.0, + "reward": 0.578125, + "reward_std": 0.1691262274980545, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 350 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3123.0, + "completions/mean_length": 838.6361694335938, + "completions/mean_terminated_length": 658.3109741210938, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 3.2798833819241984, + "grad_norm": 0.12156125158071518, + "learning_rate": 1e-06, + "loss": -0.011, + "num_tokens": 214010266.0, + "reward": 0.5770089626312256, + "reward_std": 0.15811371803283691, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 351 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3470.0, + "completions/mean_length": 898.7600708007812, + "completions/mean_terminated_length": 725.73291015625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 3.2892128279883384, + "grad_norm": 0.12426517903804779, + "learning_rate": 1e-06, + "loss": -0.0251, + "num_tokens": 214721787.0, + "reward": 0.5915178656578064, + "reward_std": 0.17329145967960358, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 352 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3852.0, + "completions/mean_length": 803.9063110351562, + "completions/mean_terminated_length": 625.745849609375, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 3.298542274052478, + "grad_norm": 0.14308196306228638, + "learning_rate": 1e-06, + "loss": -0.016, + "num_tokens": 215347191.0, + "reward": 0.6595982313156128, + "reward_std": 0.21647492051124573, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 353 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 868.4330444335938, + "completions/mean_terminated_length": 689.7573852539062, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 3.307871720116618, + "grad_norm": 0.14486242830753326, + "learning_rate": 1e-06, + "loss": 0.0092, + "num_tokens": 216023715.0, + "reward": 0.5680803656578064, + "reward_std": 0.20107145607471466, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200122833252, + "step": 354 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1989.0, + "completions/mean_length": 644.638427734375, + "completions/mean_terminated_length": 569.8654174804688, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 3.317201166180758, + "grad_norm": 0.13110017776489258, + "learning_rate": 1e-06, + "loss": -0.0004, + "num_tokens": 216600855.0, + "reward": 0.7087053656578064, + "reward_std": 0.13947896659374237, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 355 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3109.0, + "completions/mean_length": 795.8114013671875, + "completions/mean_terminated_length": 645.6278076171875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 3.326530612244898, + "grad_norm": 0.14633680880069733, + "learning_rate": 1e-06, + "loss": 0.0041, + "num_tokens": 217244918.0, + "reward": 0.6439732313156128, + "reward_std": 0.22289685904979706, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909069061279297, + "step": 356 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3065.0, + "completions/mean_length": 857.1350708007812, + "completions/mean_terminated_length": 701.821044921875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.335860058309038, + "grad_norm": 0.11338292062282562, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 217942079.0, + "reward": 0.5602678656578064, + "reward_std": 0.16682574152946472, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 357 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3731.0, + "completions/mean_length": 751.0703735351562, + "completions/mean_terminated_length": 615.0975341796875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 3.345189504373178, + "grad_norm": 0.13032779097557068, + "learning_rate": 1e-06, + "loss": 0.0041, + "num_tokens": 218555446.0, + "reward": 0.637276828289032, + "reward_std": 0.18626421689987183, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3676.0, + "completions/mean_length": 848.3973388671875, + "completions/mean_terminated_length": 684.6846313476562, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 3.354518950437318, + "grad_norm": 0.1246708482503891, + "learning_rate": 1e-06, + "loss": 0.0027, + "num_tokens": 219234202.0, + "reward": 0.5926339626312256, + "reward_std": 0.17415907979011536, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 359 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2835.0, + "completions/mean_length": 760.5625610351562, + "completions/mean_terminated_length": 629.0023193359375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 3.363848396501458, + "grad_norm": 0.13155436515808105, + "learning_rate": 1e-06, + "loss": 0.0061, + "num_tokens": 219867314.0, + "reward": 0.6227678656578064, + "reward_std": 0.17033086717128754, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 360 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2779.0, + "completions/mean_length": 768.6551513671875, + "completions/mean_terminated_length": 600.922607421875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 3.373177842565598, + "grad_norm": 0.1308959573507309, + "learning_rate": 1e-06, + "loss": -0.0068, + "num_tokens": 220480621.0, + "reward": 0.6584821939468384, + "reward_std": 0.15687911212444305, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 361 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 718.0870971679688, + "completions/mean_terminated_length": 584.8515014648438, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 3.3825072886297374, + "grad_norm": 0.1537025421857834, + "learning_rate": 1e-06, + "loss": -0.01, + "num_tokens": 221070915.0, + "reward": 0.6417410969734192, + "reward_std": 0.19739368557929993, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 362 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4068.0, + "completions/mean_length": 792.9531860351562, + "completions/mean_terminated_length": 662.6705322265625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 3.3918367346938774, + "grad_norm": 0.1273297369480133, + "learning_rate": 1e-06, + "loss": -0.0165, + "num_tokens": 221732233.0, + "reward": 0.5736607313156128, + "reward_std": 0.19155670702457428, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4095.0, + "completions/mean_length": 810.4799194335938, + "completions/mean_terminated_length": 628.5960083007812, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.4011661807580174, + "grad_norm": 0.12025588005781174, + "learning_rate": 1e-06, + "loss": -0.0125, + "num_tokens": 222363759.0, + "reward": 0.590401828289032, + "reward_std": 0.15605242550373077, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 364 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 783.779052734375, + "completions/mean_terminated_length": 641.110595703125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 3.4104956268221573, + "grad_norm": 0.12685884535312653, + "learning_rate": 1e-06, + "loss": -0.004, + "num_tokens": 222999433.0, + "reward": 0.6328125, + "reward_std": 0.18753305077552795, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 365 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 792.6517944335938, + "completions/mean_terminated_length": 678.2171020507812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.4198250728862973, + "grad_norm": 0.1262684017419815, + "learning_rate": 1e-06, + "loss": 0.0252, + "num_tokens": 223688769.0, + "reward": 0.566964328289032, + "reward_std": 0.16773614287376404, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 366 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3580.0, + "completions/mean_length": 757.9933471679688, + "completions/mean_terminated_length": 650.315673828125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.4291545189504373, + "grad_norm": 0.1352754533290863, + "learning_rate": 1e-06, + "loss": -0.0125, + "num_tokens": 224351451.0, + "reward": 0.6127232313156128, + "reward_std": 0.16506798565387726, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 367 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3375.0, + "completions/mean_length": 694.8717041015625, + "completions/mean_terminated_length": 601.2625732421875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 3.4384839650145773, + "grad_norm": 0.13980157673358917, + "learning_rate": 1e-06, + "loss": 0.0013, + "num_tokens": 224959832.0, + "reward": 0.7020089626312256, + "reward_std": 0.17600169777870178, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763102173805237, + "step": 368 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3873.0, + "completions/mean_length": 799.099365234375, + "completions/mean_terminated_length": 612.4822998046875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 3.4478134110787173, + "grad_norm": 0.13098278641700745, + "learning_rate": 1e-06, + "loss": -0.0243, + "num_tokens": 225579569.0, + "reward": 0.5837053656578064, + "reward_std": 0.162964329123497, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3454.0, + "completions/mean_length": 841.9609985351562, + "completions/mean_terminated_length": 653.7107543945312, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 3.4571428571428573, + "grad_norm": 0.13110192120075226, + "learning_rate": 1e-06, + "loss": -0.0279, + "num_tokens": 226219126.0, + "reward": 0.598214328289032, + "reward_std": 0.16946394741535187, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 370 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3737.0, + "completions/mean_length": 725.5535888671875, + "completions/mean_terminated_length": 604.7630004882812, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 3.466472303206997, + "grad_norm": 0.12862740457057953, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 226827454.0, + "reward": 0.660714328289032, + "reward_std": 0.17833246290683746, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 371 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3340.0, + "completions/mean_length": 822.5736694335938, + "completions/mean_terminated_length": 620.8933715820312, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 3.4758017492711373, + "grad_norm": 0.1471925973892212, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 227447480.0, + "reward": 0.629464328289032, + "reward_std": 0.20872971415519714, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 372 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 819.9933471679688, + "completions/mean_terminated_length": 638.6360473632812, + "completions/min_length": 194.0, + "completions/min_terminated_length": 194.0, + "epoch": 3.485131195335277, + "grad_norm": 0.12016690522432327, + "learning_rate": 1e-06, + "loss": 0.012, + "num_tokens": 228079050.0, + "reward": 0.6171875, + "reward_std": 0.14447791874408722, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3905.0, + "completions/mean_length": 801.450927734375, + "completions/mean_terminated_length": 639.4238891601562, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 3.494460641399417, + "grad_norm": 0.13443051278591156, + "learning_rate": 1e-06, + "loss": -0.0245, + "num_tokens": 228722702.0, + "reward": 0.5993303656578064, + "reward_std": 0.17769235372543335, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 374 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2972.0, + "completions/mean_length": 775.9688110351562, + "completions/mean_terminated_length": 636.99072265625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 3.503790087463557, + "grad_norm": 0.11227922886610031, + "learning_rate": 1e-06, + "loss": -0.0183, + "num_tokens": 229348490.0, + "reward": 0.566964328289032, + "reward_std": 0.1491014063358307, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 375 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3847.0, + "completions/mean_length": 850.6864013671875, + "completions/mean_terminated_length": 675.0576171875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.513119533527697, + "grad_norm": 0.14165233075618744, + "learning_rate": 1e-06, + "loss": -0.0233, + "num_tokens": 230010705.0, + "reward": 0.6272321939468384, + "reward_std": 0.22120323777198792, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 376 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3570.0, + "completions/mean_length": 860.3058471679688, + "completions/mean_terminated_length": 677.1533203125, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 3.522448979591837, + "grad_norm": 0.13399231433868408, + "learning_rate": 1e-06, + "loss": -0.0435, + "num_tokens": 230681227.0, + "reward": 0.5926339626312256, + "reward_std": 0.20519572496414185, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 377 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3750.0, + "completions/mean_length": 864.5178833007812, + "completions/mean_terminated_length": 632.59326171875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 3.5317784256559768, + "grad_norm": 0.14105992019176483, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 231304931.0, + "reward": 0.6082589626312256, + "reward_std": 0.1909932643175125, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 378 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3969.0, + "completions/mean_length": 821.0881958007812, + "completions/mean_terminated_length": 643.8576049804688, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.5411078717201168, + "grad_norm": 0.1384781301021576, + "learning_rate": 1e-06, + "loss": -0.0208, + "num_tokens": 231936850.0, + "reward": 0.6328125, + "reward_std": 0.19772250950336456, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 379 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2813.0, + "completions/mean_length": 790.8504638671875, + "completions/mean_terminated_length": 611.9835205078125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 3.5504373177842563, + "grad_norm": 0.12803971767425537, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 232557116.0, + "reward": 0.6160714626312256, + "reward_std": 0.1532345414161682, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.4866124987602234, + "step": 380 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2738.0, + "completions/mean_length": 722.622802734375, + "completions/mean_terminated_length": 589.5661010742188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 3.5597667638483967, + "grad_norm": 0.175442174077034, + "learning_rate": 1e-06, + "loss": -0.0137, + "num_tokens": 233148554.0, + "reward": 0.6484375, + "reward_std": 0.20790667831897736, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2674.0, + "completions/mean_length": 770.357177734375, + "completions/mean_terminated_length": 610.8818359375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 3.5690962099125363, + "grad_norm": 0.13977134227752686, + "learning_rate": 1e-06, + "loss": -0.0244, + "num_tokens": 233761690.0, + "reward": 0.6328125, + "reward_std": 0.18490876257419586, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 382 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3217.0, + "completions/mean_length": 729.8717041015625, + "completions/mean_terminated_length": 605.2002563476562, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 3.5784256559766763, + "grad_norm": 0.15750020742416382, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 234380591.0, + "reward": 0.6071428656578064, + "reward_std": 0.22440779209136963, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 383 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2093.0, + "completions/mean_length": 749.536865234375, + "completions/mean_terminated_length": 601.3251953125, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 3.5877551020408163, + "grad_norm": 0.1392834335565567, + "learning_rate": 1e-06, + "loss": -0.0139, + "num_tokens": 234990048.0, + "reward": 0.6428571939468384, + "reward_std": 0.17400754988193512, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 384 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2660.0, + "completions/mean_length": 729.1361694335938, + "completions/mean_terminated_length": 612.5011596679688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.5970845481049563, + "grad_norm": 0.14242801070213318, + "learning_rate": 1e-06, + "loss": -0.0178, + "num_tokens": 235613746.0, + "reward": 0.6194196939468384, + "reward_std": 0.1595052182674408, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 385 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3693.0, + "completions/mean_length": 744.544677734375, + "completions/mean_terminated_length": 644.38623046875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 3.6064139941690962, + "grad_norm": 0.12094242125749588, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 236257650.0, + "reward": 0.6026785969734192, + "reward_std": 0.15642404556274414, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 386 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3183.0, + "completions/mean_length": 796.5234985351562, + "completions/mean_terminated_length": 646.3722534179688, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.6157434402332362, + "grad_norm": 0.13153177499771118, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 236901975.0, + "reward": 0.6462053656578064, + "reward_std": 0.17405030131340027, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 387 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2754.0, + "completions/mean_length": 813.2188110351562, + "completions/mean_terminated_length": 651.7704467773438, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 3.6250728862973762, + "grad_norm": 0.14037121832370758, + "learning_rate": 1e-06, + "loss": -0.0103, + "num_tokens": 237547923.0, + "reward": 0.6015625, + "reward_std": 0.20850148797035217, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2965.0, + "completions/mean_length": 677.7600708007812, + "completions/mean_terminated_length": 563.4244384765625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 3.6344023323615158, + "grad_norm": 0.13413682579994202, + "learning_rate": 1e-06, + "loss": -0.0105, + "num_tokens": 238125732.0, + "reward": 0.676339328289032, + "reward_std": 0.13421861827373505, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 389 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4077.0, + "completions/mean_length": 820.6529541015625, + "completions/mean_terminated_length": 651.5035400390625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 3.643731778425656, + "grad_norm": 0.12404172122478485, + "learning_rate": 1e-06, + "loss": -0.016, + "num_tokens": 238761045.0, + "reward": 0.5613839626312256, + "reward_std": 0.1713072657585144, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 390 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 849.677490234375, + "completions/mean_terminated_length": 633.2559814453125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 3.6530612244897958, + "grad_norm": 0.13872458040714264, + "learning_rate": 1e-06, + "loss": -0.0196, + "num_tokens": 239379524.0, + "reward": 0.6484375, + "reward_std": 0.18359535932540894, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 391 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 807.9989013671875, + "completions/mean_terminated_length": 658.369873046875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.6623906705539357, + "grad_norm": 0.12471380829811096, + "learning_rate": 1e-06, + "loss": -0.026, + "num_tokens": 240042731.0, + "reward": 0.5926339626312256, + "reward_std": 0.1489473432302475, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161848425865173, + "step": 392 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4031.0, + "completions/mean_length": 718.3080444335938, + "completions/mean_terminated_length": 601.2979125976562, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 3.6717201166180757, + "grad_norm": 0.13518400490283966, + "learning_rate": 1e-06, + "loss": -0.0227, + "num_tokens": 240658639.0, + "reward": 0.6908482313156128, + "reward_std": 0.17446216940879822, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 393 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3633.0, + "completions/mean_length": 769.1842041015625, + "completions/mean_terminated_length": 657.9065551757812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.6810495626822157, + "grad_norm": 0.13596130907535553, + "learning_rate": 1e-06, + "loss": -0.0087, + "num_tokens": 241321572.0, + "reward": 0.5948660969734192, + "reward_std": 0.19411608576774597, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3765.0, + "completions/mean_length": 773.4085083007812, + "completions/mean_terminated_length": 642.35498046875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 3.6903790087463557, + "grad_norm": 0.11431754380464554, + "learning_rate": 1e-06, + "loss": -0.0073, + "num_tokens": 241959506.0, + "reward": 0.5792410969734192, + "reward_std": 0.14297270774841309, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 395 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3726.0, + "completions/mean_length": 807.1295166015625, + "completions/mean_terminated_length": 600.3558959960938, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 3.6997084548104957, + "grad_norm": 0.15255998075008392, + "learning_rate": 1e-06, + "loss": -0.0298, + "num_tokens": 242554550.0, + "reward": 0.6350446939468384, + "reward_std": 0.20069167017936707, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 396 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3499.0, + "completions/mean_length": 685.5714721679688, + "completions/mean_terminated_length": 579.6087646484375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 3.7090379008746357, + "grad_norm": 0.13583225011825562, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 243144590.0, + "reward": 0.660714328289032, + "reward_std": 0.17592571675777435, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 397 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3924.0, + "completions/mean_length": 854.1272583007812, + "completions/mean_terminated_length": 674.6596069335938, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.7183673469387752, + "grad_norm": 0.12747155129909515, + "learning_rate": 1e-06, + "loss": 0.0025, + "num_tokens": 243807992.0, + "reward": 0.5647321939468384, + "reward_std": 0.19001688063144684, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2888.0, + "completions/mean_length": 690.1785888671875, + "completions/mean_terminated_length": 580.3133544921875, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 3.7276967930029157, + "grad_norm": 0.12292210012674332, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 244391896.0, + "reward": 0.6953125596046448, + "reward_std": 0.133317768573761, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 399 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 773.1506958007812, + "completions/mean_terminated_length": 630.0244750976562, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.7370262390670552, + "grad_norm": 0.13916349411010742, + "learning_rate": 1e-06, + "loss": -0.0233, + "num_tokens": 245013071.0, + "reward": 0.6417410969734192, + "reward_std": 0.20583400130271912, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 400 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3731.0, + "completions/mean_length": 793.6830444335938, + "completions/mean_terminated_length": 623.140869140625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 3.746355685131195, + "grad_norm": 0.13125135004520416, + "learning_rate": 1e-06, + "loss": -0.0032, + "num_tokens": 245625651.0, + "reward": 0.660714328289032, + "reward_std": 0.15424484014511108, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 401 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2477.0, + "completions/mean_length": 780.3660888671875, + "completions/mean_terminated_length": 653.5805053710938, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 3.755685131195335, + "grad_norm": 0.12869560718536377, + "learning_rate": 1e-06, + "loss": -0.0067, + "num_tokens": 246274891.0, + "reward": 0.6462053656578064, + "reward_std": 0.17690464854240417, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 402 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 747.2455444335938, + "completions/mean_terminated_length": 639.22119140625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.765014577259475, + "grad_norm": 0.1576901376247406, + "learning_rate": 1e-06, + "loss": -0.0294, + "num_tokens": 246910479.0, + "reward": 0.6328125, + "reward_std": 0.22128063440322876, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 403 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2720.0, + "completions/mean_length": 816.6339721679688, + "completions/mean_terminated_length": 639.162353515625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.774344023323615, + "grad_norm": 0.13532805442810059, + "learning_rate": 1e-06, + "loss": -0.0245, + "num_tokens": 247538351.0, + "reward": 0.6439732313156128, + "reward_std": 0.16244368255138397, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909069061279297, + "step": 404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3952.0, + "completions/mean_length": 765.7545166015625, + "completions/mean_terminated_length": 650.3880004882812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.783673469387755, + "grad_norm": 0.13426661491394043, + "learning_rate": 1e-06, + "loss": -0.0058, + "num_tokens": 248202899.0, + "reward": 0.6305803656578064, + "reward_std": 0.18520158529281616, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 405 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3827.0, + "completions/mean_length": 779.7824096679688, + "completions/mean_terminated_length": 644.9767456054688, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 3.793002915451895, + "grad_norm": 0.1290045529603958, + "learning_rate": 1e-06, + "loss": -0.0152, + "num_tokens": 248838088.0, + "reward": 0.5837053656578064, + "reward_std": 0.1660016030073166, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 406 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3890.0, + "completions/mean_length": 769.943115234375, + "completions/mean_terminated_length": 626.6787109375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.8023323615160347, + "grad_norm": 0.1332543045282364, + "learning_rate": 1e-06, + "loss": -0.0, + "num_tokens": 249458237.0, + "reward": 0.6629464626312256, + "reward_std": 0.18096178770065308, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 407 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2842.0, + "completions/mean_length": 729.2767944335938, + "completions/mean_terminated_length": 584.2608032226562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.811661807580175, + "grad_norm": 0.1329224556684494, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 250039189.0, + "reward": 0.65625, + "reward_std": 0.15608635544776917, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 408 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3882.0, + "completions/mean_length": 897.7779541015625, + "completions/mean_terminated_length": 696.7034301757812, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 3.8209912536443147, + "grad_norm": 0.13379347324371338, + "learning_rate": 1e-06, + "loss": -0.0027, + "num_tokens": 250712974.0, + "reward": 0.5379464626312256, + "reward_std": 0.1887405961751938, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 409 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3477.0, + "completions/mean_length": 816.0658569335938, + "completions/mean_terminated_length": 658.782470703125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 3.8303206997084547, + "grad_norm": 0.1351667046546936, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 251367753.0, + "reward": 0.559151828289032, + "reward_std": 0.14981679618358612, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 410 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 746.6049194335938, + "completions/mean_terminated_length": 606.3976440429688, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.8396501457725947, + "grad_norm": 0.1125982329249382, + "learning_rate": 1e-06, + "loss": -0.0072, + "num_tokens": 251974071.0, + "reward": 0.6395089626312256, + "reward_std": 0.14421828091144562, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 411 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3120.0, + "completions/mean_length": 747.4152221679688, + "completions/mean_terminated_length": 615.33642578125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 3.8489795918367347, + "grad_norm": 0.12228783220052719, + "learning_rate": 1e-06, + "loss": 0.006, + "num_tokens": 252589067.0, + "reward": 0.6205357313156128, + "reward_std": 0.15511175990104675, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 412 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3629.0, + "completions/mean_length": 723.5201416015625, + "completions/mean_terminated_length": 570.0466918945312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.8583090379008746, + "grad_norm": 0.1554757058620453, + "learning_rate": 1e-06, + "loss": 0.0092, + "num_tokens": 253156877.0, + "reward": 0.7198660969734192, + "reward_std": 0.17220626771450043, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3540.0, + "completions/mean_length": 750.9107666015625, + "completions/mean_terminated_length": 635.030029296875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 3.8676384839650146, + "grad_norm": 0.13665427267551422, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 253802485.0, + "reward": 0.640625, + "reward_std": 0.19888296723365784, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 414 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2531.0, + "completions/mean_length": 668.0870971679688, + "completions/mean_terminated_length": 577.7755126953125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 3.8769679300291546, + "grad_norm": 0.14581741392612457, + "learning_rate": 1e-06, + "loss": 0.021, + "num_tokens": 254390843.0, + "reward": 0.684151828289032, + "reward_std": 0.18193678557872772, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 946.7031860351562, + "completions/mean_terminated_length": 675.6727294921875, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 3.8862973760932946, + "grad_norm": 0.12511715292930603, + "learning_rate": 1e-06, + "loss": -0.0328, + "num_tokens": 255044377.0, + "reward": 0.5636160969734192, + "reward_std": 0.16514213383197784, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 416 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3167.0, + "completions/mean_length": 786.529052734375, + "completions/mean_terminated_length": 623.7681274414062, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 3.8956268221574346, + "grad_norm": 0.1355588585138321, + "learning_rate": 1e-06, + "loss": -0.0147, + "num_tokens": 255670283.0, + "reward": 0.6149553656578064, + "reward_std": 0.18306472897529602, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2597.0, + "completions/mean_length": 714.9285888671875, + "completions/mean_terminated_length": 573.3953247070312, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 3.904956268221574, + "grad_norm": 0.12665985524654388, + "learning_rate": 1e-06, + "loss": -0.0168, + "num_tokens": 256254915.0, + "reward": 0.6462053656578064, + "reward_std": 0.14507634937763214, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 418 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3521.0, + "completions/mean_length": 825.068115234375, + "completions/mean_terminated_length": 643.9917602539062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 3.914285714285714, + "grad_norm": 0.13853387534618378, + "learning_rate": 1e-06, + "loss": -0.0059, + "num_tokens": 256886872.0, + "reward": 0.6305803656578064, + "reward_std": 0.19125288724899292, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 419 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3779.0, + "completions/mean_length": 714.802490234375, + "completions/mean_terminated_length": 581.4373168945312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.923615160349854, + "grad_norm": 0.14705388247966766, + "learning_rate": 1e-06, + "loss": -0.0251, + "num_tokens": 257475983.0, + "reward": 0.6595982313156128, + "reward_std": 0.19302275776863098, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 420 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3025.0, + "completions/mean_length": 812.0189819335938, + "completions/mean_terminated_length": 638.3654174804688, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.932944606413994, + "grad_norm": 0.14642846584320068, + "learning_rate": 1e-06, + "loss": -0.0125, + "num_tokens": 258112496.0, + "reward": 0.5837053656578064, + "reward_std": 0.19892574846744537, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 421 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 807.3471069335938, + "completions/mean_terminated_length": 617.0944213867188, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 3.942274052478134, + "grad_norm": 0.13945406675338745, + "learning_rate": 1e-06, + "loss": -0.0174, + "num_tokens": 258721183.0, + "reward": 0.6283482313156128, + "reward_std": 0.1816350817680359, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 422 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3032.0, + "completions/mean_length": 801.9620971679688, + "completions/mean_terminated_length": 644.0023193359375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.951603498542274, + "grad_norm": 0.13283725082874298, + "learning_rate": 1e-06, + "loss": -0.0174, + "num_tokens": 259362965.0, + "reward": 0.582589328289032, + "reward_std": 0.1793452799320221, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2691.0, + "completions/mean_length": 691.6808471679688, + "completions/mean_terminated_length": 577.810791015625, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 3.960932944606414, + "grad_norm": 0.14981742203235626, + "learning_rate": 1e-06, + "loss": -0.0154, + "num_tokens": 259946439.0, + "reward": 0.6662946939468384, + "reward_std": 0.20531336963176727, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 424 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 759.7299194335938, + "completions/mean_terminated_length": 656.0713500976562, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 3.970262390670554, + "grad_norm": 0.11969601362943649, + "learning_rate": 1e-06, + "loss": 0.0039, + "num_tokens": 260605157.0, + "reward": 0.621651828289032, + "reward_std": 0.15248315036296844, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 425 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3725.0, + "completions/mean_length": 705.3717041015625, + "completions/mean_terminated_length": 579.7928466796875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.979591836734694, + "grad_norm": 0.13710664212703705, + "learning_rate": 1e-06, + "loss": -0.0084, + "num_tokens": 261196378.0, + "reward": 0.6774553656578064, + "reward_std": 0.15935185551643372, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2558.0, + "completions/mean_length": 817.3683471679688, + "completions/mean_terminated_length": 611.2384643554688, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 3.9889212827988336, + "grad_norm": 0.11956760287284851, + "learning_rate": 1e-06, + "loss": -0.0141, + "num_tokens": 261802020.0, + "reward": 0.6595982313156128, + "reward_std": 0.14913280308246613, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 427 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022727272727272707, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3933.0, + "completions/mean_length": 682.8579711914062, + "completions/mean_terminated_length": 603.4825439453125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 3.9982507288629736, + "grad_norm": 0.1347777247428894, + "learning_rate": 1e-06, + "loss": -0.0132, + "num_tokens": 262418970.0, + "reward": 0.6662946939468384, + "reward_std": 0.1560114622116089, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 428 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3126.0, + "completions/mean_length": 751.8616333007812, + "completions/mean_terminated_length": 583.2825317382812, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 4.0093294460641395, + "grad_norm": 0.12279509007930756, + "learning_rate": 1e-06, + "loss": -0.0134, + "num_tokens": 263010606.0, + "reward": 0.6651785969734192, + "reward_std": 0.14868730306625366, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 429 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3668.0, + "completions/mean_length": 725.6585083007812, + "completions/mean_terminated_length": 608.9030151367188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 4.01865889212828, + "grad_norm": 0.1311519891023636, + "learning_rate": 1e-06, + "loss": -0.0143, + "num_tokens": 263622652.0, + "reward": 0.6361607313156128, + "reward_std": 0.1578957587480545, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 430 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3932.0, + "completions/mean_length": 799.4910888671875, + "completions/mean_terminated_length": 625.175048828125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 4.0279883381924195, + "grad_norm": 0.1415524035692215, + "learning_rate": 1e-06, + "loss": -0.0144, + "num_tokens": 264233068.0, + "reward": 0.6283482313156128, + "reward_std": 0.1713828295469284, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 431 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4082.0, + "completions/mean_length": 780.7935791015625, + "completions/mean_terminated_length": 654.0242919921875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.03731778425656, + "grad_norm": 0.1419406235218048, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 264883475.0, + "reward": 0.6707589626312256, + "reward_std": 0.1945943534374237, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 432 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2114.0, + "completions/mean_length": 810.6172485351562, + "completions/mean_terminated_length": 636.8895263671875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 4.0466472303206995, + "grad_norm": 0.15439121425151825, + "learning_rate": 1e-06, + "loss": -0.0024, + "num_tokens": 265510524.0, + "reward": 0.5993303656578064, + "reward_std": 0.17720521986484528, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 433 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3521.0, + "completions/mean_length": 834.6194458007812, + "completions/mean_terminated_length": 633.6812744140625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 4.05597667638484, + "grad_norm": 0.1669454276561737, + "learning_rate": 1e-06, + "loss": -0.0287, + "num_tokens": 266145223.0, + "reward": 0.637276828289032, + "reward_std": 0.23326674103736877, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 434 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3977.0, + "completions/mean_length": 671.0379638671875, + "completions/mean_terminated_length": 556.4774780273438, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 4.0653061224489795, + "grad_norm": 0.1450282335281372, + "learning_rate": 1e-06, + "loss": -0.0044, + "num_tokens": 266711689.0, + "reward": 0.6897321939468384, + "reward_std": 0.14289532601833344, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 435 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 756.732177734375, + "completions/mean_terminated_length": 612.8987426757812, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 4.07463556851312, + "grad_norm": 0.12799072265625, + "learning_rate": 1e-06, + "loss": -0.0235, + "num_tokens": 267328473.0, + "reward": 0.6595982313156128, + "reward_std": 0.1720554232597351, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 436 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2899.0, + "completions/mean_length": 827.2444458007812, + "completions/mean_terminated_length": 634.0555419921875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 4.0839650145772595, + "grad_norm": 0.13765114545822144, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 267954540.0, + "reward": 0.5959821939468384, + "reward_std": 0.17190389335155487, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 437 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3522.0, + "completions/mean_length": 872.2500610351562, + "completions/mean_terminated_length": 669.5706176757812, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 4.093294460641399, + "grad_norm": 0.13964110612869263, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 268600196.0, + "reward": 0.582589328289032, + "reward_std": 0.19956304132938385, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 438 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3300.0, + "completions/mean_length": 815.3995971679688, + "completions/mean_terminated_length": 662.1004638671875, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 4.1026239067055394, + "grad_norm": 0.12797996401786804, + "learning_rate": 1e-06, + "loss": -0.0006, + "num_tokens": 269257554.0, + "reward": 0.598214328289032, + "reward_std": 0.17946292459964752, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 439 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 824.0725708007812, + "completions/mean_terminated_length": 675.175048828125, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 4.111953352769679, + "grad_norm": 0.1300782859325409, + "learning_rate": 1e-06, + "loss": -0.0064, + "num_tokens": 269918267.0, + "reward": 0.5636160969734192, + "reward_std": 0.1803651601076126, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 440 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3767.0, + "completions/mean_length": 777.2824096679688, + "completions/mean_terminated_length": 654.3668823242188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.121282798833819, + "grad_norm": 0.13807030022144318, + "learning_rate": 1e-06, + "loss": 0.0169, + "num_tokens": 270567360.0, + "reward": 0.6462053656578064, + "reward_std": 0.17013472318649292, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 441 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3827.0, + "completions/mean_length": 771.0167846679688, + "completions/mean_terminated_length": 659.8004150390625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 4.130612244897959, + "grad_norm": 0.11922546476125717, + "learning_rate": 1e-06, + "loss": 0.0049, + "num_tokens": 271229703.0, + "reward": 0.6205357313156128, + "reward_std": 0.15555571019649506, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 442 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 785.0245971679688, + "completions/mean_terminated_length": 630.3060302734375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.139941690962099, + "grad_norm": 0.14290083944797516, + "learning_rate": 1e-06, + "loss": -0.0095, + "num_tokens": 271867605.0, + "reward": 0.6205357313156128, + "reward_std": 0.18878155946731567, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 443 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3378.0, + "completions/mean_length": 732.6864013671875, + "completions/mean_terminated_length": 595.96630859375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.149271137026239, + "grad_norm": 0.1580958366394043, + "learning_rate": 1e-06, + "loss": -0.027, + "num_tokens": 272477836.0, + "reward": 0.6473214626312256, + "reward_std": 0.1798744946718216, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2716.0, + "completions/mean_length": 761.0234985351562, + "completions/mean_terminated_length": 597.0081787109375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 4.158600583090379, + "grad_norm": 0.1331312358379364, + "learning_rate": 1e-06, + "loss": -0.0211, + "num_tokens": 273078689.0, + "reward": 0.6595982313156128, + "reward_std": 0.1692011058330536, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3704.0, + "completions/mean_length": 785.107177734375, + "completions/mean_terminated_length": 638.4708862304688, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 4.167930029154519, + "grad_norm": 0.13704122602939606, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 273712521.0, + "reward": 0.660714328289032, + "reward_std": 0.19730742275714874, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4061.0, + "completions/mean_length": 833.6752319335938, + "completions/mean_terminated_length": 628.5706176757812, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 4.1772594752186585, + "grad_norm": 0.14558477699756622, + "learning_rate": 1e-06, + "loss": -0.0149, + "num_tokens": 274332814.0, + "reward": 0.6205357313156128, + "reward_std": 0.1950504183769226, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 447 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3078.0, + "completions/mean_length": 763.6875610351562, + "completions/mean_terminated_length": 628.2276000976562, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 4.186588921282799, + "grad_norm": 0.13810819387435913, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 274969822.0, + "reward": 0.6395089626312256, + "reward_std": 0.19866472482681274, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 448 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3627.0, + "completions/mean_length": 851.575927734375, + "completions/mean_terminated_length": 647.5966796875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.1959183673469385, + "grad_norm": 0.14547944068908691, + "learning_rate": 1e-06, + "loss": -0.019, + "num_tokens": 275604866.0, + "reward": 0.6082589626312256, + "reward_std": 0.1976444274187088, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 449 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4018.0, + "completions/mean_length": 778.5379638671875, + "completions/mean_terminated_length": 611.3035888671875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 4.205247813411079, + "grad_norm": 0.1418633908033371, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 276214644.0, + "reward": 0.640625, + "reward_std": 0.16108639538288116, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 450 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2922.0, + "completions/mean_length": 774.700927734375, + "completions/mean_terminated_length": 639.688720703125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.214577259475218, + "grad_norm": 0.1449412852525711, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 276856936.0, + "reward": 0.645089328289032, + "reward_std": 0.1911141276359558, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 451 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3241.0, + "completions/mean_length": 775.5022583007812, + "completions/mean_terminated_length": 595.8046875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 4.223906705539359, + "grad_norm": 0.1295343041419983, + "learning_rate": 1e-06, + "loss": -0.0236, + "num_tokens": 277450338.0, + "reward": 0.6160714626312256, + "reward_std": 0.1602526754140854, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 452 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4081.0, + "completions/mean_length": 856.4408569335938, + "completions/mean_terminated_length": 677.101318359375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 4.233236151603498, + "grad_norm": 0.13079296052455902, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 278119077.0, + "reward": 0.5814732313156128, + "reward_std": 0.18411709368228912, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 453 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3469.0, + "completions/mean_length": 788.154052734375, + "completions/mean_terminated_length": 637.6219482421875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 4.242565597667639, + "grad_norm": 0.13982364535331726, + "learning_rate": 1e-06, + "loss": -0.0117, + "num_tokens": 278762287.0, + "reward": 0.6484375, + "reward_std": 0.18310680985450745, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 454 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3498.0, + "completions/mean_length": 781.599365234375, + "completions/mean_terminated_length": 626.7207641601562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 4.251895043731778, + "grad_norm": 0.14324288070201874, + "learning_rate": 1e-06, + "loss": -0.0147, + "num_tokens": 279383744.0, + "reward": 0.6696428656578064, + "reward_std": 0.18539661169052124, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 455 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3100.0, + "completions/mean_length": 784.9308471679688, + "completions/mean_terminated_length": 605.7435302734375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 4.261224489795918, + "grad_norm": 0.13799770176410675, + "learning_rate": 1e-06, + "loss": -0.0069, + "num_tokens": 279983458.0, + "reward": 0.6506696939468384, + "reward_std": 0.16848501563072205, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 456 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3605.0, + "completions/mean_length": 858.4397583007812, + "completions/mean_terminated_length": 630.224609375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 4.270553935860058, + "grad_norm": 0.14292599260807037, + "learning_rate": 1e-06, + "loss": -0.0274, + "num_tokens": 280596588.0, + "reward": 0.5915178656578064, + "reward_std": 0.17528633773326874, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3108.0, + "completions/mean_length": 764.8660888671875, + "completions/mean_terminated_length": 641.49072265625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 4.279883381924198, + "grad_norm": 0.13144493103027344, + "learning_rate": 1e-06, + "loss": -0.0076, + "num_tokens": 281236508.0, + "reward": 0.6049107313156128, + "reward_std": 0.17566722631454468, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914289474487305, + "step": 458 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2166.0, + "completions/mean_length": 783.1250610351562, + "completions/mean_terminated_length": 616.1218872070312, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 4.289212827988338, + "grad_norm": 0.1330137550830841, + "learning_rate": 1e-06, + "loss": -0.0095, + "num_tokens": 281850836.0, + "reward": 0.6350446939468384, + "reward_std": 0.16679435968399048, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 459 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2706.0, + "completions/mean_length": 835.6160888671875, + "completions/mean_terminated_length": 675.269287109375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.298542274052478, + "grad_norm": 0.12440939247608185, + "learning_rate": 1e-06, + "loss": 0.0039, + "num_tokens": 282527740.0, + "reward": 0.5758928656578064, + "reward_std": 0.17382346093654633, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 460 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2978.0, + "completions/mean_length": 827.8951416015625, + "completions/mean_terminated_length": 663.1488647460938, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 4.307871720116618, + "grad_norm": 0.14425607025623322, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 283184758.0, + "reward": 0.5915178656578064, + "reward_std": 0.1941513866186142, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2617.0, + "completions/mean_length": 745.8326416015625, + "completions/mean_terminated_length": 605.593017578125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 4.317201166180758, + "grad_norm": 0.1216999962925911, + "learning_rate": 1e-06, + "loss": 0.0067, + "num_tokens": 283800224.0, + "reward": 0.6741071939468384, + "reward_std": 0.1309078186750412, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 462 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2374.0, + "completions/mean_length": 796.9285888671875, + "completions/mean_terminated_length": 614.2944946289062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.326530612244898, + "grad_norm": 0.1325484812259674, + "learning_rate": 1e-06, + "loss": -0.0225, + "num_tokens": 284412288.0, + "reward": 0.6037946939468384, + "reward_std": 0.16032865643501282, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938119411468506, + "step": 463 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2564.0, + "completions/mean_length": 755.8248291015625, + "completions/mean_terminated_length": 628.1007690429688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.335860058309038, + "grad_norm": 0.12685535848140717, + "learning_rate": 1e-06, + "loss": -0.0147, + "num_tokens": 285041155.0, + "reward": 0.6852678656578064, + "reward_std": 0.16337548196315765, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 464 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2365.0, + "completions/mean_length": 718.0547485351562, + "completions/mean_terminated_length": 568.44873046875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.345189504373177, + "grad_norm": 0.15057477355003357, + "learning_rate": 1e-06, + "loss": 0.0172, + "num_tokens": 285615828.0, + "reward": 0.7008928656578064, + "reward_std": 0.14943771064281464, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 465 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 778.9241333007812, + "completions/mean_terminated_length": 587.0271606445312, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.354518950437318, + "grad_norm": 0.12886711955070496, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 286200688.0, + "reward": 0.6975446939468384, + "reward_std": 0.14774522185325623, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3146.0, + "completions/mean_length": 716.2355346679688, + "completions/mean_terminated_length": 574.7569580078125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 4.363848396501457, + "grad_norm": 0.13811829686164856, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 286782443.0, + "reward": 0.6540178656578064, + "reward_std": 0.16766197979450226, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 788.4285888671875, + "completions/mean_terminated_length": 617.6150512695312, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 4.373177842565598, + "grad_norm": 0.13392016291618347, + "learning_rate": 1e-06, + "loss": -0.0044, + "num_tokens": 287394275.0, + "reward": 0.6540178656578064, + "reward_std": 0.17434383928775787, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 468 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4012.0, + "completions/mean_length": 830.255615234375, + "completions/mean_terminated_length": 637.2446899414062, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 4.382507288629737, + "grad_norm": 0.1388266235589981, + "learning_rate": 1e-06, + "loss": -0.0137, + "num_tokens": 288023304.0, + "reward": 0.6160714626312256, + "reward_std": 0.16308267414569855, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 469 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3750.0, + "completions/mean_length": 825.2020263671875, + "completions/mean_terminated_length": 656.28759765625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 4.391836734693878, + "grad_norm": 0.13713406026363373, + "learning_rate": 1e-06, + "loss": -0.0311, + "num_tokens": 288678477.0, + "reward": 0.5770089626312256, + "reward_std": 0.17791494727134705, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099617958069, + "step": 470 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 866.1417846679688, + "completions/mean_terminated_length": 654.9143676757812, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.401166180758017, + "grad_norm": 0.13769063353538513, + "learning_rate": 1e-06, + "loss": -0.0215, + "num_tokens": 289326892.0, + "reward": 0.6227678656578064, + "reward_std": 0.18130694329738617, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 471 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3425.0, + "completions/mean_length": 835.146240234375, + "completions/mean_terminated_length": 605.2891235351562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.410495626822158, + "grad_norm": 0.14611920714378357, + "learning_rate": 1e-06, + "loss": -0.0139, + "num_tokens": 289924967.0, + "reward": 0.629464328289032, + "reward_std": 0.18708838522434235, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 472 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2645.0, + "completions/mean_length": 792.5881958007812, + "completions/mean_terminated_length": 572.3607177734375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 4.419825072886297, + "grad_norm": 0.16382291913032532, + "learning_rate": 1e-06, + "loss": -0.0028, + "num_tokens": 290497574.0, + "reward": 0.6439732313156128, + "reward_std": 0.18900977075099945, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 473 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2519.0, + "completions/mean_length": 882.0469360351562, + "completions/mean_terminated_length": 618.0989990234375, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 4.429154518950437, + "grad_norm": 0.14638003706932068, + "learning_rate": 1e-06, + "loss": -0.0443, + "num_tokens": 291100328.0, + "reward": 0.6205357313156128, + "reward_std": 0.19561424851417542, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 474 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4046.0, + "completions/mean_length": 990.79248046875, + "completions/mean_terminated_length": 665.3390502929688, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 4.438483965014577, + "grad_norm": 0.12503956258296967, + "learning_rate": 1e-06, + "loss": -0.0169, + "num_tokens": 291731166.0, + "reward": 0.5457589626312256, + "reward_std": 0.13696233928203583, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981797933578491, + "step": 475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2723.0, + "completions/mean_length": 815.4453735351562, + "completions/mean_terminated_length": 592.5709228515625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 4.447813411078717, + "grad_norm": 0.14756697416305542, + "learning_rate": 1e-06, + "loss": -0.0212, + "num_tokens": 292317341.0, + "reward": 0.6651785969734192, + "reward_std": 0.15462210774421692, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 476 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4033.0, + "completions/mean_length": 789.5346069335938, + "completions/mean_terminated_length": 594.1170043945312, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 4.457142857142857, + "grad_norm": 0.1497909426689148, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 292905804.0, + "reward": 0.6886160969734192, + "reward_std": 0.16909484565258026, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 477 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4072.0, + "completions/mean_length": 842.0279541015625, + "completions/mean_terminated_length": 637.4484252929688, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 4.466472303206997, + "grad_norm": 0.13087114691734314, + "learning_rate": 1e-06, + "loss": -0.0112, + "num_tokens": 293530669.0, + "reward": 0.6640625, + "reward_std": 0.1670207679271698, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 478 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3688.0, + "completions/mean_length": 869.3359985351562, + "completions/mean_terminated_length": 604.3441772460938, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 4.475801749271137, + "grad_norm": 0.16542375087738037, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 294121850.0, + "reward": 0.6551339626312256, + "reward_std": 0.19587568938732147, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 479 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3998.0, + "completions/mean_length": 908.169677734375, + "completions/mean_terminated_length": 654.6795043945312, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.485131195335277, + "grad_norm": 0.1321999430656433, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 294759058.0, + "reward": 0.6004464626312256, + "reward_std": 0.16615313291549683, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 480 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3792.0, + "completions/mean_length": 799.552490234375, + "completions/mean_terminated_length": 579.789306640625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 4.494460641399417, + "grad_norm": 0.15272067487239838, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 295332857.0, + "reward": 0.6350446939468384, + "reward_std": 0.17397361993789673, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3684.0, + "completions/mean_length": 860.7678833007812, + "completions/mean_terminated_length": 620.259033203125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 4.503790087463557, + "grad_norm": 0.1289597749710083, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 295942393.0, + "reward": 0.6194196939468384, + "reward_std": 0.16078399121761322, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 482 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2667.0, + "completions/mean_length": 752.2991333007812, + "completions/mean_terminated_length": 558.8618774414062, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 4.513119533527696, + "grad_norm": 0.1325407475233078, + "learning_rate": 1e-06, + "loss": -0.0043, + "num_tokens": 296495917.0, + "reward": 0.6986607313156128, + "reward_std": 0.14135579764842987, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 483 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 857.3772583007812, + "completions/mean_terminated_length": 665.96923828125, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 4.522448979591837, + "grad_norm": 0.12853114306926727, + "learning_rate": 1e-06, + "loss": -0.0288, + "num_tokens": 297141623.0, + "reward": 0.598214328289032, + "reward_std": 0.1748412549495697, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3001.0, + "completions/mean_length": 929.9766235351562, + "completions/mean_terminated_length": 653.333740234375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 4.531778425655976, + "grad_norm": 0.13542483747005463, + "learning_rate": 1e-06, + "loss": -0.0104, + "num_tokens": 297772850.0, + "reward": 0.5714285969734192, + "reward_std": 0.1537216752767563, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514806270599365, + "step": 485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3209.0, + "completions/mean_length": 744.8426513671875, + "completions/mean_terminated_length": 592.3395385742188, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 4.541107871720117, + "grad_norm": 0.12116781622171402, + "learning_rate": 1e-06, + "loss": -0.0197, + "num_tokens": 298358837.0, + "reward": 0.6696428656578064, + "reward_std": 0.14478282630443573, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 486 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3558.0, + "completions/mean_length": 846.4531860351562, + "completions/mean_terminated_length": 642.15185546875, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 4.550437317784256, + "grad_norm": 0.13275185227394104, + "learning_rate": 1e-06, + "loss": -0.0217, + "num_tokens": 298986179.0, + "reward": 0.6696428656578064, + "reward_std": 0.15639054775238037, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 487 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2621.0, + "completions/mean_length": 767.2589721679688, + "completions/mean_terminated_length": 570.5248413085938, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 4.559766763848397, + "grad_norm": 0.1290135681629181, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 299554739.0, + "reward": 0.7321428656578064, + "reward_std": 0.1534174680709839, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 488 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 925.091552734375, + "completions/mean_terminated_length": 660.5296630859375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 4.569096209912536, + "grad_norm": 0.13341110944747925, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 300194437.0, + "reward": 0.6283482313156128, + "reward_std": 0.17712965607643127, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 489 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2330.0, + "completions/mean_length": 887.5089721679688, + "completions/mean_terminated_length": 636.5439453125, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.578425655976677, + "grad_norm": 0.13713081181049347, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 300820957.0, + "reward": 0.5770089626312256, + "reward_std": 0.16901704668998718, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 490 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2256.0, + "completions/mean_length": 774.6094360351562, + "completions/mean_terminated_length": 598.9776611328125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.587755102040816, + "grad_norm": 0.12140315771102905, + "learning_rate": 1e-06, + "loss": -0.0183, + "num_tokens": 301416527.0, + "reward": 0.6897321939468384, + "reward_std": 0.13121342658996582, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 491 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3708.0, + "completions/mean_length": 764.3248291015625, + "completions/mean_terminated_length": 604.5602416992188, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 4.597084548104956, + "grad_norm": 0.1476830393075943, + "learning_rate": 1e-06, + "loss": -0.0282, + "num_tokens": 302021730.0, + "reward": 0.6506696939468384, + "reward_std": 0.1975356638431549, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 492 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3132.0, + "completions/mean_length": 842.5335083007812, + "completions/mean_terminated_length": 609.0310668945312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 4.606413994169096, + "grad_norm": 0.14848507940769196, + "learning_rate": 1e-06, + "loss": -0.0385, + "num_tokens": 302616672.0, + "reward": 0.6149553656578064, + "reward_std": 0.1726602166891098, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 493 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3831.0, + "completions/mean_length": 835.6975708007812, + "completions/mean_terminated_length": 638.921875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 4.615743440233236, + "grad_norm": 0.1602596938610077, + "learning_rate": 1e-06, + "loss": -0.0325, + "num_tokens": 303235945.0, + "reward": 0.6540178656578064, + "reward_std": 0.20834991335868835, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 494 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3141.0, + "completions/mean_length": 790.6875610351562, + "completions/mean_terminated_length": 591.1952514648438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 4.625072886297376, + "grad_norm": 0.13296261429786682, + "learning_rate": 1e-06, + "loss": -0.0034, + "num_tokens": 303821137.0, + "reward": 0.684151828289032, + "reward_std": 0.16078399121761322, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 495 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3589.0, + "completions/mean_length": 861.169677734375, + "completions/mean_terminated_length": 637.2792358398438, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 4.634402332361516, + "grad_norm": 0.13468492031097412, + "learning_rate": 1e-06, + "loss": -0.0006, + "num_tokens": 304446937.0, + "reward": 0.6037946939468384, + "reward_std": 0.160858154296875, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 496 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3223.0, + "completions/mean_length": 699.6250610351562, + "completions/mean_terminated_length": 594.0989379882812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 4.643731778425656, + "grad_norm": 0.1275583803653717, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 305046041.0, + "reward": 0.6986607313156128, + "reward_std": 0.158384308218956, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 497 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 936.1138916015625, + "completions/mean_terminated_length": 672.4716186523438, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 4.653061224489796, + "grad_norm": 0.1379220336675644, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 305691711.0, + "reward": 0.6127232313156128, + "reward_std": 0.16661031544208527, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3874.0, + "completions/mean_length": 811.732177734375, + "completions/mean_terminated_length": 625.8302001953125, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 4.662390670553936, + "grad_norm": 0.12915872037410736, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 306313447.0, + "reward": 0.707589328289032, + "reward_std": 0.15492630004882812, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 499 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 784.638427734375, + "completions/mean_terminated_length": 637.9813842773438, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 4.671720116618076, + "grad_norm": 0.1429835557937622, + "learning_rate": 1e-06, + "loss": -0.0143, + "num_tokens": 306951979.0, + "reward": 0.6361607313156128, + "reward_std": 0.15852878987789154, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 500 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3818.0, + "completions/mean_length": 820.997802734375, + "completions/mean_terminated_length": 655.9038696289062, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 4.681049562682215, + "grad_norm": 0.1325433999300003, + "learning_rate": 1e-06, + "loss": -0.001, + "num_tokens": 307606665.0, + "reward": 0.6116071939468384, + "reward_std": 0.16319255530834198, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 501 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3590.0, + "completions/mean_length": 861.6272583007812, + "completions/mean_terminated_length": 666.4165649414062, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 4.690379008746356, + "grad_norm": 0.13213413953781128, + "learning_rate": 1e-06, + "loss": -0.0138, + "num_tokens": 308268363.0, + "reward": 0.5680803656578064, + "reward_std": 0.18873737752437592, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3969.0, + "completions/mean_length": 950.4364013671875, + "completions/mean_terminated_length": 637.809814453125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 4.699708454810495, + "grad_norm": 0.1319461613893509, + "learning_rate": 1e-06, + "loss": -0.0294, + "num_tokens": 308882458.0, + "reward": 0.5959821939468384, + "reward_std": 0.17596031725406647, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3957.0, + "completions/mean_length": 933.9342041015625, + "completions/mean_terminated_length": 670.1100463867188, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 4.709037900874636, + "grad_norm": 0.12873594462871552, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 309528231.0, + "reward": 0.6127232313156128, + "reward_std": 0.1703290343284607, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 504 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2476.0, + "completions/mean_length": 894.0045166015625, + "completions/mean_terminated_length": 639.387939453125, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.718367346938775, + "grad_norm": 0.13246700167655945, + "learning_rate": 1e-06, + "loss": -0.0351, + "num_tokens": 310153691.0, + "reward": 0.6037946939468384, + "reward_std": 0.16484157741069794, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2955.0, + "completions/mean_length": 861.2377319335938, + "completions/mean_terminated_length": 649.6896362304688, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.727696793002916, + "grad_norm": 0.12984134256839752, + "learning_rate": 1e-06, + "loss": -0.0152, + "num_tokens": 310789592.0, + "reward": 0.5959821939468384, + "reward_std": 0.13763564825057983, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 506 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3595.0, + "completions/mean_length": 791.2344360351562, + "completions/mean_terminated_length": 604.1721801757812, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.737026239067055, + "grad_norm": 0.1601531207561493, + "learning_rate": 1e-06, + "loss": -0.0167, + "num_tokens": 311400042.0, + "reward": 0.668526828289032, + "reward_std": 0.17220766842365265, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 507 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3078.0, + "completions/mean_length": 977.1172485351562, + "completions/mean_terminated_length": 679.7176513671875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 4.746355685131196, + "grad_norm": 0.12345530837774277, + "learning_rate": 1e-06, + "loss": -0.0271, + "num_tokens": 312050763.0, + "reward": 0.5580357313156128, + "reward_std": 0.17487584054470062, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 508 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3606.0, + "completions/mean_length": 799.372802734375, + "completions/mean_terminated_length": 592.1115112304688, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 4.755685131195335, + "grad_norm": 0.16433003544807434, + "learning_rate": 1e-06, + "loss": -0.0096, + "num_tokens": 312643537.0, + "reward": 0.6350446939468384, + "reward_std": 0.1663813591003418, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 509 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4022.0, + "completions/mean_length": 840.7076416015625, + "completions/mean_terminated_length": 623.6881103515625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.765014577259475, + "grad_norm": 0.14345619082450867, + "learning_rate": 1e-06, + "loss": -0.0313, + "num_tokens": 313252035.0, + "reward": 0.7220982313156128, + "reward_std": 0.17288027703762054, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 510 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4091.0, + "completions/mean_length": 904.0256958007812, + "completions/mean_terminated_length": 666.7326049804688, + "completions/min_length": 209.0, + "completions/min_terminated_length": 209.0, + "epoch": 4.774344023323615, + "grad_norm": 0.11636216193437576, + "learning_rate": 1e-06, + "loss": -0.025, + "num_tokens": 313911074.0, + "reward": 0.6160714626312256, + "reward_std": 0.17682726681232452, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 511 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4093.0, + "completions/mean_length": 796.8314819335938, + "completions/mean_terminated_length": 555.8143920898438, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 4.783673469387755, + "grad_norm": 0.14117732644081116, + "learning_rate": 1e-06, + "loss": -0.007, + "num_tokens": 314460227.0, + "reward": 0.6618303656578064, + "reward_std": 0.15492448210716248, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3302.0, + "completions/mean_length": 824.9029541015625, + "completions/mean_terminated_length": 627.4757080078125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.793002915451895, + "grad_norm": 0.1553492546081543, + "learning_rate": 1e-06, + "loss": -0.017, + "num_tokens": 315082756.0, + "reward": 0.6037946939468384, + "reward_std": 0.16769729554653168, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 513 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3761.0, + "completions/mean_length": 861.6127319335938, + "completions/mean_terminated_length": 621.1666870117188, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 4.802332361516035, + "grad_norm": 0.1396396905183792, + "learning_rate": 1e-06, + "loss": -0.0193, + "num_tokens": 315681697.0, + "reward": 0.6227678656578064, + "reward_std": 0.17400752007961273, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 514 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2528.0, + "completions/mean_length": 873.8370971679688, + "completions/mean_terminated_length": 625.9783935546875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 4.811661807580175, + "grad_norm": 0.1418338418006897, + "learning_rate": 1e-06, + "loss": -0.0276, + "num_tokens": 316299287.0, + "reward": 0.5948660969734192, + "reward_std": 0.17002595961093903, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 515 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3122.0, + "completions/mean_length": 887.2969360351562, + "completions/mean_terminated_length": 636.3153076171875, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 4.820991253644315, + "grad_norm": 0.15696078538894653, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 316922161.0, + "reward": 0.5770089626312256, + "reward_std": 0.22053135931491852, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 516 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3438.0, + "completions/mean_length": 900.1473388671875, + "completions/mean_terminated_length": 650.1708984375, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.830320699708455, + "grad_norm": 0.14191314578056335, + "learning_rate": 1e-06, + "loss": -0.0064, + "num_tokens": 317551325.0, + "reward": 0.5948660969734192, + "reward_std": 0.18498431146144867, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 517 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 892.0960083007812, + "completions/mean_terminated_length": 620.5787353515625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 4.839650145772595, + "grad_norm": 0.12528476119041443, + "learning_rate": 1e-06, + "loss": -0.0089, + "num_tokens": 318146235.0, + "reward": 0.6618303656578064, + "reward_std": 0.13981416821479797, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3280.0, + "completions/mean_length": 893.6339721679688, + "completions/mean_terminated_length": 634.81787109375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 4.848979591836734, + "grad_norm": 0.12319959700107574, + "learning_rate": 1e-06, + "loss": -0.022, + "num_tokens": 318755603.0, + "reward": 0.6183035969734192, + "reward_std": 0.14240817725658417, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 519 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4029.0, + "completions/mean_length": 916.6920166015625, + "completions/mean_terminated_length": 668.0096435546875, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 4.858309037900875, + "grad_norm": 0.12649740278720856, + "learning_rate": 1e-06, + "loss": -0.0111, + "num_tokens": 319398551.0, + "reward": 0.5625, + "reward_std": 0.16224895417690277, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 520 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3595.0, + "completions/mean_length": 867.388427734375, + "completions/mean_terminated_length": 639.8040161132812, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.867638483965014, + "grad_norm": 0.14901678264141083, + "learning_rate": 1e-06, + "loss": -0.0114, + "num_tokens": 320026955.0, + "reward": 0.6428571939468384, + "reward_std": 0.17029766738414764, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 521 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 764.8348388671875, + "completions/mean_terminated_length": 576.2783203125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.876967930029155, + "grad_norm": 0.15150505304336548, + "learning_rate": 1e-06, + "loss": -0.0349, + "num_tokens": 320600607.0, + "reward": 0.7064732313156128, + "reward_std": 0.17123058438301086, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 522 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 848.5145263671875, + "completions/mean_terminated_length": 632.0155029296875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "epoch": 4.886297376093294, + "grad_norm": 0.16764754056930542, + "learning_rate": 1e-06, + "loss": -0.018, + "num_tokens": 321227900.0, + "reward": 0.5770089626312256, + "reward_std": 0.19899991154670715, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 523 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3416.0, + "completions/mean_length": 848.9844360351562, + "completions/mean_terminated_length": 586.5596923828125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.895626822157435, + "grad_norm": 0.1350231170654297, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 321805462.0, + "reward": 0.6205357313156128, + "reward_std": 0.14789676666259766, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 524 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2659.0, + "completions/mean_length": 1015.8739013671875, + "completions/mean_terminated_length": 688.8482055664062, + "completions/min_length": 198.0, + "completions/min_terminated_length": 198.0, + "epoch": 4.904956268221574, + "grad_norm": 0.1300228089094162, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 322459605.0, + "reward": 0.5602678656578064, + "reward_std": 0.1781495362520218, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 525 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3899.0, + "completions/mean_length": 999.6094360351562, + "completions/mean_terminated_length": 628.04248046875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 4.914285714285715, + "grad_norm": 0.14187702536582947, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 323053855.0, + "reward": 0.5837053656578064, + "reward_std": 0.191293865442276, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 526 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3691.0, + "completions/mean_length": 911.6250610351562, + "completions/mean_terminated_length": 603.7111206054688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.923615160349854, + "grad_norm": 0.14171014726161957, + "learning_rate": 1e-06, + "loss": -0.0213, + "num_tokens": 323636551.0, + "reward": 0.6004464626312256, + "reward_std": 0.16213877499103546, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 527 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3565.0, + "completions/mean_length": 953.7489013671875, + "completions/mean_terminated_length": 641.4515380859375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 4.932944606413994, + "grad_norm": 0.1304120421409607, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 324248366.0, + "reward": 0.6361607313156128, + "reward_std": 0.172769695520401, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 528 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3349.0, + "completions/mean_length": 846.3683471679688, + "completions/mean_terminated_length": 600.5978393554688, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 4.942274052478134, + "grad_norm": 0.15002506971359253, + "learning_rate": 1e-06, + "loss": -0.0327, + "num_tokens": 324828776.0, + "reward": 0.6551339626312256, + "reward_std": 0.20012785494327545, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 529 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2298.0, + "completions/mean_length": 828.1551513671875, + "completions/mean_terminated_length": 597.8052368164062, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 4.9516034985422746, + "grad_norm": 0.1383824646472931, + "learning_rate": 1e-06, + "loss": -0.0257, + "num_tokens": 325418947.0, + "reward": 0.6741071939468384, + "reward_std": 0.14744281768798828, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 530 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3979.0, + "completions/mean_length": 826.5045166015625, + "completions/mean_terminated_length": 553.717041015625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 4.960932944606414, + "grad_norm": 0.12861518561840057, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 325961119.0, + "reward": 0.7165178656578064, + "reward_std": 0.13601915538311005, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 531 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2416.0, + "completions/mean_length": 906.8672485351562, + "completions/mean_terminated_length": 607.0341796875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 4.970262390670554, + "grad_norm": 0.15313492715358734, + "learning_rate": 1e-06, + "loss": -0.0415, + "num_tokens": 326537968.0, + "reward": 0.6618303656578064, + "reward_std": 0.17908243834972382, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 532 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3499.0, + "completions/mean_length": 912.76123046875, + "completions/mean_terminated_length": 621.9659423828125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 4.979591836734694, + "grad_norm": 0.13408954441547394, + "learning_rate": 1e-06, + "loss": -0.0133, + "num_tokens": 327141306.0, + "reward": 0.6127232313156128, + "reward_std": 0.14414413273334503, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3083.0, + "completions/mean_length": 917.38623046875, + "completions/mean_terminated_length": 605.7573852539062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.988921282798834, + "grad_norm": 0.15228623151779175, + "learning_rate": 1e-06, + "loss": -0.0245, + "num_tokens": 327727812.0, + "reward": 0.5959821939468384, + "reward_std": 0.15649932622909546, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.06818181818181823, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2333.0, + "completions/mean_length": 805.5341186523438, + "completions/mean_terminated_length": 564.7682495117188, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 4.998250728862974, + "grad_norm": 0.1271304488182068, + "learning_rate": 1e-06, + "loss": -0.026, + "num_tokens": 328313761.0, + "reward": 0.6160714626312256, + "reward_std": 0.14038008451461792, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 535 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 863.4141235351562, + "completions/mean_terminated_length": 589.4661254882812, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 5.0093294460641395, + "grad_norm": 0.14470799267292023, + "learning_rate": 1e-06, + "loss": -0.0331, + "num_tokens": 328878252.0, + "reward": 0.691964328289032, + "reward_std": 0.16792160272598267, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 536 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3960.0, + "completions/mean_length": 925.5480346679688, + "completions/mean_terminated_length": 656.8656616210938, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.01865889212828, + "grad_norm": 0.13118726015090942, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 329502543.0, + "reward": 0.598214328289032, + "reward_std": 0.18490944802761078, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 537 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 980.33935546875, + "completions/mean_terminated_length": 658.029541015625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.0279883381924195, + "grad_norm": 0.15592887997627258, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 330127223.0, + "reward": 0.6004464626312256, + "reward_std": 0.1675853133201599, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 538 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 868.8850708007812, + "completions/mean_terminated_length": 603.8562622070312, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 5.03731778425656, + "grad_norm": 0.13814255595207214, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 330719824.0, + "reward": 0.629464328289032, + "reward_std": 0.154136061668396, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 833.0145263671875, + "completions/mean_terminated_length": 631.9775390625, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 5.0466472303206995, + "grad_norm": 0.13625772297382355, + "learning_rate": 1e-06, + "loss": -0.0345, + "num_tokens": 331336325.0, + "reward": 0.6674107313156128, + "reward_std": 0.18074172735214233, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 540 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3050.0, + "completions/mean_length": 841.7176513671875, + "completions/mean_terminated_length": 603.9796752929688, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 5.05597667638484, + "grad_norm": 0.15228751301765442, + "learning_rate": 1e-06, + "loss": -0.0325, + "num_tokens": 331923576.0, + "reward": 0.637276828289032, + "reward_std": 0.1720554232597351, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 541 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3485.0, + "completions/mean_length": 853.9230346679688, + "completions/mean_terminated_length": 600.3309326171875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 5.0653061224489795, + "grad_norm": 0.14871995151042938, + "learning_rate": 1e-06, + "loss": -0.025, + "num_tokens": 332504987.0, + "reward": 0.6462053656578064, + "reward_std": 0.18580886721611023, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 542 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3389.0, + "completions/mean_length": 850.8605346679688, + "completions/mean_terminated_length": 575.8486938476562, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.07463556851312, + "grad_norm": 0.13397301733493805, + "learning_rate": 1e-06, + "loss": -0.0079, + "num_tokens": 333068086.0, + "reward": 0.6305803656578064, + "reward_std": 0.14203797280788422, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 543 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 831.1127319335938, + "completions/mean_terminated_length": 588.3992919921875, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 5.0839650145772595, + "grad_norm": 0.1399272233247757, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 333652315.0, + "reward": 0.6540178656578064, + "reward_std": 0.15082639455795288, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 544 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2549.0, + "completions/mean_length": 952.5402221679688, + "completions/mean_terminated_length": 614.491943359375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.093294460641399, + "grad_norm": 0.12391702085733414, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 334248695.0, + "reward": 0.5558035969734192, + "reward_std": 0.14015227556228638, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715372920036316, + "step": 545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 810.7846069335938, + "completions/mean_terminated_length": 587.5935668945312, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 5.1026239067055394, + "grad_norm": 0.14505510032176971, + "learning_rate": 1e-06, + "loss": -0.0426, + "num_tokens": 334824654.0, + "reward": 0.6595982313156128, + "reward_std": 0.19058029353618622, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 546 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3529.0, + "completions/mean_length": 805.1027221679688, + "completions/mean_terminated_length": 585.7095336914062, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.111953352769679, + "grad_norm": 0.1331314742565155, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 335406666.0, + "reward": 0.6875000596046448, + "reward_std": 0.16033712029457092, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3445.0, + "completions/mean_length": 781.4888916015625, + "completions/mean_terminated_length": 577.2772827148438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.121282798833819, + "grad_norm": 0.13322143256664276, + "learning_rate": 1e-06, + "loss": -0.0246, + "num_tokens": 335975856.0, + "reward": 0.7265625596046448, + "reward_std": 0.1385032683610916, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 548 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4018.0, + "completions/mean_length": 831.0279541015625, + "completions/mean_terminated_length": 642.1452026367188, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 5.130612244897959, + "grad_norm": 0.12781395018100739, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 336610745.0, + "reward": 0.65625, + "reward_std": 0.13947968184947968, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 549 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3529.0, + "completions/mean_length": 972.4531860351562, + "completions/mean_terminated_length": 606.3516235351562, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 5.139941690962099, + "grad_norm": 0.1296423226594925, + "learning_rate": 1e-06, + "loss": -0.0243, + "num_tokens": 337196295.0, + "reward": 0.5323660969734192, + "reward_std": 0.14782078564167023, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 550 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3478.0, + "completions/mean_length": 789.4207763671875, + "completions/mean_terminated_length": 568.982177734375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 5.149271137026239, + "grad_norm": 0.13754911720752716, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 337762576.0, + "reward": 0.6785714626312256, + "reward_std": 0.14879578351974487, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 551 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3965.0, + "completions/mean_length": 877.2288208007812, + "completions/mean_terminated_length": 662.64404296875, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 5.158600583090379, + "grad_norm": 0.14513398706912994, + "learning_rate": 1e-06, + "loss": -0.0259, + "num_tokens": 338406477.0, + "reward": 0.6462053656578064, + "reward_std": 0.20610612630844116, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 552 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2901.0, + "completions/mean_length": 853.2957763671875, + "completions/mean_terminated_length": 582.74365234375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.167930029154519, + "grad_norm": 0.12653854489326477, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 338974630.0, + "reward": 0.640625, + "reward_std": 0.1569557934999466, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 553 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2789.0, + "completions/mean_length": 912.3873291015625, + "completions/mean_terminated_length": 642.589599609375, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 5.1772594752186585, + "grad_norm": 0.1305101215839386, + "learning_rate": 1e-06, + "loss": -0.0219, + "num_tokens": 339592825.0, + "reward": 0.598214328289032, + "reward_std": 0.16029362380504608, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3868.0, + "completions/mean_length": 873.7455444335938, + "completions/mean_terminated_length": 646.6093139648438, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 5.186588921282799, + "grad_norm": 0.14321625232696533, + "learning_rate": 1e-06, + "loss": -0.0207, + "num_tokens": 340219941.0, + "reward": 0.6506696939468384, + "reward_std": 0.18930260837078094, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 555 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3865.0, + "completions/mean_length": 768.2935791015625, + "completions/mean_terminated_length": 612.793212890625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.1959183673469385, + "grad_norm": 0.141605406999588, + "learning_rate": 1e-06, + "loss": -0.016, + "num_tokens": 340835604.0, + "reward": 0.7087053656578064, + "reward_std": 0.16909119486808777, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 556 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3796.0, + "completions/mean_length": 1025.8973388671875, + "completions/mean_terminated_length": 653.1815185546875, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 5.205247813411079, + "grad_norm": 0.13470318913459778, + "learning_rate": 1e-06, + "loss": -0.021, + "num_tokens": 341446464.0, + "reward": 0.6149553656578064, + "reward_std": 0.15893998742103577, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 557 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2573.0, + "completions/mean_length": 891.2076416015625, + "completions/mean_terminated_length": 598.443359375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 5.214577259475218, + "grad_norm": 0.14519642293453217, + "learning_rate": 1e-06, + "loss": -0.0235, + "num_tokens": 342033314.0, + "reward": 0.6473214626312256, + "reward_std": 0.17735788226127625, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 558 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3818.0, + "completions/mean_length": 949.864990234375, + "completions/mean_terminated_length": 598.5595703125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.223906705539359, + "grad_norm": 0.15106196701526642, + "learning_rate": 1e-06, + "loss": -0.098, + "num_tokens": 342601065.0, + "reward": 0.6584821939468384, + "reward_std": 0.19118830561637878, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 559 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3128.0, + "completions/mean_length": 992.7623291015625, + "completions/mean_terminated_length": 616.0238037109375, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 5.233236151603498, + "grad_norm": 0.12691587209701538, + "learning_rate": 1e-06, + "loss": -0.029, + "num_tokens": 343182988.0, + "reward": 0.5837053656578064, + "reward_std": 0.14538057148456573, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 560 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 957.6953735351562, + "completions/mean_terminated_length": 598.5858154296875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.242565597667639, + "grad_norm": 0.11801715940237045, + "learning_rate": 1e-06, + "loss": -0.0113, + "num_tokens": 343755379.0, + "reward": 0.668526828289032, + "reward_std": 0.11490915715694427, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 561 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3912.0, + "completions/mean_length": 996.5513916015625, + "completions/mean_terminated_length": 637.5865478515625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 5.251895043731778, + "grad_norm": 0.14140519499778748, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 344352681.0, + "reward": 0.6361607313156128, + "reward_std": 0.1584521383047104, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3803.0, + "completions/mean_length": 937.9230346679688, + "completions/mean_terminated_length": 615.5116577148438, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.261224489795918, + "grad_norm": 0.16505105793476105, + "learning_rate": 1e-06, + "loss": -0.0808, + "num_tokens": 344938364.0, + "reward": 0.6127232313156128, + "reward_std": 0.21027132868766785, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 563 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3324.0, + "completions/mean_length": 964.7734985351562, + "completions/mean_terminated_length": 674.5621948242188, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.270553935860058, + "grad_norm": 0.13483592867851257, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 345591153.0, + "reward": 0.5625, + "reward_std": 0.19460533559322357, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 564 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2277.0, + "completions/mean_length": 872.5892944335938, + "completions/mean_terminated_length": 595.1806030273438, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 5.279883381924198, + "grad_norm": 0.14715257287025452, + "learning_rate": 1e-06, + "loss": -0.0267, + "num_tokens": 346173185.0, + "reward": 0.59375, + "reward_std": 0.18054990470409393, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 565 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4003.0, + "completions/mean_length": 845.1395263671875, + "completions/mean_terminated_length": 590.8604125976562, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 5.289212827988338, + "grad_norm": 0.13903407752513885, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 346750094.0, + "reward": 0.6540178656578064, + "reward_std": 0.15282267332077026, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4015.0, + "completions/mean_length": 941.3527221679688, + "completions/mean_terminated_length": 615.0098266601562, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 5.298542274052478, + "grad_norm": 0.140536829829216, + "learning_rate": 1e-06, + "loss": -0.0368, + "num_tokens": 347334946.0, + "reward": 0.6428571939468384, + "reward_std": 0.17589251697063446, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3127.0, + "completions/mean_length": 907.17529296875, + "completions/mean_terminated_length": 611.6256103515625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.307871720116618, + "grad_norm": 0.1325594037771225, + "learning_rate": 1e-06, + "loss": -0.0314, + "num_tokens": 347923239.0, + "reward": 0.625, + "reward_std": 0.14969733357429504, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 568 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4091.0, + "completions/mean_length": 962.1451416015625, + "completions/mean_terminated_length": 637.9531860351562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 5.317201166180758, + "grad_norm": 0.1273690015077591, + "learning_rate": 1e-06, + "loss": -0.0082, + "num_tokens": 348533513.0, + "reward": 0.6238839626312256, + "reward_std": 0.12805670499801636, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 569 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 928.4107666015625, + "completions/mean_terminated_length": 617.86279296875, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 5.326530612244898, + "grad_norm": 0.1530476212501526, + "learning_rate": 1e-06, + "loss": -0.0284, + "num_tokens": 349122657.0, + "reward": 0.6171875, + "reward_std": 0.1736719310283661, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 570 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3463.0, + "completions/mean_length": 896.8839721679688, + "completions/mean_terminated_length": 613.1227416992188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.335860058309038, + "grad_norm": 0.14444182813167572, + "learning_rate": 1e-06, + "loss": -0.0102, + "num_tokens": 349709481.0, + "reward": 0.6729910969734192, + "reward_std": 0.1595052033662796, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 571 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4033.0, + "completions/mean_length": 899.0748291015625, + "completions/mean_terminated_length": 615.5079345703125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 5.345189504373177, + "grad_norm": 0.12167084217071533, + "learning_rate": 1e-06, + "loss": -0.0368, + "num_tokens": 350308732.0, + "reward": 0.5814732313156128, + "reward_std": 0.15804095566272736, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 572 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2504.0, + "completions/mean_length": 881.7678833007812, + "completions/mean_terminated_length": 588.1412963867188, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "epoch": 5.354518950437318, + "grad_norm": 0.15102258324623108, + "learning_rate": 1e-06, + "loss": -0.0666, + "num_tokens": 350886172.0, + "reward": 0.6595982313156128, + "reward_std": 0.16093485057353973, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 573 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3569.0, + "completions/mean_length": 826.3248291015625, + "completions/mean_terminated_length": 557.8007202148438, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 5.363848396501457, + "grad_norm": 0.15547627210617065, + "learning_rate": 1e-06, + "loss": -0.031, + "num_tokens": 351438583.0, + "reward": 0.6350446939468384, + "reward_std": 0.1589820384979248, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 574 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3774.0, + "completions/mean_length": 987.6741333007812, + "completions/mean_terminated_length": 666.1231689453125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.373177842565598, + "grad_norm": 0.13818617165088654, + "learning_rate": 1e-06, + "loss": -0.0171, + "num_tokens": 352071955.0, + "reward": 0.6116071939468384, + "reward_std": 0.16728109121322632, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 575 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3811.0, + "completions/mean_length": 878.8850708007812, + "completions/mean_terminated_length": 602.0181884765625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.382507288629737, + "grad_norm": 0.12957793474197388, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 352665164.0, + "reward": 0.6484375, + "reward_std": 0.1345556080341339, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 576 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4053.0, + "completions/mean_length": 944.3772583007812, + "completions/mean_terminated_length": 656.47021484375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 5.391836734693878, + "grad_norm": 0.11640162765979767, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 353290702.0, + "reward": 0.6662946939468384, + "reward_std": 0.13621418178081512, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2960.0, + "completions/mean_length": 894.20654296875, + "completions/mean_terminated_length": 610.2078247070312, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.401166180758017, + "grad_norm": 0.12326303869485855, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 353888527.0, + "reward": 0.6462053656578064, + "reward_std": 0.12328417599201202, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 578 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 1066.86279296875, + "completions/mean_terminated_length": 647.3252563476562, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 5.410495626822158, + "grad_norm": 0.14486698806285858, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 354492012.0, + "reward": 0.5792410969734192, + "reward_std": 0.17547360062599182, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 579 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3511.0, + "completions/mean_length": 851.958740234375, + "completions/mean_terminated_length": 635.6893310546875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 5.419825072886297, + "grad_norm": 0.1359625905752182, + "learning_rate": 1e-06, + "loss": -0.0004, + "num_tokens": 355111727.0, + "reward": 0.6662946939468384, + "reward_std": 0.15698717534542084, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 580 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3601.0, + "completions/mean_length": 881.1719360351562, + "completions/mean_terminated_length": 574.6234741210938, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.429154518950437, + "grad_norm": 0.1265256404876709, + "learning_rate": 1e-06, + "loss": -0.0276, + "num_tokens": 355667993.0, + "reward": 0.6395089626312256, + "reward_std": 0.12561675906181335, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 581 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3102.0, + "completions/mean_length": 940.0513916015625, + "completions/mean_terminated_length": 639.1173706054688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.438483965014577, + "grad_norm": 0.12790414690971375, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 356272319.0, + "reward": 0.6629464626312256, + "reward_std": 0.15544694662094116, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 582 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3826.0, + "completions/mean_length": 879.7645263671875, + "completions/mean_terminated_length": 632.36181640625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.447813411078717, + "grad_norm": 0.13306255638599396, + "learning_rate": 1e-06, + "loss": -0.054, + "num_tokens": 356881332.0, + "reward": 0.6830357313156128, + "reward_std": 0.1675432324409485, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 583 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2826.0, + "completions/mean_length": 888.4029541015625, + "completions/mean_terminated_length": 603.8894653320312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.457142857142857, + "grad_norm": 0.14992226660251617, + "learning_rate": 1e-06, + "loss": -0.0473, + "num_tokens": 357462973.0, + "reward": 0.6328125, + "reward_std": 0.1885533183813095, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3238.0, + "completions/mean_length": 1020.11279296875, + "completions/mean_terminated_length": 616.2083129882812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 5.466472303206997, + "grad_norm": 0.13353706896305084, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 358047978.0, + "reward": 0.6104910969734192, + "reward_std": 0.14271056652069092, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 585 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2578.0, + "completions/mean_length": 893.99560546875, + "completions/mean_terminated_length": 588.6699829101562, + "completions/min_length": 192.0, + "completions/min_terminated_length": 192.0, + "epoch": 5.475801749271137, + "grad_norm": 0.15376587212085724, + "learning_rate": 1e-06, + "loss": -0.0344, + "num_tokens": 358617246.0, + "reward": 0.645089328289032, + "reward_std": 0.17957280576229095, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 586 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3793.0, + "completions/mean_length": 977.6563110351562, + "completions/mean_terminated_length": 655.0689697265625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.485131195335277, + "grad_norm": 0.1412321925163269, + "learning_rate": 1e-06, + "loss": -0.0368, + "num_tokens": 359247778.0, + "reward": 0.5859375, + "reward_std": 0.17341090738773346, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 587 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3510.0, + "completions/mean_length": 994.4620971679688, + "completions/mean_terminated_length": 639.5596923828125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.494460641399417, + "grad_norm": 0.1472453474998474, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 359859656.0, + "reward": 0.5915178656578064, + "reward_std": 0.16070912778377533, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 588 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3516.0, + "completions/mean_length": 886.232177734375, + "completions/mean_terminated_length": 618.4281005859375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.503790087463557, + "grad_norm": 0.14182664453983307, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 360456752.0, + "reward": 0.6651785969734192, + "reward_std": 0.15687909722328186, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 589 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3090.0, + "completions/mean_length": 806.6563110351562, + "completions/mean_terminated_length": 578.9928588867188, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 5.513119533527696, + "grad_norm": 0.15292948484420776, + "learning_rate": 1e-06, + "loss": -0.0103, + "num_tokens": 361025668.0, + "reward": 0.6462053656578064, + "reward_std": 0.16935335099697113, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 590 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3808.0, + "completions/mean_length": 976.122802734375, + "completions/mean_terminated_length": 627.7493896484375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 5.522448979591837, + "grad_norm": 0.15715806186199188, + "learning_rate": 1e-06, + "loss": -0.0177, + "num_tokens": 361627714.0, + "reward": 0.6361607313156128, + "reward_std": 0.16368110477924347, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 591 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3077.0, + "completions/mean_length": 877.677490234375, + "completions/mean_terminated_length": 621.7626342773438, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.531778425655976, + "grad_norm": 0.12807443737983704, + "learning_rate": 1e-06, + "loss": -0.0201, + "num_tokens": 362245009.0, + "reward": 0.6328125, + "reward_std": 0.14109407365322113, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 592 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3820.0, + "completions/mean_length": 955.97998046875, + "completions/mean_terminated_length": 592.3162841796875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 5.541107871720117, + "grad_norm": 0.14723558723926544, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 362816959.0, + "reward": 0.6540178656578064, + "reward_std": 0.17333491146564484, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 593 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3927.0, + "completions/mean_length": 1091.693115234375, + "completions/mean_terminated_length": 684.264892578125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 5.550437317784256, + "grad_norm": 0.1476609855890274, + "learning_rate": 1e-06, + "loss": -0.0731, + "num_tokens": 363441420.0, + "reward": 0.5803571939468384, + "reward_std": 0.17998328804969788, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 594 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4033.0, + "completions/mean_length": 1016.8795166015625, + "completions/mean_terminated_length": 655.9850463867188, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 5.559766763848397, + "grad_norm": 0.17093658447265625, + "learning_rate": 1e-06, + "loss": -0.0362, + "num_tokens": 364074560.0, + "reward": 0.5647321939468384, + "reward_std": 0.1983177363872528, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606892466545105, + "step": 595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3781.0, + "completions/mean_length": 930.9029541015625, + "completions/mean_terminated_length": 633.3297119140625, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 5.569096209912536, + "grad_norm": 0.14461827278137207, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 364669057.0, + "reward": 0.6328125, + "reward_std": 0.1841912716627121, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 596 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4090.0, + "completions/mean_length": 840.8717041015625, + "completions/mean_terminated_length": 623.8630981445312, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 5.578425655976677, + "grad_norm": 0.11666166037321091, + "learning_rate": 1e-06, + "loss": -0.0211, + "num_tokens": 365270686.0, + "reward": 0.65625, + "reward_std": 0.11599431931972504, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 597 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3237.0, + "completions/mean_length": 914.8761596679688, + "completions/mean_terminated_length": 620.0402221679688, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 5.587755102040816, + "grad_norm": 0.1359519064426422, + "learning_rate": 1e-06, + "loss": -0.0465, + "num_tokens": 365863631.0, + "reward": 0.5993303656578064, + "reward_std": 0.1713072657585144, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 598 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3632.0, + "completions/mean_length": 1019.1629638671875, + "completions/mean_terminated_length": 649.9425048828125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.597084548104956, + "grad_norm": 0.1442059427499771, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 366475177.0, + "reward": 0.59375, + "reward_std": 0.1817174106836319, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 599 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4021.0, + "completions/mean_length": 995.5826416015625, + "completions/mean_terminated_length": 623.532470703125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.606413994169096, + "grad_norm": 0.157792866230011, + "learning_rate": 1e-06, + "loss": -0.0248, + "num_tokens": 367061315.0, + "reward": 0.5959821939468384, + "reward_std": 0.17359314858913422, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 600 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3730.0, + "completions/mean_length": 831.3092041015625, + "completions/mean_terminated_length": 630.1671142578125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.615743440233236, + "grad_norm": 0.15126554667949677, + "learning_rate": 1e-06, + "loss": -0.0244, + "num_tokens": 367681616.0, + "reward": 0.6696428656578064, + "reward_std": 0.2040313333272934, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 601 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3341.0, + "completions/mean_length": 813.552490234375, + "completions/mean_terminated_length": 611.3163452148438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 5.625072886297376, + "grad_norm": 0.136699840426445, + "learning_rate": 1e-06, + "loss": -0.04, + "num_tokens": 368277167.0, + "reward": 0.65625, + "reward_std": 0.16355982422828674, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 602 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3628.0, + "completions/mean_length": 865.9207763671875, + "completions/mean_terminated_length": 625.7949829101562, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.634402332361516, + "grad_norm": 0.1896849274635315, + "learning_rate": 1e-06, + "loss": -0.0125, + "num_tokens": 368893856.0, + "reward": 0.684151828289032, + "reward_std": 0.15683633089065552, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 603 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3299.0, + "completions/mean_length": 892.7957763671875, + "completions/mean_terminated_length": 650.53662109375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 5.643731778425656, + "grad_norm": 0.15675078332424164, + "learning_rate": 1e-06, + "loss": -0.0289, + "num_tokens": 369522193.0, + "reward": 0.6584821939468384, + "reward_std": 0.19088521599769592, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 604 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3277.0, + "completions/mean_length": 894.27685546875, + "completions/mean_terminated_length": 635.5126953125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 5.653061224489796, + "grad_norm": 0.13467352092266083, + "learning_rate": 1e-06, + "loss": -0.0175, + "num_tokens": 370138521.0, + "reward": 0.6428571939468384, + "reward_std": 0.14661546051502228, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2968.0, + "completions/mean_length": 903.70654296875, + "completions/mean_terminated_length": 595.0269165039062, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.662390670553936, + "grad_norm": 0.15189126133918762, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 370711842.0, + "reward": 0.6595982313156128, + "reward_std": 0.18558135628700256, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 606 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2084.0, + "completions/mean_length": 881.8917846679688, + "completions/mean_terminated_length": 601.04736328125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 5.671720116618076, + "grad_norm": 0.14081822335720062, + "learning_rate": 1e-06, + "loss": -0.0229, + "num_tokens": 371293897.0, + "reward": 0.6651785969734192, + "reward_std": 0.15308114886283875, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 607 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2640.0, + "completions/mean_length": 735.9152221679688, + "completions/mean_terminated_length": 570.6651000976562, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 5.681049562682215, + "grad_norm": 0.1424606740474701, + "learning_rate": 1e-06, + "loss": -0.0206, + "num_tokens": 371875309.0, + "reward": 0.7120535969734192, + "reward_std": 0.1468009203672409, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 608 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3414.0, + "completions/mean_length": 980.39404296875, + "completions/mean_terminated_length": 636.7893676757812, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 5.690379008746356, + "grad_norm": 0.14881321787834167, + "learning_rate": 1e-06, + "loss": -0.0301, + "num_tokens": 372478582.0, + "reward": 0.5970982313156128, + "reward_std": 0.17972436547279358, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.4907552897930145, + "step": 609 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3137.0, + "completions/mean_length": 769.6517944335938, + "completions/mean_terminated_length": 577.2183837890625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.699708454810495, + "grad_norm": 0.14888311922550201, + "learning_rate": 1e-06, + "loss": -0.0104, + "num_tokens": 373058974.0, + "reward": 0.6886160969734192, + "reward_std": 0.13023702800273895, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 610 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3141.0, + "completions/mean_length": 868.497802734375, + "completions/mean_terminated_length": 624.4009399414062, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 5.709037900874636, + "grad_norm": 0.15976490080356598, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 373674500.0, + "reward": 0.6506696939468384, + "reward_std": 0.18904298543930054, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 611 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 1077.009033203125, + "completions/mean_terminated_length": 632.47119140625, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 5.718367346938775, + "grad_norm": 0.16654469072818756, + "learning_rate": 1e-06, + "loss": -0.0576, + "num_tokens": 374268108.0, + "reward": 0.629464328289032, + "reward_std": 0.20331665873527527, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 612 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3129.0, + "completions/mean_length": 952.7589721679688, + "completions/mean_terminated_length": 593.0845336914062, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 5.727696793002916, + "grad_norm": 0.15046453475952148, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 374834604.0, + "reward": 0.6484375, + "reward_std": 0.13665924966335297, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 613 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3612.0, + "completions/mean_length": 902.69091796875, + "completions/mean_terminated_length": 593.9130859375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 5.737026239067055, + "grad_norm": 0.1479462832212448, + "learning_rate": 1e-06, + "loss": -0.0382, + "num_tokens": 375413567.0, + "reward": 0.6116071939468384, + "reward_std": 0.15800705552101135, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.4876568913459778, + "step": 614 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3977.0, + "completions/mean_length": 953.8013916015625, + "completions/mean_terminated_length": 620.1852416992188, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 5.746355685131196, + "grad_norm": 0.14168311655521393, + "learning_rate": 1e-06, + "loss": -0.0232, + "num_tokens": 376004413.0, + "reward": 0.59375, + "reward_std": 0.16570919752120972, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 615 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3523.0, + "completions/mean_length": 846.5625610351562, + "completions/mean_terminated_length": 621.6611328125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 5.755685131195335, + "grad_norm": 0.12374944239854813, + "learning_rate": 1e-06, + "loss": -0.0434, + "num_tokens": 376618197.0, + "reward": 0.6718750596046448, + "reward_std": 0.1447739452123642, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 616 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2804.0, + "completions/mean_length": 860.2567138671875, + "completions/mean_terminated_length": 594.519287109375, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 5.765014577259475, + "grad_norm": 0.160252645611763, + "learning_rate": 1e-06, + "loss": -0.02, + "num_tokens": 377196795.0, + "reward": 0.6506696939468384, + "reward_std": 0.1742025464773178, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 617 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2952.0, + "completions/mean_length": 987.1897583007812, + "completions/mean_terminated_length": 674.0172119140625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 5.774344023323615, + "grad_norm": 0.14358121156692505, + "learning_rate": 1e-06, + "loss": -0.0263, + "num_tokens": 377829733.0, + "reward": 0.65625, + "reward_std": 0.17585016787052155, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 618 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3669.0, + "completions/mean_length": 844.825927734375, + "completions/mean_terminated_length": 594.735595703125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 5.783673469387755, + "grad_norm": 0.12825733423233032, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 378414057.0, + "reward": 0.6350446939468384, + "reward_std": 0.1277536004781723, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 619 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3822.0, + "completions/mean_length": 804.2142944335938, + "completions/mean_terminated_length": 588.9369506835938, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 5.793002915451895, + "grad_norm": 0.13255034387111664, + "learning_rate": 1e-06, + "loss": -0.0332, + "num_tokens": 379006233.0, + "reward": 0.6830357313156128, + "reward_std": 0.1640915721654892, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 620 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4055.0, + "completions/mean_length": 948.0614013671875, + "completions/mean_terminated_length": 652.1013793945312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 5.802332361516035, + "grad_norm": 0.14350251853466034, + "learning_rate": 1e-06, + "loss": -0.037, + "num_tokens": 379627384.0, + "reward": 0.6506696939468384, + "reward_std": 0.14628097414970398, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 621 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3452.0, + "completions/mean_length": 905.6428833007812, + "completions/mean_terminated_length": 664.3553466796875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 5.811661807580175, + "grad_norm": 0.1307872086763382, + "learning_rate": 1e-06, + "loss": -0.0197, + "num_tokens": 380265000.0, + "reward": 0.6651785969734192, + "reward_std": 0.17063285410404205, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 622 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3543.0, + "completions/mean_length": 999.4844360351562, + "completions/mean_terminated_length": 691.7324829101562, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 5.820991253644315, + "grad_norm": 0.16919715702533722, + "learning_rate": 1e-06, + "loss": -0.0447, + "num_tokens": 380929162.0, + "reward": 0.625, + "reward_std": 0.17322616279125214, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4052.0, + "completions/mean_length": 939.6451416015625, + "completions/mean_terminated_length": 663.8471069335938, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 5.830320699708455, + "grad_norm": 0.12789836525917053, + "learning_rate": 1e-06, + "loss": -0.0252, + "num_tokens": 381577948.0, + "reward": 0.5892857313156128, + "reward_std": 0.15480753779411316, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 624 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2909.0, + "completions/mean_length": 856.5089721679688, + "completions/mean_terminated_length": 594.6923828125, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 5.839650145772595, + "grad_norm": 0.1532348096370697, + "learning_rate": 1e-06, + "loss": -0.0253, + "num_tokens": 382150580.0, + "reward": 0.6897321939468384, + "reward_std": 0.16172830760478973, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 625 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4006.0, + "completions/mean_length": 837.177490234375, + "completions/mean_terminated_length": 607.4635620117188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.848979591836734, + "grad_norm": 0.12481541186571121, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 382739003.0, + "reward": 0.652901828289032, + "reward_std": 0.12261268496513367, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 626 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 877.5938110351562, + "completions/mean_terminated_length": 617.4813232421875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 5.858309037900875, + "grad_norm": 0.1397131234407425, + "learning_rate": 1e-06, + "loss": -0.0418, + "num_tokens": 383345855.0, + "reward": 0.6707589626312256, + "reward_std": 0.1529742032289505, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 627 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2534.0, + "completions/mean_length": 923.6998291015625, + "completions/mean_terminated_length": 625.4493408203125, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 5.867638483965014, + "grad_norm": 0.1381191462278366, + "learning_rate": 1e-06, + "loss": -0.0311, + "num_tokens": 383938402.0, + "reward": 0.6283482313156128, + "reward_std": 0.1447400450706482, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 628 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3131.0, + "completions/mean_length": 959.6529541015625, + "completions/mean_terminated_length": 647.9423217773438, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 5.876967930029155, + "grad_norm": 0.139832004904747, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 384564595.0, + "reward": 0.6417410969734192, + "reward_std": 0.14628097414970398, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 629 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 933.3125610351562, + "completions/mean_terminated_length": 627.4957275390625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 5.886297376093294, + "grad_norm": 0.1456081122159958, + "learning_rate": 1e-06, + "loss": -0.0152, + "num_tokens": 385169291.0, + "reward": 0.6718750596046448, + "reward_std": 0.15022796392440796, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 630 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 909.4810791015625, + "completions/mean_terminated_length": 597.0772094726562, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 5.895626822157435, + "grad_norm": 0.14111317694187164, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 385744194.0, + "reward": 0.637276828289032, + "reward_std": 0.14966341853141785, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 631 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3519.0, + "completions/mean_length": 927.8460083007812, + "completions/mean_terminated_length": 663.513916015625, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 5.904956268221574, + "grad_norm": 0.1582006961107254, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 386389872.0, + "reward": 0.5848214626312256, + "reward_std": 0.19813229143619537, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 632 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2702.0, + "completions/mean_length": 850.9520263671875, + "completions/mean_terminated_length": 609.7134399414062, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 5.914285714285715, + "grad_norm": 0.126397505402565, + "learning_rate": 1e-06, + "loss": -0.0334, + "num_tokens": 386979381.0, + "reward": 0.6551339626312256, + "reward_std": 0.15273529291152954, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 633 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2512.0, + "completions/mean_length": 826.7366333007812, + "completions/mean_terminated_length": 612.9321899414062, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 5.923615160349854, + "grad_norm": 0.14366699755191803, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 387587097.0, + "reward": 0.6584821939468384, + "reward_std": 0.17355993390083313, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 634 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3223.0, + "completions/mean_length": 1015.529052734375, + "completions/mean_terminated_length": 641.554443359375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 5.932944606413994, + "grad_norm": 0.15155993402004242, + "learning_rate": 1e-06, + "loss": -0.063, + "num_tokens": 388197099.0, + "reward": 0.613839328289032, + "reward_std": 0.19697365164756775, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714011907577515, + "step": 635 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3900.0, + "completions/mean_length": 943.7969360351562, + "completions/mean_terminated_length": 647.4359130859375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 5.942274052478134, + "grad_norm": 0.13637402653694153, + "learning_rate": 1e-06, + "loss": -0.0345, + "num_tokens": 388814469.0, + "reward": 0.6272321939468384, + "reward_std": 0.11986783146858215, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3611.0, + "completions/mean_length": 945.7288208007812, + "completions/mean_terminated_length": 619.8386840820312, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.9516034985422746, + "grad_norm": 0.13491137325763702, + "learning_rate": 1e-06, + "loss": -0.0101, + "num_tokens": 389413506.0, + "reward": 0.598214328289032, + "reward_std": 0.15251845121383667, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 637 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 970.0569458007812, + "completions/mean_terminated_length": 684.4957885742188, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 5.960932944606414, + "grad_norm": 0.13851997256278992, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 390071213.0, + "reward": 0.6238839626312256, + "reward_std": 0.16845403611660004, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 638 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3853.0, + "completions/mean_length": 818.927490234375, + "completions/mean_terminated_length": 592.1134033203125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.970262390670554, + "grad_norm": 0.13052687048912048, + "learning_rate": 1e-06, + "loss": -0.0337, + "num_tokens": 390658620.0, + "reward": 0.6305803656578064, + "reward_std": 0.14638976752758026, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 639 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2962.0, + "completions/mean_length": 837.6283569335938, + "completions/mean_terminated_length": 548.6112060546875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.979591836734694, + "grad_norm": 0.1723228394985199, + "learning_rate": 1e-06, + "loss": -0.0344, + "num_tokens": 391190367.0, + "reward": 0.6506696939468384, + "reward_std": 0.14162610471248627, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 640 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2989.0, + "completions/mean_length": 1043.232177734375, + "completions/mean_terminated_length": 642.3636474609375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 5.988921282798834, + "grad_norm": 0.1537492275238037, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 391799087.0, + "reward": 0.621651828289032, + "reward_std": 0.17757610976696014, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 641 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.06534090909090906, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2963.0, + "completions/mean_length": 905.102294921875, + "completions/mean_terminated_length": 682.0303955078125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.998250728862974, + "grad_norm": 0.13764840364456177, + "learning_rate": 1e-06, + "loss": -0.0343, + "num_tokens": 392405329.0, + "reward": 0.6729910969734192, + "reward_std": 0.13868872821331024, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 642 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3857.0, + "completions/mean_length": 923.958740234375, + "completions/mean_terminated_length": 638.3978271484375, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.0093294460641395, + "grad_norm": 0.14158585667610168, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 393035036.0, + "reward": 0.6149553656578064, + "reward_std": 0.15007779002189636, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 643 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2789.0, + "completions/mean_length": 798.8761596679688, + "completions/mean_terminated_length": 562.2404174804688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 6.01865889212828, + "grad_norm": 0.1399795264005661, + "learning_rate": 1e-06, + "loss": -0.0274, + "num_tokens": 393582685.0, + "reward": 0.6975446939468384, + "reward_std": 0.16465751826763153, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 644 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2469.0, + "completions/mean_length": 933.51904296875, + "completions/mean_terminated_length": 619.2110595703125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 6.0279883381924195, + "grad_norm": 0.162126362323761, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 394178870.0, + "reward": 0.6640625, + "reward_std": 0.19099467992782593, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 645 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3825.0, + "completions/mean_length": 966.0535888671875, + "completions/mean_terminated_length": 642.2659912109375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 6.03731778425656, + "grad_norm": 0.1526937335729599, + "learning_rate": 1e-06, + "loss": -0.0494, + "num_tokens": 394801222.0, + "reward": 0.5814732313156128, + "reward_std": 0.17611299455165863, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 646 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3258.0, + "completions/mean_length": 969.9598388671875, + "completions/mean_terminated_length": 709.1414794921875, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 6.0466472303206995, + "grad_norm": 0.14036892354488373, + "learning_rate": 1e-06, + "loss": -0.0402, + "num_tokens": 395472018.0, + "reward": 0.625, + "reward_std": 0.17430922389030457, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 647 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3772.0, + "completions/mean_length": 883.6373291015625, + "completions/mean_terminated_length": 607.1793823242188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 6.05597667638484, + "grad_norm": 0.16802679002285004, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 396060477.0, + "reward": 0.6584821939468384, + "reward_std": 0.18765361607074738, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 648 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3419.0, + "completions/mean_length": 930.7545166015625, + "completions/mean_terminated_length": 654.1796264648438, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 6.0653061224489795, + "grad_norm": 0.13883596658706665, + "learning_rate": 1e-06, + "loss": -0.0284, + "num_tokens": 396687241.0, + "reward": 0.6305803656578064, + "reward_std": 0.16116377711296082, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 649 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2177.0, + "completions/mean_length": 822.1551513671875, + "completions/mean_terminated_length": 582.988037109375, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 6.07463556851312, + "grad_norm": 0.14306329190731049, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 397265372.0, + "reward": 0.645089328289032, + "reward_std": 0.1659344881772995, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 650 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3128.0, + "completions/mean_length": 742.6563110351562, + "completions/mean_terminated_length": 585.9579467773438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.0839650145772595, + "grad_norm": 0.15480206906795502, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 397857944.0, + "reward": 0.6383928656578064, + "reward_std": 0.17096774280071259, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 651 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3221.0, + "completions/mean_length": 807.1339721679688, + "completions/mean_terminated_length": 592.0475463867188, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.093294460641399, + "grad_norm": 0.14246529340744019, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 398444032.0, + "reward": 0.6540178656578064, + "reward_std": 0.14120467007160187, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 652 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2717.0, + "completions/mean_length": 840.1484985351562, + "completions/mean_terminated_length": 618.9523315429688, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 6.1026239067055394, + "grad_norm": 0.13658668100833893, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 399054125.0, + "reward": 0.6462053656578064, + "reward_std": 0.14989235997200012, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 653 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3798.0, + "completions/mean_length": 956.7801513671875, + "completions/mean_terminated_length": 644.7840576171875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 6.111953352769679, + "grad_norm": 0.1408751904964447, + "learning_rate": 1e-06, + "loss": -0.0278, + "num_tokens": 399676840.0, + "reward": 0.6350446939468384, + "reward_std": 0.1657087653875351, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 654 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3113.0, + "completions/mean_length": 865.1730346679688, + "completions/mean_terminated_length": 570.0304565429688, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.121282798833819, + "grad_norm": 0.15640050172805786, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 400226979.0, + "reward": 0.6584821939468384, + "reward_std": 0.16442786157131195, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 655 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2838.0, + "completions/mean_length": 827.4185791015625, + "completions/mean_terminated_length": 580.21484375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 6.130612244897959, + "grad_norm": 0.15364786982536316, + "learning_rate": 1e-06, + "loss": -0.0474, + "num_tokens": 400787530.0, + "reward": 0.7131696939468384, + "reward_std": 0.1642010509967804, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 656 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 874.4799194335938, + "completions/mean_terminated_length": 626.6707153320312, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 6.139941690962099, + "grad_norm": 0.1450071930885315, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 401393920.0, + "reward": 0.6540178656578064, + "reward_std": 0.18205510079860687, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 657 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3379.0, + "completions/mean_length": 934.6473388671875, + "completions/mean_terminated_length": 598.99755859375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 6.149271137026239, + "grad_norm": 0.143891841173172, + "learning_rate": 1e-06, + "loss": -0.0092, + "num_tokens": 401972916.0, + "reward": 0.6283482313156128, + "reward_std": 0.13147233426570892, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 658 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3322.0, + "completions/mean_length": 796.9051513671875, + "completions/mean_terminated_length": 568.5668334960938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 6.158600583090379, + "grad_norm": 0.1359974443912506, + "learning_rate": 1e-06, + "loss": -0.0178, + "num_tokens": 402534447.0, + "reward": 0.7500000596046448, + "reward_std": 0.12967249751091003, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 659 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2839.0, + "completions/mean_length": 912.6317138671875, + "completions/mean_terminated_length": 617.5878295898438, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 6.167930029154519, + "grad_norm": 0.14906616508960724, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 403127413.0, + "reward": 0.6361607313156128, + "reward_std": 0.16349565982818604, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 660 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3903.0, + "completions/mean_length": 1067.2054443359375, + "completions/mean_terminated_length": 669.48486328125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 6.1772594752186585, + "grad_norm": 0.15464231371879578, + "learning_rate": 1e-06, + "loss": -0.0402, + "num_tokens": 403759925.0, + "reward": 0.5424107313156128, + "reward_std": 0.19016914069652557, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 661 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3561.0, + "completions/mean_length": 804.9163208007812, + "completions/mean_terminated_length": 626.810546875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 6.186588921282799, + "grad_norm": 0.1159672886133194, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 404375642.0, + "reward": 0.6830357313156128, + "reward_std": 0.1319269835948944, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 662 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3932.0, + "completions/mean_length": 858.0000610351562, + "completions/mean_terminated_length": 592.0772705078125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 6.1959183673469385, + "grad_norm": 0.14473634958267212, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 404959298.0, + "reward": 0.6819196939468384, + "reward_std": 0.1619861125946045, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 663 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2976.0, + "completions/mean_length": 855.5781860351562, + "completions/mean_terminated_length": 572.4344482421875, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 6.205247813411079, + "grad_norm": 0.14569254219532013, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 405512112.0, + "reward": 0.6194196939468384, + "reward_std": 0.14958924055099487, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580071330070496, + "step": 664 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 959.193115234375, + "completions/mean_terminated_length": 643.2002563476562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.214577259475218, + "grad_norm": 0.13061484694480896, + "learning_rate": 1e-06, + "loss": -0.0183, + "num_tokens": 406134557.0, + "reward": 0.5647321939468384, + "reward_std": 0.12395747750997543, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 665 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2272.0, + "completions/mean_length": 807.224365234375, + "completions/mean_terminated_length": 532.8283081054688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.223906705539359, + "grad_norm": 0.14673306047916412, + "learning_rate": 1e-06, + "loss": -0.0389, + "num_tokens": 406653582.0, + "reward": 0.691964328289032, + "reward_std": 0.14338389039039612, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 666 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4093.0, + "completions/mean_length": 917.33154296875, + "completions/mean_terminated_length": 668.6991577148438, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 6.233236151603498, + "grad_norm": 0.15444384515285492, + "learning_rate": 1e-06, + "loss": -0.0292, + "num_tokens": 407291391.0, + "reward": 0.6428571939468384, + "reward_std": 0.1678142249584198, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 667 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3245.0, + "completions/mean_length": 1007.15185546875, + "completions/mean_terminated_length": 636.489990234375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 6.242565597667639, + "grad_norm": 0.13743498921394348, + "learning_rate": 1e-06, + "loss": -0.0308, + "num_tokens": 407896367.0, + "reward": 0.625, + "reward_std": 0.13929422199726105, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 668 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3550.0, + "completions/mean_length": 868.232177734375, + "completions/mean_terminated_length": 628.2781982421875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.251895043731778, + "grad_norm": 0.14347991347312927, + "learning_rate": 1e-06, + "loss": -0.0186, + "num_tokens": 408511703.0, + "reward": 0.6272321939468384, + "reward_std": 0.17171843349933624, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 669 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3620.0, + "completions/mean_length": 856.5469360351562, + "completions/mean_terminated_length": 573.4878540039062, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.261224489795918, + "grad_norm": 0.13810473680496216, + "learning_rate": 1e-06, + "loss": -0.0071, + "num_tokens": 409074489.0, + "reward": 0.6785714626312256, + "reward_std": 0.14083263278007507, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728575229644775, + "step": 670 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2964.0, + "completions/mean_length": 876.2723388671875, + "completions/mean_terminated_length": 582.1437377929688, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 6.270553935860058, + "grad_norm": 0.15649516880512238, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 409641469.0, + "reward": 0.668526828289032, + "reward_std": 0.1605243682861328, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 671 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3685.0, + "completions/mean_length": 946.2935791015625, + "completions/mean_terminated_length": 654.3695068359375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 6.279883381924198, + "grad_norm": 0.13064932823181152, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 410272356.0, + "reward": 0.6662946939468384, + "reward_std": 0.16495807468891144, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 877.6473388671875, + "completions/mean_terminated_length": 592.1798706054688, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 6.289212827988338, + "grad_norm": 0.1501125693321228, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 410859448.0, + "reward": 0.6272321939468384, + "reward_std": 0.18333503603935242, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 673 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3199.0, + "completions/mean_length": 791.2142944335938, + "completions/mean_terminated_length": 636.7850341796875, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 6.298542274052478, + "grad_norm": 0.13228359818458557, + "learning_rate": 1e-06, + "loss": -0.0158, + "num_tokens": 411491376.0, + "reward": 0.6707589626312256, + "reward_std": 0.17299653589725494, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3417.0, + "completions/mean_length": 936.0904541015625, + "completions/mean_terminated_length": 604.90380859375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 6.307871720116618, + "grad_norm": 0.14179129898548126, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 412057697.0, + "reward": 0.6495535969734192, + "reward_std": 0.1608920693397522, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 675 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3303.0, + "completions/mean_length": 999.37060546875, + "completions/mean_terminated_length": 640.7322387695312, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 6.317201166180758, + "grad_norm": 0.1590924859046936, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 412659309.0, + "reward": 0.6205357313156128, + "reward_std": 0.18904118239879608, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 676 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3902.0, + "completions/mean_length": 946.8292846679688, + "completions/mean_terminated_length": 629.5908813476562, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.326530612244898, + "grad_norm": 0.1469426304101944, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 413263156.0, + "reward": 0.6383928656578064, + "reward_std": 0.1629982441663742, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 677 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4089.0, + "completions/mean_length": 932.6719360351562, + "completions/mean_terminated_length": 592.4869995117188, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 6.335860058309038, + "grad_norm": 0.12550754845142365, + "learning_rate": 1e-06, + "loss": -0.0382, + "num_tokens": 413845526.0, + "reward": 0.6127232313156128, + "reward_std": 0.1274493932723999, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 678 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 910.4129638671875, + "completions/mean_terminated_length": 606.65283203125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 6.345189504373177, + "grad_norm": 0.14531581103801727, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 414430264.0, + "reward": 0.6517857313156128, + "reward_std": 0.18306542932987213, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 679 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3730.0, + "completions/mean_length": 921.97998046875, + "completions/mean_terminated_length": 661.3115844726562, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 6.354518950437318, + "grad_norm": 0.14274469017982483, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 415063694.0, + "reward": 0.6305803656578064, + "reward_std": 0.1825340837240219, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 680 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4038.0, + "completions/mean_length": 1006.2254638671875, + "completions/mean_terminated_length": 678.1753540039062, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 6.363848396501457, + "grad_norm": 0.13987290859222412, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 415703816.0, + "reward": 0.606026828289032, + "reward_std": 0.14969801902770996, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 681 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 906.9006958007812, + "completions/mean_terminated_length": 632.44482421875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 6.373177842565598, + "grad_norm": 0.136633038520813, + "learning_rate": 1e-06, + "loss": -0.0362, + "num_tokens": 416320927.0, + "reward": 0.6316964626312256, + "reward_std": 0.1664562225341797, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.482613742351532, + "step": 682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 896.70654296875, + "completions/mean_terminated_length": 608.6921997070312, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 6.382507288629737, + "grad_norm": 0.14189589023590088, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 416913872.0, + "reward": 0.7053571939468384, + "reward_std": 0.1504889577627182, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 683 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3625.0, + "completions/mean_length": 993.0781860351562, + "completions/mean_terminated_length": 650.8723754882812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.391836734693878, + "grad_norm": 0.15703332424163818, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 417529446.0, + "reward": 0.5870535969734192, + "reward_std": 0.18070964515209198, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263834953308105, + "step": 684 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2871.0, + "completions/mean_length": 876.4006958007812, + "completions/mean_terminated_length": 560.7537231445312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.401166180758017, + "grad_norm": 0.15196703374385834, + "learning_rate": 1e-06, + "loss": -0.0565, + "num_tokens": 418078669.0, + "reward": 0.7254464626312256, + "reward_std": 0.17904219031333923, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 685 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3305.0, + "completions/mean_length": 980.3560791015625, + "completions/mean_terminated_length": 606.478759765625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 6.410495626822158, + "grad_norm": 0.1479184329509735, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 418651356.0, + "reward": 0.6383928656578064, + "reward_std": 0.1532677263021469, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 686 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 968.357177734375, + "completions/mean_terminated_length": 644.807861328125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 6.419825072886297, + "grad_norm": 0.1460193246603012, + "learning_rate": 1e-06, + "loss": -0.0377, + "num_tokens": 419259372.0, + "reward": 0.6495535969734192, + "reward_std": 0.17299720644950867, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.47737622261047363, + "step": 687 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3936.0, + "completions/mean_length": 967.22216796875, + "completions/mean_terminated_length": 673.0635375976562, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.429154518950437, + "grad_norm": 0.13802941143512726, + "learning_rate": 1e-06, + "loss": -0.0185, + "num_tokens": 419898163.0, + "reward": 0.629464328289032, + "reward_std": 0.15996097028255463, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 688 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 967.2310791015625, + "completions/mean_terminated_length": 630.7626342773438, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.438483965014577, + "grad_norm": 0.13665814697742462, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 420499314.0, + "reward": 0.6261160969734192, + "reward_std": 0.1519550383090973, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 689 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3241.0, + "completions/mean_length": 943.3136596679688, + "completions/mean_terminated_length": 634.2267456054688, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 6.447813411078717, + "grad_norm": 0.1733466535806656, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 421101675.0, + "reward": 0.6473214626312256, + "reward_std": 0.20812352001667023, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 690 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 1010.22998046875, + "completions/mean_terminated_length": 648.55615234375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 6.457142857142857, + "grad_norm": 0.16263236105442047, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 421710289.0, + "reward": 0.6238839626312256, + "reward_std": 0.20151540637016296, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 691 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3612.0, + "completions/mean_length": 847.0145263671875, + "completions/mean_terminated_length": 605.4832153320312, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 6.466472303206997, + "grad_norm": 0.14803138375282288, + "learning_rate": 1e-06, + "loss": -0.0226, + "num_tokens": 422305550.0, + "reward": 0.6551339626312256, + "reward_std": 0.1765669286251068, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 692 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3020.0, + "completions/mean_length": 955.7176513671875, + "completions/mean_terminated_length": 596.3818359375, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 6.475801749271137, + "grad_norm": 0.15864917635917664, + "learning_rate": 1e-06, + "loss": -0.0641, + "num_tokens": 422867433.0, + "reward": 0.6785714626312256, + "reward_std": 0.19347740709781647, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 693 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 834.5736694335938, + "completions/mean_terminated_length": 587.9111328125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 6.485131195335277, + "grad_norm": 0.13870617747306824, + "learning_rate": 1e-06, + "loss": -0.0163, + "num_tokens": 423448731.0, + "reward": 0.6595982313156128, + "reward_std": 0.1334686130285263, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 694 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3923.0, + "completions/mean_length": 963.0625610351562, + "completions/mean_terminated_length": 626.1458740234375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 6.494460641399417, + "grad_norm": 0.16906289756298065, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 424042643.0, + "reward": 0.6462053656578064, + "reward_std": 0.17826010286808014, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3490.0, + "completions/mean_length": 971.302490234375, + "completions/mean_terminated_length": 635.2719116210938, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.503790087463557, + "grad_norm": 0.1475287228822708, + "learning_rate": 1e-06, + "loss": -0.024, + "num_tokens": 424650170.0, + "reward": 0.5970982313156128, + "reward_std": 0.16814985871315002, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.4907552897930145, + "step": 696 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3707.0, + "completions/mean_length": 1071.118408203125, + "completions/mean_terminated_length": 638.9923095703125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 6.513119533527696, + "grad_norm": 0.15114979445934296, + "learning_rate": 1e-06, + "loss": -0.0548, + "num_tokens": 425247812.0, + "reward": 0.5993303656578064, + "reward_std": 0.17641498148441315, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3772.0, + "completions/mean_length": 983.82373046875, + "completions/mean_terminated_length": 619.0548706054688, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 6.522448979591837, + "grad_norm": 0.15000033378601074, + "learning_rate": 1e-06, + "loss": -0.0444, + "num_tokens": 425844046.0, + "reward": 0.6551339626312256, + "reward_std": 0.1713828444480896, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 698 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2950.0, + "completions/mean_length": 903.763427734375, + "completions/mean_terminated_length": 560.4697265625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.531778425655976, + "grad_norm": 0.1456417441368103, + "learning_rate": 1e-06, + "loss": -0.0553, + "num_tokens": 426387194.0, + "reward": 0.7120535969734192, + "reward_std": 0.164730966091156, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 699 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2657.0, + "completions/mean_length": 914.0792846679688, + "completions/mean_terminated_length": 576.2457275390625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 6.541107871720117, + "grad_norm": 0.11200274527072906, + "learning_rate": 1e-06, + "loss": -0.0158, + "num_tokens": 426934137.0, + "reward": 0.6651785969734192, + "reward_std": 0.09923429787158966, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 700 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 961.568115234375, + "completions/mean_terminated_length": 641.5707397460938, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 6.550437317784256, + "grad_norm": 0.1343359798192978, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 427540870.0, + "reward": 0.65625, + "reward_std": 0.15600824356079102, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 701 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4031.0, + "completions/mean_length": 859.8873291015625, + "completions/mean_terminated_length": 598.3438110351562, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.559766763848397, + "grad_norm": 0.13791272044181824, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 428121201.0, + "reward": 0.6595982313156128, + "reward_std": 0.14248304069042206, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 702 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 1159.1741943359375, + "completions/mean_terminated_length": 678.6026000976562, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 6.569096209912536, + "grad_norm": 0.12637297809123993, + "learning_rate": 1e-06, + "loss": -0.0812, + "num_tokens": 428720069.0, + "reward": 0.6149553656578064, + "reward_std": 0.1504140943288803, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 703 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3918.0, + "completions/mean_length": 874.036865234375, + "completions/mean_terminated_length": 613.6369018554688, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 6.578425655976677, + "grad_norm": 0.13937804102897644, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 429314326.0, + "reward": 0.6886160969734192, + "reward_std": 0.15642446279525757, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 704 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3819.0, + "completions/mean_length": 897.896240234375, + "completions/mean_terminated_length": 651.8882446289062, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 6.587755102040816, + "grad_norm": 0.1512024700641632, + "learning_rate": 1e-06, + "loss": -0.0264, + "num_tokens": 429937857.0, + "reward": 0.6863839626312256, + "reward_std": 0.17739178240299225, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 705 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3628.0, + "completions/mean_length": 1037.453125, + "completions/mean_terminated_length": 661.8421020507812, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 6.597084548104956, + "grad_norm": 0.15871959924697876, + "learning_rate": 1e-06, + "loss": -0.047, + "num_tokens": 430559063.0, + "reward": 0.6305803656578064, + "reward_std": 0.18881294131278992, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 706 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 828.9520263671875, + "completions/mean_terminated_length": 611.1488037109375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 6.606413994169096, + "grad_norm": 0.1267891377210617, + "learning_rate": 1e-06, + "loss": -0.0235, + "num_tokens": 431163820.0, + "reward": 0.640625, + "reward_std": 0.14740711450576782, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 707 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2135.0, + "completions/mean_length": 868.3895263671875, + "completions/mean_terminated_length": 577.8260498046875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 6.615743440233236, + "grad_norm": 0.15745864808559418, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 431728921.0, + "reward": 0.6707589626312256, + "reward_std": 0.1829138696193695, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 708 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3596.0, + "completions/mean_length": 1067.5570068359375, + "completions/mean_terminated_length": 621.6273803710938, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 6.625072886297376, + "grad_norm": 0.1713033765554428, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 432303708.0, + "reward": 0.640625, + "reward_std": 0.1825447827577591, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 709 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3884.0, + "completions/mean_length": 1034.798095703125, + "completions/mean_terminated_length": 632.8219604492188, + "completions/min_length": 197.0, + "completions/min_terminated_length": 197.0, + "epoch": 6.634402332361516, + "grad_norm": 0.13735699653625488, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 432904695.0, + "reward": 0.668526828289032, + "reward_std": 0.14263640344142914, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056781768799, + "step": 710 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2790.0, + "completions/mean_length": 931.4219360351562, + "completions/mean_terminated_length": 633.8974609375, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 6.643731778425656, + "grad_norm": 0.13081882894039154, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 433507713.0, + "reward": 0.6484375, + "reward_std": 0.15428510308265686, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 711 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3449.0, + "completions/mean_length": 1001.9688110351562, + "completions/mean_terminated_length": 630.6849975585938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.653061224489796, + "grad_norm": 0.1479046493768692, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 434108629.0, + "reward": 0.606026828289032, + "reward_std": 0.1705554574728012, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 712 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3057.0, + "completions/mean_length": 942.3527221679688, + "completions/mean_terminated_length": 603.2089233398438, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 6.662390670553936, + "grad_norm": 0.14598913490772247, + "learning_rate": 1e-06, + "loss": -0.0169, + "num_tokens": 434687641.0, + "reward": 0.65625, + "reward_std": 0.13046273589134216, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 713 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3565.0, + "completions/mean_length": 990.7500610351562, + "completions/mean_terminated_length": 626.7930297851562, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 6.671720116618076, + "grad_norm": 0.12860062718391418, + "learning_rate": 1e-06, + "loss": -0.0391, + "num_tokens": 435289441.0, + "reward": 0.6171875, + "reward_std": 0.11997590214014053, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 714 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 4014.0, + "completions/mean_length": 1056.54248046875, + "completions/mean_terminated_length": 622.3341674804688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.681049562682215, + "grad_norm": 0.1592123955488205, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 435868087.0, + "reward": 0.6473214626312256, + "reward_std": 0.16281278431415558, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 715 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3352.0, + "completions/mean_length": 1017.450927734375, + "completions/mean_terminated_length": 635.04638671875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 6.690379008746356, + "grad_norm": 0.14399969577789307, + "learning_rate": 1e-06, + "loss": -0.0576, + "num_tokens": 436464971.0, + "reward": 0.6082589626312256, + "reward_std": 0.18994088470935822, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 716 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 1069.6640625, + "completions/mean_terminated_length": 654.8870239257812, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 6.699708454810495, + "grad_norm": 0.13264822959899902, + "learning_rate": 1e-06, + "loss": -0.0719, + "num_tokens": 437070030.0, + "reward": 0.6026785969734192, + "reward_std": 0.15311436355113983, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 717 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2946.0, + "completions/mean_length": 933.1094360351562, + "completions/mean_terminated_length": 601.6103515625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 6.709037900874636, + "grad_norm": 0.15438160300254822, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 437642256.0, + "reward": 0.6662946939468384, + "reward_std": 0.15169289708137512, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 718 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4088.0, + "completions/mean_length": 920.818115234375, + "completions/mean_terminated_length": 605.2478637695312, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 6.718367346938775, + "grad_norm": 0.14090892672538757, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 438234413.0, + "reward": 0.6852678656578064, + "reward_std": 0.1369616538286209, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 719 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3112.0, + "completions/mean_length": 975.7154541015625, + "completions/mean_terminated_length": 614.3374633789062, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.727696793002916, + "grad_norm": 0.14739517867565155, + "learning_rate": 1e-06, + "loss": -0.0464, + "num_tokens": 438816646.0, + "reward": 0.6238839626312256, + "reward_std": 0.16131463646888733, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 720 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2812.0, + "completions/mean_length": 896.2857666015625, + "completions/mean_terminated_length": 620.9163818359375, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 6.737026239067055, + "grad_norm": 0.14392498135566711, + "learning_rate": 1e-06, + "loss": -0.0464, + "num_tokens": 439409822.0, + "reward": 0.6640625, + "reward_std": 0.17532657086849213, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 721 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2842.0, + "completions/mean_length": 912.3817138671875, + "completions/mean_terminated_length": 565.6510009765625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "epoch": 6.746355685131196, + "grad_norm": 0.14372050762176514, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 439969644.0, + "reward": 0.6484375, + "reward_std": 0.1350851058959961, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 722 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3879.0, + "completions/mean_length": 957.3928833007812, + "completions/mean_terminated_length": 606.9280395507812, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.755685131195335, + "grad_norm": 0.13563469052314758, + "learning_rate": 1e-06, + "loss": -0.0652, + "num_tokens": 440546764.0, + "reward": 0.6328125, + "reward_std": 0.1422245353460312, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 723 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3991.0, + "completions/mean_length": 849.232177734375, + "completions/mean_terminated_length": 574.0823364257812, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 6.765014577259475, + "grad_norm": 0.1479085236787796, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 441101988.0, + "reward": 0.6886160969734192, + "reward_std": 0.17622952163219452, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 724 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3999.0, + "completions/mean_length": 1020.1082763671875, + "completions/mean_terminated_length": 638.0338745117188, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 6.774344023323615, + "grad_norm": 0.15377195179462433, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 441696149.0, + "reward": 0.6194196939468384, + "reward_std": 0.17641469836235046, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580071330070496, + "step": 725 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3398.0, + "completions/mean_length": 964.0881958007812, + "completions/mean_terminated_length": 627.2817993164062, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.783673469387755, + "grad_norm": 0.1564217507839203, + "learning_rate": 1e-06, + "loss": -0.0682, + "num_tokens": 442291068.0, + "reward": 0.6473214626312256, + "reward_std": 0.19231371581554413, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 726 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4060.0, + "completions/mean_length": 1033.1953125, + "completions/mean_terminated_length": 639.7367553710938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 6.793002915451895, + "grad_norm": 0.17613981664180756, + "learning_rate": 1e-06, + "loss": -0.035, + "num_tokens": 442889419.0, + "reward": 0.5881696939468384, + "reward_std": 0.1970803141593933, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 727 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3906.0, + "completions/mean_length": 989.1339721679688, + "completions/mean_terminated_length": 655.02099609375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "epoch": 6.802332361516035, + "grad_norm": 0.15963493287563324, + "learning_rate": 1e-06, + "loss": -0.0396, + "num_tokens": 443503339.0, + "reward": 0.5881696939468384, + "reward_std": 0.1958770900964737, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 728 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 942.12060546875, + "completions/mean_terminated_length": 641.3839111328125, + "completions/min_length": 13.0, + "completions/min_terminated_length": 13.0, + "epoch": 6.811661807580175, + "grad_norm": 0.1958419531583786, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 444109367.0, + "reward": 0.6428571939468384, + "reward_std": 0.18606920540332794, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 729 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 1009.857177734375, + "completions/mean_terminated_length": 630.8571166992188, + "completions/min_length": 22.0, + "completions/min_terminated_length": 22.0, + "epoch": 6.820991253644315, + "grad_norm": 0.18248870968818665, + "learning_rate": 1e-06, + "loss": -0.0442, + "num_tokens": 444689327.0, + "reward": 0.5758928656578064, + "reward_std": 0.18600277602672577, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 730 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3845.0, + "completions/mean_length": 1059.735595703125, + "completions/mean_terminated_length": 634.81298828125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 6.830320699708455, + "grad_norm": 0.15374544262886047, + "learning_rate": 1e-06, + "loss": -0.026, + "num_tokens": 445277090.0, + "reward": 0.6116071939468384, + "reward_std": 0.1599598526954651, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.4876568913459778, + "step": 731 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4089.0, + "completions/mean_length": 921.9420166015625, + "completions/mean_terminated_length": 657.1173095703125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.839650145772595, + "grad_norm": 0.14172156155109406, + "learning_rate": 1e-06, + "loss": -0.0343, + "num_tokens": 445911110.0, + "reward": 0.6439732313156128, + "reward_std": 0.16476628184318542, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 732 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3452.0, + "completions/mean_length": 923.560302734375, + "completions/mean_terminated_length": 591.0604248046875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 6.848979591836734, + "grad_norm": 0.15090593695640564, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 446491548.0, + "reward": 0.6573660969734192, + "reward_std": 0.16721759736537933, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 733 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2433.0, + "completions/mean_length": 929.9810791015625, + "completions/mean_terminated_length": 602.4617919921875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 6.858309037900875, + "grad_norm": 0.13988660275936127, + "learning_rate": 1e-06, + "loss": -0.0618, + "num_tokens": 447069459.0, + "reward": 0.645089328289032, + "reward_std": 0.1465412974357605, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 734 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3668.0, + "completions/mean_length": 1077.0726318359375, + "completions/mean_terminated_length": 614.7142944335938, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 6.867638483965014, + "grad_norm": 0.16139984130859375, + "learning_rate": 1e-06, + "loss": -0.0665, + "num_tokens": 447646876.0, + "reward": 0.6160714626312256, + "reward_std": 0.13996683061122894, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2852.0, + "completions/mean_length": 1014.2410888671875, + "completions/mean_terminated_length": 596.3092651367188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 6.876967930029155, + "grad_norm": 0.14299577474594116, + "learning_rate": 1e-06, + "loss": -0.0561, + "num_tokens": 448217500.0, + "reward": 0.6595982313156128, + "reward_std": 0.13617070019245148, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 736 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2525.0, + "completions/mean_length": 862.4129638671875, + "completions/mean_terminated_length": 579.8665161132812, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 6.886297376093294, + "grad_norm": 0.11797186732292175, + "learning_rate": 1e-06, + "loss": -0.0234, + "num_tokens": 448794326.0, + "reward": 0.6941964626312256, + "reward_std": 0.11922591924667358, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 737 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 989.216552734375, + "completions/mean_terminated_length": 672.0418090820312, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 6.895626822157435, + "grad_norm": 0.1549098640680313, + "learning_rate": 1e-06, + "loss": -0.0284, + "num_tokens": 449434768.0, + "reward": 0.6149553656578064, + "reward_std": 0.1636369228363037, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 738 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 912.7355346679688, + "completions/mean_terminated_length": 579.10107421875, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 6.904956268221574, + "grad_norm": 0.14883829653263092, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 449985603.0, + "reward": 0.6551339626312256, + "reward_std": 0.15169291198253632, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3245.0, + "completions/mean_length": 1001.7991333007812, + "completions/mean_terminated_length": 639.13720703125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 6.914285714285715, + "grad_norm": 0.12948399782180786, + "learning_rate": 1e-06, + "loss": -0.0302, + "num_tokens": 450592591.0, + "reward": 0.5948660969734192, + "reward_std": 0.1337614506483078, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 740 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3968.0, + "completions/mean_length": 851.8638916015625, + "completions/mean_terminated_length": 581.1922607421875, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 6.923615160349854, + "grad_norm": 0.15998977422714233, + "learning_rate": 1e-06, + "loss": -0.0512, + "num_tokens": 451153725.0, + "reward": 0.7120535969734192, + "reward_std": 0.19448630511760712, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 741 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2984.0, + "completions/mean_length": 918.7131958007812, + "completions/mean_terminated_length": 602.9337158203125, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 6.932944606413994, + "grad_norm": 0.1429097205400467, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 451729620.0, + "reward": 0.6584821939468384, + "reward_std": 0.15169358253479004, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 742 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3919.0, + "completions/mean_length": 830.9141235351562, + "completions/mean_terminated_length": 583.9747924804688, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 6.942274052478134, + "grad_norm": 0.14885066449642181, + "learning_rate": 1e-06, + "loss": -0.0336, + "num_tokens": 452311039.0, + "reward": 0.6540178656578064, + "reward_std": 0.14469726383686066, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4050.0, + "completions/mean_length": 1033.8203125, + "completions/mean_terminated_length": 662.0663452148438, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.9516034985422746, + "grad_norm": 0.20462048053741455, + "learning_rate": 1e-06, + "loss": -0.0529, + "num_tokens": 452926334.0, + "reward": 0.5970982313156128, + "reward_std": 0.14797276258468628, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3846.0, + "completions/mean_length": 979.1563110351562, + "completions/mean_terminated_length": 631.12158203125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.960932944606414, + "grad_norm": 0.1431466042995453, + "learning_rate": 1e-06, + "loss": -0.0692, + "num_tokens": 453525578.0, + "reward": 0.6495535969734192, + "reward_std": 0.17400752007961273, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 745 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 986.2031860351562, + "completions/mean_terminated_length": 599.9171752929688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 6.970262390670554, + "grad_norm": 0.14980660378932953, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 454094632.0, + "reward": 0.660714328289032, + "reward_std": 0.156576007604599, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 746 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 961.2969360351562, + "completions/mean_terminated_length": 589.5155639648438, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.979591836734694, + "grad_norm": 0.1564895510673523, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 454653346.0, + "reward": 0.6194196939468384, + "reward_std": 0.17081506550312042, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 747 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3631.0, + "completions/mean_length": 973.3303833007812, + "completions/mean_terminated_length": 624.6451416015625, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 6.988921282798834, + "grad_norm": 0.13402433693408966, + "learning_rate": 1e-06, + "loss": -0.0175, + "num_tokens": 455237186.0, + "reward": 0.6852678656578064, + "reward_std": 0.11468092352151871, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 748 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.13068181818181823, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2105.0, + "completions/mean_length": 1031.70458984375, + "completions/mean_terminated_length": 571.058837890625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.998250728862974, + "grad_norm": 0.15120480954647064, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 455789856.0, + "reward": 0.6082589626312256, + "reward_std": 0.1374509036540985, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 749 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2845.0, + "completions/mean_length": 1026.2489013671875, + "completions/mean_terminated_length": 601.0863647460938, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.0093294460641395, + "grad_norm": 0.1724964827299118, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 456358799.0, + "reward": 0.6428571939468384, + "reward_std": 0.18479041755199432, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 750 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3597.0, + "completions/mean_length": 910.23779296875, + "completions/mean_terminated_length": 585.0, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 7.01865889212828, + "grad_norm": 0.16510646045207977, + "learning_rate": 1e-06, + "loss": -0.0469, + "num_tokens": 456925588.0, + "reward": 0.6651785969734192, + "reward_std": 0.15650072693824768, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 751 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3010.0, + "completions/mean_length": 1020.1574096679688, + "completions/mean_terminated_length": 589.6959228515625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 7.0279883381924195, + "grad_norm": 0.14759129285812378, + "learning_rate": 1e-06, + "loss": -0.0536, + "num_tokens": 457473401.0, + "reward": 0.6551339626312256, + "reward_std": 0.15037837624549866, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 752 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3612.0, + "completions/mean_length": 1010.2935791015625, + "completions/mean_terminated_length": 648.627197265625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 7.03731778425656, + "grad_norm": 0.15122123062610626, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 458079256.0, + "reward": 0.645089328289032, + "reward_std": 0.1657833456993103, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 753 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3690.0, + "completions/mean_length": 1047.5592041015625, + "completions/mean_terminated_length": 603.1572875976562, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.0466472303206995, + "grad_norm": 0.1405087411403656, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 458637829.0, + "reward": 0.6618303656578064, + "reward_std": 0.13647380471229553, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 754 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2927.0, + "completions/mean_length": 1016.2545166015625, + "completions/mean_terminated_length": 642.3679809570312, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 7.05597667638484, + "grad_norm": 0.14581283926963806, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 459251761.0, + "reward": 0.625, + "reward_std": 0.14744029939174652, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 755 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3745.0, + "completions/mean_length": 960.5703735351562, + "completions/mean_terminated_length": 648.950927734375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.0653061224489795, + "grad_norm": 0.16463600099086761, + "learning_rate": 1e-06, + "loss": -0.0216, + "num_tokens": 459869728.0, + "reward": 0.6595982313156128, + "reward_std": 0.14688828587532043, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2517.0, + "completions/mean_length": 911.763427734375, + "completions/mean_terminated_length": 573.6839599609375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.07463556851312, + "grad_norm": 0.13060037791728973, + "learning_rate": 1e-06, + "loss": -0.0026, + "num_tokens": 460421652.0, + "reward": 0.6875000596046448, + "reward_std": 0.10295554995536804, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 757 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2671.0, + "completions/mean_length": 982.70654296875, + "completions/mean_terminated_length": 591.5891723632812, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 7.0839650145772595, + "grad_norm": 0.14986397325992584, + "learning_rate": 1e-06, + "loss": -0.0647, + "num_tokens": 460980549.0, + "reward": 0.6651785969734192, + "reward_std": 0.16799576580524445, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 758 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2811.0, + "completions/mean_length": 1005.8381958007812, + "completions/mean_terminated_length": 568.8853759765625, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 7.093294460641399, + "grad_norm": 0.15561595559120178, + "learning_rate": 1e-06, + "loss": -0.0749, + "num_tokens": 461519324.0, + "reward": 0.637276828289032, + "reward_std": 0.17356175184249878, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 759 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3806.0, + "completions/mean_length": 950.8225708007812, + "completions/mean_terminated_length": 625.4593505859375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.1026239067055394, + "grad_norm": 0.1461905837059021, + "learning_rate": 1e-06, + "loss": -0.031, + "num_tokens": 462119117.0, + "reward": 0.6439732313156128, + "reward_std": 0.14902400970458984, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 760 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3537.0, + "completions/mean_length": 994.786865234375, + "completions/mean_terminated_length": 574.2167358398438, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 7.111953352769679, + "grad_norm": 0.14319172501564026, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 462654902.0, + "reward": 0.645089328289032, + "reward_std": 0.13403314352035522, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3699.0, + "completions/mean_length": 1089.7578125, + "completions/mean_terminated_length": 660.2946166992188, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 7.121282798833819, + "grad_norm": 0.15229927003383636, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 463262917.0, + "reward": 0.6227678656578064, + "reward_std": 0.1388823539018631, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644601345062, + "step": 762 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 1020.075927734375, + "completions/mean_terminated_length": 629.298095703125, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 7.130612244897959, + "grad_norm": 0.1739799678325653, + "learning_rate": 1e-06, + "loss": -0.0425, + "num_tokens": 463857961.0, + "reward": 0.6261160969734192, + "reward_std": 0.20489785075187683, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 763 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4037.0, + "completions/mean_length": 919.8660888671875, + "completions/mean_terminated_length": 633.9367065429688, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 7.139941690962099, + "grad_norm": 0.13902339339256287, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 464459257.0, + "reward": 0.6875000596046448, + "reward_std": 0.14173556864261627, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 764 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3244.0, + "completions/mean_length": 1117.0390625, + "completions/mean_terminated_length": 634.0687255859375, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 7.149271137026239, + "grad_norm": 0.14938311278820038, + "learning_rate": 1e-06, + "loss": -0.0649, + "num_tokens": 465047596.0, + "reward": 0.640625, + "reward_std": 0.15270251035690308, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 4034.0, + "completions/mean_length": 1020.6027221679688, + "completions/mean_terminated_length": 581.2601928710938, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 7.158600583090379, + "grad_norm": 0.15528525412082672, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 465586256.0, + "reward": 0.6741071939468384, + "reward_std": 0.13925214111804962, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 766 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3763.0, + "completions/mean_length": 924.0022583007812, + "completions/mean_terminated_length": 646.83740234375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 7.167930029154519, + "grad_norm": 0.12575432658195496, + "learning_rate": 1e-06, + "loss": -0.013, + "num_tokens": 466212634.0, + "reward": 0.6082589626312256, + "reward_std": 0.14304685592651367, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.4884119927883148, + "step": 767 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2874.0, + "completions/mean_length": 929.0256958007812, + "completions/mean_terminated_length": 584.107666015625, + "completions/min_length": 194.0, + "completions/min_terminated_length": 194.0, + "epoch": 7.1772594752186585, + "grad_norm": 0.14837078750133514, + "learning_rate": 1e-06, + "loss": -0.068, + "num_tokens": 466770673.0, + "reward": 0.691964328289032, + "reward_std": 0.15920916199684143, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392156600952, + "step": 768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3504.0, + "completions/mean_length": 899.2991333007812, + "completions/mean_terminated_length": 624.1890869140625, + "completions/min_length": 195.0, + "completions/min_terminated_length": 195.0, + "epoch": 7.186588921282799, + "grad_norm": 0.14967486262321472, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 467373669.0, + "reward": 0.6852678656578064, + "reward_std": 0.18547075986862183, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 769 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2592.0, + "completions/mean_length": 1030.08154296875, + "completions/mean_terminated_length": 596.5567016601562, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.1959183673469385, + "grad_norm": 0.1560906171798706, + "learning_rate": 1e-06, + "loss": -0.0659, + "num_tokens": 467934486.0, + "reward": 0.6417410969734192, + "reward_std": 0.14846131205558777, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 770 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4011.0, + "completions/mean_length": 1075.2388916015625, + "completions/mean_terminated_length": 648.099365234375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 7.205247813411079, + "grad_norm": 0.1507735550403595, + "learning_rate": 1e-06, + "loss": -0.0271, + "num_tokens": 468528676.0, + "reward": 0.613839328289032, + "reward_std": 0.1463979184627533, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 771 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2812.0, + "completions/mean_length": 971.9375610351562, + "completions/mean_terminated_length": 605.7755737304688, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.214577259475218, + "grad_norm": 0.13557055592536926, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 469107900.0, + "reward": 0.6863839626312256, + "reward_std": 0.1357244998216629, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 772 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3604.0, + "completions/mean_length": 997.388427734375, + "completions/mean_terminated_length": 550.2069091796875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.223906705539359, + "grad_norm": 0.17375889420509338, + "learning_rate": 1e-06, + "loss": -0.0812, + "num_tokens": 469633592.0, + "reward": 0.606026828289032, + "reward_std": 0.1640166938304901, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 773 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4060.0, + "completions/mean_length": 940.03466796875, + "completions/mean_terminated_length": 609.2614135742188, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.233236151603498, + "grad_norm": 0.14711524546146393, + "learning_rate": 1e-06, + "loss": -0.0118, + "num_tokens": 470219519.0, + "reward": 0.640625, + "reward_std": 0.14214855432510376, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 774 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3198.0, + "completions/mean_length": 1053.8460693359375, + "completions/mean_terminated_length": 663.040283203125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 7.242565597667639, + "grad_norm": 0.1405188888311386, + "learning_rate": 1e-06, + "loss": -0.0667, + "num_tokens": 470837261.0, + "reward": 0.668526828289032, + "reward_std": 0.1637124866247177, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 775 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 1063.1473388671875, + "completions/mean_terminated_length": 651.847900390625, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 7.251895043731778, + "grad_norm": 0.15649832785129547, + "learning_rate": 1e-06, + "loss": -0.0607, + "num_tokens": 471437553.0, + "reward": 0.6540178656578064, + "reward_std": 0.14928685128688812, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3452.0, + "completions/mean_length": 907.5547485351562, + "completions/mean_terminated_length": 577.7155151367188, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 7.261224489795918, + "grad_norm": 0.13619768619537354, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 471988922.0, + "reward": 0.6886160969734192, + "reward_std": 0.15026184916496277, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 777 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 949.4810791015625, + "completions/mean_terminated_length": 649.4462280273438, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 7.270553935860058, + "grad_norm": 0.14237339794635773, + "learning_rate": 1e-06, + "loss": -0.0526, + "num_tokens": 472605641.0, + "reward": 0.6741071939468384, + "reward_std": 0.15361180901527405, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 778 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2924.0, + "completions/mean_length": 1091.9129638671875, + "completions/mean_terminated_length": 697.4368896484375, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 7.279883381924198, + "grad_norm": 0.17100362479686737, + "learning_rate": 1e-06, + "loss": -0.0293, + "num_tokens": 473250547.0, + "reward": 0.5993303656578064, + "reward_std": 0.1982492059469223, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 779 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 909.9207763671875, + "completions/mean_terminated_length": 601.8421020507812, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 7.289212827988338, + "grad_norm": 0.15483121573925018, + "learning_rate": 1e-06, + "loss": -0.0242, + "num_tokens": 473830476.0, + "reward": 0.6339285969734192, + "reward_std": 0.17097026109695435, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 780 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3814.0, + "completions/mean_length": 907.6428833007812, + "completions/mean_terminated_length": 590.76318359375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.298542274052478, + "grad_norm": 0.14821454882621765, + "learning_rate": 1e-06, + "loss": -0.0387, + "num_tokens": 474394252.0, + "reward": 0.6930803656578064, + "reward_std": 0.14398936927318573, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 781 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4092.0, + "completions/mean_length": 1057.7421875, + "completions/mean_terminated_length": 641.3311767578125, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 7.307871720116618, + "grad_norm": 0.1600455492734909, + "learning_rate": 1e-06, + "loss": -0.077, + "num_tokens": 474993029.0, + "reward": 0.6160714626312256, + "reward_std": 0.1810038536787033, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 782 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3665.0, + "completions/mean_length": 989.4029541015625, + "completions/mean_terminated_length": 625.2880249023438, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 7.317201166180758, + "grad_norm": 0.14733344316482544, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 475579854.0, + "reward": 0.6752232313156128, + "reward_std": 0.16022127866744995, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 783 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 920.8817138671875, + "completions/mean_terminated_length": 566.3399658203125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.326530612244898, + "grad_norm": 0.15291845798492432, + "learning_rate": 1e-06, + "loss": -0.0857, + "num_tokens": 476126268.0, + "reward": 0.6863839626312256, + "reward_std": 0.15285545587539673, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 784 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3022.0, + "completions/mean_length": 929.0234985351562, + "completions/mean_terminated_length": 601.4051513671875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 7.335860058309038, + "grad_norm": 0.14991307258605957, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 476696249.0, + "reward": 0.6662946939468384, + "reward_std": 0.15007638931274414, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 785 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3840.0, + "completions/mean_length": 1021.325927734375, + "completions/mean_terminated_length": 626.342529296875, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.345189504373177, + "grad_norm": 0.13734778761863708, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 477283549.0, + "reward": 0.606026828289032, + "reward_std": 0.14091113209724426, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 786 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 1034.755615234375, + "completions/mean_terminated_length": 650.1771240234375, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 7.354518950437318, + "grad_norm": 0.16111509501934052, + "learning_rate": 1e-06, + "loss": -0.0525, + "num_tokens": 477881346.0, + "reward": 0.65625, + "reward_std": 0.13042178750038147, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 787 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 934.5223388671875, + "completions/mean_terminated_length": 670.747314453125, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.363848396501457, + "grad_norm": 0.13426731526851654, + "learning_rate": 1e-06, + "loss": -0.0221, + "num_tokens": 478531622.0, + "reward": 0.6808035969734192, + "reward_std": 0.1428300142288208, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 788 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3587.0, + "completions/mean_length": 953.122802734375, + "completions/mean_terminated_length": 619.4345703125, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.373177842565598, + "grad_norm": 0.14411106705665588, + "learning_rate": 1e-06, + "loss": -0.0371, + "num_tokens": 479119204.0, + "reward": 0.7087053656578064, + "reward_std": 0.1360633224248886, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 789 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3287.0, + "completions/mean_length": 824.443115234375, + "completions/mean_terminated_length": 593.83154296875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.382507288629737, + "grad_norm": 0.15481586754322052, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 479706241.0, + "reward": 0.676339328289032, + "reward_std": 0.17938736081123352, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 790 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3903.0, + "completions/mean_length": 925.86279296875, + "completions/mean_terminated_length": 602.2201538085938, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 7.391836734693878, + "grad_norm": 0.14867083728313446, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 480286422.0, + "reward": 0.6975446939468384, + "reward_std": 0.1306488960981369, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 791 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2666.0, + "completions/mean_length": 863.3482666015625, + "completions/mean_terminated_length": 559.4237060546875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 7.401166180758017, + "grad_norm": 0.15872924029827118, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 480842262.0, + "reward": 0.6462053656578064, + "reward_std": 0.15240898728370667, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3198.0, + "completions/mean_length": 964.70654296875, + "completions/mean_terminated_length": 571.327880859375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.410495626822158, + "grad_norm": 0.14340198040008545, + "learning_rate": 1e-06, + "loss": -0.0253, + "num_tokens": 481386087.0, + "reward": 0.6819196939468384, + "reward_std": 0.12166907638311386, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 793 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 880.7935791015625, + "completions/mean_terminated_length": 641.7733764648438, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 7.419825072886297, + "grad_norm": 0.19853998720645905, + "learning_rate": 1e-06, + "loss": -0.0173, + "num_tokens": 482007174.0, + "reward": 0.6808035969734192, + "reward_std": 0.15431898832321167, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 794 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3523.0, + "completions/mean_length": 878.8359985351562, + "completions/mean_terminated_length": 572.0648193359375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 7.429154518950437, + "grad_norm": 0.15817876160144806, + "learning_rate": 1e-06, + "loss": -0.0458, + "num_tokens": 482555395.0, + "reward": 0.7031250596046448, + "reward_std": 0.13662534952163696, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 795 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3699.0, + "completions/mean_length": 1139.6796875, + "completions/mean_terminated_length": 725.9453125, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 7.438483965014577, + "grad_norm": 0.140898197889328, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 483225604.0, + "reward": 0.5870535969734192, + "reward_std": 0.16953669488430023, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 796 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3069.0, + "completions/mean_length": 1163.546875, + "completions/mean_terminated_length": 647.8661499023438, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.447813411078717, + "grad_norm": 0.16424517333507538, + "learning_rate": 1e-06, + "loss": -0.084, + "num_tokens": 483806214.0, + "reward": 0.59375, + "reward_std": 0.19643910229206085, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 797 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3591.0, + "completions/mean_length": 1056.661865234375, + "completions/mean_terminated_length": 644.48291015625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 7.457142857142857, + "grad_norm": 0.15525534749031067, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 484410879.0, + "reward": 0.6651785969734192, + "reward_std": 0.16991646587848663, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 798 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4079.0, + "completions/mean_length": 979.09716796875, + "completions/mean_terminated_length": 626.7515258789062, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 7.466472303206997, + "grad_norm": 0.1439310908317566, + "learning_rate": 1e-06, + "loss": -0.0479, + "num_tokens": 485007070.0, + "reward": 0.6930803656578064, + "reward_std": 0.1469549834728241, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 799 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3260.0, + "completions/mean_length": 932.7489013671875, + "completions/mean_terminated_length": 635.3492431640625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 7.475801749271137, + "grad_norm": 0.1319865733385086, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 485618653.0, + "reward": 0.6774553656578064, + "reward_std": 0.13624626398086548, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 800 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 1058.0770263671875, + "completions/mean_terminated_length": 628.5108642578125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 7.485131195335277, + "grad_norm": 0.17439578473567963, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 486194834.0, + "reward": 0.6361607313156128, + "reward_std": 0.18791463971138, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 801 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3046.0, + "completions/mean_length": 1086.14404296875, + "completions/mean_terminated_length": 625.175048828125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 7.494460641399417, + "grad_norm": 0.1541268229484558, + "learning_rate": 1e-06, + "loss": -0.069, + "num_tokens": 486765475.0, + "reward": 0.6540178656578064, + "reward_std": 0.164872944355011, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 802 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3808.0, + "completions/mean_length": 1109.700927734375, + "completions/mean_terminated_length": 647.9020385742188, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.503790087463557, + "grad_norm": 0.14485584199428558, + "learning_rate": 1e-06, + "loss": -0.0446, + "num_tokens": 487362919.0, + "reward": 0.6428571939468384, + "reward_std": 0.13996681571006775, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 803 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3594.0, + "completions/mean_length": 1082.977783203125, + "completions/mean_terminated_length": 608.056884765625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 7.513119533527696, + "grad_norm": 0.18010494112968445, + "learning_rate": 1e-06, + "loss": -0.0716, + "num_tokens": 487925395.0, + "reward": 0.6484375, + "reward_std": 0.1855827420949936, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3636.0, + "completions/mean_length": 1021.4453735351562, + "completions/mean_terminated_length": 630.8414916992188, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 7.522448979591837, + "grad_norm": 0.15672233700752258, + "learning_rate": 1e-06, + "loss": -0.0813, + "num_tokens": 488512066.0, + "reward": 0.6662946939468384, + "reward_std": 0.18678346276283264, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 805 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2837.0, + "completions/mean_length": 1024.31591796875, + "completions/mean_terminated_length": 625.3455200195312, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 7.531778425655976, + "grad_norm": 0.13961265981197357, + "learning_rate": 1e-06, + "loss": -0.0253, + "num_tokens": 489087253.0, + "reward": 0.6417410969734192, + "reward_std": 0.10915841907262802, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 806 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 1130.915283203125, + "completions/mean_terminated_length": 609.4960327148438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.541107871720117, + "grad_norm": 0.16179832816123962, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 489639329.0, + "reward": 0.5870535969734192, + "reward_std": 0.12253670394420624, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 807 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 994.7969360351562, + "completions/mean_terminated_length": 678.19189453125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 7.550437317784256, + "grad_norm": 0.12515318393707275, + "learning_rate": 1e-06, + "loss": -0.0474, + "num_tokens": 490270091.0, + "reward": 0.5915178656578064, + "reward_std": 0.1443263590335846, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 808 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 1075.321533203125, + "completions/mean_terminated_length": 665.6729736328125, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 7.559766763848397, + "grad_norm": 0.15811415016651154, + "learning_rate": 1e-06, + "loss": -0.0883, + "num_tokens": 490883131.0, + "reward": 0.629464328289032, + "reward_std": 0.1891588270664215, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 809 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2694.0, + "completions/mean_length": 979.1004638671875, + "completions/mean_terminated_length": 600.703369140625, + "completions/min_length": 198.0, + "completions/min_terminated_length": 198.0, + "epoch": 7.569096209912536, + "grad_norm": 0.16665959358215332, + "learning_rate": 1e-06, + "loss": -0.0393, + "num_tokens": 491465133.0, + "reward": 0.6328125, + "reward_std": 0.17394039034843445, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 810 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2918.0, + "completions/mean_length": 1056.6015625, + "completions/mean_terminated_length": 653.1416015625, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 7.578425655976677, + "grad_norm": 0.14888350665569305, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 492067320.0, + "reward": 0.6261160969734192, + "reward_std": 0.1754724681377411, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 811 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 1046.505615234375, + "completions/mean_terminated_length": 637.3328857421875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.587755102040816, + "grad_norm": 0.15476591885089874, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 492663789.0, + "reward": 0.613839328289032, + "reward_std": 0.16751113533973694, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 812 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3923.0, + "completions/mean_length": 1062.5692138671875, + "completions/mean_terminated_length": 624.7943725585938, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 7.597084548104956, + "grad_norm": 0.1265028417110443, + "learning_rate": 1e-06, + "loss": -0.0847, + "num_tokens": 493240459.0, + "reward": 0.6741071939468384, + "reward_std": 0.13162527978420258, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 813 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2685.0, + "completions/mean_length": 807.7500610351562, + "completions/mean_terminated_length": 541.9927978515625, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 7.606413994169096, + "grad_norm": 0.17390672862529755, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 493778947.0, + "reward": 0.7332589626312256, + "reward_std": 0.15353690087795258, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425028860569, + "step": 814 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2613.0, + "completions/mean_length": 1092.364990234375, + "completions/mean_terminated_length": 591.7591552734375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 7.615743440233236, + "grad_norm": 0.14996708929538727, + "learning_rate": 1e-06, + "loss": -0.0481, + "num_tokens": 494318954.0, + "reward": 0.6696428656578064, + "reward_std": 0.11276020854711533, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3920.0, + "completions/mean_length": 1029.2467041015625, + "completions/mean_terminated_length": 648.307373046875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.625072886297376, + "grad_norm": 0.16492515802383423, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 494921807.0, + "reward": 0.6049107313156128, + "reward_std": 0.19456438720226288, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 816 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2582.0, + "completions/mean_length": 1058.930908203125, + "completions/mean_terminated_length": 598.295654296875, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 7.634402332361516, + "grad_norm": 0.13783925771713257, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 495472913.0, + "reward": 0.6316964626312256, + "reward_std": 0.1241518035531044, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 817 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3297.0, + "completions/mean_length": 1077.78125, + "completions/mean_terminated_length": 624.46728515625, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 7.643731778425656, + "grad_norm": 0.14262299239635468, + "learning_rate": 1e-06, + "loss": -0.0805, + "num_tokens": 496050789.0, + "reward": 0.6127232313156128, + "reward_std": 0.1150946095585823, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3792.0, + "completions/mean_length": 1025.185302734375, + "completions/mean_terminated_length": 665.2643432617188, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 7.653061224489796, + "grad_norm": 0.179700568318367, + "learning_rate": 1e-06, + "loss": -0.0663, + "num_tokens": 496681899.0, + "reward": 0.6462053656578064, + "reward_std": 0.19621339440345764, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3769.0, + "completions/mean_length": 867.5513916015625, + "completions/mean_terminated_length": 585.4539184570312, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.662390670553936, + "grad_norm": 0.13440033793449402, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 497246881.0, + "reward": 0.7287946939468384, + "reward_std": 0.1191510558128357, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 820 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 1038.3114013671875, + "completions/mean_terminated_length": 692.6596069335938, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 7.671720116618076, + "grad_norm": 0.16319260001182556, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 497883560.0, + "reward": 0.6383928656578064, + "reward_std": 0.18257686495780945, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4082.0, + "completions/mean_length": 1162.907470703125, + "completions/mean_terminated_length": 647.1141967773438, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 7.681049562682215, + "grad_norm": 0.1864021271467209, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 498472957.0, + "reward": 0.582589328289032, + "reward_std": 0.16871324181556702, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 822 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2861.0, + "completions/mean_length": 1001.0201416015625, + "completions/mean_terminated_length": 581.2952880859375, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 7.690379008746356, + "grad_norm": 0.17237922549247742, + "learning_rate": 1e-06, + "loss": -0.0565, + "num_tokens": 499017711.0, + "reward": 0.6674107313156128, + "reward_std": 0.15991777181625366, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140389680862427, + "step": 823 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4072.0, + "completions/mean_length": 1097.649658203125, + "completions/mean_terminated_length": 602.4733276367188, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 7.699708454810495, + "grad_norm": 0.16699232161045074, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 499567957.0, + "reward": 0.6127232313156128, + "reward_std": 0.13565146923065186, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 824 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3463.0, + "completions/mean_length": 1012.4185791015625, + "completions/mean_terminated_length": 633.7330932617188, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 7.709037900874636, + "grad_norm": 0.1604796200990677, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 500168724.0, + "reward": 0.6462053656578064, + "reward_std": 0.17299792170524597, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 825 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2477.0, + "completions/mean_length": 1059.7757568359375, + "completions/mean_terminated_length": 603.7573852539062, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 7.718367346938775, + "grad_norm": 0.16491541266441345, + "learning_rate": 1e-06, + "loss": -0.0648, + "num_tokens": 500733163.0, + "reward": 0.6595982313156128, + "reward_std": 0.14812389016151428, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 826 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3093.0, + "completions/mean_length": 1029.9140625, + "completions/mean_terminated_length": 618.5151977539062, + "completions/min_length": 202.0, + "completions/min_terminated_length": 202.0, + "epoch": 7.727696793002916, + "grad_norm": 0.15837045013904572, + "learning_rate": 1e-06, + "loss": -0.0479, + "num_tokens": 501317390.0, + "reward": 0.645089328289032, + "reward_std": 0.16461192071437836, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 827 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3114.0, + "completions/mean_length": 1072.04248046875, + "completions/mean_terminated_length": 670.6321411132812, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 7.737026239067055, + "grad_norm": 0.14271746575832367, + "learning_rate": 1e-06, + "loss": -0.0562, + "num_tokens": 501940604.0, + "reward": 0.6082589626312256, + "reward_std": 0.13996751606464386, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 828 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3696.0, + "completions/mean_length": 987.7188110351562, + "completions/mean_terminated_length": 601.62109375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 7.746355685131196, + "grad_norm": 0.16480091214179993, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 502510240.0, + "reward": 0.7142857313156128, + "reward_std": 0.14147454500198364, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 829 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3593.0, + "completions/mean_length": 1021.4676513671875, + "completions/mean_terminated_length": 656.8226928710938, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 7.755685131195335, + "grad_norm": 0.1681492030620575, + "learning_rate": 1e-06, + "loss": -0.0335, + "num_tokens": 503129171.0, + "reward": 0.6495535969734192, + "reward_std": 0.15567445755004883, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 830 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2689.0, + "completions/mean_length": 1121.786865234375, + "completions/mean_terminated_length": 652.9832153320312, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 7.765014577259475, + "grad_norm": 0.1567305028438568, + "learning_rate": 1e-06, + "loss": -0.0799, + "num_tokens": 503721316.0, + "reward": 0.6674107313156128, + "reward_std": 0.16691270470619202, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 831 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3431.0, + "completions/mean_length": 1183.7489013671875, + "completions/mean_terminated_length": 676.1088256835938, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 7.774344023323615, + "grad_norm": 0.14212025701999664, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 504318731.0, + "reward": 0.5926339626312256, + "reward_std": 0.1441006362438202, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3618.0, + "completions/mean_length": 897.48779296875, + "completions/mean_terminated_length": 643.148193359375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 7.783673469387755, + "grad_norm": 0.15433727204799652, + "learning_rate": 1e-06, + "loss": -0.0141, + "num_tokens": 504940736.0, + "reward": 0.6428571939468384, + "reward_std": 0.14556559920310974, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.47942501306533813, + "step": 833 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 978.8795166015625, + "completions/mean_terminated_length": 596.0751953125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 7.793002915451895, + "grad_norm": 0.15658137202262878, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 505513500.0, + "reward": 0.7031250596046448, + "reward_std": 0.14864607155323029, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 834 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3422.0, + "completions/mean_length": 1131.747802734375, + "completions/mean_terminated_length": 677.76318359375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.802332361516035, + "grad_norm": 0.15244808793067932, + "learning_rate": 1e-06, + "loss": -0.0189, + "num_tokens": 506132762.0, + "reward": 0.6261160969734192, + "reward_std": 0.14838463068008423, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 835 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3523.0, + "completions/mean_length": 1018.4609985351562, + "completions/mean_terminated_length": 609.9380493164062, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.811661807580175, + "grad_norm": 0.14682023227214813, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 506715895.0, + "reward": 0.6328125, + "reward_std": 0.14342664182186127, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 836 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3556.0, + "completions/mean_length": 1073.087158203125, + "completions/mean_terminated_length": 632.4066162109375, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 7.820991253644315, + "grad_norm": 0.13297942280769348, + "learning_rate": 1e-06, + "loss": -0.0778, + "num_tokens": 507296341.0, + "reward": 0.6707589626312256, + "reward_std": 0.13778719305992126, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 837 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3708.0, + "completions/mean_length": 1040.8382568359375, + "completions/mean_terminated_length": 599.9271850585938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.830320699708455, + "grad_norm": 0.12285968661308289, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 507853228.0, + "reward": 0.65625, + "reward_std": 0.09630580246448517, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2163.0, + "completions/mean_length": 946.7902221679688, + "completions/mean_terminated_length": 599.4795532226562, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 7.839650145772595, + "grad_norm": 0.16421256959438324, + "learning_rate": 1e-06, + "loss": -0.0755, + "num_tokens": 508432288.0, + "reward": 0.668526828289032, + "reward_std": 0.18483206629753113, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 839 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3049.0, + "completions/mean_length": 991.1473388671875, + "completions/mean_terminated_length": 631.556640625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.848979591836734, + "grad_norm": 0.14624477922916412, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 509024308.0, + "reward": 0.6439732313156128, + "reward_std": 0.14707262814044952, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 840 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3407.0, + "completions/mean_length": 863.888427734375, + "completions/mean_terminated_length": 581.4708862304688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 7.858309037900875, + "grad_norm": 0.16527561843395233, + "learning_rate": 1e-06, + "loss": -0.0455, + "num_tokens": 509600840.0, + "reward": 0.6953125596046448, + "reward_std": 0.1464664340019226, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3342.0, + "completions/mean_length": 995.10498046875, + "completions/mean_terminated_length": 665.8740844726562, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.867638483965014, + "grad_norm": 0.14894086122512817, + "learning_rate": 1e-06, + "loss": -0.059, + "num_tokens": 510224766.0, + "reward": 0.637276828289032, + "reward_std": 0.1710016429424286, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 842 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3342.0, + "completions/mean_length": 888.997802734375, + "completions/mean_terminated_length": 583.1956176757812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 7.876967930029155, + "grad_norm": 0.14595289528369904, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 510790988.0, + "reward": 0.6674107313156128, + "reward_std": 0.13973930478096008, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 843 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 916.6808471679688, + "completions/mean_terminated_length": 544.0424194335938, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.886297376093294, + "grad_norm": 0.13176091015338898, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 511310462.0, + "reward": 0.707589328289032, + "reward_std": 0.10885602235794067, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 844 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2973.0, + "completions/mean_length": 898.08154296875, + "completions/mean_terminated_length": 614.426513671875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 7.895626822157435, + "grad_norm": 0.1341967135667801, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 511913487.0, + "reward": 0.65625, + "reward_std": 0.146052747964859, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 845 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3089.0, + "completions/mean_length": 950.8125610351562, + "completions/mean_terminated_length": 621.1688842773438, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 7.904956268221574, + "grad_norm": 0.15034164488315582, + "learning_rate": 1e-06, + "loss": -0.0341, + "num_tokens": 512513919.0, + "reward": 0.6540178656578064, + "reward_std": 0.146052747964859, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 846 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2751.0, + "completions/mean_length": 963.4754638671875, + "completions/mean_terminated_length": 656.365234375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 7.914285714285715, + "grad_norm": 0.15172356367111206, + "learning_rate": 1e-06, + "loss": -0.0232, + "num_tokens": 513150345.0, + "reward": 0.6875000596046448, + "reward_std": 0.1837814897298813, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 847 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2917.0, + "completions/mean_length": 1067.719970703125, + "completions/mean_terminated_length": 608.417724609375, + "completions/min_length": 192.0, + "completions/min_terminated_length": 192.0, + "epoch": 7.923615160349854, + "grad_norm": 0.1684914082288742, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 513715222.0, + "reward": 0.6361607313156128, + "reward_std": 0.14920878410339355, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 848 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4043.0, + "completions/mean_length": 1003.7210083007812, + "completions/mean_terminated_length": 645.5865478515625, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 7.932944606413994, + "grad_norm": 0.13501520454883575, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 514324860.0, + "reward": 0.65625, + "reward_std": 0.137190580368042, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 849 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3537.0, + "completions/mean_length": 1103.036865234375, + "completions/mean_terminated_length": 635.7484130859375, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 7.942274052478134, + "grad_norm": 0.16624946892261505, + "learning_rate": 1e-06, + "loss": -0.1007, + "num_tokens": 514912341.0, + "reward": 0.6194196939468384, + "reward_std": 0.16465610265731812, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580071330070496, + "step": 850 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3974.0, + "completions/mean_length": 963.2991333007812, + "completions/mean_terminated_length": 656.1715698242188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 7.9516034985422746, + "grad_norm": 0.15555180609226227, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 515533929.0, + "reward": 0.6283482313156128, + "reward_std": 0.172096386551857, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 851 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4006.0, + "completions/mean_length": 1144.5748291015625, + "completions/mean_terminated_length": 657.1482543945312, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.960932944606414, + "grad_norm": 0.18935686349868774, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 516128092.0, + "reward": 0.598214328289032, + "reward_std": 0.1679982841014862, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 852 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2854.0, + "completions/mean_length": 926.00341796875, + "completions/mean_terminated_length": 649.0133666992188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 7.970262390670554, + "grad_norm": 0.15786395967006683, + "learning_rate": 1e-06, + "loss": -0.0538, + "num_tokens": 516743751.0, + "reward": 0.707589328289032, + "reward_std": 0.1711532026529312, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 853 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3564.0, + "completions/mean_length": 977.70654296875, + "completions/mean_terminated_length": 603.51123046875, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 7.979591836734694, + "grad_norm": 0.1329893171787262, + "learning_rate": 1e-06, + "loss": -0.0199, + "num_tokens": 517312008.0, + "reward": 0.707589328289032, + "reward_std": 0.12542173266410828, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3417.0, + "completions/mean_length": 1025.9185791015625, + "completions/mean_terminated_length": 670.3549194335938, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 7.988921282798834, + "grad_norm": 0.15467685461044312, + "learning_rate": 1e-06, + "loss": -0.0606, + "num_tokens": 517951415.0, + "reward": 0.6149553656578064, + "reward_std": 0.16111889481544495, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 855 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.10511363636363635, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3768.0, + "completions/mean_length": 1108.477294921875, + "completions/mean_terminated_length": 757.5618896484375, + "completions/min_length": 231.0, + "completions/min_terminated_length": 231.0, + "epoch": 7.998250728862974, + "grad_norm": 0.1504327952861786, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 518547114.0, + "reward": 0.6517857313156128, + "reward_std": 0.13324150443077087, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 856 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3454.0, + "completions/mean_length": 1041.661865234375, + "completions/mean_terminated_length": 675.1412353515625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 8.00932944606414, + "grad_norm": 0.1281849443912506, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 519181643.0, + "reward": 0.6283482313156128, + "reward_std": 0.14560836553573608, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 857 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2888.0, + "completions/mean_length": 1023.0803833007812, + "completions/mean_terminated_length": 637.03515625, + "completions/min_length": 198.0, + "completions/min_terminated_length": 198.0, + "epoch": 8.018658892128279, + "grad_norm": 0.14101547002792358, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 519784579.0, + "reward": 0.6908482313156128, + "reward_std": 0.15180166065692902, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 858 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3584.0, + "completions/mean_length": 1123.53125, + "completions/mean_terminated_length": 646.0880737304688, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 8.02798833819242, + "grad_norm": 0.13459783792495728, + "learning_rate": 1e-06, + "loss": -0.0282, + "num_tokens": 520367951.0, + "reward": 0.6160714626312256, + "reward_std": 0.1164814680814743, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.4866124987602234, + "step": 859 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3709.0, + "completions/mean_length": 1063.2366943359375, + "completions/mean_terminated_length": 703.5455322265625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 8.03731778425656, + "grad_norm": 0.14415781199932098, + "learning_rate": 1e-06, + "loss": -0.0559, + "num_tokens": 521012059.0, + "reward": 0.6551339626312256, + "reward_std": 0.16484296321868896, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 860 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3998.0, + "completions/mean_length": 985.091552734375, + "completions/mean_terminated_length": 624.7994995117188, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.0466472303207, + "grad_norm": 0.1543746292591095, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 521598821.0, + "reward": 0.6774553656578064, + "reward_std": 0.1519104540348053, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 861 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3893.0, + "completions/mean_length": 1037.724365234375, + "completions/mean_terminated_length": 644.8475952148438, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.055976676384839, + "grad_norm": 0.1572313755750656, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 522192262.0, + "reward": 0.6886160969734192, + "reward_std": 0.16191446781158447, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 862 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4015.0, + "completions/mean_length": 1053.443115234375, + "completions/mean_terminated_length": 645.2012939453125, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 8.06530612244898, + "grad_norm": 0.14160212874412537, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 522806307.0, + "reward": 0.6339285969734192, + "reward_std": 0.14158472418785095, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 863 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3677.0, + "completions/mean_length": 994.9129638671875, + "completions/mean_terminated_length": 665.6617431640625, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 8.07463556851312, + "grad_norm": 0.13232599198818207, + "learning_rate": 1e-06, + "loss": -0.0439, + "num_tokens": 523431381.0, + "reward": 0.6986607313156128, + "reward_std": 0.11791391670703888, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 864 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3664.0, + "completions/mean_length": 1088.2054443359375, + "completions/mean_terminated_length": 614.1085205078125, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 8.08396501457726, + "grad_norm": 0.1556416004896164, + "learning_rate": 1e-06, + "loss": -0.0253, + "num_tokens": 524004125.0, + "reward": 0.6205357313156128, + "reward_std": 0.161197692155838, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3066.0, + "completions/mean_length": 934.4888916015625, + "completions/mean_terminated_length": 607.4359741210938, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 8.093294460641399, + "grad_norm": 0.16776452958583832, + "learning_rate": 1e-06, + "loss": -0.0452, + "num_tokens": 524583003.0, + "reward": 0.6696428656578064, + "reward_std": 0.16698937118053436, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 866 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3467.0, + "completions/mean_length": 1013.8683471679688, + "completions/mean_terminated_length": 678.1906127929688, + "completions/min_length": 202.0, + "completions/min_terminated_length": 202.0, + "epoch": 8.102623906705539, + "grad_norm": 0.14194448292255402, + "learning_rate": 1e-06, + "loss": -0.0366, + "num_tokens": 525211685.0, + "reward": 0.6316964626312256, + "reward_std": 0.15352734923362732, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 973.5457763671875, + "completions/mean_terminated_length": 658.998779296875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 8.11195335276968, + "grad_norm": 0.16999202966690063, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 525845910.0, + "reward": 0.6629464626312256, + "reward_std": 0.1806240826845169, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 868 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3410.0, + "completions/mean_length": 1045.3951416015625, + "completions/mean_terminated_length": 653.5037231445312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 8.12128279883382, + "grad_norm": 0.14440327882766724, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 526446952.0, + "reward": 0.660714328289032, + "reward_std": 0.15300700068473816, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 869 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 1011.40966796875, + "completions/mean_terminated_length": 662.7167358398438, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.130612244897959, + "grad_norm": 0.14697590470314026, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 527065823.0, + "reward": 0.6205357313156128, + "reward_std": 0.15597867965698242, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 870 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3392.0, + "completions/mean_length": 1015.286865234375, + "completions/mean_terminated_length": 645.6012573242188, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 8.139941690962099, + "grad_norm": 0.1197003424167633, + "learning_rate": 1e-06, + "loss": -0.0443, + "num_tokens": 527677616.0, + "reward": 0.6729910969734192, + "reward_std": 0.11539769917726517, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 871 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 967.677490234375, + "completions/mean_terminated_length": 639.8014526367188, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 8.14927113702624, + "grad_norm": 0.17388927936553955, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 528292895.0, + "reward": 0.6674107313156128, + "reward_std": 0.15108489990234375, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 872 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3690.0, + "completions/mean_length": 973.161865234375, + "completions/mean_terminated_length": 628.7596435546875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 8.15860058309038, + "grad_norm": 0.15422604978084564, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 528893016.0, + "reward": 0.6964285969734192, + "reward_std": 0.17156758904457092, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 873 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3953.0, + "completions/mean_length": 1122.3404541015625, + "completions/mean_terminated_length": 675.7188720703125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 8.167930029154519, + "grad_norm": 0.16426236927509308, + "learning_rate": 1e-06, + "loss": -0.0536, + "num_tokens": 529511961.0, + "reward": 0.6484375, + "reward_std": 0.16157424449920654, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 874 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3935.0, + "completions/mean_length": 970.1272583007812, + "completions/mean_terminated_length": 621.0843505859375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 8.177259475218658, + "grad_norm": 0.14696228504180908, + "learning_rate": 1e-06, + "loss": -0.0616, + "num_tokens": 530093051.0, + "reward": 0.668526828289032, + "reward_std": 0.14872092008590698, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2971.0, + "completions/mean_length": 950.3370971679688, + "completions/mean_terminated_length": 559.5960083007812, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 8.186588921282798, + "grad_norm": 0.1521543264389038, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 530631577.0, + "reward": 0.6941964626312256, + "reward_std": 0.1464325189590454, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 876 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2943.0, + "completions/mean_length": 957.1763916015625, + "completions/mean_terminated_length": 619.626708984375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 8.19591836734694, + "grad_norm": 0.16165433824062347, + "learning_rate": 1e-06, + "loss": -0.0465, + "num_tokens": 531232575.0, + "reward": 0.6718750596046448, + "reward_std": 0.17010192573070526, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 877 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 1080.52685546875, + "completions/mean_terminated_length": 640.9309692382812, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 8.205247813411079, + "grad_norm": 0.1745336651802063, + "learning_rate": 1e-06, + "loss": -0.0727, + "num_tokens": 531827103.0, + "reward": 0.6495535969734192, + "reward_std": 0.15352876484394073, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 878 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2929.0, + "completions/mean_length": 935.3438110351562, + "completions/mean_terminated_length": 646.6114501953125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.214577259475218, + "grad_norm": 0.14890854060649872, + "learning_rate": 1e-06, + "loss": -0.0098, + "num_tokens": 532450347.0, + "reward": 0.6662946939468384, + "reward_std": 0.13894835114479065, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 1087.28125, + "completions/mean_terminated_length": 692.1969604492188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.223906705539358, + "grad_norm": 0.15094119310379028, + "learning_rate": 1e-06, + "loss": -0.0465, + "num_tokens": 533087719.0, + "reward": 0.59375, + "reward_std": 0.16506868600845337, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 880 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2684.0, + "completions/mean_length": 925.794677734375, + "completions/mean_terminated_length": 631.970703125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 8.2332361516035, + "grad_norm": 0.15419861674308777, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 533701359.0, + "reward": 0.65625, + "reward_std": 0.17521215975284576, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 881 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3938.0, + "completions/mean_length": 1071.513427734375, + "completions/mean_terminated_length": 643.84716796875, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 8.242565597667639, + "grad_norm": 0.1561041623353958, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 534310291.0, + "reward": 0.6104910969734192, + "reward_std": 0.17322544753551483, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 882 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 961.6741333007812, + "completions/mean_terminated_length": 616.0049438476562, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 8.251895043731778, + "grad_norm": 0.18054859340190887, + "learning_rate": 1e-06, + "loss": -0.0291, + "num_tokens": 534897119.0, + "reward": 0.6339285969734192, + "reward_std": 0.1587563306093216, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3047.0, + "completions/mean_length": 1010.0770263671875, + "completions/mean_terminated_length": 644.0811157226562, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.261224489795918, + "grad_norm": 0.12483751028776169, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 535495828.0, + "reward": 0.640625, + "reward_std": 0.13440406322479248, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 884 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 1054.0748291015625, + "completions/mean_terminated_length": 632.7661743164062, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.270553935860057, + "grad_norm": 0.1630391925573349, + "learning_rate": 1e-06, + "loss": -0.061, + "num_tokens": 536091255.0, + "reward": 0.640625, + "reward_std": 0.1556745022535324, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 885 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2993.0, + "completions/mean_length": 1008.9230346679688, + "completions/mean_terminated_length": 651.3909912109375, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 8.279883381924199, + "grad_norm": 0.14400123059749603, + "learning_rate": 1e-06, + "loss": -0.0066, + "num_tokens": 536705106.0, + "reward": 0.6428571939468384, + "reward_std": 0.1310279667377472, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.47942501306533813, + "step": 886 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3232.0, + "completions/mean_length": 914.4777221679688, + "completions/mean_terminated_length": 619.6048583984375, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 8.289212827988338, + "grad_norm": 0.1641090363264084, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 537297718.0, + "reward": 0.6774553656578064, + "reward_std": 0.19467133283615112, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 887 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3310.0, + "completions/mean_length": 914.794677734375, + "completions/mean_terminated_length": 619.9512329101562, + "completions/min_length": 202.0, + "completions/min_terminated_length": 202.0, + "epoch": 8.298542274052478, + "grad_norm": 0.15631234645843506, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 537896078.0, + "reward": 0.6551339626312256, + "reward_std": 0.14620289206504822, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 888 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2598.0, + "completions/mean_length": 954.6574096679688, + "completions/mean_terminated_length": 629.6908569335938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.307871720116617, + "grad_norm": 0.1582319438457489, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 538493875.0, + "reward": 0.7098214626312256, + "reward_std": 0.14324188232421875, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 889 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3435.0, + "completions/mean_length": 990.7801513671875, + "completions/mean_terminated_length": 635.4564819335938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 8.317201166180759, + "grad_norm": 0.1555420458316803, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 539098374.0, + "reward": 0.6964285969734192, + "reward_std": 0.12700362503528595, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 890 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 1056.364990234375, + "completions/mean_terminated_length": 630.9707641601562, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 8.326530612244898, + "grad_norm": 0.25550377368927, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 539690261.0, + "reward": 0.629464328289032, + "reward_std": 0.16071052849292755, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 891 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2905.0, + "completions/mean_length": 909.872802734375, + "completions/mean_terminated_length": 593.2147216796875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 8.335860058309038, + "grad_norm": 0.14263087511062622, + "learning_rate": 1e-06, + "loss": -0.0237, + "num_tokens": 540261715.0, + "reward": 0.6395089626312256, + "reward_std": 0.14571575820446014, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 892 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3723.0, + "completions/mean_length": 894.8058471679688, + "completions/mean_terminated_length": 644.4115600585938, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 8.345189504373177, + "grad_norm": 0.17056600749492645, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 540878357.0, + "reward": 0.6662946939468384, + "reward_std": 0.1889663189649582, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3472.0, + "completions/mean_length": 970.8984985351562, + "completions/mean_terminated_length": 613.2997436523438, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 8.354518950437317, + "grad_norm": 0.18080608546733856, + "learning_rate": 1e-06, + "loss": -0.0331, + "num_tokens": 541464946.0, + "reward": 0.6506696939468384, + "reward_std": 0.14489202201366425, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 894 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3953.0, + "completions/mean_length": 1053.421875, + "completions/mean_terminated_length": 688.3125, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.363848396501458, + "grad_norm": 0.143610879778862, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 542099068.0, + "reward": 0.621651828289032, + "reward_std": 0.16694408655166626, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 895 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3614.0, + "completions/mean_length": 1019.86279296875, + "completions/mean_terminated_length": 650.7262573242188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.373177842565598, + "grad_norm": 0.15640057623386383, + "learning_rate": 1e-06, + "loss": -0.0583, + "num_tokens": 542703033.0, + "reward": 0.691964328289032, + "reward_std": 0.1497408002614975, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 896 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3837.0, + "completions/mean_length": 907.3638916015625, + "completions/mean_terminated_length": 581.8327026367188, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.382507288629737, + "grad_norm": 0.15166081488132477, + "learning_rate": 1e-06, + "loss": -0.0803, + "num_tokens": 543276783.0, + "reward": 0.6897321939468384, + "reward_std": 0.16543565690517426, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 897 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 931.1563110351562, + "completions/mean_terminated_length": 625.1309814453125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 8.391836734693877, + "grad_norm": 0.1535331904888153, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 543864531.0, + "reward": 0.7287946939468384, + "reward_std": 0.13065001368522644, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 898 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3598.0, + "completions/mean_length": 883.3192138671875, + "completions/mean_terminated_length": 581.2722778320312, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 8.401166180758018, + "grad_norm": 0.1459299474954605, + "learning_rate": 1e-06, + "loss": -0.0106, + "num_tokens": 544432697.0, + "reward": 0.6852678656578064, + "reward_std": 0.13290520012378693, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 899 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3848.0, + "completions/mean_length": 891.98779296875, + "completions/mean_terminated_length": 586.470703125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.410495626822158, + "grad_norm": 0.19860893487930298, + "learning_rate": 1e-06, + "loss": -0.0242, + "num_tokens": 544992830.0, + "reward": 0.7243303656578064, + "reward_std": 0.1822744905948639, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 900 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4006.0, + "completions/mean_length": 984.35498046875, + "completions/mean_terminated_length": 645.462890625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.419825072886297, + "grad_norm": 0.16229510307312012, + "learning_rate": 1e-06, + "loss": -0.0643, + "num_tokens": 545597484.0, + "reward": 0.6774553656578064, + "reward_std": 0.1874253898859024, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 901 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2895.0, + "completions/mean_length": 828.7377319335938, + "completions/mean_terminated_length": 585.8477172851562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 8.429154518950437, + "grad_norm": 0.13886448740959167, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 546167209.0, + "reward": 0.7254464626312256, + "reward_std": 0.12926062941551208, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3234.0, + "completions/mean_length": 949.51123046875, + "completions/mean_terminated_length": 619.731201171875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 8.438483965014576, + "grad_norm": 0.159769669175148, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 546757243.0, + "reward": 0.6785714626312256, + "reward_std": 0.1511283665895462, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 903 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3728.0, + "completions/mean_length": 947.2288208007812, + "completions/mean_terminated_length": 659.582275390625, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 8.447813411078718, + "grad_norm": 0.14023955166339874, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 547395560.0, + "reward": 0.6729910969734192, + "reward_std": 0.13290591537952423, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2821.0, + "completions/mean_length": 898.2053833007812, + "completions/mean_terminated_length": 623.001220703125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 8.457142857142857, + "grad_norm": 0.16417650878429413, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 548005400.0, + "reward": 0.6774553656578064, + "reward_std": 0.1587124466896057, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3877.0, + "completions/mean_length": 969.3214721679688, + "completions/mean_terminated_length": 611.5422973632812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 8.466472303206997, + "grad_norm": 0.1593242734670639, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 548583000.0, + "reward": 0.6986607313156128, + "reward_std": 0.13207922875881195, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 906 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3351.0, + "completions/mean_length": 1042.243408203125, + "completions/mean_terminated_length": 654.28173828125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 8.475801749271136, + "grad_norm": 0.1524815559387207, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 549188602.0, + "reward": 0.6417410969734192, + "reward_std": 0.12610389292240143, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 907 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3177.0, + "completions/mean_length": 1148.8929443359375, + "completions/mean_terminated_length": 662.179443359375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.485131195335278, + "grad_norm": 0.18232788145542145, + "learning_rate": 1e-06, + "loss": -0.069, + "num_tokens": 549781050.0, + "reward": 0.6517857313156128, + "reward_std": 0.1669459044933319, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 908 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 918.5614013671875, + "completions/mean_terminated_length": 632.5145874023438, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.494460641399417, + "grad_norm": 0.14517763257026672, + "learning_rate": 1e-06, + "loss": -0.0206, + "num_tokens": 550387033.0, + "reward": 0.65625, + "reward_std": 0.14086836576461792, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3756.0, + "completions/mean_length": 980.107177734375, + "completions/mean_terminated_length": 657.7733764648438, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 8.503790087463557, + "grad_norm": 0.16488566994667053, + "learning_rate": 1e-06, + "loss": -0.0434, + "num_tokens": 551021209.0, + "reward": 0.6517857313156128, + "reward_std": 0.15206310153007507, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 910 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2943.0, + "completions/mean_length": 796.7344360351562, + "completions/mean_terminated_length": 555.710205078125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 8.513119533527696, + "grad_norm": 0.15691713988780975, + "learning_rate": 1e-06, + "loss": -0.0366, + "num_tokens": 551574083.0, + "reward": 0.7220982313156128, + "reward_std": 0.13760285079479218, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 911 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3972.0, + "completions/mean_length": 1059.30029296875, + "completions/mean_terminated_length": 682.0941162109375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 8.522448979591836, + "grad_norm": 0.15728360414505005, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 552211912.0, + "reward": 0.566964328289032, + "reward_std": 0.14309103786945343, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 912 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 1040.5904541015625, + "completions/mean_terminated_length": 669.6583251953125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 8.531778425655977, + "grad_norm": 0.1604510396718979, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 552832257.0, + "reward": 0.6328125, + "reward_std": 0.1799853891134262, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 913 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2722.0, + "completions/mean_length": 938.3069458007812, + "completions/mean_terminated_length": 568.2032470703125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 8.541107871720117, + "grad_norm": 0.16543976962566376, + "learning_rate": 1e-06, + "loss": -0.0653, + "num_tokens": 553382220.0, + "reward": 0.7031250596046448, + "reward_std": 0.1338823139667511, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 914 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3344.0, + "completions/mean_length": 874.4688110351562, + "completions/mean_terminated_length": 605.6832275390625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 8.550437317784256, + "grad_norm": 0.1473904699087143, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 553971696.0, + "reward": 0.7120535969734192, + "reward_std": 0.13981598615646362, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 915 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 1065.8695068359375, + "completions/mean_terminated_length": 744.15185546875, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 8.559766763848396, + "grad_norm": 0.14510010182857513, + "learning_rate": 1e-06, + "loss": -0.0387, + "num_tokens": 554673475.0, + "reward": 0.574776828289032, + "reward_std": 0.1414317786693573, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3618.0, + "completions/mean_length": 1033.57373046875, + "completions/mean_terminated_length": 635.8057861328125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.569096209912537, + "grad_norm": 0.1547669768333435, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 555266077.0, + "reward": 0.629464328289032, + "reward_std": 0.14958995580673218, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 917 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2662.0, + "completions/mean_length": 932.8449096679688, + "completions/mean_terminated_length": 626.9828491210938, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 8.578425655976677, + "grad_norm": 0.16214434802532196, + "learning_rate": 1e-06, + "loss": -0.0265, + "num_tokens": 555872298.0, + "reward": 0.640625, + "reward_std": 0.14816778898239136, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 918 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2974.0, + "completions/mean_length": 839.763427734375, + "completions/mean_terminated_length": 576.593505859375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.587755102040816, + "grad_norm": 0.1655416637659073, + "learning_rate": 1e-06, + "loss": -0.0484, + "num_tokens": 556445814.0, + "reward": 0.7165178656578064, + "reward_std": 0.16172832250595093, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 919 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3134.0, + "completions/mean_length": 961.5859985351562, + "completions/mean_terminated_length": 675.2509155273438, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.597084548104956, + "grad_norm": 0.14155803620815277, + "learning_rate": 1e-06, + "loss": -0.0173, + "num_tokens": 557082923.0, + "reward": 0.6149553656578064, + "reward_std": 0.14917626976966858, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 920 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2920.0, + "completions/mean_length": 959.739990234375, + "completions/mean_terminated_length": 600.8644409179688, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 8.606413994169095, + "grad_norm": 0.1766148805618286, + "learning_rate": 1e-06, + "loss": -0.0337, + "num_tokens": 557648698.0, + "reward": 0.6875000596046448, + "reward_std": 0.1504889726638794, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 921 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3246.0, + "completions/mean_length": 896.1295166015625, + "completions/mean_terminated_length": 666.4736328125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 8.615743440233237, + "grad_norm": 0.15933699905872345, + "learning_rate": 1e-06, + "loss": -0.024, + "num_tokens": 558289982.0, + "reward": 0.6506696939468384, + "reward_std": 0.16368041932582855, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 922 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2674.0, + "completions/mean_length": 1005.1551513671875, + "completions/mean_terminated_length": 634.2537231445312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 8.625072886297376, + "grad_norm": 0.1732131987810135, + "learning_rate": 1e-06, + "loss": -0.0742, + "num_tokens": 558893569.0, + "reward": 0.6796875596046448, + "reward_std": 0.1779584288597107, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 923 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4087.0, + "completions/mean_length": 945.92529296875, + "completions/mean_terminated_length": 602.8477783203125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 8.634402332361516, + "grad_norm": 0.17209410667419434, + "learning_rate": 1e-06, + "loss": -0.0707, + "num_tokens": 559472990.0, + "reward": 0.6752232313156128, + "reward_std": 0.15511246025562286, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3556.0, + "completions/mean_length": 993.716552734375, + "completions/mean_terminated_length": 638.7288208007812, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 8.643731778425655, + "grad_norm": 0.16843537986278534, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 560078504.0, + "reward": 0.652901828289032, + "reward_std": 0.16239909827709198, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 925 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3466.0, + "completions/mean_length": 936.8348388671875, + "completions/mean_terminated_length": 618.5896606445312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 8.653061224489797, + "grad_norm": 0.160786971449852, + "learning_rate": 1e-06, + "loss": -0.0538, + "num_tokens": 560669916.0, + "reward": 0.6863839626312256, + "reward_std": 0.16897355020046234, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 926 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3447.0, + "completions/mean_length": 847.3047485351562, + "completions/mean_terminated_length": 580.5036010742188, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 8.662390670553936, + "grad_norm": 0.12944777309894562, + "learning_rate": 1e-06, + "loss": -0.0156, + "num_tokens": 561241789.0, + "reward": 0.6517857313156128, + "reward_std": 0.09615495055913925, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 927 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3919.0, + "completions/mean_length": 1026.333740234375, + "completions/mean_terminated_length": 649.3571166992188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.671720116618076, + "grad_norm": 0.17108920216560364, + "learning_rate": 1e-06, + "loss": -0.0709, + "num_tokens": 561849080.0, + "reward": 0.590401828289032, + "reward_std": 0.17250937223434448, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 928 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3197.0, + "completions/mean_length": 944.7489013671875, + "completions/mean_terminated_length": 640.0379638671875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 8.681049562682215, + "grad_norm": 0.16505981981754303, + "learning_rate": 1e-06, + "loss": -0.0298, + "num_tokens": 562460687.0, + "reward": 0.6674107313156128, + "reward_std": 0.1601106822490692, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 929 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3878.0, + "completions/mean_length": 1026.344970703125, + "completions/mean_terminated_length": 618.8685302734375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.690379008746355, + "grad_norm": 0.17340941727161407, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 563036524.0, + "reward": 0.6651785969734192, + "reward_std": 0.14771132171154022, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 930 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3892.0, + "completions/mean_length": 1091.484375, + "completions/mean_terminated_length": 675.3569946289062, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 8.699708454810496, + "grad_norm": 0.16088353097438812, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 563661470.0, + "reward": 0.6194196939468384, + "reward_std": 0.1681487262248993, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2490.0, + "completions/mean_length": 959.5881958007812, + "completions/mean_terminated_length": 647.8711547851562, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 8.709037900874636, + "grad_norm": 0.15052680671215057, + "learning_rate": 1e-06, + "loss": -0.0249, + "num_tokens": 564283093.0, + "reward": 0.6383928656578064, + "reward_std": 0.1403045356273651, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 932 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3959.0, + "completions/mean_length": 1056.3035888671875, + "completions/mean_terminated_length": 687.2791137695312, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 8.718367346938775, + "grad_norm": 0.16645358502864838, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 564920917.0, + "reward": 0.65625, + "reward_std": 0.1550350785255432, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 933 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 1070.3404541015625, + "completions/mean_terminated_length": 690.232421875, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 8.727696793002915, + "grad_norm": 0.16387112438678741, + "learning_rate": 1e-06, + "loss": -0.0079, + "num_tokens": 565556510.0, + "reward": 0.6160714626312256, + "reward_std": 0.1564226597547531, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 934 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3985.0, + "completions/mean_length": 961.8303833007812, + "completions/mean_terminated_length": 603.1940307617188, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 8.737026239067056, + "grad_norm": 0.1628786325454712, + "learning_rate": 1e-06, + "loss": -0.0749, + "num_tokens": 566120942.0, + "reward": 0.6718750596046448, + "reward_std": 0.1592409610748291, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 935 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3378.0, + "completions/mean_length": 869.2902221679688, + "completions/mean_terminated_length": 625.2532958984375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.746355685131196, + "grad_norm": 0.170728400349617, + "learning_rate": 1e-06, + "loss": -0.0338, + "num_tokens": 566729890.0, + "reward": 0.6473214626312256, + "reward_std": 0.1675506979227066, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 936 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2456.0, + "completions/mean_length": 906.0614013671875, + "completions/mean_terminated_length": 614.6541137695312, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.755685131195335, + "grad_norm": 0.1713971644639969, + "learning_rate": 1e-06, + "loss": -0.0362, + "num_tokens": 567327737.0, + "reward": 0.6863839626312256, + "reward_std": 0.16453874111175537, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 937 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3724.0, + "completions/mean_length": 999.1641235351562, + "completions/mean_terminated_length": 661.8848876953125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 8.765014577259475, + "grad_norm": 0.17029625177383423, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 567949212.0, + "reward": 0.6785714626312256, + "reward_std": 0.14473934471607208, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 938 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2817.0, + "completions/mean_length": 767.677490234375, + "completions/mean_terminated_length": 545.789306640625, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 8.774344023323614, + "grad_norm": 0.14980624616146088, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 568491651.0, + "reward": 0.7254464626312256, + "reward_std": 0.14207187294960022, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 939 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 1074.49560546875, + "completions/mean_terminated_length": 629.5877075195312, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.783673469387756, + "grad_norm": 0.17293629050254822, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 569077039.0, + "reward": 0.6339285969734192, + "reward_std": 0.15390461683273315, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 940 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3494.0, + "completions/mean_length": 829.3917846679688, + "completions/mean_terminated_length": 607.46484375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.793002915451895, + "grad_norm": 0.15138781070709229, + "learning_rate": 1e-06, + "loss": -0.0418, + "num_tokens": 569665766.0, + "reward": 0.7120535969734192, + "reward_std": 0.14996832609176636, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 941 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3891.0, + "completions/mean_length": 992.7723388671875, + "completions/mean_terminated_length": 616.0350341796875, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 8.802332361516035, + "grad_norm": 0.16286368668079376, + "learning_rate": 1e-06, + "loss": -0.0106, + "num_tokens": 570243074.0, + "reward": 0.6350446939468384, + "reward_std": 0.14894984662532806, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 942 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 906.7969360351562, + "completions/mean_terminated_length": 602.6919555664062, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 8.811661807580174, + "grad_norm": 0.1696043610572815, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 570826212.0, + "reward": 0.7098214626312256, + "reward_std": 0.15097612142562866, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 943 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3213.0, + "completions/mean_length": 951.2623291015625, + "completions/mean_terminated_length": 659.7987670898438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 8.820991253644316, + "grad_norm": 0.11802743375301361, + "learning_rate": 1e-06, + "loss": -0.0352, + "num_tokens": 571452839.0, + "reward": 0.6629464626312256, + "reward_std": 0.13601915538311005, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 944 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 1012.1529541015625, + "completions/mean_terminated_length": 624.7349243164062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.830320699708455, + "grad_norm": 0.1597726047039032, + "learning_rate": 1e-06, + "loss": -0.0252, + "num_tokens": 572034000.0, + "reward": 0.6484375, + "reward_std": 0.13767842948436737, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 945 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 918.6350708007812, + "completions/mean_terminated_length": 589.9420776367188, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 8.839650145772595, + "grad_norm": 0.18028230965137482, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 572596777.0, + "reward": 0.6819196939468384, + "reward_std": 0.16442857682704926, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 946 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3385.0, + "completions/mean_length": 941.6451416015625, + "completions/mean_terminated_length": 632.3946533203125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 8.848979591836734, + "grad_norm": 0.1487460881471634, + "learning_rate": 1e-06, + "loss": -0.0219, + "num_tokens": 573195331.0, + "reward": 0.6551339626312256, + "reward_std": 0.12985864281654358, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 947 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3945.0, + "completions/mean_length": 1013.7120971679688, + "completions/mean_terminated_length": 686.4568481445312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 8.858309037900874, + "grad_norm": 0.14884889125823975, + "learning_rate": 1e-06, + "loss": -0.0612, + "num_tokens": 573841481.0, + "reward": 0.6629464626312256, + "reward_std": 0.1504134237766266, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 948 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3585.0, + "completions/mean_length": 968.1741333007812, + "completions/mean_terminated_length": 601.5711059570312, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 8.867638483965015, + "grad_norm": 0.17088302969932556, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 574410709.0, + "reward": 0.6595982313156128, + "reward_std": 0.1497747004032135, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 949 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3131.0, + "completions/mean_length": 1037.688720703125, + "completions/mean_terminated_length": 591.8478393554688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 8.876967930029155, + "grad_norm": 0.16182148456573486, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 574963590.0, + "reward": 0.6584821939468384, + "reward_std": 0.11979115009307861, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 950 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2928.0, + "completions/mean_length": 1056.55029296875, + "completions/mean_terminated_length": 635.58447265625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 8.886297376093294, + "grad_norm": 0.1759294718503952, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 575557835.0, + "reward": 0.6316964626312256, + "reward_std": 0.15469878911972046, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3753.0, + "completions/mean_length": 975.974365234375, + "completions/mean_terminated_length": 627.5843505859375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 8.895626822157434, + "grad_norm": 0.15874795615673065, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 576153716.0, + "reward": 0.652901828289032, + "reward_std": 0.1577788144350052, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 952 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 862.3013916015625, + "completions/mean_terminated_length": 609.3646240234375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.904956268221575, + "grad_norm": 0.17351357638835907, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 576751274.0, + "reward": 0.6618303656578064, + "reward_std": 0.1762627214193344, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 953 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3106.0, + "completions/mean_length": 917.1563110351562, + "completions/mean_terminated_length": 618.2906494140625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.914285714285715, + "grad_norm": 0.17315226793289185, + "learning_rate": 1e-06, + "loss": -0.0401, + "num_tokens": 577348886.0, + "reward": 0.6808035969734192, + "reward_std": 0.1712619662284851, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 954 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3196.0, + "completions/mean_length": 1004.9241333007812, + "completions/mean_terminated_length": 612.2213745117188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.923615160349854, + "grad_norm": 0.20299702882766724, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 577919322.0, + "reward": 0.6741071939468384, + "reward_std": 0.18408571183681488, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 955 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3038.0, + "completions/mean_length": 997.677490234375, + "completions/mean_terminated_length": 625.8787231445312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.932944606413994, + "grad_norm": 0.14476415514945984, + "learning_rate": 1e-06, + "loss": -0.0046, + "num_tokens": 578510257.0, + "reward": 0.6328125, + "reward_std": 0.10393194854259491, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 956 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4050.0, + "completions/mean_length": 1154.4632568359375, + "completions/mean_terminated_length": 668.669677734375, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 8.942274052478133, + "grad_norm": 0.17567642033100128, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 579116144.0, + "reward": 0.5680803656578064, + "reward_std": 0.16258454322814941, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 957 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 1015.1417846679688, + "completions/mean_terminated_length": 666.8707885742188, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 8.951603498542275, + "grad_norm": 0.15093271434307098, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 579748223.0, + "reward": 0.6183035969734192, + "reward_std": 0.15924237668514252, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 958 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3779.0, + "completions/mean_length": 1059.5748291015625, + "completions/mean_terminated_length": 621.3677978515625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.960932944606414, + "grad_norm": 0.1546793133020401, + "learning_rate": 1e-06, + "loss": -0.05, + "num_tokens": 580325962.0, + "reward": 0.6640625, + "reward_std": 0.1260286271572113, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 959 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3806.0, + "completions/mean_length": 1076.0145263671875, + "completions/mean_terminated_length": 679.4507446289062, + "completions/min_length": 197.0, + "completions/min_terminated_length": 197.0, + "epoch": 8.970262390670554, + "grad_norm": 0.16840921342372894, + "learning_rate": 1e-06, + "loss": -0.0753, + "num_tokens": 580960807.0, + "reward": 0.652901828289032, + "reward_std": 0.16589029133319855, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 960 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3538.0, + "completions/mean_length": 895.0982666015625, + "completions/mean_terminated_length": 606.9391479492188, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.979591836734693, + "grad_norm": 0.14716318249702454, + "learning_rate": 1e-06, + "loss": -0.0383, + "num_tokens": 581548271.0, + "reward": 0.7243303656578064, + "reward_std": 0.12230847775936127, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3686.0, + "completions/mean_length": 1048.3046875, + "completions/mean_terminated_length": 643.743408203125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 8.988921282798835, + "grad_norm": 0.158160999417305, + "learning_rate": 1e-06, + "loss": -0.0451, + "num_tokens": 582154744.0, + "reward": 0.59375, + "reward_std": 0.1725100725889206, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 962 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.08806818181818177, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2611.0, + "completions/mean_length": 963.0568237304688, + "completions/mean_terminated_length": 660.4984130859375, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 8.998250728862974, + "grad_norm": 0.1551835685968399, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 582796379.0, + "reward": 0.6171875, + "reward_std": 0.17074090242385864, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 963 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3567.0, + "completions/mean_length": 1075.719970703125, + "completions/mean_terminated_length": 696.2876586914062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.00932944606414, + "grad_norm": 0.17357894778251648, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 583442832.0, + "reward": 0.6551339626312256, + "reward_std": 0.18359534442424774, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 964 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4023.0, + "completions/mean_length": 943.8638916015625, + "completions/mean_terminated_length": 574.4114990234375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 9.018658892128279, + "grad_norm": 0.1535881757736206, + "learning_rate": 1e-06, + "loss": -0.0146, + "num_tokens": 583989494.0, + "reward": 0.6930803656578064, + "reward_std": 0.11445268988609314, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 965 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2672.0, + "completions/mean_length": 913.6350708007812, + "completions/mean_terminated_length": 605.91552734375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 9.02798833819242, + "grad_norm": 0.15367503464221954, + "learning_rate": 1e-06, + "loss": -0.049, + "num_tokens": 584581199.0, + "reward": 0.6752232313156128, + "reward_std": 0.1781042218208313, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 966 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2673.0, + "completions/mean_length": 1067.841552734375, + "completions/mean_terminated_length": 648.4396362304688, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.03731778425656, + "grad_norm": 0.16338209807872772, + "learning_rate": 1e-06, + "loss": -0.0812, + "num_tokens": 585181921.0, + "reward": 0.6473214626312256, + "reward_std": 0.1717950999736786, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 967 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3597.0, + "completions/mean_length": 1065.376220703125, + "completions/mean_terminated_length": 667.4154052734375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 9.0466472303207, + "grad_norm": 0.16495244204998016, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 585816362.0, + "reward": 0.609375, + "reward_std": 0.1637563705444336, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 968 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3891.0, + "completions/mean_length": 979.1886596679688, + "completions/mean_terminated_length": 613.8765869140625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.055976676384839, + "grad_norm": 0.15480107069015503, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 586393331.0, + "reward": 0.6964285969734192, + "reward_std": 0.15631386637687683, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 969 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2841.0, + "completions/mean_length": 846.8516235351562, + "completions/mean_terminated_length": 567.2279052734375, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 9.06530612244898, + "grad_norm": 0.15223339200019836, + "learning_rate": 1e-06, + "loss": -0.0323, + "num_tokens": 586945822.0, + "reward": 0.7198660969734192, + "reward_std": 0.13193020224571228, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 970 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3711.0, + "completions/mean_length": 1049.376220703125, + "completions/mean_terminated_length": 657.9962158203125, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 9.07463556851312, + "grad_norm": 0.15771889686584473, + "learning_rate": 1e-06, + "loss": -0.0285, + "num_tokens": 587560679.0, + "reward": 0.625, + "reward_std": 0.14496827125549316, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 971 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3058.0, + "completions/mean_length": 870.4598388671875, + "completions/mean_terminated_length": 613.9710693359375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 9.08396501457726, + "grad_norm": 0.1466354876756668, + "learning_rate": 1e-06, + "loss": -0.0389, + "num_tokens": 588153195.0, + "reward": 0.6964285969734192, + "reward_std": 0.14481490850448608, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3716.0, + "completions/mean_length": 959.4453735351562, + "completions/mean_terminated_length": 626.4284057617188, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.093294460641399, + "grad_norm": 0.1576017290353775, + "learning_rate": 1e-06, + "loss": -0.0469, + "num_tokens": 588748626.0, + "reward": 0.6897321939468384, + "reward_std": 0.1539480984210968, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 973 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3221.0, + "completions/mean_length": 1120.8828125, + "completions/mean_terminated_length": 669.6439819335938, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 9.102623906705539, + "grad_norm": 0.1499904990196228, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 589349721.0, + "reward": 0.6160714626312256, + "reward_std": 0.1361714005470276, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 974 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3254.0, + "completions/mean_length": 1097.8035888671875, + "completions/mean_terminated_length": 643.0642700195312, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 9.11195335276968, + "grad_norm": 0.17377203702926636, + "learning_rate": 1e-06, + "loss": -0.0699, + "num_tokens": 589955537.0, + "reward": 0.6049107313156128, + "reward_std": 0.1645801067352295, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4077.0, + "completions/mean_length": 1063.610595703125, + "completions/mean_terminated_length": 621.548583984375, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 9.12128279883382, + "grad_norm": 0.1703103482723236, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 590533764.0, + "reward": 0.6071428656578064, + "reward_std": 0.13842841982841492, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 976 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2921.0, + "completions/mean_length": 1112.9654541015625, + "completions/mean_terminated_length": 678.0984497070312, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.130612244897959, + "grad_norm": 0.1722087413072586, + "learning_rate": 1e-06, + "loss": -0.0685, + "num_tokens": 591156989.0, + "reward": 0.609375, + "reward_std": 0.19054891169071198, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 977 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2800.0, + "completions/mean_length": 1054.8538818359375, + "completions/mean_terminated_length": 624.8318481445312, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 9.139941690962099, + "grad_norm": 0.15863892436027527, + "learning_rate": 1e-06, + "loss": -0.065, + "num_tokens": 591736250.0, + "reward": 0.652901828289032, + "reward_std": 0.1472145915031433, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2974.0, + "completions/mean_length": 1022.419677734375, + "completions/mean_terminated_length": 640.6323852539062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 9.14927113702624, + "grad_norm": 0.14767885208129883, + "learning_rate": 1e-06, + "loss": -0.0626, + "num_tokens": 592341626.0, + "reward": 0.6752232313156128, + "reward_std": 0.14011839032173157, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2632.0, + "completions/mean_length": 1121.891845703125, + "completions/mean_terminated_length": 612.6000366210938, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 9.15860058309038, + "grad_norm": 0.14965596795082092, + "learning_rate": 1e-06, + "loss": -0.0794, + "num_tokens": 592901465.0, + "reward": 0.6551339626312256, + "reward_std": 0.17138142883777618, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 980 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2515.0, + "completions/mean_length": 914.122802734375, + "completions/mean_terminated_length": 606.450439453125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 9.167930029154519, + "grad_norm": 0.1620972603559494, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 593494359.0, + "reward": 0.6584821939468384, + "reward_std": 0.1645801067352295, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 981 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3590.0, + "completions/mean_length": 948.9631958007812, + "completions/mean_terminated_length": 601.8922119140625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.177259475218658, + "grad_norm": 0.14093415439128876, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 594059542.0, + "reward": 0.6696428656578064, + "reward_std": 0.1346304714679718, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 982 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2735.0, + "completions/mean_length": 865.7142944335938, + "completions/mean_terminated_length": 604.6417236328125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.186588921282798, + "grad_norm": 0.12272296100854874, + "learning_rate": 1e-06, + "loss": -0.0773, + "num_tokens": 594646142.0, + "reward": 0.7399553656578064, + "reward_std": 0.14286141097545624, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 983 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3964.0, + "completions/mean_length": 1021.22216796875, + "completions/mean_terminated_length": 639.2860717773438, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 9.19591836734694, + "grad_norm": 0.15324079990386963, + "learning_rate": 1e-06, + "loss": -0.0327, + "num_tokens": 595245405.0, + "reward": 0.6707589626312256, + "reward_std": 0.14676813781261444, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 984 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2819.0, + "completions/mean_length": 858.1060791015625, + "completions/mean_terminated_length": 587.9552612304688, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 9.205247813411079, + "grad_norm": 0.15378931164741516, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 595803316.0, + "reward": 0.738839328289032, + "reward_std": 0.12065806984901428, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2483.0, + "completions/mean_length": 800.0379638671875, + "completions/mean_terminated_length": 576.1168212890625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 9.214577259475218, + "grad_norm": 0.14740386605262756, + "learning_rate": 1e-06, + "loss": -0.048, + "num_tokens": 596372262.0, + "reward": 0.7287946939468384, + "reward_std": 0.1233583390712738, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 986 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 1108.1875, + "completions/mean_terminated_length": 676.9961547851562, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 9.223906705539358, + "grad_norm": 0.1522056609392166, + "learning_rate": 1e-06, + "loss": -0.068, + "num_tokens": 597001806.0, + "reward": 0.6194196939468384, + "reward_std": 0.13087642192840576, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 987 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3695.0, + "completions/mean_length": 1029.610595703125, + "completions/mean_terminated_length": 704.043212890625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 9.2332361516035, + "grad_norm": 0.15282049775123596, + "learning_rate": 1e-06, + "loss": -0.0249, + "num_tokens": 597671961.0, + "reward": 0.629464328289032, + "reward_std": 0.17878004908561707, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 988 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2499.0, + "completions/mean_length": 935.6629638671875, + "completions/mean_terminated_length": 600.1210327148438, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 9.242565597667639, + "grad_norm": 0.1368328332901001, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 598241691.0, + "reward": 0.7120535969734192, + "reward_std": 0.11265284568071365, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 989 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3921.0, + "completions/mean_length": 971.911865234375, + "completions/mean_terminated_length": 635.9468383789062, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 9.251895043731778, + "grad_norm": 0.17850814759731293, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 598853556.0, + "reward": 0.6629464626312256, + "reward_std": 0.20267587900161743, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 990 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2447.0, + "completions/mean_length": 921.1864013671875, + "completions/mean_terminated_length": 618.4535522460938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 9.261224489795918, + "grad_norm": 0.14040787518024445, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 599461579.0, + "reward": 0.6383928656578064, + "reward_std": 0.1288476437330246, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 991 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2509.0, + "completions/mean_length": 999.724365234375, + "completions/mean_terminated_length": 636.8192138671875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 9.270553935860057, + "grad_norm": 0.14651288092136383, + "learning_rate": 1e-06, + "loss": -0.0372, + "num_tokens": 600058644.0, + "reward": 0.6350446939468384, + "reward_std": 0.1253136545419693, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 992 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2805.0, + "completions/mean_length": 952.8638916015625, + "completions/mean_terminated_length": 640.478515625, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 9.279883381924199, + "grad_norm": 0.13161633908748627, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 600679842.0, + "reward": 0.691964328289032, + "reward_std": 0.12471451610326767, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 993 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2948.0, + "completions/mean_length": 942.80810546875, + "completions/mean_terminated_length": 642.136962890625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.289212827988338, + "grad_norm": 0.14655812084674835, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 601288286.0, + "reward": 0.6964285969734192, + "reward_std": 0.1597309112548828, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2822.0, + "completions/mean_length": 990.450927734375, + "completions/mean_terminated_length": 635.0895385742188, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 9.298542274052478, + "grad_norm": 0.15559905767440796, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 601891442.0, + "reward": 0.6674107313156128, + "reward_std": 0.1550350934267044, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 995 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3974.0, + "completions/mean_length": 999.51123046875, + "completions/mean_terminated_length": 606.1207275390625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 9.307871720116617, + "grad_norm": 0.16221782565116882, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 602465748.0, + "reward": 0.6830357313156128, + "reward_std": 0.11317459493875504, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 996 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3369.0, + "completions/mean_length": 915.6272583007812, + "completions/mean_terminated_length": 612.3643188476562, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.317201166180759, + "grad_norm": 0.17097395658493042, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 603053990.0, + "reward": 0.668526828289032, + "reward_std": 0.17521145939826965, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 997 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2651.0, + "completions/mean_length": 885.4732666015625, + "completions/mean_terminated_length": 613.3947143554688, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 9.326530612244898, + "grad_norm": 0.1613817811012268, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 603638366.0, + "reward": 0.6897321939468384, + "reward_std": 0.16308379173278809, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 998 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2974.0, + "completions/mean_length": 851.1730346679688, + "completions/mean_terminated_length": 580.4437866210938, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 9.335860058309038, + "grad_norm": 0.16307564079761505, + "learning_rate": 1e-06, + "loss": -0.0355, + "num_tokens": 604209769.0, + "reward": 0.6863839626312256, + "reward_std": 0.16149896383285522, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 999 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3464.0, + "completions/mean_length": 1016.1217041015625, + "completions/mean_terminated_length": 616.0870361328125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 9.345189504373177, + "grad_norm": 0.13637185096740723, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 604793046.0, + "reward": 0.6551339626312256, + "reward_std": 0.12467173486948013, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 1000 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3012.0, + "completions/mean_length": 1028.419677734375, + "completions/mean_terminated_length": 607.9898071289062, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 9.354518950437317, + "grad_norm": 0.15392091870307922, + "learning_rate": 1e-06, + "loss": -0.0823, + "num_tokens": 605365990.0, + "reward": 0.6808035969734192, + "reward_std": 0.15672573447227478, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2266.0, + "completions/mean_length": 922.6194458007812, + "completions/mean_terminated_length": 572.6431274414062, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.363848396501458, + "grad_norm": 0.1472814381122589, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 605910817.0, + "reward": 0.6752232313156128, + "reward_std": 0.09153397381305695, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1002 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3679.0, + "completions/mean_length": 1002.6105346679688, + "completions/mean_terminated_length": 618.3626098632812, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.373177842565598, + "grad_norm": 0.16151034832000732, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 606491340.0, + "reward": 0.6462053656578064, + "reward_std": 0.1381234973669052, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1003 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3622.0, + "completions/mean_length": 926.2120971679688, + "completions/mean_terminated_length": 623.9584350585938, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 9.382507288629737, + "grad_norm": 0.1649787873029709, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 607093194.0, + "reward": 0.6462053656578064, + "reward_std": 0.17649024724960327, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1004 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 1013.5089721679688, + "completions/mean_terminated_length": 634.9573974609375, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 9.391836734693877, + "grad_norm": 0.15088605880737305, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 607684922.0, + "reward": 0.6696428656578064, + "reward_std": 0.12369534373283386, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1005 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3121.0, + "completions/mean_length": 987.2511596679688, + "completions/mean_terminated_length": 678.283447265625, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 9.401166180758018, + "grad_norm": 0.15360213816165924, + "learning_rate": 1e-06, + "loss": -0.0336, + "num_tokens": 608326331.0, + "reward": 0.6629464626312256, + "reward_std": 0.15297238528728485, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1006 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3543.0, + "completions/mean_length": 986.15185546875, + "completions/mean_terminated_length": 591.0641479492188, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.410495626822158, + "grad_norm": 0.14688071608543396, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 608884675.0, + "reward": 0.6930803656578064, + "reward_std": 0.13636502623558044, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1007 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3300.0, + "completions/mean_length": 985.6105346679688, + "completions/mean_terminated_length": 638.2965087890625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.419825072886297, + "grad_norm": 0.14192089438438416, + "learning_rate": 1e-06, + "loss": -0.04, + "num_tokens": 609487726.0, + "reward": 0.6127232313156128, + "reward_std": 0.13342514634132385, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 1008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2320.0, + "completions/mean_length": 926.54248046875, + "completions/mean_terminated_length": 657.9443359375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 9.429154518950437, + "grad_norm": 0.14036203920841217, + "learning_rate": 1e-06, + "loss": -0.0379, + "num_tokens": 610119020.0, + "reward": 0.6808035969734192, + "reward_std": 0.13883888721466064, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1009 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3672.0, + "completions/mean_length": 1002.7076416015625, + "completions/mean_terminated_length": 596.5177001953125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.438483965014576, + "grad_norm": 0.1588844507932663, + "learning_rate": 1e-06, + "loss": -0.0389, + "num_tokens": 610688742.0, + "reward": 0.6484375, + "reward_std": 0.1463116705417633, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1010 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3221.0, + "completions/mean_length": 940.87841796875, + "completions/mean_terminated_length": 618.7687377929688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 9.447813411078718, + "grad_norm": 0.14721238613128662, + "learning_rate": 1e-06, + "loss": -0.0559, + "num_tokens": 611282289.0, + "reward": 0.7042410969734192, + "reward_std": 0.148458793759346, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2689.0, + "completions/mean_length": 916.1741333007812, + "completions/mean_terminated_length": 565.4869995117188, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 9.457142857142857, + "grad_norm": 0.17061950266361237, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 611828373.0, + "reward": 0.6986607313156128, + "reward_std": 0.13835172355175018, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1012 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2452.0, + "completions/mean_length": 905.169677734375, + "completions/mean_terminated_length": 617.9172973632812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.466472303206997, + "grad_norm": 0.1752035915851593, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 612443141.0, + "reward": 0.6328125, + "reward_std": 0.1632649153470993, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1013 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3795.0, + "completions/mean_length": 1090.075927734375, + "completions/mean_terminated_length": 712.4472045898438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.475801749271136, + "grad_norm": 0.15727031230926514, + "learning_rate": 1e-06, + "loss": -0.0385, + "num_tokens": 613093729.0, + "reward": 0.6104910969734192, + "reward_std": 0.1594706028699875, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 1014 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3332.0, + "completions/mean_length": 1075.4576416015625, + "completions/mean_terminated_length": 700.2584838867188, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 9.485131195335278, + "grad_norm": 0.15654075145721436, + "learning_rate": 1e-06, + "loss": -0.0335, + "num_tokens": 613741923.0, + "reward": 0.5647321939468384, + "reward_std": 0.1702541708946228, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 1015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4069.0, + "completions/mean_length": 894.3839721679688, + "completions/mean_terminated_length": 593.3773193359375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 9.494460641399417, + "grad_norm": 0.1495983600616455, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 614307971.0, + "reward": 0.7020089626312256, + "reward_std": 0.1140422374010086, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1016 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3553.0, + "completions/mean_length": 1007.6797485351562, + "completions/mean_terminated_length": 615.3270263671875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.503790087463557, + "grad_norm": 0.15612642467021942, + "learning_rate": 1e-06, + "loss": -0.045, + "num_tokens": 614879468.0, + "reward": 0.699776828289032, + "reward_std": 0.1381234973669052, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1017 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3540.0, + "completions/mean_length": 958.232177734375, + "completions/mean_terminated_length": 616.4950561523438, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.513119533527696, + "grad_norm": 0.17223958671092987, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 615470988.0, + "reward": 0.6852678656578064, + "reward_std": 0.17217238247394562, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1018 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3682.0, + "completions/mean_length": 917.2277221679688, + "completions/mean_terminated_length": 626.8404541015625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 9.522448979591836, + "grad_norm": 0.1533408910036087, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 616066824.0, + "reward": 0.7176339626312256, + "reward_std": 0.17461301386356354, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1019 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3232.0, + "completions/mean_length": 973.87060546875, + "completions/mean_terminated_length": 638.1162109375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 9.531778425655977, + "grad_norm": 0.15724606812000275, + "learning_rate": 1e-06, + "loss": -0.0245, + "num_tokens": 616667908.0, + "reward": 0.6674107313156128, + "reward_std": 0.14421899616718292, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1020 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3696.0, + "completions/mean_length": 1000.5870971679688, + "completions/mean_terminated_length": 620.4486083984375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 9.541107871720117, + "grad_norm": 0.12718215584754944, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 617250818.0, + "reward": 0.676339328289032, + "reward_std": 0.1176849827170372, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1021 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3039.0, + "completions/mean_length": 1051.359375, + "completions/mean_terminated_length": 634.0736083984375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.550437317784256, + "grad_norm": 0.14777742326259613, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 617838620.0, + "reward": 0.6964285969734192, + "reward_std": 0.1374502032995224, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1022 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3545.0, + "completions/mean_length": 1105.6473388671875, + "completions/mean_terminated_length": 746.8049926757812, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 9.559766763848396, + "grad_norm": 0.14828738570213318, + "learning_rate": 1e-06, + "loss": -0.0753, + "num_tokens": 618537208.0, + "reward": 0.6116071939468384, + "reward_std": 0.16101223230361938, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 1023 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2577.0, + "completions/mean_length": 875.0949096679688, + "completions/mean_terminated_length": 635.651123046875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.569096209912537, + "grad_norm": 0.158096581697464, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 619166709.0, + "reward": 0.7053571939468384, + "reward_std": 0.1504133939743042, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 1024 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3686.0, + "completions/mean_length": 1024.938720703125, + "completions/mean_terminated_length": 647.7907104492188, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 9.578425655976677, + "grad_norm": 0.13805274665355682, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 619766750.0, + "reward": 0.6774553656578064, + "reward_std": 0.12651577591896057, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1025 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3483.0, + "completions/mean_length": 948.3906860351562, + "completions/mean_terminated_length": 575.07861328125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 9.587755102040816, + "grad_norm": 0.13969606161117554, + "learning_rate": 1e-06, + "loss": -0.0415, + "num_tokens": 620324668.0, + "reward": 0.684151828289032, + "reward_std": 0.10505989193916321, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1026 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3338.0, + "completions/mean_length": 933.9453735351562, + "completions/mean_terminated_length": 611.1279296875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 9.597084548104956, + "grad_norm": 0.11436095833778381, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 620907499.0, + "reward": 0.723214328289032, + "reward_std": 0.10205401480197906, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.4476594924926758, + "step": 1027 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 1061.7098388671875, + "completions/mean_terminated_length": 693.3416748046875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.606413994169095, + "grad_norm": 0.15719173848628998, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 621555911.0, + "reward": 0.613839328289032, + "reward_std": 0.1708935797214508, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1028 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2652.0, + "completions/mean_length": 872.9910888671875, + "completions/mean_terminated_length": 569.97314453125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 9.615743440233237, + "grad_norm": 0.16108962893486023, + "learning_rate": 1e-06, + "loss": -0.0192, + "num_tokens": 622117487.0, + "reward": 0.668526828289032, + "reward_std": 0.1215910017490387, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1029 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2228.0, + "completions/mean_length": 863.154052734375, + "completions/mean_terminated_length": 576.4010009765625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 9.625072886297376, + "grad_norm": 0.130886048078537, + "learning_rate": 1e-06, + "loss": -0.0751, + "num_tokens": 622677217.0, + "reward": 0.7444196939468384, + "reward_std": 0.12677650153636932, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1030 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3420.0, + "completions/mean_length": 986.5402221679688, + "completions/mean_terminated_length": 652.1483154296875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.634402332361516, + "grad_norm": 0.15525905787944794, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 623294469.0, + "reward": 0.6540178656578064, + "reward_std": 0.1435782015323639, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1031 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3081.0, + "completions/mean_length": 1017.1272583007812, + "completions/mean_terminated_length": 647.6624755859375, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 9.643731778425655, + "grad_norm": 0.15184278786182404, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 623906839.0, + "reward": 0.6819196939468384, + "reward_std": 0.16431981325149536, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1032 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3006.0, + "completions/mean_length": 996.3035888671875, + "completions/mean_terminated_length": 602.5056762695312, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 9.653061224489797, + "grad_norm": 0.1531890630722046, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 624469775.0, + "reward": 0.7031250596046448, + "reward_std": 0.1356828510761261, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3280.0, + "completions/mean_length": 1031.4554443359375, + "completions/mean_terminated_length": 650.7904663085938, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 9.662390670553936, + "grad_norm": 0.15431952476501465, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 625078079.0, + "reward": 0.6595982313156128, + "reward_std": 0.12189549207687378, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1034 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3946.0, + "completions/mean_length": 903.677490234375, + "completions/mean_terminated_length": 607.8036499023438, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 9.671720116618076, + "grad_norm": 0.15675632655620575, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 625661590.0, + "reward": 0.6696428656578064, + "reward_std": 0.16067594289779663, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1035 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3691.0, + "completions/mean_length": 969.3370971679688, + "completions/mean_terminated_length": 611.5596923828125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 9.681049562682215, + "grad_norm": 0.17152757942676544, + "learning_rate": 1e-06, + "loss": -0.0687, + "num_tokens": 626242276.0, + "reward": 0.6640625, + "reward_std": 0.1575499027967453, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1036 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3885.0, + "completions/mean_length": 994.7188110351562, + "completions/mean_terminated_length": 648.4218139648438, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 9.690379008746355, + "grad_norm": 0.15449100732803345, + "learning_rate": 1e-06, + "loss": -0.0466, + "num_tokens": 626844176.0, + "reward": 0.7053571939468384, + "reward_std": 0.14602065086364746, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1037 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3598.0, + "completions/mean_length": 1010.8482666015625, + "completions/mean_terminated_length": 670.6022338867188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 9.699708454810496, + "grad_norm": 0.17662253975868225, + "learning_rate": 1e-06, + "loss": -0.0265, + "num_tokens": 627469016.0, + "reward": 0.6305803656578064, + "reward_std": 0.18498291075229645, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4031.0, + "completions/mean_length": 844.6295166015625, + "completions/mean_terminated_length": 623.73779296875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 9.709037900874636, + "grad_norm": 0.16566677391529083, + "learning_rate": 1e-06, + "loss": -0.019, + "num_tokens": 628081644.0, + "reward": 0.691964328289032, + "reward_std": 0.16209669411182404, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1039 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2892.0, + "completions/mean_length": 1054.6920166015625, + "completions/mean_terminated_length": 633.4688720703125, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 9.718367346938775, + "grad_norm": 0.1413305401802063, + "learning_rate": 1e-06, + "loss": -0.0235, + "num_tokens": 628665344.0, + "reward": 0.6462053656578064, + "reward_std": 0.12151571363210678, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1040 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3107.0, + "completions/mean_length": 863.1953735351562, + "completions/mean_terminated_length": 576.4459228515625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 9.727696793002915, + "grad_norm": 0.16033810377120972, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 629230319.0, + "reward": 0.7087053656578064, + "reward_std": 0.15117114782333374, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3069.0, + "completions/mean_length": 858.1920166015625, + "completions/mean_terminated_length": 600.7277221679688, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 9.737026239067056, + "grad_norm": 0.13776952028274536, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 629811307.0, + "reward": 0.7254464626312256, + "reward_std": 0.13267767429351807, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1042 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3617.0, + "completions/mean_length": 1069.3695068359375, + "completions/mean_terminated_length": 619.255126953125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 9.746355685131196, + "grad_norm": 0.14992396533489227, + "learning_rate": 1e-06, + "loss": -0.0538, + "num_tokens": 630381254.0, + "reward": 0.6540178656578064, + "reward_std": 0.12899848818778992, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1043 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2844.0, + "completions/mean_length": 968.8850708007812, + "completions/mean_terminated_length": 636.8704223632812, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 9.755685131195335, + "grad_norm": 0.1591418832540512, + "learning_rate": 1e-06, + "loss": -0.0654, + "num_tokens": 630982847.0, + "reward": 0.6886160969734192, + "reward_std": 0.15018518269062042, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1044 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3728.0, + "completions/mean_length": 954.8292846679688, + "completions/mean_terminated_length": 599.7403564453125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 9.765014577259475, + "grad_norm": 0.1500181406736374, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 631540606.0, + "reward": 0.6930803656578064, + "reward_std": 0.13711318373680115, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1045 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 1027.9051513671875, + "completions/mean_terminated_length": 731.2350463867188, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 9.774344023323614, + "grad_norm": 0.15682144463062286, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 632232321.0, + "reward": 0.6674107313156128, + "reward_std": 0.18344520032405853, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1046 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3091.0, + "completions/mean_length": 1128.544677734375, + "completions/mean_terminated_length": 678.4678344726562, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 9.783673469387756, + "grad_norm": 0.15192003548145294, + "learning_rate": 1e-06, + "loss": -0.0607, + "num_tokens": 632857577.0, + "reward": 0.6194196939468384, + "reward_std": 0.13519500195980072, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 1047 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3623.0, + "completions/mean_length": 1026.7421875, + "completions/mean_terminated_length": 675.5335693359375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 9.793002915451895, + "grad_norm": 0.14906039834022522, + "learning_rate": 1e-06, + "loss": -0.0431, + "num_tokens": 633494130.0, + "reward": 0.6696428656578064, + "reward_std": 0.13598594069480896, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1048 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3550.0, + "completions/mean_length": 965.2589721679688, + "completions/mean_terminated_length": 641.38916015625, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 9.802332361516035, + "grad_norm": 0.1628960818052292, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 634102778.0, + "reward": 0.6194196939468384, + "reward_std": 0.1832476705312729, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 1049 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2185.0, + "completions/mean_length": 947.47216796875, + "completions/mean_terminated_length": 621.7623291015625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 9.811661807580174, + "grad_norm": 0.16928373277187347, + "learning_rate": 1e-06, + "loss": -0.0301, + "num_tokens": 634695489.0, + "reward": 0.7500000596046448, + "reward_std": 0.14579172432422638, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1050 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3358.0, + "completions/mean_length": 992.6094360351562, + "completions/mean_terminated_length": 624.5418090820312, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 9.820991253644316, + "grad_norm": 0.182173490524292, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 635276315.0, + "reward": 0.684151828289032, + "reward_std": 0.1749507337808609, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1051 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3876.0, + "completions/mean_length": 1172.7601318359375, + "completions/mean_terminated_length": 698.8236083984375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 9.830320699708455, + "grad_norm": 0.13864195346832275, + "learning_rate": 1e-06, + "loss": -0.0693, + "num_tokens": 635909356.0, + "reward": 0.6328125, + "reward_std": 0.12505152821540833, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1052 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3337.0, + "completions/mean_length": 1001.6551513671875, + "completions/mean_terminated_length": 617.28857421875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 9.839650145772595, + "grad_norm": 0.14914987981319427, + "learning_rate": 1e-06, + "loss": -0.0586, + "num_tokens": 636484175.0, + "reward": 0.6796875596046448, + "reward_std": 0.12238264828920364, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1053 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3370.0, + "completions/mean_length": 1054.5648193359375, + "completions/mean_terminated_length": 642.1013793945312, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 9.848979591836734, + "grad_norm": 0.15453556180000305, + "learning_rate": 1e-06, + "loss": -0.0266, + "num_tokens": 637080849.0, + "reward": 0.621651828289032, + "reward_std": 0.1303137093782425, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 1054 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2204.0, + "completions/mean_length": 1002.6964721679688, + "completions/mean_terminated_length": 605.3198852539062, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 9.858309037900874, + "grad_norm": 0.15499663352966309, + "learning_rate": 1e-06, + "loss": -0.064, + "num_tokens": 637663625.0, + "reward": 0.6707589626312256, + "reward_std": 0.12779568135738373, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3460.0, + "completions/mean_length": 1049.083740234375, + "completions/mean_terminated_length": 683.4537353515625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 9.867638483965015, + "grad_norm": 0.13593731820583344, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 638304596.0, + "reward": 0.6473214626312256, + "reward_std": 0.1352698653936386, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 1056 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2919.0, + "completions/mean_length": 1067.0692138671875, + "completions/mean_terminated_length": 634.3648071289062, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 9.876967930029155, + "grad_norm": 0.150171160697937, + "learning_rate": 1e-06, + "loss": -0.0407, + "num_tokens": 638889042.0, + "reward": 0.6584821939468384, + "reward_std": 0.13069167733192444, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1057 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3939.0, + "completions/mean_length": 1020.2957763671875, + "completions/mean_terminated_length": 642.5776977539062, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 9.886297376093294, + "grad_norm": 0.15999186038970947, + "learning_rate": 1e-06, + "loss": -0.0579, + "num_tokens": 639495667.0, + "reward": 0.6618303656578064, + "reward_std": 0.16619592905044556, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1058 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3659.0, + "completions/mean_length": 1047.9609375, + "completions/mean_terminated_length": 652.061767578125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 9.895626822157434, + "grad_norm": 0.1646793782711029, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 640099264.0, + "reward": 0.6316964626312256, + "reward_std": 0.17156758904457092, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1059 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3322.0, + "completions/mean_length": 962.232177734375, + "completions/mean_terminated_length": 586.1799926757812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.904956268221575, + "grad_norm": 0.14414063096046448, + "learning_rate": 1e-06, + "loss": -0.0132, + "num_tokens": 640654392.0, + "reward": 0.6808035969734192, + "reward_std": 0.092888742685318, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1060 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3160.0, + "completions/mean_length": 774.5313110351562, + "completions/mean_terminated_length": 557.312744140625, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.914285714285715, + "grad_norm": 0.15139399468898773, + "learning_rate": 1e-06, + "loss": -0.0244, + "num_tokens": 641203716.0, + "reward": 0.7421875596046448, + "reward_std": 0.11265214532613754, + "rewards/verify_math_reward/mean": 0.7421875, + "rewards/verify_math_reward/std": 0.43767455220222473, + "step": 1061 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3058.0, + "completions/mean_length": 905.9285888671875, + "completions/mean_terminated_length": 622.9696655273438, + "completions/min_length": 207.0, + "completions/min_terminated_length": 207.0, + "epoch": 9.923615160349854, + "grad_norm": 0.13992974162101746, + "learning_rate": 1e-06, + "loss": -0.0313, + "num_tokens": 641805276.0, + "reward": 0.7042410969734192, + "reward_std": 0.12148109823465347, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1062 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 932.1328735351562, + "completions/mean_terminated_length": 574.478271484375, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.932944606413994, + "grad_norm": 0.16572162508964539, + "learning_rate": 1e-06, + "loss": -0.0338, + "num_tokens": 642363331.0, + "reward": 0.6785714626312256, + "reward_std": 0.12482328712940216, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1063 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3439.0, + "completions/mean_length": 1039.891845703125, + "completions/mean_terminated_length": 711.2373046875, + "completions/min_length": 217.0, + "completions/min_terminated_length": 217.0, + "epoch": 9.942274052478133, + "grad_norm": 0.161685973405838, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 643018562.0, + "reward": 0.6897321939468384, + "reward_std": 0.1583433598279953, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1064 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3772.0, + "completions/mean_length": 1000.2567138671875, + "completions/mean_terminated_length": 646.0173950195312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 9.951603498542275, + "grad_norm": 0.15626482665538788, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 643629864.0, + "reward": 0.6875000596046448, + "reward_std": 0.13511762022972107, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1065 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4000.0, + "completions/mean_length": 910.5725708007812, + "completions/mean_terminated_length": 615.3378295898438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.960932944606414, + "grad_norm": 0.14039351046085358, + "learning_rate": 1e-06, + "loss": -0.0359, + "num_tokens": 644225617.0, + "reward": 0.6852678656578064, + "reward_std": 0.1278284788131714, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.4646684527397156, + "step": 1066 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2472.0, + "completions/mean_length": 1004.35498046875, + "completions/mean_terminated_length": 641.9925537109375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.970262390670554, + "grad_norm": 0.1683109551668167, + "learning_rate": 1e-06, + "loss": -0.0494, + "num_tokens": 644833279.0, + "reward": 0.6595982313156128, + "reward_std": 0.1755887120962143, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1067 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3040.0, + "completions/mean_length": 851.1373291015625, + "completions/mean_terminated_length": 609.9124755859375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 9.979591836734693, + "grad_norm": 0.1568983495235443, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 645420890.0, + "reward": 0.7555803656578064, + "reward_std": 0.13038787245750427, + "rewards/verify_math_reward/mean": 0.7555803656578064, + "rewards/verify_math_reward/std": 0.42998260259628296, + "step": 1068 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2420.0, + "completions/mean_length": 1032.6273193359375, + "completions/mean_terminated_length": 643.4440307617188, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.988921282798835, + "grad_norm": 0.16326849162578583, + "learning_rate": 1e-06, + "loss": -0.0608, + "num_tokens": 646036516.0, + "reward": 0.6383928656578064, + "reward_std": 0.15537913143634796, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 1069 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.17045454545454541, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2543.0, + "completions/mean_length": 1248.866455078125, + "completions/mean_terminated_length": 663.8390502929688, + "completions/min_length": 221.0, + "completions/min_terminated_length": 221.0, + "epoch": 9.998250728862974, + "grad_norm": 0.15602833032608032, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 646637390.0, + "reward": 0.7020089626312256, + "reward_std": 0.14574895799160004, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1070 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3991.0, + "completions/mean_length": 1052.3114013671875, + "completions/mean_terminated_length": 643.917724609375, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 10.00932944606414, + "grad_norm": 0.18460336327552795, + "learning_rate": 1e-06, + "loss": -0.0736, + "num_tokens": 647238589.0, + "reward": 0.625, + "reward_std": 0.17086149752140045, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 1071 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3022.0, + "completions/mean_length": 974.3214721679688, + "completions/mean_terminated_length": 655.6260375976562, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 10.018658892128279, + "grad_norm": 0.1680677831172943, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 647865021.0, + "reward": 0.629464328289032, + "reward_std": 0.15605314075946808, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1072 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3301.0, + "completions/mean_length": 1047.950927734375, + "completions/mean_terminated_length": 638.97216796875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 10.02798833819242, + "grad_norm": 0.1526758223772049, + "learning_rate": 1e-06, + "loss": -0.0788, + "num_tokens": 648460737.0, + "reward": 0.6674107313156128, + "reward_std": 0.1469210684299469, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1073 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3904.0, + "completions/mean_length": 1009.3516235351562, + "completions/mean_terminated_length": 630.2894897460938, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 10.03731778425656, + "grad_norm": 0.1646697074174881, + "learning_rate": 1e-06, + "loss": -0.0606, + "num_tokens": 649049660.0, + "reward": 0.684151828289032, + "reward_std": 0.1620207130908966, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1074 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4093.0, + "completions/mean_length": 1050.563720703125, + "completions/mean_terminated_length": 689.3695068359375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 10.0466472303207, + "grad_norm": 0.15181021392345428, + "learning_rate": 1e-06, + "loss": -0.0704, + "num_tokens": 649697277.0, + "reward": 0.6573660969734192, + "reward_std": 0.14507634937763214, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1075 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 896.8772583007812, + "completions/mean_terminated_length": 621.5587768554688, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 10.055976676384839, + "grad_norm": 0.15000467002391815, + "learning_rate": 1e-06, + "loss": -0.0569, + "num_tokens": 650286639.0, + "reward": 0.7533482313156128, + "reward_std": 0.13714781403541565, + "rewards/verify_math_reward/mean": 0.7533482313156128, + "rewards/verify_math_reward/std": 0.4313030242919922, + "step": 1076 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3429.0, + "completions/mean_length": 1108.6317138671875, + "completions/mean_terminated_length": 642.216796875, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 10.06530612244898, + "grad_norm": 0.15487568080425262, + "learning_rate": 1e-06, + "loss": -0.0827, + "num_tokens": 650876701.0, + "reward": 0.6886160969734192, + "reward_std": 0.13929352164268494, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1077 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3924.0, + "completions/mean_length": 1101.8092041015625, + "completions/mean_terminated_length": 687.11181640625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 10.07463556851312, + "grad_norm": 0.1741434931755066, + "learning_rate": 1e-06, + "loss": -0.0699, + "num_tokens": 651504938.0, + "reward": 0.6618303656578064, + "reward_std": 0.1892682909965515, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1078 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3902.0, + "completions/mean_length": 1098.8013916015625, + "completions/mean_terminated_length": 653.0641479492188, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 10.08396501457726, + "grad_norm": 0.16074298322200775, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 652117112.0, + "reward": 0.6462053656578064, + "reward_std": 0.1419203132390976, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1079 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 1060.25, + "completions/mean_terminated_length": 674.5761108398438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.093294460641399, + "grad_norm": 0.14189216494560242, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 652755272.0, + "reward": 0.684151828289032, + "reward_std": 0.1442515105009079, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1080 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 1005.5178833007812, + "completions/mean_terminated_length": 630.3279418945312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 10.102623906705539, + "grad_norm": 0.1739742010831833, + "learning_rate": 1e-06, + "loss": -0.0383, + "num_tokens": 653351568.0, + "reward": 0.6629464626312256, + "reward_std": 0.16825789213180542, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2823.0, + "completions/mean_length": 970.3683471679688, + "completions/mean_terminated_length": 621.3523559570312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 10.11195335276968, + "grad_norm": 0.17674899101257324, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 653947650.0, + "reward": 0.6495535969734192, + "reward_std": 0.1453045904636383, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1082 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4061.0, + "completions/mean_length": 953.341552734375, + "completions/mean_terminated_length": 649.46142578125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 10.12128279883382, + "grad_norm": 0.1479760855436325, + "learning_rate": 1e-06, + "loss": -0.0796, + "num_tokens": 654578276.0, + "reward": 0.6662946939468384, + "reward_std": 0.1575152724981308, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1083 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2731.0, + "completions/mean_length": 997.9944458007812, + "completions/mean_terminated_length": 677.5110473632812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.130612244897959, + "grad_norm": 0.16130521893501282, + "learning_rate": 1e-06, + "loss": -0.0391, + "num_tokens": 655222855.0, + "reward": 0.6194196939468384, + "reward_std": 0.13940368592739105, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 1084 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3313.0, + "completions/mean_length": 1030.5067138671875, + "completions/mean_terminated_length": 671.20947265625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 10.139941690962099, + "grad_norm": 0.1279752552509308, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 655844813.0, + "reward": 0.6785714626312256, + "reward_std": 0.11858651787042618, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1085 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3103.0, + "completions/mean_length": 853.3850708007812, + "completions/mean_terminated_length": 561.4708251953125, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 10.14927113702624, + "grad_norm": 0.15580108761787415, + "learning_rate": 1e-06, + "loss": -0.0528, + "num_tokens": 656389894.0, + "reward": 0.7868303656578064, + "reward_std": 0.1168951541185379, + "rewards/verify_math_reward/mean": 0.7868303656578064, + "rewards/verify_math_reward/std": 0.4097752273082733, + "step": 1086 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4050.0, + "completions/mean_length": 998.2511596679688, + "completions/mean_terminated_length": 626.521240234375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 10.15860058309038, + "grad_norm": 0.1457132250070572, + "learning_rate": 1e-06, + "loss": -0.0328, + "num_tokens": 656986199.0, + "reward": 0.6886160969734192, + "reward_std": 0.11753595620393753, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3383.0, + "completions/mean_length": 1092.73779296875, + "completions/mean_terminated_length": 681.1231079101562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 10.167930029154519, + "grad_norm": 0.1617908775806427, + "learning_rate": 1e-06, + "loss": -0.0176, + "num_tokens": 657607948.0, + "reward": 0.684151828289032, + "reward_std": 0.1563466638326645, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 1112.1976318359375, + "completions/mean_terminated_length": 711.8392333984375, + "completions/min_length": 202.0, + "completions/min_terminated_length": 202.0, + "epoch": 10.177259475218658, + "grad_norm": 0.13057444989681244, + "learning_rate": 1e-06, + "loss": -0.0431, + "num_tokens": 658261861.0, + "reward": 0.6428571939468384, + "reward_std": 0.12099438905715942, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1089 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 883.5803833007812, + "completions/mean_terminated_length": 602.8834838867188, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 10.186588921282798, + "grad_norm": 0.1505209058523178, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 658844061.0, + "reward": 0.7455357313156128, + "reward_std": 0.1307237595319748, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1090 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3926.0, + "completions/mean_length": 880.7667846679688, + "completions/mean_terminated_length": 582.76953125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 10.19591836734694, + "grad_norm": 0.19235415756702423, + "learning_rate": 1e-06, + "loss": -0.0418, + "num_tokens": 659406084.0, + "reward": 0.7198660969734192, + "reward_std": 0.17243380844593048, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1091 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3177.0, + "completions/mean_length": 1098.2388916015625, + "completions/mean_terminated_length": 678.704833984375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 10.205247813411079, + "grad_norm": 0.18200965225696564, + "learning_rate": 1e-06, + "loss": -0.042, + "num_tokens": 660031442.0, + "reward": 0.6183035969734192, + "reward_std": 0.16570918262004852, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 1092 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4075.0, + "completions/mean_length": 953.6082763671875, + "completions/mean_terminated_length": 662.3621826171875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 10.214577259475218, + "grad_norm": 0.14221099019050598, + "learning_rate": 1e-06, + "loss": -0.0294, + "num_tokens": 660657771.0, + "reward": 0.6819196939468384, + "reward_std": 0.12820754945278168, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1093 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 1083.5357666015625, + "completions/mean_terminated_length": 670.659912109375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 10.223906705539358, + "grad_norm": 0.1637217104434967, + "learning_rate": 1e-06, + "loss": -0.0621, + "num_tokens": 661282835.0, + "reward": 0.6484375, + "reward_std": 0.1703290492296219, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1094 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3873.0, + "completions/mean_length": 975.1763916015625, + "completions/mean_terminated_length": 643.8296508789062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 10.2332361516035, + "grad_norm": 0.15690085291862488, + "learning_rate": 1e-06, + "loss": -0.0163, + "num_tokens": 661888313.0, + "reward": 0.6662946939468384, + "reward_std": 0.11422768235206604, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3734.0, + "completions/mean_length": 1146.0045166015625, + "completions/mean_terminated_length": 672.1709594726562, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 10.242565597667639, + "grad_norm": 0.16998103260993958, + "learning_rate": 1e-06, + "loss": -0.0812, + "num_tokens": 662497989.0, + "reward": 0.6662946939468384, + "reward_std": 0.17367054522037506, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1096 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3975.0, + "completions/mean_length": 959.1763916015625, + "completions/mean_terminated_length": 613.2317504882812, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.251895043731778, + "grad_norm": 0.13692548871040344, + "learning_rate": 1e-06, + "loss": -0.0656, + "num_tokens": 663079347.0, + "reward": 0.7087053656578064, + "reward_std": 0.12956301867961884, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1097 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3943.0, + "completions/mean_length": 1013.2467041015625, + "completions/mean_terminated_length": 643.3162231445312, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 10.261224489795918, + "grad_norm": 0.15446586906909943, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 663680168.0, + "reward": 0.6986607313156128, + "reward_std": 0.13606150448322296, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1098 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2962.0, + "completions/mean_length": 1048.00341796875, + "completions/mean_terminated_length": 594.7115478515625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 10.270553935860057, + "grad_norm": 0.18502353131771088, + "learning_rate": 1e-06, + "loss": -0.0665, + "num_tokens": 664237363.0, + "reward": 0.6517857313156128, + "reward_std": 0.15774603188037872, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1099 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3784.0, + "completions/mean_length": 927.7410888671875, + "completions/mean_terminated_length": 625.63330078125, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 10.279883381924199, + "grad_norm": 0.1612381637096405, + "learning_rate": 1e-06, + "loss": -0.0258, + "num_tokens": 664850587.0, + "reward": 0.629464328289032, + "reward_std": 0.1337621510028839, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1100 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2885.0, + "completions/mean_length": 839.0971069335938, + "completions/mean_terminated_length": 605.3480834960938, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 10.289212827988338, + "grad_norm": 0.16346973180770874, + "learning_rate": 1e-06, + "loss": -0.0166, + "num_tokens": 665446594.0, + "reward": 0.6852678656578064, + "reward_std": 0.1456729769706726, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1101 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3736.0, + "completions/mean_length": 987.5178833007812, + "completions/mean_terminated_length": 665.9507446289062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.298542274052478, + "grad_norm": 0.15040259063243866, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 666079994.0, + "reward": 0.6584821939468384, + "reward_std": 0.15139050781726837, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1102 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 1012.4464721679688, + "completions/mean_terminated_length": 646.7315673828125, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 10.307871720116617, + "grad_norm": 0.14104214310646057, + "learning_rate": 1e-06, + "loss": -0.0287, + "num_tokens": 666691042.0, + "reward": 0.6707589626312256, + "reward_std": 0.12474842369556427, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1103 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3559.0, + "completions/mean_length": 1050.540283203125, + "completions/mean_terminated_length": 680.8160400390625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.317201166180759, + "grad_norm": 0.1439608782529831, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 667319358.0, + "reward": 0.6361607313156128, + "reward_std": 0.10746845602989197, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3646.0, + "completions/mean_length": 957.76123046875, + "completions/mean_terminated_length": 645.862548828125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 10.326530612244898, + "grad_norm": 0.16249814629554749, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 667930424.0, + "reward": 0.6863839626312256, + "reward_std": 0.1633773297071457, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3228.0, + "completions/mean_length": 978.1038208007812, + "completions/mean_terminated_length": 655.5628051757812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 10.335860058309038, + "grad_norm": 0.15066829323768616, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 668547517.0, + "reward": 0.660714328289032, + "reward_std": 0.14267031848430634, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1106 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2457.0, + "completions/mean_length": 797.7600708007812, + "completions/mean_terminated_length": 582.0606079101562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 10.345189504373177, + "grad_norm": 0.16265735030174255, + "learning_rate": 1e-06, + "loss": -0.0243, + "num_tokens": 669123614.0, + "reward": 0.6863839626312256, + "reward_std": 0.12061669677495956, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 1107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 952.9285888671875, + "completions/mean_terminated_length": 610.6138305664062, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 10.354518950437317, + "grad_norm": 0.17049048840999603, + "learning_rate": 1e-06, + "loss": -0.066, + "num_tokens": 669700902.0, + "reward": 0.7042410969734192, + "reward_std": 0.1596214473247528, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2682.0, + "completions/mean_length": 796.6920166015625, + "completions/mean_terminated_length": 589.2621459960938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 10.363848396501458, + "grad_norm": 0.13095861673355103, + "learning_rate": 1e-06, + "loss": -0.0308, + "num_tokens": 670285042.0, + "reward": 0.7410714626312256, + "reward_std": 0.11294566094875336, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1109 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 1026.9832763671875, + "completions/mean_terminated_length": 645.7628784179688, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 10.373177842565598, + "grad_norm": 0.14463892579078674, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 670884811.0, + "reward": 0.6707589626312256, + "reward_std": 0.1314416527748108, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1110 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3369.0, + "completions/mean_length": 994.3147583007812, + "completions/mean_terminated_length": 630.7755737304688, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.382507288629737, + "grad_norm": 0.15866541862487793, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 671464445.0, + "reward": 0.65625, + "reward_std": 0.14447720348834991, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3549.0, + "completions/mean_length": 892.4141235351562, + "completions/mean_terminated_length": 608.2564086914062, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 10.391836734693877, + "grad_norm": 0.14958196878433228, + "learning_rate": 1e-06, + "loss": -0.0344, + "num_tokens": 672050560.0, + "reward": 0.754464328289032, + "reward_std": 0.12921854853630066, + "rewards/verify_math_reward/mean": 0.7544642686843872, + "rewards/verify_math_reward/std": 0.43064478039741516, + "step": 1112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3337.0, + "completions/mean_length": 1150.2310791015625, + "completions/mean_terminated_length": 672.6420288085938, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 10.401166180758018, + "grad_norm": 0.14654992520809174, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 672668919.0, + "reward": 0.6104910969734192, + "reward_std": 0.12042984366416931, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 1113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3357.0, + "completions/mean_length": 876.6116333007812, + "completions/mean_terminated_length": 612.2174072265625, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 10.410495626822158, + "grad_norm": 0.12034684419631958, + "learning_rate": 1e-06, + "loss": -0.0504, + "num_tokens": 673270203.0, + "reward": 0.7433035969734192, + "reward_std": 0.11126275360584259, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1114 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3316.0, + "completions/mean_length": 1035.78125, + "completions/mean_terminated_length": 611.93896484375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 10.419825072886297, + "grad_norm": 0.1664595603942871, + "learning_rate": 1e-06, + "loss": -0.0763, + "num_tokens": 673831095.0, + "reward": 0.7020089626312256, + "reward_std": 0.1554897278547287, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3208.0, + "completions/mean_length": 1114.919677734375, + "completions/mean_terminated_length": 631.605712890625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 10.429154518950437, + "grad_norm": 0.17589278519153595, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 674410319.0, + "reward": 0.606026828289032, + "reward_std": 0.16345034539699554, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 1116 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3903.0, + "completions/mean_length": 810.6975708007812, + "completions/mean_terminated_length": 649.125244140625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 10.438483965014576, + "grad_norm": 0.1613699197769165, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 675044560.0, + "reward": 0.7265625596046448, + "reward_std": 0.16243509948253632, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1117 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3036.0, + "completions/mean_length": 990.950927734375, + "completions/mean_terminated_length": 652.7772216796875, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.447813411078718, + "grad_norm": 0.15767882764339447, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 675663276.0, + "reward": 0.6852678656578064, + "reward_std": 0.15511175990104675, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2819.0, + "completions/mean_length": 1018.247802734375, + "completions/mean_terminated_length": 657.5137329101562, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.457142857142857, + "grad_norm": 0.15579132735729218, + "learning_rate": 1e-06, + "loss": -0.025, + "num_tokens": 676283754.0, + "reward": 0.5970982313156128, + "reward_std": 0.13391369581222534, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 1119 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2636.0, + "completions/mean_length": 1041.25, + "completions/mean_terminated_length": 613.740478515625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 10.466472303206997, + "grad_norm": 0.1636897474527359, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 676857530.0, + "reward": 0.6796875596046448, + "reward_std": 0.14211323857307434, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1120 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3175.0, + "completions/mean_length": 935.755615234375, + "completions/mean_terminated_length": 638.6386108398438, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 10.475801749271136, + "grad_norm": 0.1545189917087555, + "learning_rate": 1e-06, + "loss": -0.0561, + "num_tokens": 677471551.0, + "reward": 0.6819196939468384, + "reward_std": 0.1334686130285263, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1121 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2984.0, + "completions/mean_length": 996.8370971679688, + "completions/mean_terminated_length": 576.5449829101562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 10.485131195335278, + "grad_norm": 0.18272235989570618, + "learning_rate": 1e-06, + "loss": -0.0657, + "num_tokens": 678007077.0, + "reward": 0.699776828289032, + "reward_std": 0.14128093421459198, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1122 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3729.0, + "completions/mean_length": 966.0848388671875, + "completions/mean_terminated_length": 650.7862548828125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.494460641399417, + "grad_norm": 0.15357354283332825, + "learning_rate": 1e-06, + "loss": -0.0641, + "num_tokens": 678620913.0, + "reward": 0.6774553656578064, + "reward_std": 0.14135509729385376, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1123 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3157.0, + "completions/mean_length": 894.8438110351562, + "completions/mean_terminated_length": 615.131103515625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 10.503790087463557, + "grad_norm": 0.15871113538742065, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 679213685.0, + "reward": 0.6662946939468384, + "reward_std": 0.13241735100746155, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1124 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 850.6417846679688, + "completions/mean_terminated_length": 588.3510131835938, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 10.513119533527696, + "grad_norm": 0.15624141693115234, + "learning_rate": 1e-06, + "loss": -0.0363, + "num_tokens": 679781652.0, + "reward": 0.7087053656578064, + "reward_std": 0.13200436532497406, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1125 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2380.0, + "completions/mean_length": 929.8739013671875, + "completions/mean_terminated_length": 563.186767578125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 10.522448979591836, + "grad_norm": 0.1677175760269165, + "learning_rate": 1e-06, + "loss": -0.0369, + "num_tokens": 680339675.0, + "reward": 0.7053571939468384, + "reward_std": 0.1344464272260666, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 1126 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2253.0, + "completions/mean_length": 871.9241333007812, + "completions/mean_terminated_length": 551.4944458007812, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 10.531778425655977, + "grad_norm": 0.178900808095932, + "learning_rate": 1e-06, + "loss": -0.023, + "num_tokens": 680880983.0, + "reward": 0.7187500596046448, + "reward_std": 0.13984806835651398, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1127 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2851.0, + "completions/mean_length": 933.958740234375, + "completions/mean_terminated_length": 567.7446899414062, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 10.541107871720117, + "grad_norm": 0.1565646529197693, + "learning_rate": 1e-06, + "loss": -0.0333, + "num_tokens": 681438370.0, + "reward": 0.684151828289032, + "reward_std": 0.10942055284976959, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2629.0, + "completions/mean_length": 890.15185546875, + "completions/mean_terminated_length": 588.7472534179688, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.550437317784256, + "grad_norm": 0.15974071621894836, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 682009338.0, + "reward": 0.65625, + "reward_std": 0.13673411309719086, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1129 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2735.0, + "completions/mean_length": 946.3326416015625, + "completions/mean_terminated_length": 611.9234619140625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 10.559766763848396, + "grad_norm": 0.1527777463197708, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 682591908.0, + "reward": 0.715401828289032, + "reward_std": 0.14594466984272003, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1130 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2332.0, + "completions/mean_length": 839.0949096679688, + "completions/mean_terminated_length": 592.7742919921875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 10.569096209912537, + "grad_norm": 0.16545908153057098, + "learning_rate": 1e-06, + "loss": -0.0187, + "num_tokens": 683166153.0, + "reward": 0.6830357313156128, + "reward_std": 0.12828493118286133, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1131 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2772.0, + "completions/mean_length": 1189.87060546875, + "completions/mean_terminated_length": 665.312255859375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.578425655976677, + "grad_norm": 0.16721579432487488, + "learning_rate": 1e-06, + "loss": -0.0883, + "num_tokens": 683764973.0, + "reward": 0.6383928656578064, + "reward_std": 0.15875522792339325, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1132 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3871.0, + "completions/mean_length": 857.3750610351562, + "completions/mean_terminated_length": 624.9378051757812, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.587755102040816, + "grad_norm": 0.14830856025218964, + "learning_rate": 1e-06, + "loss": -0.0194, + "num_tokens": 684386341.0, + "reward": 0.7031250596046448, + "reward_std": 0.12651759386062622, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1133 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3153.0, + "completions/mean_length": 1000.8516235351562, + "completions/mean_terminated_length": 650.9652099609375, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 10.597084548104956, + "grad_norm": 0.15503476560115814, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 685000200.0, + "reward": 0.6328125, + "reward_std": 0.14075776934623718, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1134 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3690.0, + "completions/mean_length": 888.6585083007812, + "completions/mean_terminated_length": 578.5238647460938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.606413994169095, + "grad_norm": 0.17902444303035736, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 685558062.0, + "reward": 0.7377232313156128, + "reward_std": 0.15229134261608124, + "rewards/verify_math_reward/mean": 0.7377232313156128, + "rewards/verify_math_reward/std": 0.4401180148124695, + "step": 1135 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2762.0, + "completions/mean_length": 934.3817138671875, + "completions/mean_terminated_length": 576.9813842773438, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 10.615743440233237, + "grad_norm": 0.16346406936645508, + "learning_rate": 1e-06, + "loss": -0.0851, + "num_tokens": 686110260.0, + "reward": 0.6908482313156128, + "reward_std": 0.147819384932518, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1136 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2515.0, + "completions/mean_length": 1025.3092041015625, + "completions/mean_terminated_length": 591.1095581054688, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 10.625072886297376, + "grad_norm": 0.2029131054878235, + "learning_rate": 1e-06, + "loss": -0.0639, + "num_tokens": 686671081.0, + "reward": 0.6595982313156128, + "reward_std": 0.16875603795051575, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1137 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 893.138427734375, + "completions/mean_terminated_length": 609.0449829101562, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 10.634402332361516, + "grad_norm": 0.14421267807483673, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 687264261.0, + "reward": 0.7287946939468384, + "reward_std": 0.11178633570671082, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1138 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2668.0, + "completions/mean_length": 917.5881958007812, + "completions/mean_terminated_length": 623.003662109375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.643731778425655, + "grad_norm": 0.151288241147995, + "learning_rate": 1e-06, + "loss": -0.0255, + "num_tokens": 687852260.0, + "reward": 0.6886160969734192, + "reward_std": 0.1359841227531433, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1139 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3071.0, + "completions/mean_length": 1052.0513916015625, + "completions/mean_terminated_length": 647.9873657226562, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 10.653061224489797, + "grad_norm": 0.162523090839386, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 688463394.0, + "reward": 0.6517857313156128, + "reward_std": 0.13906781375408173, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1140 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 1101.118408203125, + "completions/mean_terminated_length": 646.8817749023438, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.662390670553936, + "grad_norm": 0.17056100070476532, + "learning_rate": 1e-06, + "loss": -0.085, + "num_tokens": 689050604.0, + "reward": 0.65625, + "reward_std": 0.15037909150123596, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1141 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3051.0, + "completions/mean_length": 968.9676513671875, + "completions/mean_terminated_length": 628.3997192382812, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 10.671720116618076, + "grad_norm": 0.14288674294948578, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 689652943.0, + "reward": 0.7198660969734192, + "reward_std": 0.1219707801938057, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1142 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3984.0, + "completions/mean_length": 986.19873046875, + "completions/mean_terminated_length": 582.2774047851562, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.681049562682215, + "grad_norm": 0.17460224032402039, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 690213561.0, + "reward": 0.6785714626312256, + "reward_std": 0.13147304952144623, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1143 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2849.0, + "completions/mean_length": 908.8605346679688, + "completions/mean_terminated_length": 617.7089233398438, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 10.690379008746355, + "grad_norm": 0.13765406608581543, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 690810356.0, + "reward": 0.6629464626312256, + "reward_std": 0.12587566673755646, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1144 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2463.0, + "completions/mean_length": 953.1975708007812, + "completions/mean_terminated_length": 615.2200317382812, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 10.699708454810496, + "grad_norm": 0.16693969070911407, + "learning_rate": 1e-06, + "loss": -0.0839, + "num_tokens": 691397125.0, + "reward": 0.6819196939468384, + "reward_std": 0.1585690677165985, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1145 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3089.0, + "completions/mean_length": 942.4933471679688, + "completions/mean_terminated_length": 629.0772705078125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 10.709037900874636, + "grad_norm": 0.14636832475662231, + "learning_rate": 1e-06, + "loss": -0.0446, + "num_tokens": 691991519.0, + "reward": 0.6584821939468384, + "reward_std": 0.1571401208639145, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1146 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3242.0, + "completions/mean_length": 986.2254638671875, + "completions/mean_terminated_length": 621.7381591796875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 10.718367346938775, + "grad_norm": 0.13918933272361755, + "learning_rate": 1e-06, + "loss": -0.0368, + "num_tokens": 692579833.0, + "reward": 0.7008928656578064, + "reward_std": 0.11283829063177109, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3907.0, + "completions/mean_length": 1017.70654296875, + "completions/mean_terminated_length": 630.9861450195312, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 10.727696793002915, + "grad_norm": 0.16253246366977692, + "learning_rate": 1e-06, + "loss": -0.0393, + "num_tokens": 693167786.0, + "reward": 0.6595982313156128, + "reward_std": 0.14210325479507446, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1148 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4062.0, + "completions/mean_length": 930.05029296875, + "completions/mean_terminated_length": 632.3968505859375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.737026239067056, + "grad_norm": 0.15924592316150665, + "learning_rate": 1e-06, + "loss": -0.0704, + "num_tokens": 693771247.0, + "reward": 0.7287946939468384, + "reward_std": 0.14770174026489258, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1149 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3442.0, + "completions/mean_length": 1046.6507568359375, + "completions/mean_terminated_length": 646.2310791015625, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 10.746355685131196, + "grad_norm": 0.15888065099716187, + "learning_rate": 1e-06, + "loss": -0.0646, + "num_tokens": 694374342.0, + "reward": 0.6272321939468384, + "reward_std": 0.17344370484352112, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1150 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4019.0, + "completions/mean_length": 972.58935546875, + "completions/mean_terminated_length": 636.6971435546875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 10.755685131195335, + "grad_norm": 0.15894484519958496, + "learning_rate": 1e-06, + "loss": -0.0379, + "num_tokens": 694979950.0, + "reward": 0.621651828289032, + "reward_std": 0.145303875207901, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 1151 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3627.0, + "completions/mean_length": 1081.4676513671875, + "completions/mean_terminated_length": 633.152587890625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 10.765014577259475, + "grad_norm": 0.15034615993499756, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 695567729.0, + "reward": 0.6305803656578064, + "reward_std": 0.14722095429897308, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3079.0, + "completions/mean_length": 952.6920166015625, + "completions/mean_terminated_length": 597.3614501953125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 10.774344023323614, + "grad_norm": 0.16728702187538147, + "learning_rate": 1e-06, + "loss": -0.0606, + "num_tokens": 696140021.0, + "reward": 0.7087053656578064, + "reward_std": 0.14117074012756348, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3968.0, + "completions/mean_length": 980.482177734375, + "completions/mean_terminated_length": 623.9801025390625, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 10.783673469387756, + "grad_norm": 0.1673901528120041, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 696730101.0, + "reward": 0.699776828289032, + "reward_std": 0.15590018033981323, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1154 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3332.0, + "completions/mean_length": 954.3594360351562, + "completions/mean_terminated_length": 625.0875244140625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.793002915451895, + "grad_norm": 0.15242673456668854, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 697324351.0, + "reward": 0.640625, + "reward_std": 0.15248385071754456, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 1155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3945.0, + "completions/mean_length": 981.66748046875, + "completions/mean_terminated_length": 655.2576904296875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.802332361516035, + "grad_norm": 0.15549147129058838, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 697946629.0, + "reward": 0.6752232313156128, + "reward_std": 0.13659143447875977, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1156 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3469.0, + "completions/mean_length": 1054.188720703125, + "completions/mean_terminated_length": 615.204345703125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 10.811661807580174, + "grad_norm": 0.16542983055114746, + "learning_rate": 1e-06, + "loss": -0.0822, + "num_tokens": 698526854.0, + "reward": 0.6941964626312256, + "reward_std": 0.15578323602676392, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 1057.3226318359375, + "completions/mean_terminated_length": 627.6497192382812, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 10.820991253644316, + "grad_norm": 0.17541861534118652, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 699109343.0, + "reward": 0.6718750596046448, + "reward_std": 0.15529470145702362, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1158 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4045.0, + "completions/mean_length": 781.0558471679688, + "completions/mean_terminated_length": 547.3858642578125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 10.830320699708455, + "grad_norm": 0.1647026389837265, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 699652657.0, + "reward": 0.8024553656578064, + "reward_std": 0.10817524790763855, + "rewards/verify_math_reward/mean": 0.8024553656578064, + "rewards/verify_math_reward/std": 0.3983690142631531, + "step": 1159 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3876.0, + "completions/mean_length": 1030.8326416015625, + "completions/mean_terminated_length": 645.7612915039062, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 10.839650145772595, + "grad_norm": 0.1713884472846985, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 700260491.0, + "reward": 0.6796875596046448, + "reward_std": 0.1737472116947174, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1160 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3724.0, + "completions/mean_length": 873.5670166015625, + "completions/mean_terminated_length": 579.1912841796875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.848979591836734, + "grad_norm": 0.1826779991388321, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 700818055.0, + "reward": 0.7109375596046448, + "reward_std": 0.1509779393672943, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1161 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2199.0, + "completions/mean_length": 972.4453735351562, + "completions/mean_terminated_length": 619.3478393554688, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 10.858309037900874, + "grad_norm": 0.1627926379442215, + "learning_rate": 1e-06, + "loss": -0.0667, + "num_tokens": 701404254.0, + "reward": 0.6886160969734192, + "reward_std": 0.16717414557933807, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1162 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 848.239990234375, + "completions/mean_terminated_length": 598.4122924804688, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 10.867638483965015, + "grad_norm": 0.17906469106674194, + "learning_rate": 1e-06, + "loss": -0.0337, + "num_tokens": 701991813.0, + "reward": 0.738839328289032, + "reward_std": 0.15097863972187042, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1163 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3925.0, + "completions/mean_length": 1101.38623046875, + "completions/mean_terminated_length": 669.2132568359375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 10.876967930029155, + "grad_norm": 0.17050032317638397, + "learning_rate": 1e-06, + "loss": -0.0427, + "num_tokens": 702598143.0, + "reward": 0.6316964626312256, + "reward_std": 0.13147304952144623, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1164 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 1037.6328125, + "completions/mean_terminated_length": 591.7838745117188, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 10.886297376093294, + "grad_norm": 0.1833752989768982, + "learning_rate": 1e-06, + "loss": -0.0762, + "num_tokens": 703146134.0, + "reward": 0.7098214626312256, + "reward_std": 0.15315786004066467, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3746.0, + "completions/mean_length": 935.4386596679688, + "completions/mean_terminated_length": 608.4840087890625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 10.895626822157434, + "grad_norm": 0.15420681238174438, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 703728695.0, + "reward": 0.7444196939468384, + "reward_std": 0.13523708283901215, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1166 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3185.0, + "completions/mean_length": 930.8795166015625, + "completions/mean_terminated_length": 624.8274536132812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.904956268221575, + "grad_norm": 0.1642269343137741, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 704327027.0, + "reward": 0.731026828289032, + "reward_std": 0.14661867916584015, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1167 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3135.0, + "completions/mean_length": 1023.3147583007812, + "completions/mean_terminated_length": 548.1572265625, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 10.914285714285715, + "grad_norm": 0.16227683424949646, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 704844149.0, + "reward": 0.7042410969734192, + "reward_std": 0.1000252291560173, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3465.0, + "completions/mean_length": 1114.774658203125, + "completions/mean_terminated_length": 740.2487182617188, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 10.923615160349854, + "grad_norm": 0.1442495584487915, + "learning_rate": 1e-06, + "loss": -0.0345, + "num_tokens": 705519483.0, + "reward": 0.6183035969734192, + "reward_std": 0.13106118142604828, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 1169 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3631.0, + "completions/mean_length": 1007.4397583007812, + "completions/mean_terminated_length": 597.453857421875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 10.932944606413994, + "grad_norm": 0.18583931028842926, + "learning_rate": 1e-06, + "loss": -0.0617, + "num_tokens": 706083317.0, + "reward": 0.660714328289032, + "reward_std": 0.17540396749973297, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1170 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 1024.9910888671875, + "completions/mean_terminated_length": 639.1859130859375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 10.942274052478133, + "grad_norm": 0.15888862311840057, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 706688533.0, + "reward": 0.6037946939468384, + "reward_std": 0.1430903524160385, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 1171 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3884.0, + "completions/mean_length": 1144.587158203125, + "completions/mean_terminated_length": 679.3772583007812, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 10.951603498542275, + "grad_norm": 0.1542436182498932, + "learning_rate": 1e-06, + "loss": -0.0807, + "num_tokens": 707304755.0, + "reward": 0.621651828289032, + "reward_std": 0.1274939775466919, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 1172 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3657.0, + "completions/mean_length": 1071.61279296875, + "completions/mean_terminated_length": 657.102783203125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.960932944606414, + "grad_norm": 0.1752346009016037, + "learning_rate": 1e-06, + "loss": -0.0667, + "num_tokens": 707904080.0, + "reward": 0.6473214626312256, + "reward_std": 0.17299720644950867, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 1173 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3129.0, + "completions/mean_length": 1063.5592041015625, + "completions/mean_terminated_length": 647.9454345703125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.970262390670554, + "grad_norm": 0.1515476256608963, + "learning_rate": 1e-06, + "loss": -0.0861, + "num_tokens": 708500077.0, + "reward": 0.6383928656578064, + "reward_std": 0.15665017068386078, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 1174 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3753.0, + "completions/mean_length": 834.0636596679688, + "completions/mean_terminated_length": 633.0912475585938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 10.979591836734693, + "grad_norm": 0.1750250905752182, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 709127406.0, + "reward": 0.7555803656578064, + "reward_std": 0.1519550383090973, + "rewards/verify_math_reward/mean": 0.7555803656578064, + "rewards/verify_math_reward/std": 0.42998257279396057, + "step": 1175 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2986.0, + "completions/mean_length": 970.29248046875, + "completions/mean_terminated_length": 612.6243896484375, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 10.988921282798835, + "grad_norm": 0.17585763335227966, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 709709836.0, + "reward": 0.7287946939468384, + "reward_std": 0.16139131784439087, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1176 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.10795454545454541, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2754.0, + "completions/mean_length": 1061.8011474609375, + "completions/mean_terminated_length": 694.6051025390625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.998250728862974, + "grad_norm": 0.15547995269298553, + "learning_rate": 1e-06, + "loss": -0.0263, + "num_tokens": 710339164.0, + "reward": 0.6205357313156128, + "reward_std": 0.11501862853765488, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1177 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2581.0, + "completions/mean_length": 931.5067138671875, + "completions/mean_terminated_length": 659.1685180664062, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 11.00932944606414, + "grad_norm": 0.14872713387012482, + "learning_rate": 1e-06, + "loss": -0.06, + "num_tokens": 710976674.0, + "reward": 0.7020089626312256, + "reward_std": 0.14740893244743347, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1178 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3588.0, + "completions/mean_length": 1179.7679443359375, + "completions/mean_terminated_length": 728.8040771484375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 11.018658892128279, + "grad_norm": 0.15998126566410065, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 711633698.0, + "reward": 0.609375, + "reward_std": 0.15597756206989288, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 1179 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2585.0, + "completions/mean_length": 1049.69091796875, + "completions/mean_terminated_length": 596.6500244140625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.02798833819242, + "grad_norm": 0.16757667064666748, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 712180453.0, + "reward": 0.6629464626312256, + "reward_std": 0.16473238170146942, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1180 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3572.0, + "completions/mean_length": 1004.279052734375, + "completions/mean_terminated_length": 607.1057739257812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.03731778425656, + "grad_norm": 0.16762277483940125, + "learning_rate": 1e-06, + "loss": -0.078, + "num_tokens": 712748647.0, + "reward": 0.6808035969734192, + "reward_std": 0.15608635544776917, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1181 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 938.44091796875, + "completions/mean_terminated_length": 649.9915161132812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 11.0466472303207, + "grad_norm": 0.15891775488853455, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 713368754.0, + "reward": 0.6830357313156128, + "reward_std": 0.16032683849334717, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1182 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2878.0, + "completions/mean_length": 916.239990234375, + "completions/mean_terminated_length": 625.7625122070312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.055976676384839, + "grad_norm": 0.14600877463817596, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 713971889.0, + "reward": 0.6986607313156128, + "reward_std": 0.11802128702402115, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1183 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3359.0, + "completions/mean_length": 1092.568115234375, + "completions/mean_terminated_length": 645.9038696289062, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 11.06530612244898, + "grad_norm": 0.14944584667682648, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 714571630.0, + "reward": 0.6573660969734192, + "reward_std": 0.1328292340040207, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1184 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3289.0, + "completions/mean_length": 1031.2054443359375, + "completions/mean_terminated_length": 624.3742065429688, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.07463556851312, + "grad_norm": 0.17703565955162048, + "learning_rate": 1e-06, + "loss": -0.0934, + "num_tokens": 715155062.0, + "reward": 0.6506696939468384, + "reward_std": 0.1618363857269287, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1185 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1607142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 1179.536865234375, + "completions/mean_terminated_length": 621.0651245117188, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 11.08396501457726, + "grad_norm": 0.19329240918159485, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 715717031.0, + "reward": 0.6026785969734192, + "reward_std": 0.17566610872745514, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 1186 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2515.0, + "completions/mean_length": 973.349365234375, + "completions/mean_terminated_length": 576.6351928710938, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 11.093294460641399, + "grad_norm": 0.17535319924354553, + "learning_rate": 1e-06, + "loss": -0.0776, + "num_tokens": 716268528.0, + "reward": 0.6897321939468384, + "reward_std": 0.15649932622909546, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1187 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3204.0, + "completions/mean_length": 998.8359985351562, + "completions/mean_terminated_length": 635.8267211914062, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 11.102623906705539, + "grad_norm": 0.1688397377729416, + "learning_rate": 1e-06, + "loss": -0.0825, + "num_tokens": 716878469.0, + "reward": 0.6350446939468384, + "reward_std": 0.16555652022361755, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1188 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3219.0, + "completions/mean_length": 956.2332763671875, + "completions/mean_terminated_length": 557.3446655273438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 11.11195335276968, + "grad_norm": 0.16133227944374084, + "learning_rate": 1e-06, + "loss": -0.0653, + "num_tokens": 717396902.0, + "reward": 0.7098214626312256, + "reward_std": 0.13421748578548431, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1189 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3606.0, + "completions/mean_length": 1134.30810546875, + "completions/mean_terminated_length": 613.4855346679688, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 11.12128279883382, + "grad_norm": 0.15135248005390167, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 717959154.0, + "reward": 0.6808035969734192, + "reward_std": 0.11693934351205826, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1190 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3105.0, + "completions/mean_length": 926.7467041015625, + "completions/mean_terminated_length": 568.4832153320312, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.130612244897959, + "grad_norm": 0.13976995646953583, + "learning_rate": 1e-06, + "loss": -0.0198, + "num_tokens": 718493855.0, + "reward": 0.7031250596046448, + "reward_std": 0.09394001215696335, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1191 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2716.0, + "completions/mean_length": 1129.1953125, + "completions/mean_terminated_length": 670.4110717773438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 11.139941690962099, + "grad_norm": 0.18943127989768982, + "learning_rate": 1e-06, + "loss": -0.052, + "num_tokens": 719117126.0, + "reward": 0.5680803656578064, + "reward_std": 0.15352554619312286, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 1192 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3081.0, + "completions/mean_length": 995.0457763671875, + "completions/mean_terminated_length": 622.9312133789062, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 11.14927113702624, + "grad_norm": 0.16766230762004852, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 719699495.0, + "reward": 0.7098214626312256, + "reward_std": 0.131475567817688, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1193 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3329.0, + "completions/mean_length": 1019.068115234375, + "completions/mean_terminated_length": 588.4542236328125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 11.15860058309038, + "grad_norm": 0.15402153134346008, + "learning_rate": 1e-06, + "loss": -0.0701, + "num_tokens": 720246516.0, + "reward": 0.7087053656578064, + "reward_std": 0.12061528861522675, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1194 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2794.0, + "completions/mean_length": 956.50341796875, + "completions/mean_terminated_length": 584.153564453125, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 11.167930029154519, + "grad_norm": 0.1751590371131897, + "learning_rate": 1e-06, + "loss": -0.061, + "num_tokens": 720816391.0, + "reward": 0.668526828289032, + "reward_std": 0.12651829421520233, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3932.0, + "completions/mean_length": 1132.688720703125, + "completions/mean_terminated_length": 625.2457885742188, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.177259475218658, + "grad_norm": 0.1672411561012268, + "learning_rate": 1e-06, + "loss": -0.0657, + "num_tokens": 721373448.0, + "reward": 0.6629464626312256, + "reward_std": 0.13978277146816254, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2343.0, + "completions/mean_length": 854.2176513671875, + "completions/mean_terminated_length": 562.3783569335938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.186588921282798, + "grad_norm": 0.16484569013118744, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 721922379.0, + "reward": 0.7109375596046448, + "reward_std": 0.1349353790283203, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1197 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2601.0, + "completions/mean_length": 920.3694458007812, + "completions/mean_terminated_length": 630.2692260742188, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 11.19591836734694, + "grad_norm": 0.1329880952835083, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 722524246.0, + "reward": 0.6986607313156128, + "reward_std": 0.09739913791418076, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960144996643, + "step": 1198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3234.0, + "completions/mean_length": 1060.9710693359375, + "completions/mean_terminated_length": 600.645263671875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.205247813411079, + "grad_norm": 0.13303211331367493, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 723087260.0, + "reward": 0.6819196939468384, + "reward_std": 0.1017170175909996, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1199 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 889.958740234375, + "completions/mean_terminated_length": 622.465576171875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 11.214577259475218, + "grad_norm": 0.15574702620506287, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 723697247.0, + "reward": 0.7120535969734192, + "reward_std": 0.1455221176147461, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1200 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3817.0, + "completions/mean_length": 968.8125610351562, + "completions/mean_terminated_length": 645.3103637695312, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 11.223906705539358, + "grad_norm": 0.15969006717205048, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 724303935.0, + "reward": 0.6785714626312256, + "reward_std": 0.14222271740436554, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1201 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3110.0, + "completions/mean_length": 931.5781860351562, + "completions/mean_terminated_length": 659.24609375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.2332361516035, + "grad_norm": 0.13475307822227478, + "learning_rate": 1e-06, + "loss": -0.0289, + "num_tokens": 724934901.0, + "reward": 0.7109375596046448, + "reward_std": 0.1170462816953659, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1202 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3203.0, + "completions/mean_length": 887.38623046875, + "completions/mean_terminated_length": 568.4932250976562, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.242565597667639, + "grad_norm": 0.1430935561656952, + "learning_rate": 1e-06, + "loss": -0.049, + "num_tokens": 725484655.0, + "reward": 0.7533482313156128, + "reward_std": 0.12437751144170761, + "rewards/verify_math_reward/mean": 0.7533482313156128, + "rewards/verify_math_reward/std": 0.4313030242919922, + "step": 1203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2292.0, + "completions/mean_length": 889.6506958007812, + "completions/mean_terminated_length": 583.9107666015625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.251895043731778, + "grad_norm": 0.14268061518669128, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 726055446.0, + "reward": 0.7533482313156128, + "reward_std": 0.11426863074302673, + "rewards/verify_math_reward/mean": 0.7533482313156128, + "rewards/verify_math_reward/std": 0.4313030242919922, + "step": 1204 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4043.0, + "completions/mean_length": 1027.79248046875, + "completions/mean_terminated_length": 655.306640625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.261224489795918, + "grad_norm": 0.1561819463968277, + "learning_rate": 1e-06, + "loss": -0.0264, + "num_tokens": 726688596.0, + "reward": 0.5970982313156128, + "reward_std": 0.13203644752502441, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 1205 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3806.0, + "completions/mean_length": 846.3739013671875, + "completions/mean_terminated_length": 545.1890258789062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.270553935860057, + "grad_norm": 0.16859789192676544, + "learning_rate": 1e-06, + "loss": -0.0397, + "num_tokens": 727226587.0, + "reward": 0.6752232313156128, + "reward_std": 0.15555500984191895, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1206 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3603.0, + "completions/mean_length": 953.45654296875, + "completions/mean_terminated_length": 624.0900268554688, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 11.279883381924199, + "grad_norm": 0.15560832619667053, + "learning_rate": 1e-06, + "loss": -0.0803, + "num_tokens": 727818852.0, + "reward": 0.6808035969734192, + "reward_std": 0.16096875071525574, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1207 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3903.0, + "completions/mean_length": 905.1785888671875, + "completions/mean_terminated_length": 588.053955078125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 11.289212827988338, + "grad_norm": 0.12583661079406738, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 728391084.0, + "reward": 0.6718750596046448, + "reward_std": 0.10908354818820953, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3655.0, + "completions/mean_length": 852.5167846679688, + "completions/mean_terminated_length": 590.3775634765625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 11.298542274052478, + "grad_norm": 0.143217071890831, + "learning_rate": 1e-06, + "loss": -0.0369, + "num_tokens": 728968195.0, + "reward": 0.6930803656578064, + "reward_std": 0.09841014444828033, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 1209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2773.0, + "completions/mean_length": 844.8080444335938, + "completions/mean_terminated_length": 543.47802734375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.307871720116617, + "grad_norm": 0.1402239054441452, + "learning_rate": 1e-06, + "loss": -0.0186, + "num_tokens": 729493103.0, + "reward": 0.7645089626312256, + "reward_std": 0.08067300915718079, + "rewards/verify_math_reward/mean": 0.7645089030265808, + "rewards/verify_math_reward/std": 0.42454230785369873, + "step": 1210 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1696428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2438.0, + "completions/mean_length": 1224.149658203125, + "completions/mean_terminated_length": 637.4274291992188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.317201166180759, + "grad_norm": 0.18407954275608063, + "learning_rate": 1e-06, + "loss": -0.0679, + "num_tokens": 730067565.0, + "reward": 0.606026828289032, + "reward_std": 0.16558974981307983, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 1211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2587.0, + "completions/mean_length": 948.5647583007812, + "completions/mean_terminated_length": 584.0422973632812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 11.326530612244898, + "grad_norm": 0.16666607558727264, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 730623055.0, + "reward": 0.6941964626312256, + "reward_std": 0.1431998312473297, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1212 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2293.0, + "completions/mean_length": 859.6886596679688, + "completions/mean_terminated_length": 572.6282348632812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.335860058309038, + "grad_norm": 0.13089048862457275, + "learning_rate": 1e-06, + "loss": -0.0319, + "num_tokens": 731171656.0, + "reward": 0.7477678656578064, + "reward_std": 0.09063425660133362, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2888.0, + "completions/mean_length": 965.044677734375, + "completions/mean_terminated_length": 584.941162109375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 11.345189504373177, + "grad_norm": 0.1596495509147644, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 731727240.0, + "reward": 0.6383928656578064, + "reward_std": 0.12381549179553986, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1214 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2284.0, + "completions/mean_length": 965.7310791015625, + "completions/mean_terminated_length": 629.1013793945312, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 11.354518950437317, + "grad_norm": 0.15813496708869934, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 732330639.0, + "reward": 0.715401828289032, + "reward_std": 0.15995915234088898, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2881.0, + "completions/mean_length": 1054.8504638671875, + "completions/mean_terminated_length": 624.8280639648438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.363848396501458, + "grad_norm": 0.1552128940820694, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 732902953.0, + "reward": 0.6506696939468384, + "reward_std": 0.12463965266942978, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1216 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3611.0, + "completions/mean_length": 1056.7388916015625, + "completions/mean_terminated_length": 595.7711791992188, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.373177842565598, + "grad_norm": 0.16231252253055573, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 733466711.0, + "reward": 0.6808035969734192, + "reward_std": 0.13403314352035522, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1217 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2360.0, + "completions/mean_length": 917.9520263671875, + "completions/mean_terminated_length": 619.1611938476562, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.382507288629737, + "grad_norm": 0.16147896647453308, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 734068820.0, + "reward": 0.7031250596046448, + "reward_std": 0.15331009030342102, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1218 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3944.0, + "completions/mean_length": 876.1652221679688, + "completions/mean_terminated_length": 569.139404296875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 11.391836734693877, + "grad_norm": 0.16741862893104553, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 734624400.0, + "reward": 0.6930803656578064, + "reward_std": 0.14684370160102844, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3256.0, + "completions/mean_length": 930.39404296875, + "completions/mean_terminated_length": 607.2139892578125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 11.401166180758018, + "grad_norm": 0.16763009130954742, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 735205969.0, + "reward": 0.6953125596046448, + "reward_std": 0.1680731475353241, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1220 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3747.0, + "completions/mean_length": 1020.7734985351562, + "completions/mean_terminated_length": 643.114013671875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.410495626822158, + "grad_norm": 0.17864224314689636, + "learning_rate": 1e-06, + "loss": -0.0738, + "num_tokens": 735801574.0, + "reward": 0.6227678656578064, + "reward_std": 0.16679365932941437, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 1221 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4084.0, + "completions/mean_length": 995.3013916015625, + "completions/mean_terminated_length": 627.5530395507812, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 11.419825072886297, + "grad_norm": 0.1446545571088791, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 736391636.0, + "reward": 0.7031250596046448, + "reward_std": 0.1317761391401291, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1222 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 1016.5201416015625, + "completions/mean_terminated_length": 693.7632446289062, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 11.429154518950437, + "grad_norm": 0.1537589728832245, + "learning_rate": 1e-06, + "loss": -0.0133, + "num_tokens": 737034382.0, + "reward": 0.668526828289032, + "reward_std": 0.14556488394737244, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2297.0, + "completions/mean_length": 764.3973388671875, + "completions/mean_terminated_length": 550.7316284179688, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 11.438483965014576, + "grad_norm": 0.16585788130760193, + "learning_rate": 1e-06, + "loss": -0.0678, + "num_tokens": 737582242.0, + "reward": 0.7500000596046448, + "reward_std": 0.13129010796546936, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1224 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2606.0, + "completions/mean_length": 1060.091552734375, + "completions/mean_terminated_length": 626.3903198242188, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 11.447813411078718, + "grad_norm": 0.1416480541229248, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 738165292.0, + "reward": 0.6383928656578064, + "reward_std": 0.12287301570177078, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1225 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2388.0, + "completions/mean_length": 945.841552734375, + "completions/mean_terminated_length": 594.0868530273438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.457142857142857, + "grad_norm": 0.15614363551139832, + "learning_rate": 1e-06, + "loss": -0.0448, + "num_tokens": 738720542.0, + "reward": 0.6908482313156128, + "reward_std": 0.1535690277814865, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1226 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3001.0, + "completions/mean_length": 954.1797485351562, + "completions/mean_terminated_length": 577.1612548828125, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 11.466472303206997, + "grad_norm": 0.15307436883449554, + "learning_rate": 1e-06, + "loss": -0.0821, + "num_tokens": 739264279.0, + "reward": 0.6774553656578064, + "reward_std": 0.13560545444488525, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1227 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2871.0, + "completions/mean_length": 996.4732666015625, + "completions/mean_terminated_length": 602.6968383789062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 11.475801749271136, + "grad_norm": 0.19294264912605286, + "learning_rate": 1e-06, + "loss": -0.0605, + "num_tokens": 739828023.0, + "reward": 0.6674107313156128, + "reward_std": 0.1803184598684311, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1228 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 954.1886596679688, + "completions/mean_terminated_length": 559.4887084960938, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.485131195335278, + "grad_norm": 0.16445766389369965, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 740345448.0, + "reward": 0.7198660969734192, + "reward_std": 0.1117110475897789, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931530952453613, + "step": 1229 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3309.0, + "completions/mean_length": 967.78466796875, + "completions/mean_terminated_length": 652.6572265625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 11.494460641399417, + "grad_norm": 0.18219566345214844, + "learning_rate": 1e-06, + "loss": -0.0739, + "num_tokens": 740968567.0, + "reward": 0.7008928656578064, + "reward_std": 0.2081684023141861, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1230 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2991.0, + "completions/mean_length": 1003.1038208007812, + "completions/mean_terminated_length": 623.2744140625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.503790087463557, + "grad_norm": 0.16143383085727692, + "learning_rate": 1e-06, + "loss": -0.0856, + "num_tokens": 741559204.0, + "reward": 0.6640625, + "reward_std": 0.16510257124900818, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1231 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3719.0, + "completions/mean_length": 841.4230346679688, + "completions/mean_terminated_length": 612.0083618164062, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 11.513119533527696, + "grad_norm": 0.17548957467079163, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 742162751.0, + "reward": 0.6595982313156128, + "reward_std": 0.15210728347301483, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1232 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3645.0, + "completions/mean_length": 876.7433471679688, + "completions/mean_terminated_length": 582.6577758789062, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 11.522448979591836, + "grad_norm": 0.13381995260715485, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 742723553.0, + "reward": 0.7064732313156128, + "reward_std": 0.10588474571704865, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1233 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3214.0, + "completions/mean_length": 1078.0234375, + "completions/mean_terminated_length": 638.0626831054688, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 11.531778425655977, + "grad_norm": 0.1594962626695633, + "learning_rate": 1e-06, + "loss": -0.084, + "num_tokens": 743313998.0, + "reward": 0.6618303656578064, + "reward_std": 0.16273680329322815, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3345.0, + "completions/mean_length": 1116.313720703125, + "completions/mean_terminated_length": 628.7285766601562, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 11.541107871720117, + "grad_norm": 0.16360148787498474, + "learning_rate": 1e-06, + "loss": -0.0512, + "num_tokens": 743886247.0, + "reward": 0.6674107313156128, + "reward_std": 0.14094392955303192, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3894.0, + "completions/mean_length": 1009.2042846679688, + "completions/mean_terminated_length": 625.775390625, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 11.550437317784256, + "grad_norm": 0.15397955477237701, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 744469222.0, + "reward": 0.7120535969734192, + "reward_std": 0.12561605870723724, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3810.0, + "completions/mean_length": 1049.5179443359375, + "completions/mean_terminated_length": 649.4747314453125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 11.559766763848396, + "grad_norm": 0.1630489081144333, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 745088654.0, + "reward": 0.6540178656578064, + "reward_std": 0.14199630916118622, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1237 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2357.0, + "completions/mean_length": 1058.3671875, + "completions/mean_terminated_length": 624.4196166992188, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 11.569096209912537, + "grad_norm": 0.16160880029201508, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 745666399.0, + "reward": 0.6640625, + "reward_std": 0.13632294535636902, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1238 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3503.0, + "completions/mean_length": 881.24560546875, + "completions/mean_terminated_length": 621.42822265625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 11.578425655976677, + "grad_norm": 0.17230761051177979, + "learning_rate": 1e-06, + "loss": -0.0292, + "num_tokens": 746269011.0, + "reward": 0.7299107313156128, + "reward_std": 0.1424834430217743, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1239 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3984.0, + "completions/mean_length": 1062.9632568359375, + "completions/mean_terminated_length": 580.3453979492188, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.587755102040816, + "grad_norm": 0.14987659454345703, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 746811970.0, + "reward": 0.676339328289032, + "reward_std": 0.11103634536266327, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1240 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3377.0, + "completions/mean_length": 1102.1820068359375, + "completions/mean_terminated_length": 607.7542114257812, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.597084548104956, + "grad_norm": 0.19255205988883972, + "learning_rate": 1e-06, + "loss": -0.0559, + "num_tokens": 747384501.0, + "reward": 0.640625, + "reward_std": 0.14894802868366241, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 1241 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2733.0, + "completions/mean_length": 1017.060302734375, + "completions/mean_terminated_length": 638.9448852539062, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 11.606413994169095, + "grad_norm": 0.15114101767539978, + "learning_rate": 1e-06, + "loss": -0.0439, + "num_tokens": 747996859.0, + "reward": 0.6339285969734192, + "reward_std": 0.1306827962398529, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1242 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3820.0, + "completions/mean_length": 975.1261596679688, + "completions/mean_terminated_length": 652.277099609375, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 11.615743440233237, + "grad_norm": 0.1555086225271225, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 748615068.0, + "reward": 0.6852678656578064, + "reward_std": 0.11896629631519318, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1243 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2302.0, + "completions/mean_length": 916.6920166015625, + "completions/mean_terminated_length": 592.1131591796875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 11.625072886297376, + "grad_norm": 0.16252316534519196, + "learning_rate": 1e-06, + "loss": -0.0292, + "num_tokens": 749184664.0, + "reward": 0.7633928656578064, + "reward_std": 0.11840218305587769, + "rewards/verify_math_reward/mean": 0.7633928656578064, + "rewards/verify_math_reward/std": 0.42523646354675293, + "step": 1244 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3999.0, + "completions/mean_length": 986.1763916015625, + "completions/mean_terminated_length": 617.3458251953125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.634402332361516, + "grad_norm": 0.16937656700611115, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 749766830.0, + "reward": 0.6808035969734192, + "reward_std": 0.1407584697008133, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1245 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3679.0, + "completions/mean_length": 999.9832763671875, + "completions/mean_terminated_length": 649.9987182617188, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.643731778425655, + "grad_norm": 0.1853286325931549, + "learning_rate": 1e-06, + "loss": -0.0448, + "num_tokens": 750378423.0, + "reward": 0.668526828289032, + "reward_std": 0.14022713899612427, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1246 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3245.0, + "completions/mean_length": 978.87841796875, + "completions/mean_terminated_length": 596.0739135742188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 11.653061224489797, + "grad_norm": 0.17841339111328125, + "learning_rate": 1e-06, + "loss": -0.0713, + "num_tokens": 750943274.0, + "reward": 0.7042410969734192, + "reward_std": 0.1328292191028595, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1247 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2524.0, + "completions/mean_length": 1038.84375, + "completions/mean_terminated_length": 588.686279296875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 11.662390670553936, + "grad_norm": 0.1635752022266388, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 751483150.0, + "reward": 0.7053571939468384, + "reward_std": 0.1323067545890808, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1248 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2604.0, + "completions/mean_length": 1083.110595703125, + "completions/mean_terminated_length": 612.7109985351562, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.671720116618076, + "grad_norm": 0.1981825828552246, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 752042585.0, + "reward": 0.6796875596046448, + "reward_std": 0.16946211457252502, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1249 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3924.0, + "completions/mean_length": 1095.9085693359375, + "completions/mean_terminated_length": 614.0285034179688, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 11.681049562682215, + "grad_norm": 0.1804428994655609, + "learning_rate": 1e-06, + "loss": -0.069, + "num_tokens": 752608287.0, + "reward": 0.6428571939468384, + "reward_std": 0.169614776968956, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.47942501306533813, + "step": 1250 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3392.0, + "completions/mean_length": 1035.0257568359375, + "completions/mean_terminated_length": 689.00244140625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 11.690379008746355, + "grad_norm": 0.154035747051239, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 753254926.0, + "reward": 0.6395089626312256, + "reward_std": 0.1656678169965744, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 1251 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3401.0, + "completions/mean_length": 957.4051513671875, + "completions/mean_terminated_length": 589.5399169921875, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 11.699708454810496, + "grad_norm": 0.16448703408241272, + "learning_rate": 1e-06, + "loss": -0.0755, + "num_tokens": 753824177.0, + "reward": 0.7120535969734192, + "reward_std": 0.147146075963974, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1252 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3890.0, + "completions/mean_length": 1064.3582763671875, + "completions/mean_terminated_length": 644.4739379882812, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.709037900874636, + "grad_norm": 0.13975155353546143, + "learning_rate": 1e-06, + "loss": -0.0735, + "num_tokens": 754418306.0, + "reward": 0.6875000596046448, + "reward_std": 0.11606848239898682, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1253 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3444.0, + "completions/mean_length": 1033.3851318359375, + "completions/mean_terminated_length": 586.9168701171875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.718367346938775, + "grad_norm": 0.19333618879318237, + "learning_rate": 1e-06, + "loss": -0.0667, + "num_tokens": 754966867.0, + "reward": 0.6707589626312256, + "reward_std": 0.14692038297653198, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1254 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3599.0, + "completions/mean_length": 1159.64404296875, + "completions/mean_terminated_length": 652.3154296875, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 11.727696793002915, + "grad_norm": 0.17620694637298584, + "learning_rate": 1e-06, + "loss": -0.0839, + "num_tokens": 755560348.0, + "reward": 0.6718750596046448, + "reward_std": 0.17574025690555573, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1255 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3360.0, + "completions/mean_length": 946.05029296875, + "completions/mean_terminated_length": 585.6082153320312, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 11.737026239067056, + "grad_norm": 0.18818381428718567, + "learning_rate": 1e-06, + "loss": -0.0637, + "num_tokens": 756123209.0, + "reward": 0.6964285969734192, + "reward_std": 0.1242605671286583, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1256 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3126.0, + "completions/mean_length": 1015.9754638671875, + "completions/mean_terminated_length": 633.3876953125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 11.746355685131196, + "grad_norm": 0.14721885323524475, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 756718867.0, + "reward": 0.6729910969734192, + "reward_std": 0.11843497306108475, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1257 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1685267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3276.0, + "completions/mean_length": 1170.4598388671875, + "completions/mean_terminated_length": 577.4979858398438, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 11.755685131195335, + "grad_norm": 0.19740772247314453, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 757244039.0, + "reward": 0.621651828289032, + "reward_std": 0.16713841259479523, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 1258 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3864.0, + "completions/mean_length": 1039.0592041015625, + "completions/mean_terminated_length": 615.6708984375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 11.765014577259475, + "grad_norm": 0.19369196891784668, + "learning_rate": 1e-06, + "loss": -0.0659, + "num_tokens": 757820004.0, + "reward": 0.6395089626312256, + "reward_std": 0.15939393639564514, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 1259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2894.0, + "completions/mean_length": 878.6217041015625, + "completions/mean_terminated_length": 580.4255981445312, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.774344023323614, + "grad_norm": 0.15944312512874603, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 758380193.0, + "reward": 0.7243303656578064, + "reward_std": 0.10814176499843597, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1260 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 906.1016235351562, + "completions/mean_terminated_length": 532.2232055664062, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 11.783673469387756, + "grad_norm": 0.2016732543706894, + "learning_rate": 1e-06, + "loss": -0.1057, + "num_tokens": 758890092.0, + "reward": 0.6953125596046448, + "reward_std": 0.12317540496587753, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1261 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2700.0, + "completions/mean_length": 1033.1484375, + "completions/mean_terminated_length": 541.1878051757812, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 11.793002915451895, + "grad_norm": 0.15855073928833008, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 759394193.0, + "reward": 0.7299107313156128, + "reward_std": 0.12546269595623016, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1262 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2559.0, + "completions/mean_length": 1027.313720703125, + "completions/mean_terminated_length": 659.0712280273438, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 11.802332361516035, + "grad_norm": 0.14641793072223663, + "learning_rate": 1e-06, + "loss": -0.0418, + "num_tokens": 760007170.0, + "reward": 0.6618303656578064, + "reward_std": 0.12050792574882507, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1263 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3289.0, + "completions/mean_length": 1008.8047485351562, + "completions/mean_terminated_length": 581.2261352539062, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 11.811661807580174, + "grad_norm": 0.17122164368629456, + "learning_rate": 1e-06, + "loss": -0.0828, + "num_tokens": 760551675.0, + "reward": 0.7366071939468384, + "reward_std": 0.13139888644218445, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3143.0, + "completions/mean_length": 1137.3660888671875, + "completions/mean_terminated_length": 607.9263305664062, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 11.820991253644316, + "grad_norm": 0.2073400616645813, + "learning_rate": 1e-06, + "loss": -0.0876, + "num_tokens": 761102363.0, + "reward": 0.6517857313156128, + "reward_std": 0.1649945080280304, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1265 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3564.0, + "completions/mean_length": 1010.6194458007812, + "completions/mean_terminated_length": 657.56591796875, + "completions/min_length": 212.0, + "completions/min_terminated_length": 212.0, + "epoch": 11.830320699708455, + "grad_norm": 0.16307416558265686, + "learning_rate": 1e-06, + "loss": -0.0667, + "num_tokens": 761714086.0, + "reward": 0.6517857313156128, + "reward_std": 0.16969034075737, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1266 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3386.0, + "completions/mean_length": 1013.3605346679688, + "completions/mean_terminated_length": 590.8667602539062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 11.839650145772595, + "grad_norm": 0.17521661520004272, + "learning_rate": 1e-06, + "loss": -0.0439, + "num_tokens": 762272081.0, + "reward": 0.6741071939468384, + "reward_std": 0.10295554995536804, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.46896928548812866, + "step": 1267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2764.0, + "completions/mean_length": 996.5480346679688, + "completions/mean_terminated_length": 593.9710083007812, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.848979591836734, + "grad_norm": 0.16372117400169373, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 762846148.0, + "reward": 0.6540178656578064, + "reward_std": 0.14388129115104675, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1268 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3723.0, + "completions/mean_length": 862.4721069335938, + "completions/mean_terminated_length": 558.4652099609375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.858309037900874, + "grad_norm": 0.1823384314775467, + "learning_rate": 1e-06, + "loss": -0.035, + "num_tokens": 763391499.0, + "reward": 0.6741071939468384, + "reward_std": 0.1446651816368103, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1269 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2362.0, + "completions/mean_length": 938.33154296875, + "completions/mean_terminated_length": 590.0880126953125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 11.867638483965015, + "grad_norm": 0.18177086114883423, + "learning_rate": 1e-06, + "loss": -0.0448, + "num_tokens": 763959236.0, + "reward": 0.6640625, + "reward_std": 0.15526078641414642, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1270 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2271.0, + "completions/mean_length": 895.81591796875, + "completions/mean_terminated_length": 547.2809448242188, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.876967930029155, + "grad_norm": 0.15784738957881927, + "learning_rate": 1e-06, + "loss": -0.0487, + "num_tokens": 764497839.0, + "reward": 0.7276785969734192, + "reward_std": 0.1324947476387024, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 975.2366333007812, + "completions/mean_terminated_length": 587.5884399414062, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 11.886297376093294, + "grad_norm": 0.16080334782600403, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 765061243.0, + "reward": 0.6819196939468384, + "reward_std": 0.15747319161891937, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1272 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4092.0, + "completions/mean_length": 1047.294677734375, + "completions/mean_terminated_length": 616.203857421875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.895626822157434, + "grad_norm": 0.1867516040802002, + "learning_rate": 1e-06, + "loss": -0.0609, + "num_tokens": 765633643.0, + "reward": 0.652901828289032, + "reward_std": 0.1632685512304306, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 1273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2565.0, + "completions/mean_length": 992.4766235351562, + "completions/mean_terminated_length": 620.0537109375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.904956268221575, + "grad_norm": 0.164963960647583, + "learning_rate": 1e-06, + "loss": -0.0633, + "num_tokens": 766219014.0, + "reward": 0.65625, + "reward_std": 0.13211244344711304, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1274 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2826.0, + "completions/mean_length": 1123.415283203125, + "completions/mean_terminated_length": 659.3084106445312, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 11.914285714285715, + "grad_norm": 0.1873733103275299, + "learning_rate": 1e-06, + "loss": -0.0489, + "num_tokens": 766821154.0, + "reward": 0.7064732313156128, + "reward_std": 0.1466914415359497, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1275 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3558.0, + "completions/mean_length": 996.5592041015625, + "completions/mean_terminated_length": 585.1289672851562, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 11.923615160349854, + "grad_norm": 0.18284200131893158, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 767375671.0, + "reward": 0.6595982313156128, + "reward_std": 0.1528950184583664, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1276 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 976.0535888671875, + "completions/mean_terminated_length": 588.5068969726562, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 11.932944606413994, + "grad_norm": 0.18158194422721863, + "learning_rate": 1e-06, + "loss": -0.0634, + "num_tokens": 767935183.0, + "reward": 0.7276785969734192, + "reward_std": 0.13203828036785126, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1277 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2264.0, + "completions/mean_length": 1147.37060546875, + "completions/mean_terminated_length": 624.2890625, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 11.942274052478133, + "grad_norm": 0.21804052591323853, + "learning_rate": 1e-06, + "loss": -0.0713, + "num_tokens": 768494283.0, + "reward": 0.6350446939468384, + "reward_std": 0.15808303654193878, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 1278 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 983.9230346679688, + "completions/mean_terminated_length": 610.4737548828125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.951603498542275, + "grad_norm": 0.16081461310386658, + "learning_rate": 1e-06, + "loss": -0.06, + "num_tokens": 769064278.0, + "reward": 0.6729910969734192, + "reward_std": 0.12823963165283203, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1279 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2395.0, + "completions/mean_length": 889.0279541015625, + "completions/mean_terminated_length": 548.5346069335938, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 11.960932944606414, + "grad_norm": 0.1919887661933899, + "learning_rate": 1e-06, + "loss": -0.0511, + "num_tokens": 769602375.0, + "reward": 0.7008928656578064, + "reward_std": 0.12178602069616318, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1280 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 966.2176513671875, + "completions/mean_terminated_length": 577.4491577148438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 11.970262390670554, + "grad_norm": 0.16791851818561554, + "learning_rate": 1e-06, + "loss": -0.0665, + "num_tokens": 770154770.0, + "reward": 0.7209821939468384, + "reward_std": 0.13298076391220093, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1281 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3227.0, + "completions/mean_length": 1076.696533203125, + "completions/mean_terminated_length": 609.7938232421875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.979591836734693, + "grad_norm": 0.1630984991788864, + "learning_rate": 1e-06, + "loss": -0.0866, + "num_tokens": 770721498.0, + "reward": 0.6752232313156128, + "reward_std": 0.16164980828762054, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1282 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3505.0, + "completions/mean_length": 974.7667846679688, + "completions/mean_terminated_length": 656.1168212890625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 11.988921282798835, + "grad_norm": 0.18171954154968262, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 771348113.0, + "reward": 0.691964328289032, + "reward_std": 0.1393709033727646, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.12215909090909094, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2173.0, + "completions/mean_length": 1008.0426635742188, + "completions/mean_terminated_length": 578.326904296875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.998250728862974, + "grad_norm": 0.16108766198158264, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 771902179.0, + "reward": 0.7053571939468384, + "reward_std": 0.12459687143564224, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1284 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2262.0, + "completions/mean_length": 919.7154541015625, + "completions/mean_terminated_length": 565.0433959960938, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.00932944606414, + "grad_norm": 0.17704452574253082, + "learning_rate": 1e-06, + "loss": -0.0088, + "num_tokens": 772445100.0, + "reward": 0.6651785969734192, + "reward_std": 0.1348900943994522, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 1285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3168.0, + "completions/mean_length": 978.2734985351562, + "completions/mean_terminated_length": 555.4638671875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 12.018658892128279, + "grad_norm": 0.184604674577713, + "learning_rate": 1e-06, + "loss": -0.0847, + "num_tokens": 772971441.0, + "reward": 0.6953125596046448, + "reward_std": 0.1429726928472519, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1286 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 891.95654296875, + "completions/mean_terminated_length": 556.1442260742188, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 12.02798833819242, + "grad_norm": 0.1772415190935135, + "learning_rate": 1e-06, + "loss": -0.0478, + "num_tokens": 773513842.0, + "reward": 0.7421875596046448, + "reward_std": 0.13801473379135132, + "rewards/verify_math_reward/mean": 0.7421875, + "rewards/verify_math_reward/std": 0.43767455220222473, + "step": 1287 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3328.0, + "completions/mean_length": 856.6350708007812, + "completions/mean_terminated_length": 599.0469970703125, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 12.03731778425656, + "grad_norm": 0.18522457778453827, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 774099371.0, + "reward": 0.7354910969734192, + "reward_std": 0.15349596738815308, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 1288 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2372.0, + "completions/mean_length": 904.7210083007812, + "completions/mean_terminated_length": 557.1559448242188, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 12.0466472303207, + "grad_norm": 0.16180400550365448, + "learning_rate": 1e-06, + "loss": -0.0887, + "num_tokens": 774627489.0, + "reward": 0.7633928656578064, + "reward_std": 0.12651507556438446, + "rewards/verify_math_reward/mean": 0.7633928656578064, + "rewards/verify_math_reward/std": 0.42523646354675293, + "step": 1289 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3647.0, + "completions/mean_length": 997.5391235351562, + "completions/mean_terminated_length": 595.0908203125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 12.055976676384839, + "grad_norm": 0.17720985412597656, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 775185988.0, + "reward": 0.6863839626312256, + "reward_std": 0.14808152616024017, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1290 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3709.0, + "completions/mean_length": 897.6272583007812, + "completions/mean_terminated_length": 571.10205078125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 12.06530612244898, + "grad_norm": 0.1768137663602829, + "learning_rate": 1e-06, + "loss": -0.0339, + "num_tokens": 775740902.0, + "reward": 0.6707589626312256, + "reward_std": 0.12985724210739136, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1291 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3453.0, + "completions/mean_length": 1031.993408203125, + "completions/mean_terminated_length": 585.322265625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.07463556851312, + "grad_norm": 0.20356561243534088, + "learning_rate": 1e-06, + "loss": -0.0697, + "num_tokens": 776293160.0, + "reward": 0.6696428656578064, + "reward_std": 0.15687799453735352, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1292 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2554.0, + "completions/mean_length": 989.4386596679688, + "completions/mean_terminated_length": 625.3279418945312, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 12.08396501457726, + "grad_norm": 0.1500304639339447, + "learning_rate": 1e-06, + "loss": -0.0338, + "num_tokens": 776884929.0, + "reward": 0.6897321939468384, + "reward_std": 0.12241654098033905, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1293 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2184.0, + "completions/mean_length": 1029.31591796875, + "completions/mean_terminated_length": 600.1361083984375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 12.093294460641399, + "grad_norm": 0.17582301795482635, + "learning_rate": 1e-06, + "loss": -0.0768, + "num_tokens": 777451084.0, + "reward": 0.6729910969734192, + "reward_std": 0.16303987801074982, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3556.0, + "completions/mean_length": 954.3984985351562, + "completions/mean_terminated_length": 616.550048828125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.102623906705539, + "grad_norm": 0.14249971508979797, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 778042577.0, + "reward": 0.6886160969734192, + "reward_std": 0.1263660490512848, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4017.0, + "completions/mean_length": 907.0100708007812, + "completions/mean_terminated_length": 585.7604370117188, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 12.11195335276968, + "grad_norm": 0.15846046805381775, + "learning_rate": 1e-06, + "loss": -0.0439, + "num_tokens": 778602466.0, + "reward": 0.6752232313156128, + "reward_std": 0.11892351508140564, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1296 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3237.0, + "completions/mean_length": 917.5513916015625, + "completions/mean_terminated_length": 614.471923828125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 12.12128279883382, + "grad_norm": 0.16542382538318634, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 779191208.0, + "reward": 0.715401828289032, + "reward_std": 0.13542252779006958, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1297 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4080.0, + "completions/mean_length": 1065.638427734375, + "completions/mean_terminated_length": 641.5419921875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 12.130612244897959, + "grad_norm": 0.16897358000278473, + "learning_rate": 1e-06, + "loss": -0.0736, + "num_tokens": 779788996.0, + "reward": 0.6462053656578064, + "reward_std": 0.14940379559993744, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1298 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3499.0, + "completions/mean_length": 891.5078735351562, + "completions/mean_terminated_length": 594.506103515625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.139941690962099, + "grad_norm": 0.19319415092468262, + "learning_rate": 1e-06, + "loss": -0.0636, + "num_tokens": 780361811.0, + "reward": 0.731026828289032, + "reward_std": 0.1532338261604309, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1299 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3070.0, + "completions/mean_length": 1117.298095703125, + "completions/mean_terminated_length": 652.2361450195312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 12.14927113702624, + "grad_norm": 0.16154277324676514, + "learning_rate": 1e-06, + "loss": -0.0864, + "num_tokens": 780967286.0, + "reward": 0.6718750596046448, + "reward_std": 0.15308228135108948, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1300 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3265.0, + "completions/mean_length": 895.7053833007812, + "completions/mean_terminated_length": 560.2860717773438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 12.15860058309038, + "grad_norm": 0.15245497226715088, + "learning_rate": 1e-06, + "loss": -0.0545, + "num_tokens": 781506206.0, + "reward": 0.7332589626312256, + "reward_std": 0.11896559596061707, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 1301 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 960.9029541015625, + "completions/mean_terminated_length": 602.1604614257812, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 12.167930029154519, + "grad_norm": 0.16760271787643433, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 782080783.0, + "reward": 0.6941964626312256, + "reward_std": 0.1509019434452057, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3541.0, + "completions/mean_length": 989.880615234375, + "completions/mean_terminated_length": 586.4375610351562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 12.177259475218658, + "grad_norm": 0.1760021299123764, + "learning_rate": 1e-06, + "loss": -0.0259, + "num_tokens": 782628540.0, + "reward": 0.7254464626312256, + "reward_std": 0.12456367164850235, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1303 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3533.0, + "completions/mean_length": 891.9297485351562, + "completions/mean_terminated_length": 551.7444458007812, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.186588921282798, + "grad_norm": 0.15361468493938446, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 783155133.0, + "reward": 0.7366071939468384, + "reward_std": 0.0924757644534111, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1304 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 896.4531860351562, + "completions/mean_terminated_length": 629.5018310546875, + "completions/min_length": 195.0, + "completions/min_terminated_length": 195.0, + "epoch": 12.19591836734694, + "grad_norm": 0.15269605815410614, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 783766427.0, + "reward": 0.691964328289032, + "reward_std": 0.13298188149929047, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1305 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2687.0, + "completions/mean_length": 910.26904296875, + "completions/mean_terminated_length": 572.0308837890625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 12.205247813411079, + "grad_norm": 0.16092495620250702, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 784316140.0, + "reward": 0.707589328289032, + "reward_std": 0.1300102025270462, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1306 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2597.0, + "completions/mean_length": 1090.2913818359375, + "completions/mean_terminated_length": 593.8998413085938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.214577259475218, + "grad_norm": 0.14995354413986206, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 784861497.0, + "reward": 0.691964328289032, + "reward_std": 0.10641466826200485, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1307 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 1098.84716796875, + "completions/mean_terminated_length": 630.904541015625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 12.223906705539358, + "grad_norm": 0.16611260175704956, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 785436336.0, + "reward": 0.660714328289032, + "reward_std": 0.13203828036785126, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1308 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 1118.825927734375, + "completions/mean_terminated_length": 613.5613403320312, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 12.2332361516035, + "grad_norm": 0.1684187799692154, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 785993132.0, + "reward": 0.6339285969734192, + "reward_std": 0.12309802323579788, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1309 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2849.0, + "completions/mean_length": 929.34716796875, + "completions/mean_terminated_length": 593.1345825195312, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 12.242565597667639, + "grad_norm": 0.14068344235420227, + "learning_rate": 1e-06, + "loss": -0.0338, + "num_tokens": 786560843.0, + "reward": 0.7209821939468384, + "reward_std": 0.10964765399694443, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1310 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3152.0, + "completions/mean_length": 851.7299194335938, + "completions/mean_terminated_length": 568.25, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.251895043731778, + "grad_norm": 0.18744957447052002, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 787109425.0, + "reward": 0.7220982313156128, + "reward_std": 0.15172749757766724, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 1311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2660.0, + "completions/mean_length": 1037.90185546875, + "completions/mean_terminated_length": 618.7715454101562, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 12.261224489795918, + "grad_norm": 0.1908305287361145, + "learning_rate": 1e-06, + "loss": -0.0619, + "num_tokens": 787689361.0, + "reward": 0.6584821939468384, + "reward_std": 0.16453665494918823, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1312 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3699.0, + "completions/mean_length": 981.357177734375, + "completions/mean_terminated_length": 663.3800659179688, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 12.270553935860057, + "grad_norm": 0.16000255942344666, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 788314313.0, + "reward": 0.6595982313156128, + "reward_std": 0.14620430767536163, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1313 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 911.138427734375, + "completions/mean_terminated_length": 590.3046875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.279883381924199, + "grad_norm": 0.179154172539711, + "learning_rate": 1e-06, + "loss": -0.0707, + "num_tokens": 788885853.0, + "reward": 0.7109375596046448, + "reward_std": 0.15608564019203186, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1314 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3775.0, + "completions/mean_length": 871.1897583007812, + "completions/mean_terminated_length": 546.3316650390625, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 12.289212827988338, + "grad_norm": 0.18365663290023804, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 789410783.0, + "reward": 0.7254464626312256, + "reward_std": 0.14443765580654144, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1315 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2976.0, + "completions/mean_length": 991.2656860351562, + "completions/mean_terminated_length": 614.345458984375, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.298542274052478, + "grad_norm": 0.16882410645484924, + "learning_rate": 1e-06, + "loss": -0.0368, + "num_tokens": 789994317.0, + "reward": 0.6785714626312256, + "reward_std": 0.12272076308727264, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1316 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2904.0, + "completions/mean_length": 903.130615234375, + "completions/mean_terminated_length": 619.9234619140625, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 12.307871720116617, + "grad_norm": 0.16580593585968018, + "learning_rate": 1e-06, + "loss": -0.0553, + "num_tokens": 790591658.0, + "reward": 0.7098214626312256, + "reward_std": 0.12677721679210663, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1317 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3063.0, + "completions/mean_length": 1142.607177734375, + "completions/mean_terminated_length": 641.3786010742188, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 12.317201166180759, + "grad_norm": 0.14159466326236725, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 791182506.0, + "reward": 0.6439732313156128, + "reward_std": 0.12726393342018127, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3872.0, + "completions/mean_length": 853.8058471679688, + "completions/mean_terminated_length": 574.7806396484375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.326530612244898, + "grad_norm": 0.1810448169708252, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 791739988.0, + "reward": 0.7198660969734192, + "reward_std": 0.13241735100746155, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931530952453613, + "step": 1319 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3693.0, + "completions/mean_length": 1094.3348388671875, + "completions/mean_terminated_length": 647.933349609375, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 12.335860058309038, + "grad_norm": 0.1589130163192749, + "learning_rate": 1e-06, + "loss": -0.0713, + "num_tokens": 792331776.0, + "reward": 0.6540178656578064, + "reward_std": 0.12369534373283386, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1320 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 1160.415283203125, + "completions/mean_terminated_length": 657.7202758789062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 12.345189504373177, + "grad_norm": 0.18702104687690735, + "learning_rate": 1e-06, + "loss": -0.038, + "num_tokens": 792919556.0, + "reward": 0.6350446939468384, + "reward_std": 0.1349007785320282, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 1321 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2964.0, + "completions/mean_length": 972.5156860351562, + "completions/mean_terminated_length": 584.5294799804688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 12.354518950437317, + "grad_norm": 0.15318487584590912, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 793472234.0, + "reward": 0.6986607313156128, + "reward_std": 0.10626270622015, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1322 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 1973.0, + "completions/mean_length": 1044.76904296875, + "completions/mean_terminated_length": 577.4633178710938, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 12.363848396501458, + "grad_norm": 0.14388787746429443, + "learning_rate": 1e-06, + "loss": -0.0625, + "num_tokens": 794008203.0, + "reward": 0.6886160969734192, + "reward_std": 0.1033681184053421, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3668.0, + "completions/mean_length": 1075.6015625, + "completions/mean_terminated_length": 661.6383056640625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 12.373177842565598, + "grad_norm": 0.17792291939258575, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 794625670.0, + "reward": 0.606026828289032, + "reward_std": 0.14925295114517212, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 1324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3186.0, + "completions/mean_length": 1120.458740234375, + "completions/mean_terminated_length": 669.155517578125, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 12.382507288629737, + "grad_norm": 0.15607257187366486, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 795232065.0, + "reward": 0.640625, + "reward_std": 0.12433473765850067, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 1325 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 1240.5045166015625, + "completions/mean_terminated_length": 711.708984375, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 12.391836734693877, + "grad_norm": 0.18093542754650116, + "learning_rate": 1e-06, + "loss": -0.0685, + "num_tokens": 795863077.0, + "reward": 0.6004464626312256, + "reward_std": 0.1535683125257492, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 1326 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3909.0, + "completions/mean_length": 998.341552734375, + "completions/mean_terminated_length": 609.1884155273438, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 12.401166180758018, + "grad_norm": 0.14908741414546967, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 796426615.0, + "reward": 0.6908482313156128, + "reward_std": 0.12001685798168182, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1327 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2903.0, + "completions/mean_length": 932.1317138671875, + "completions/mean_terminated_length": 613.4127807617188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.410495626822158, + "grad_norm": 0.1725604236125946, + "learning_rate": 1e-06, + "loss": -0.0639, + "num_tokens": 797015125.0, + "reward": 0.723214328289032, + "reward_std": 0.15842002630233765, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2154.0, + "completions/mean_length": 892.33154296875, + "completions/mean_terminated_length": 569.6032104492188, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 12.419825072886297, + "grad_norm": 0.1555093377828598, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 797568966.0, + "reward": 0.7120535969734192, + "reward_std": 0.11144932359457016, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1329 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3333.0, + "completions/mean_length": 1057.67529296875, + "completions/mean_terminated_length": 610.2906494140625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.429154518950437, + "grad_norm": 0.17622613906860352, + "learning_rate": 1e-06, + "loss": -0.091, + "num_tokens": 798146707.0, + "reward": 0.6796875596046448, + "reward_std": 0.14943701028823853, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1330 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 1018.3404541015625, + "completions/mean_terminated_length": 587.6246948242188, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 12.438483965014576, + "grad_norm": 0.1874133199453354, + "learning_rate": 1e-06, + "loss": -0.0763, + "num_tokens": 798696132.0, + "reward": 0.7220982313156128, + "reward_std": 0.16345147788524628, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 1331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3343.0, + "completions/mean_length": 832.0703735351562, + "completions/mean_terminated_length": 559.747314453125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 12.447813411078718, + "grad_norm": 0.17387337982654572, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 799245067.0, + "reward": 0.7667410969734192, + "reward_std": 0.1385018676519394, + "rewards/verify_math_reward/mean": 0.7667410969734192, + "rewards/verify_math_reward/std": 0.4231418967247009, + "step": 1332 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3726.0, + "completions/mean_length": 1023.9788208007812, + "completions/mean_terminated_length": 607.3675537109375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 12.457142857142857, + "grad_norm": 0.16259723901748657, + "learning_rate": 1e-06, + "loss": -0.0339, + "num_tokens": 799825280.0, + "reward": 0.676339328289032, + "reward_std": 0.12610207498073578, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1333 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2765.0, + "completions/mean_length": 1042.396240234375, + "completions/mean_terminated_length": 597.24169921875, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 12.466472303206997, + "grad_norm": 0.19924825429916382, + "learning_rate": 1e-06, + "loss": -0.0663, + "num_tokens": 800384387.0, + "reward": 0.6897321939468384, + "reward_std": 0.15571653842926025, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2879.0, + "completions/mean_length": 797.6607666015625, + "completions/mean_terminated_length": 569.3746948242188, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 12.475801749271136, + "grad_norm": 0.1662532091140747, + "learning_rate": 1e-06, + "loss": -0.0402, + "num_tokens": 800942139.0, + "reward": 0.7455357313156128, + "reward_std": 0.12035498023033142, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2092.0, + "completions/mean_length": 927.92529296875, + "completions/mean_terminated_length": 591.561767578125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.485131195335278, + "grad_norm": 0.16006889939308167, + "learning_rate": 1e-06, + "loss": -0.0885, + "num_tokens": 801516072.0, + "reward": 0.7220982313156128, + "reward_std": 0.14646530151367188, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1336 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3846.0, + "completions/mean_length": 892.2288208007812, + "completions/mean_terminated_length": 595.2938842773438, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 12.494460641399417, + "grad_norm": 0.1678542047739029, + "learning_rate": 1e-06, + "loss": -0.0353, + "num_tokens": 802096565.0, + "reward": 0.7209821939468384, + "reward_std": 0.12084492295980453, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1337 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3214.0, + "completions/mean_length": 1028.69091796875, + "completions/mean_terminated_length": 639.0075073242188, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 12.503790087463557, + "grad_norm": 0.15641376376152039, + "learning_rate": 1e-06, + "loss": -0.0604, + "num_tokens": 802694352.0, + "reward": 0.668526828289032, + "reward_std": 0.11039625108242035, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 1124.415283203125, + "completions/mean_terminated_length": 651.57568359375, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 12.513119533527696, + "grad_norm": 0.15743494033813477, + "learning_rate": 1e-06, + "loss": -0.0579, + "num_tokens": 803298988.0, + "reward": 0.6517857313156128, + "reward_std": 0.1293707937002182, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1339 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3953.0, + "completions/mean_length": 1020.8058471679688, + "completions/mean_terminated_length": 664.6500244140625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.522448979591836, + "grad_norm": 0.18502841889858246, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 803922670.0, + "reward": 0.7064732313156128, + "reward_std": 0.1542416214942932, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1340 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2672.0, + "completions/mean_length": 978.1317138671875, + "completions/mean_terminated_length": 621.3606567382812, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 12.531778425655977, + "grad_norm": 0.1659466028213501, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 804505540.0, + "reward": 0.6975446939468384, + "reward_std": 0.14962133765220642, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1341 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3578.0, + "completions/mean_length": 997.57373046875, + "completions/mean_terminated_length": 630.0948486328125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 12.541107871720117, + "grad_norm": 0.15725573897361755, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 805098878.0, + "reward": 0.6383928656578064, + "reward_std": 0.13106118142604828, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3441.0, + "completions/mean_length": 1033.943115234375, + "completions/mean_terminated_length": 631.8547973632812, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 12.550437317784256, + "grad_norm": 0.2267020046710968, + "learning_rate": 1e-06, + "loss": -0.0922, + "num_tokens": 805681755.0, + "reward": 0.6964285969734192, + "reward_std": 0.19861942529678345, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1343 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2230.0, + "completions/mean_length": 845.5357666015625, + "completions/mean_terminated_length": 578.58935546875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.559766763848396, + "grad_norm": 0.17514190077781677, + "learning_rate": 1e-06, + "loss": -0.0253, + "num_tokens": 806241291.0, + "reward": 0.7455357313156128, + "reward_std": 0.10716535896062851, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358028769493103, + "step": 1344 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3598.0, + "completions/mean_length": 1042.587158203125, + "completions/mean_terminated_length": 624.0989990234375, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 12.569096209912537, + "grad_norm": 0.1883087158203125, + "learning_rate": 1e-06, + "loss": -0.056, + "num_tokens": 806826489.0, + "reward": 0.6495535969734192, + "reward_std": 0.16037283837795258, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1345 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2810.0, + "completions/mean_length": 804.0736694335938, + "completions/mean_terminated_length": 588.7871704101562, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 12.578425655976677, + "grad_norm": 0.13668419420719147, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 807415971.0, + "reward": 0.7343750596046448, + "reward_std": 0.10547357052564621, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 1346 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3737.0, + "completions/mean_length": 950.7467041015625, + "completions/mean_terminated_length": 586.4769287109375, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 12.587755102040816, + "grad_norm": 0.16991546750068665, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 807977240.0, + "reward": 0.715401828289032, + "reward_std": 0.13632294535636902, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1347 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3794.0, + "completions/mean_length": 910.466552734375, + "completions/mean_terminated_length": 576.5942993164062, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.597084548104956, + "grad_norm": 0.16234464943408966, + "learning_rate": 1e-06, + "loss": -0.0267, + "num_tokens": 808532818.0, + "reward": 0.7299107313156128, + "reward_std": 0.1177295669913292, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1348 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3961.0, + "completions/mean_length": 1011.341552734375, + "completions/mean_terminated_length": 645.49560546875, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 12.606413994169095, + "grad_norm": 0.1682359129190445, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 809147988.0, + "reward": 0.6383928656578064, + "reward_std": 0.14597606658935547, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 1349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3683.0, + "completions/mean_length": 984.0145263671875, + "completions/mean_terminated_length": 548.4949340820312, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 12.615743440233237, + "grad_norm": 0.21174253523349762, + "learning_rate": 1e-06, + "loss": -0.0811, + "num_tokens": 809672473.0, + "reward": 0.6863839626312256, + "reward_std": 0.1639835089445114, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1350 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2775.0, + "completions/mean_length": 1005.33154296875, + "completions/mean_terminated_length": 617.0565185546875, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 12.625072886297376, + "grad_norm": 0.19106462597846985, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 810253146.0, + "reward": 0.6506696939468384, + "reward_std": 0.16112029552459717, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1351 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3165.0, + "completions/mean_length": 1008.5234985351562, + "completions/mean_terminated_length": 603.0972290039062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 12.634402332361516, + "grad_norm": 0.16046831011772156, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 810818711.0, + "reward": 0.6573660969734192, + "reward_std": 0.11452937871217728, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1352 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2979.0, + "completions/mean_length": 1000.4475708007812, + "completions/mean_terminated_length": 589.5335083007812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.643731778425655, + "grad_norm": 0.17968566715717316, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 811379064.0, + "reward": 0.7142857313156128, + "reward_std": 0.13891556859016418, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1353 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2172.0, + "completions/mean_length": 849.7756958007812, + "completions/mean_terminated_length": 527.144775390625, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 12.653061224489797, + "grad_norm": 0.17325682938098907, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 811886023.0, + "reward": 0.7433035969734192, + "reward_std": 0.11167574673891068, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705445528030396, + "step": 1354 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3369.0, + "completions/mean_length": 979.11279296875, + "completions/mean_terminated_length": 635.3668212890625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 12.662390670553936, + "grad_norm": 0.1870400756597519, + "learning_rate": 1e-06, + "loss": -0.0528, + "num_tokens": 812491924.0, + "reward": 0.6808035969734192, + "reward_std": 0.16149936616420746, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1355 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 2982.0, + "completions/mean_length": 1071.8125, + "completions/mean_terminated_length": 639.7857055664062, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 12.671720116618076, + "grad_norm": 0.17932631075382233, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 813095188.0, + "reward": 0.6540178656578064, + "reward_std": 0.14564228057861328, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1356 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3583.0, + "completions/mean_length": 1052.454345703125, + "completions/mean_terminated_length": 657.1387329101562, + "completions/min_length": 194.0, + "completions/min_terminated_length": 194.0, + "epoch": 12.681049562682215, + "grad_norm": 0.15587233006954193, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 813703083.0, + "reward": 0.676339328289032, + "reward_std": 0.1476346254348755, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 1357 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3403.0, + "completions/mean_length": 912.771240234375, + "completions/mean_terminated_length": 566.0828857421875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 12.690379008746355, + "grad_norm": 0.1470702588558197, + "learning_rate": 1e-06, + "loss": -0.0479, + "num_tokens": 814246710.0, + "reward": 0.7433035969734192, + "reward_std": 0.10442119091749191, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2786.0, + "completions/mean_length": 900.04248046875, + "completions/mean_terminated_length": 608.0852661132812, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 12.699708454810496, + "grad_norm": 0.1994343101978302, + "learning_rate": 1e-06, + "loss": -0.0452, + "num_tokens": 814836004.0, + "reward": 0.7477678656578064, + "reward_std": 0.14654381573200226, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.434536337852478, + "step": 1359 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2552.0, + "completions/mean_length": 1048.6317138671875, + "completions/mean_terminated_length": 635.36376953125, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 12.709037900874636, + "grad_norm": 0.19583085179328918, + "learning_rate": 1e-06, + "loss": -0.0498, + "num_tokens": 815420586.0, + "reward": 0.6852678656578064, + "reward_std": 0.17641426622867584, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1360 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2716.0, + "completions/mean_length": 908.3527221679688, + "completions/mean_terminated_length": 604.3961181640625, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 12.718367346938775, + "grad_norm": 0.1439504623413086, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 815996278.0, + "reward": 0.7165178656578064, + "reward_std": 0.11283829063177109, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1361 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3505.0, + "completions/mean_length": 1073.703125, + "completions/mean_terminated_length": 723.6737060546875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 12.727696793002915, + "grad_norm": 0.16441071033477783, + "learning_rate": 1e-06, + "loss": -0.0397, + "num_tokens": 816668652.0, + "reward": 0.5915178656578064, + "reward_std": 0.15462207794189453, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 1362 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 1008.8058471679688, + "completions/mean_terminated_length": 634.0150146484375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 12.737026239067056, + "grad_norm": 0.17051641643047333, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 817268550.0, + "reward": 0.6662946939468384, + "reward_std": 0.15022866427898407, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3320.0, + "completions/mean_length": 961.864990234375, + "completions/mean_terminated_length": 646.1412353515625, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 12.746355685131196, + "grad_norm": 0.17612028121948242, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 817895781.0, + "reward": 0.6662946939468384, + "reward_std": 0.15364569425582886, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1364 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 4067.0, + "completions/mean_length": 857.0435791015625, + "completions/mean_terminated_length": 582.5557250976562, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.755685131195335, + "grad_norm": 0.18695229291915894, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 818457052.0, + "reward": 0.7399553656578064, + "reward_std": 0.14011836051940918, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1365 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 1015.6585083007812, + "completions/mean_terminated_length": 633.0313720703125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 12.765014577259475, + "grad_norm": 0.15656854212284088, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 819052034.0, + "reward": 0.6808035969734192, + "reward_std": 0.11953011900186539, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1366 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2720.0, + "completions/mean_length": 1076.602783203125, + "completions/mean_terminated_length": 591.6217651367188, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.774344023323614, + "grad_norm": 0.18124838173389435, + "learning_rate": 1e-06, + "loss": -0.0628, + "num_tokens": 819594854.0, + "reward": 0.6696428656578064, + "reward_std": 0.13080155849456787, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1367 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2631.0, + "completions/mean_length": 927.62841796875, + "completions/mean_terminated_length": 599.8657836914062, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.783673469387756, + "grad_norm": 0.16686727106571198, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 820168089.0, + "reward": 0.7165178656578064, + "reward_std": 0.12189479172229767, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1368 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2667.0, + "completions/mean_length": 907.0859985351562, + "completions/mean_terminated_length": 624.2296752929688, + "completions/min_length": 221.0, + "completions/min_terminated_length": 221.0, + "epoch": 12.793002915451895, + "grad_norm": 0.18051575124263763, + "learning_rate": 1e-06, + "loss": -0.0447, + "num_tokens": 820783638.0, + "reward": 0.7399553656578064, + "reward_std": 0.14902400970458984, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890368938446045, + "step": 1369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3744.0, + "completions/mean_length": 926.1819458007812, + "completions/mean_terminated_length": 623.9254760742188, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.802332361516035, + "grad_norm": 0.1778830587863922, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 821384769.0, + "reward": 0.7109375596046448, + "reward_std": 0.1708942949771881, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1370 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4093.0, + "completions/mean_length": 1093.9029541015625, + "completions/mean_terminated_length": 647.4371948242188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 12.811661807580174, + "grad_norm": 0.16026508808135986, + "learning_rate": 1e-06, + "loss": -0.0772, + "num_tokens": 821992778.0, + "reward": 0.6417410969734192, + "reward_std": 0.14282074570655823, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1371 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4038.0, + "completions/mean_length": 964.0938110351562, + "completions/mean_terminated_length": 652.8245239257812, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.820991253644316, + "grad_norm": 0.16004188358783722, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 822618230.0, + "reward": 0.6439732313156128, + "reward_std": 0.11914853006601334, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1372 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3299.0, + "completions/mean_length": 980.5982666015625, + "completions/mean_terminated_length": 632.7245483398438, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 12.830320699708455, + "grad_norm": 0.16314876079559326, + "learning_rate": 1e-06, + "loss": -0.0118, + "num_tokens": 823224918.0, + "reward": 0.691964328289032, + "reward_std": 0.11636271327733994, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3817.0, + "completions/mean_length": 929.4710083007812, + "completions/mean_terminated_length": 614.7607421875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.839650145772595, + "grad_norm": 0.14862915873527527, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 823812964.0, + "reward": 0.7243303656578064, + "reward_std": 0.11930190026760101, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1374 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4093.0, + "completions/mean_length": 934.1707763671875, + "completions/mean_terminated_length": 559.1723022460938, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 12.848979591836734, + "grad_norm": 0.17180867493152618, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 824352845.0, + "reward": 0.7455357313156128, + "reward_std": 0.12839370965957642, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1375 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3816.0, + "completions/mean_length": 970.70654296875, + "completions/mean_terminated_length": 647.4002075195312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.858309037900874, + "grad_norm": 0.18521800637245178, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 824966574.0, + "reward": 0.6796875596046448, + "reward_std": 0.16837765276432037, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1376 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3370.0, + "completions/mean_length": 1005.3828735351562, + "completions/mean_terminated_length": 677.2432250976562, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.867638483965015, + "grad_norm": 0.1676863133907318, + "learning_rate": 1e-06, + "loss": -0.0278, + "num_tokens": 825600125.0, + "reward": 0.7120535969734192, + "reward_std": 0.14038120210170746, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1377 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 1065.5614013671875, + "completions/mean_terminated_length": 697.6608276367188, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.876967930029155, + "grad_norm": 0.17156261205673218, + "learning_rate": 1e-06, + "loss": -0.0765, + "num_tokens": 826260612.0, + "reward": 0.6205357313156128, + "reward_std": 0.14515121281147003, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1378 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2480.0, + "completions/mean_length": 849.0960083007812, + "completions/mean_terminated_length": 611.8970336914062, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 12.886297376093294, + "grad_norm": 0.16720975935459137, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 826868538.0, + "reward": 0.6975446939468384, + "reward_std": 0.1314416527748108, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957791805267334, + "step": 1379 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3027.0, + "completions/mean_length": 994.341552734375, + "completions/mean_terminated_length": 630.8054809570312, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 12.895626822157434, + "grad_norm": 0.170975923538208, + "learning_rate": 1e-06, + "loss": -0.0613, + "num_tokens": 827460932.0, + "reward": 0.6964285969734192, + "reward_std": 0.14591330289840698, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1380 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2790.0, + "completions/mean_length": 1023.6741333007812, + "completions/mean_terminated_length": 642.0426635742188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.904956268221575, + "grad_norm": 0.1732223927974701, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 828070152.0, + "reward": 0.6908482313156128, + "reward_std": 0.14030523598194122, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3375.0, + "completions/mean_length": 818.6250610351562, + "completions/mean_terminated_length": 591.7899780273438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.914285714285715, + "grad_norm": 0.1687568575143814, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 828646616.0, + "reward": 0.7265625596046448, + "reward_std": 0.13046343624591827, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1382 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2679.0, + "completions/mean_length": 843.630615234375, + "completions/mean_terminated_length": 572.2720947265625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.923615160349854, + "grad_norm": 0.19517146050930023, + "learning_rate": 1e-06, + "loss": -0.0609, + "num_tokens": 829217773.0, + "reward": 0.7723214626312256, + "reward_std": 0.16517743468284607, + "rewards/verify_math_reward/mean": 0.7723214030265808, + "rewards/verify_math_reward/std": 0.41956827044487, + "step": 1383 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3864.0, + "completions/mean_length": 939.700927734375, + "completions/mean_terminated_length": 591.6084594726562, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 12.932944606413994, + "grad_norm": 0.20255514979362488, + "learning_rate": 1e-06, + "loss": -0.032, + "num_tokens": 829787625.0, + "reward": 0.699776828289032, + "reward_std": 0.15135660767555237, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1384 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2739.0, + "completions/mean_length": 968.3594360351562, + "completions/mean_terminated_length": 601.778076171875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 12.942274052478133, + "grad_norm": 0.1823241263628006, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 830359459.0, + "reward": 0.7209821939468384, + "reward_std": 0.15818998217582703, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1385 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3683.0, + "completions/mean_length": 938.708740234375, + "completions/mean_terminated_length": 620.65234375, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 12.951603498542275, + "grad_norm": 0.15472254157066345, + "learning_rate": 1e-06, + "loss": -0.0442, + "num_tokens": 830950886.0, + "reward": 0.6573660969734192, + "reward_std": 0.13481523096561432, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1386 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3923.0, + "completions/mean_length": 899.4766235351562, + "completions/mean_terminated_length": 555.7218627929688, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 12.960932944606414, + "grad_norm": 0.18460515141487122, + "learning_rate": 1e-06, + "loss": -0.0311, + "num_tokens": 831497049.0, + "reward": 0.6953125596046448, + "reward_std": 0.12790584564208984, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1387 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3798.0, + "completions/mean_length": 898.130615234375, + "completions/mean_terminated_length": 597.4761962890625, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.970262390670554, + "grad_norm": 0.16650892794132233, + "learning_rate": 1e-06, + "loss": -0.0098, + "num_tokens": 832073598.0, + "reward": 0.676339328289032, + "reward_std": 0.12392427027225494, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3658.0, + "completions/mean_length": 980.9933471679688, + "completions/mean_terminated_length": 567.4968872070312, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 12.979591836734693, + "grad_norm": 0.19185760617256165, + "learning_rate": 1e-06, + "loss": -0.0335, + "num_tokens": 832613704.0, + "reward": 0.6651785969734192, + "reward_std": 0.11509531736373901, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1389 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4066.0, + "completions/mean_length": 1063.399658203125, + "completions/mean_terminated_length": 638.9898071289062, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 12.988921282798835, + "grad_norm": 0.160457581281662, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 833204518.0, + "reward": 0.6752232313156128, + "reward_std": 0.133957177400589, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1390 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.11363636363636365, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3888.0, + "completions/mean_length": 1072.8096923828125, + "completions/mean_terminated_length": 685.22119140625, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 12.998250728862974, + "grad_norm": 0.1643132120370865, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 833804753.0, + "reward": 0.6808035969734192, + "reward_std": 0.1345216929912567, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1391 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3634.0, + "completions/mean_length": 953.075927734375, + "completions/mean_terminated_length": 610.7772216796875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 13.00932944606414, + "grad_norm": 0.17346398532390594, + "learning_rate": 1e-06, + "loss": -0.0611, + "num_tokens": 834386909.0, + "reward": 0.6651785969734192, + "reward_std": 0.14928685128688812, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 1392 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2894.0, + "completions/mean_length": 888.6763916015625, + "completions/mean_terminated_length": 633.6361694335938, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 13.018658892128279, + "grad_norm": 0.14932873845100403, + "learning_rate": 1e-06, + "loss": -0.0253, + "num_tokens": 834992195.0, + "reward": 0.738839328289032, + "reward_std": 0.12621337175369263, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1393 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2754.0, + "completions/mean_length": 1021.1138916015625, + "completions/mean_terminated_length": 656.4269409179688, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 13.02798833819242, + "grad_norm": 0.16457971930503845, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 835608521.0, + "reward": 0.6205357313156128, + "reward_std": 0.1498509645462036, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2785.0, + "completions/mean_length": 961.0982666015625, + "completions/mean_terminated_length": 576.1102905273438, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 13.03731778425656, + "grad_norm": 0.1542545109987259, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 836155785.0, + "reward": 0.7008928656578064, + "reward_std": 0.10126376897096634, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1395 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2983.0, + "completions/mean_length": 990.2388916015625, + "completions/mean_terminated_length": 595.67041015625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.0466472303207, + "grad_norm": 0.1856381595134735, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 836721383.0, + "reward": 0.691964328289032, + "reward_std": 0.13380561769008636, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1396 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3297.0, + "completions/mean_length": 833.700927734375, + "completions/mean_terminated_length": 582.7548217773438, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 13.055976676384839, + "grad_norm": 0.17601068317890167, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 837299187.0, + "reward": 0.7343750596046448, + "reward_std": 0.12926062941551208, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 1397 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3341.0, + "completions/mean_length": 843.6920166015625, + "completions/mean_terminated_length": 572.3385620117188, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 13.06530612244898, + "grad_norm": 0.1704408824443817, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 837850399.0, + "reward": 0.7600446939468384, + "reward_std": 0.13583441078662872, + "rewards/verify_math_reward/mean": 0.7600446343421936, + "rewards/verify_math_reward/std": 0.42729446291923523, + "step": 1398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4028.0, + "completions/mean_length": 1035.3013916015625, + "completions/mean_terminated_length": 602.5146484375, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 13.07463556851312, + "grad_norm": 0.24100685119628906, + "learning_rate": 1e-06, + "loss": -0.0672, + "num_tokens": 838411797.0, + "reward": 0.7087053656578064, + "reward_std": 0.11257798224687576, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1399 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3841.0, + "completions/mean_length": 946.1730346679688, + "completions/mean_terminated_length": 633.1226806640625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.08396501457726, + "grad_norm": 0.15355545282363892, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 839024728.0, + "reward": 0.7243303656578064, + "reward_std": 0.10806508362293243, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1400 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2144.0, + "completions/mean_length": 1005.9732666015625, + "completions/mean_terminated_length": 604.6204223632812, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.093294460641399, + "grad_norm": 0.194878950715065, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 839586024.0, + "reward": 0.699776828289032, + "reward_std": 0.1322651207447052, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1401 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3258.0, + "completions/mean_length": 895.685302734375, + "completions/mean_terminated_length": 607.5790405273438, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 13.102623906705539, + "grad_norm": 0.13444751501083374, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 840177678.0, + "reward": 0.7600446939468384, + "reward_std": 0.11994270235300064, + "rewards/verify_math_reward/mean": 0.7600446343421936, + "rewards/verify_math_reward/std": 0.42729443311691284, + "step": 1402 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3630.0, + "completions/mean_length": 960.70654296875, + "completions/mean_terminated_length": 606.281982421875, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 13.11195335276968, + "grad_norm": 0.17206306755542755, + "learning_rate": 1e-06, + "loss": -0.0304, + "num_tokens": 840761095.0, + "reward": 0.660714328289032, + "reward_std": 0.13516220450401306, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1403 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3259.0, + "completions/mean_length": 1089.16748046875, + "completions/mean_terminated_length": 601.6783447265625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 13.12128279883382, + "grad_norm": 0.16660507023334503, + "learning_rate": 1e-06, + "loss": -0.0788, + "num_tokens": 841316509.0, + "reward": 0.699776828289032, + "reward_std": 0.12794680893421173, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3819.0, + "completions/mean_length": 1164.313720703125, + "completions/mean_terminated_length": 719.6619262695312, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 13.130612244897959, + "grad_norm": 0.14841970801353455, + "learning_rate": 1e-06, + "loss": -0.0612, + "num_tokens": 841972326.0, + "reward": 0.613839328289032, + "reward_std": 0.1604444682598114, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1405 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3977.0, + "completions/mean_length": 987.4844360351562, + "completions/mean_terminated_length": 640.379638671875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 13.139941690962099, + "grad_norm": 0.16702315211296082, + "learning_rate": 1e-06, + "loss": -0.0799, + "num_tokens": 842582616.0, + "reward": 0.7053571939468384, + "reward_std": 0.15785367786884308, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1406 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3283.0, + "completions/mean_length": 898.89404296875, + "completions/mean_terminated_length": 581.144775390625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 13.14927113702624, + "grad_norm": 0.19033049046993256, + "learning_rate": 1e-06, + "loss": -0.0695, + "num_tokens": 843137129.0, + "reward": 0.746651828289032, + "reward_std": 0.14774522185325623, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1407 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2645.0, + "completions/mean_length": 1156.993408203125, + "completions/mean_terminated_length": 608.1192016601562, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.15860058309038, + "grad_norm": 0.18148021399974823, + "learning_rate": 1e-06, + "loss": -0.0383, + "num_tokens": 843690163.0, + "reward": 0.6506696939468384, + "reward_std": 0.13955454528331757, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1408 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3721.0, + "completions/mean_length": 1054.8304443359375, + "completions/mean_terminated_length": 598.0693359375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.167930029154519, + "grad_norm": 0.14989906549453735, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 844262651.0, + "reward": 0.6227678656578064, + "reward_std": 0.12433473765850067, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 1409 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3295.0, + "completions/mean_length": 805.2756958007812, + "completions/mean_terminated_length": 569.0992431640625, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 13.177259475218658, + "grad_norm": 0.16026248037815094, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 844832618.0, + "reward": 0.793526828289032, + "reward_std": 0.12212050706148148, + "rewards/verify_math_reward/mean": 0.7935267686843872, + "rewards/verify_math_reward/std": 0.40500015020370483, + "step": 1410 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2769.0, + "completions/mean_length": 982.1484985351562, + "completions/mean_terminated_length": 564.3405151367188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 13.186588921282798, + "grad_norm": 0.1801268607378006, + "learning_rate": 1e-06, + "loss": -0.0843, + "num_tokens": 845362831.0, + "reward": 0.7533482313156128, + "reward_std": 0.17502741515636444, + "rewards/verify_math_reward/mean": 0.7533482313156128, + "rewards/verify_math_reward/std": 0.4313030242919922, + "step": 1411 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3175.0, + "completions/mean_length": 1020.4285888671875, + "completions/mean_terminated_length": 625.3299560546875, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 13.19591836734694, + "grad_norm": 0.17634259164333344, + "learning_rate": 1e-06, + "loss": -0.0628, + "num_tokens": 845954191.0, + "reward": 0.6651785969734192, + "reward_std": 0.14676883816719055, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1412 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3678.0, + "completions/mean_length": 960.4631958007812, + "completions/mean_terminated_length": 566.551513671875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.205247813411079, + "grad_norm": 0.17410670220851898, + "learning_rate": 1e-06, + "loss": -0.0817, + "num_tokens": 846494438.0, + "reward": 0.7042410969734192, + "reward_std": 0.13226650655269623, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2934.0, + "completions/mean_length": 1141.9910888671875, + "completions/mean_terminated_length": 581.0040283203125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 13.214577259475218, + "grad_norm": 0.19327431917190552, + "learning_rate": 1e-06, + "loss": -0.0892, + "num_tokens": 847020630.0, + "reward": 0.6573660969734192, + "reward_std": 0.14274516701698303, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485536336898804, + "step": 1414 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2706.0, + "completions/mean_length": 1073.8638916015625, + "completions/mean_terminated_length": 597.5064697265625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.223906705539358, + "grad_norm": 0.1650964766740799, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 847579596.0, + "reward": 0.668526828289032, + "reward_std": 0.13229858875274658, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3567.0, + "completions/mean_length": 938.71435546875, + "completions/mean_terminated_length": 577.4328002929688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.2332361516035, + "grad_norm": 0.16018874943256378, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 848139716.0, + "reward": 0.6707589626312256, + "reward_std": 0.11599501967430115, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1416 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2709.0, + "completions/mean_length": 896.26904296875, + "completions/mean_terminated_length": 633.4891357421875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 13.242565597667639, + "grad_norm": 0.16617926955223083, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 848744485.0, + "reward": 0.7220982313156128, + "reward_std": 0.13252753019332886, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 1417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3493.0, + "completions/mean_length": 900.2500610351562, + "completions/mean_terminated_length": 625.2218017578125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.251895043731778, + "grad_norm": 0.14421969652175903, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 849353517.0, + "reward": 0.7455357313156128, + "reward_std": 0.11453007161617279, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1418 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3192.0, + "completions/mean_length": 1053.899658203125, + "completions/mean_terminated_length": 688.8474731445312, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 13.261224489795918, + "grad_norm": 0.13865505158901215, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 849993419.0, + "reward": 0.6930803656578064, + "reward_std": 0.13440224528312683, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 1419 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1940.0, + "completions/mean_length": 892.9308471679688, + "completions/mean_terminated_length": 544.0816650390625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 13.270553935860057, + "grad_norm": 0.17532752454280853, + "learning_rate": 1e-06, + "loss": -0.0765, + "num_tokens": 850527917.0, + "reward": 0.746651828289032, + "reward_std": 0.14556489884853363, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1420 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3509.0, + "completions/mean_length": 851.6105346679688, + "completions/mean_terminated_length": 546.5824584960938, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 13.279883381924199, + "grad_norm": 0.14888831973075867, + "learning_rate": 1e-06, + "loss": -0.0464, + "num_tokens": 851056416.0, + "reward": 0.6986607313156128, + "reward_std": 0.09777004271745682, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1421 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3261.0, + "completions/mean_length": 988.5335083007812, + "completions/mean_terminated_length": 580.4822998046875, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 13.289212827988338, + "grad_norm": 0.17417357861995697, + "learning_rate": 1e-06, + "loss": -0.0778, + "num_tokens": 851608230.0, + "reward": 0.715401828289032, + "reward_std": 0.14951257407665253, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1422 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2408.0, + "completions/mean_length": 836.9386596679688, + "completions/mean_terminated_length": 582.01806640625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.298542274052478, + "grad_norm": 0.13680718839168549, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 852181215.0, + "reward": 0.7399553656578064, + "reward_std": 0.11306652426719666, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3765.0, + "completions/mean_length": 980.2467041015625, + "completions/mean_terminated_length": 653.6880493164062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 13.307871720116617, + "grad_norm": 0.17567257583141327, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 852804780.0, + "reward": 0.6674107313156128, + "reward_std": 0.14635583758354187, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1424 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3773.0, + "completions/mean_length": 1001.1495971679688, + "completions/mean_terminated_length": 651.296875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 13.317201166180759, + "grad_norm": 0.1757599264383316, + "learning_rate": 1e-06, + "loss": -0.0636, + "num_tokens": 853409298.0, + "reward": 0.6361607313156128, + "reward_std": 0.16750863194465637, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1425 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2792.0, + "completions/mean_length": 957.5803833007812, + "completions/mean_terminated_length": 632.916259765625, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "epoch": 13.326530612244898, + "grad_norm": 0.1588900089263916, + "learning_rate": 1e-06, + "loss": -0.0611, + "num_tokens": 854007802.0, + "reward": 0.7354910969734192, + "reward_std": 0.13827435672283173, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131773710250854, + "step": 1426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 1095.169677734375, + "completions/mean_terminated_length": 657.7084350585938, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 13.335860058309038, + "grad_norm": 0.1701296716928482, + "learning_rate": 1e-06, + "loss": -0.0847, + "num_tokens": 854608994.0, + "reward": 0.6506696939468384, + "reward_std": 0.17250937223434448, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1427 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2462.0, + "completions/mean_length": 885.3292846679688, + "completions/mean_terminated_length": 630.0228881835938, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.345189504373177, + "grad_norm": 0.17226791381835938, + "learning_rate": 1e-06, + "loss": -0.0514, + "num_tokens": 855220193.0, + "reward": 0.7031250596046448, + "reward_std": 0.12787194550037384, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1428 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3438.0, + "completions/mean_length": 955.6428833007812, + "completions/mean_terminated_length": 569.9849853515625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 13.354518950437317, + "grad_norm": 0.1471056342124939, + "learning_rate": 1e-06, + "loss": -0.0876, + "num_tokens": 855756977.0, + "reward": 0.7433035969734192, + "reward_std": 0.13203828036785126, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1429 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3461.0, + "completions/mean_length": 1002.271240234375, + "completions/mean_terminated_length": 591.5992431640625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 13.363848396501458, + "grad_norm": 0.1863008588552475, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 856304996.0, + "reward": 0.6930803656578064, + "reward_std": 0.13151581585407257, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1430 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2640.0, + "completions/mean_length": 802.6886596679688, + "completions/mean_terminated_length": 536.5223388671875, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 13.373177842565598, + "grad_norm": 0.16983327269554138, + "learning_rate": 1e-06, + "loss": -0.0341, + "num_tokens": 856844869.0, + "reward": 0.7399553656578064, + "reward_std": 0.12756815552711487, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1431 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2443.0, + "completions/mean_length": 1055.243408203125, + "completions/mean_terminated_length": 642.8720092773438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.382507288629737, + "grad_norm": 0.16662375628948212, + "learning_rate": 1e-06, + "loss": -0.0659, + "num_tokens": 857444103.0, + "reward": 0.6752232313156128, + "reward_std": 0.14350220561027527, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1432 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2658.0, + "completions/mean_length": 997.90185546875, + "completions/mean_terminated_length": 568.8131713867188, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.391836734693877, + "grad_norm": 0.1733788251876831, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 857986231.0, + "reward": 0.7131696939468384, + "reward_std": 0.15146787464618683, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1433 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2272.0, + "completions/mean_length": 884.0413208007812, + "completions/mean_terminated_length": 556.129150390625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.401166180758018, + "grad_norm": 0.17448434233665466, + "learning_rate": 1e-06, + "loss": -0.035, + "num_tokens": 858530740.0, + "reward": 0.7343750596046448, + "reward_std": 0.10626383125782013, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 1434 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3361.0, + "completions/mean_length": 888.1350708007812, + "completions/mean_terminated_length": 582.2506103515625, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 13.410495626822158, + "grad_norm": 0.16415566205978394, + "learning_rate": 1e-06, + "loss": -0.0775, + "num_tokens": 859086229.0, + "reward": 0.7633928656578064, + "reward_std": 0.11175242811441422, + "rewards/verify_math_reward/mean": 0.7633928656578064, + "rewards/verify_math_reward/std": 0.42523646354675293, + "step": 1435 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2904.0, + "completions/mean_length": 941.43310546875, + "completions/mean_terminated_length": 584.8297729492188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.419825072886297, + "grad_norm": 0.17067740857601166, + "learning_rate": 1e-06, + "loss": -0.0717, + "num_tokens": 859655705.0, + "reward": 0.7031250596046448, + "reward_std": 0.14327649772167206, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1436 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2738.0, + "completions/mean_length": 880.6641235351562, + "completions/mean_terminated_length": 624.9867553710938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.429154518950437, + "grad_norm": 0.15706272423267365, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 860261260.0, + "reward": 0.7243303656578064, + "reward_std": 0.1296411007642746, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1437 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3175.0, + "completions/mean_length": 956.1551513671875, + "completions/mean_terminated_length": 566.1367797851562, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 13.438483965014576, + "grad_norm": 0.14836430549621582, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 860804039.0, + "reward": 0.7366071939468384, + "reward_std": 0.09964726865291595, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1438 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1584821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3724.0, + "completions/mean_length": 1184.2645263671875, + "completions/mean_terminated_length": 635.9005126953125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.447813411078718, + "grad_norm": 0.16693590581417084, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 861379044.0, + "reward": 0.6328125, + "reward_std": 0.13072557747364044, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1439 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2738.0, + "completions/mean_length": 949.8281860351562, + "completions/mean_terminated_length": 637.14111328125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.457142857142857, + "grad_norm": 0.1452966183423996, + "learning_rate": 1e-06, + "loss": -0.0615, + "num_tokens": 861981666.0, + "reward": 0.7399553656578064, + "reward_std": 0.1275695562362671, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1440 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3538.0, + "completions/mean_length": 1031.602783203125, + "completions/mean_terminated_length": 650.9560546875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.466472303206997, + "grad_norm": 0.15150801837444305, + "learning_rate": 1e-06, + "loss": -0.0436, + "num_tokens": 862592182.0, + "reward": 0.6674107313156128, + "reward_std": 0.12366325408220291, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1441 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4080.0, + "completions/mean_length": 1003.4531860351562, + "completions/mean_terminated_length": 645.2876586914062, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 13.475801749271136, + "grad_norm": 0.15310557186603546, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 863197084.0, + "reward": 0.7131696939468384, + "reward_std": 0.11945554614067078, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1442 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2603.0, + "completions/mean_length": 985.8069458007812, + "completions/mean_terminated_length": 581.8347778320312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 13.485131195335278, + "grad_norm": 0.1850886195898056, + "learning_rate": 1e-06, + "loss": -0.0633, + "num_tokens": 863758295.0, + "reward": 0.7031250596046448, + "reward_std": 0.14481602609157562, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1443 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3487.0, + "completions/mean_length": 912.3303833007812, + "completions/mean_terminated_length": 613.010986328125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 13.494460641399417, + "grad_norm": 0.152607262134552, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 864348439.0, + "reward": 0.7578125596046448, + "reward_std": 0.13591107726097107, + "rewards/verify_math_reward/mean": 0.7578125, + "rewards/verify_math_reward/std": 0.428646445274353, + "step": 1444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2739.0, + "completions/mean_length": 1003.7098388671875, + "completions/mean_terminated_length": 597.6514892578125, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 13.503790087463557, + "grad_norm": 0.1480863243341446, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 864912723.0, + "reward": 0.7042410969734192, + "reward_std": 0.1145307794213295, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3719.0, + "completions/mean_length": 985.6272583007812, + "completions/mean_terminated_length": 634.0198364257812, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 13.513119533527696, + "grad_norm": 0.14685481786727905, + "learning_rate": 1e-06, + "loss": -0.0363, + "num_tokens": 865506973.0, + "reward": 0.6674107313156128, + "reward_std": 0.10893410444259644, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3281.0, + "completions/mean_length": 995.97216796875, + "completions/mean_terminated_length": 588.897705078125, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.522448979591836, + "grad_norm": 0.18371158838272095, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 866075324.0, + "reward": 0.65625, + "reward_std": 0.13511762022972107, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1447 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3313.0, + "completions/mean_length": 1007.96435546875, + "completions/mean_terminated_length": 571.3121337890625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 13.531778425655977, + "grad_norm": 0.17786386609077454, + "learning_rate": 1e-06, + "loss": -0.0565, + "num_tokens": 866604804.0, + "reward": 0.707589328289032, + "reward_std": 0.12373880296945572, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1448 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3462.0, + "completions/mean_length": 856.0658569335938, + "completions/mean_terminated_length": 589.9843139648438, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 13.541107871720117, + "grad_norm": 0.1376257985830307, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 867182991.0, + "reward": 0.7343750596046448, + "reward_std": 0.11727311462163925, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 1449 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3789.0, + "completions/mean_length": 961.005615234375, + "completions/mean_terminated_length": 589.1897583007812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 13.550437317784256, + "grad_norm": 0.13891948759555817, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 867755420.0, + "reward": 0.6964285969734192, + "reward_std": 0.10777155309915543, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1450 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3354.0, + "completions/mean_length": 894.74560546875, + "completions/mean_terminated_length": 598.0438842773438, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 13.559766763848396, + "grad_norm": 0.15713556110858917, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 868326080.0, + "reward": 0.7254464626312256, + "reward_std": 0.1337721347808838, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1451 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2098.0, + "completions/mean_length": 948.052490234375, + "completions/mean_terminated_length": 579.091064453125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 13.569096209912537, + "grad_norm": 0.17287586629390717, + "learning_rate": 1e-06, + "loss": -0.0613, + "num_tokens": 868873743.0, + "reward": 0.7087053656578064, + "reward_std": 0.15394878387451172, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1452 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 934.5614013671875, + "completions/mean_terminated_length": 598.9024658203125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 13.578425655976677, + "grad_norm": 0.18285146355628967, + "learning_rate": 1e-06, + "loss": -0.0806, + "num_tokens": 869445102.0, + "reward": 0.7031250596046448, + "reward_std": 0.15530356764793396, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1453 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3315.0, + "completions/mean_length": 980.521240234375, + "completions/mean_terminated_length": 611.0199584960938, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 13.587755102040816, + "grad_norm": 0.20162659883499146, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 870028041.0, + "reward": 0.6651785969734192, + "reward_std": 0.13929423689842224, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 1454 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3592.0, + "completions/mean_length": 990.0045166015625, + "completions/mean_terminated_length": 634.592041015625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 13.597084548104956, + "grad_norm": 0.16121090948581696, + "learning_rate": 1e-06, + "loss": -0.0656, + "num_tokens": 870634525.0, + "reward": 0.691964328289032, + "reward_std": 0.14263570308685303, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1455 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4001.0, + "completions/mean_length": 947.6160888671875, + "completions/mean_terminated_length": 643.182373046875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 13.606413994169095, + "grad_norm": 0.1407116949558258, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 871256381.0, + "reward": 0.6886160969734192, + "reward_std": 0.11900020390748978, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1456 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2704.0, + "completions/mean_length": 1019.2824096679688, + "completions/mean_terminated_length": 637.1054077148438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 13.615743440233237, + "grad_norm": 0.18685175478458405, + "learning_rate": 1e-06, + "loss": -0.072, + "num_tokens": 871850650.0, + "reward": 0.6741071939468384, + "reward_std": 0.15289682149887085, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2851.0, + "completions/mean_length": 766.0357666015625, + "completions/mean_terminated_length": 548.2615966796875, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 13.625072886297376, + "grad_norm": 0.127822607755661, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 872400834.0, + "reward": 0.7555803656578064, + "reward_std": 0.10325934737920761, + "rewards/verify_math_reward/mean": 0.7555803656578064, + "rewards/verify_math_reward/std": 0.42998260259628296, + "step": 1458 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3882.0, + "completions/mean_length": 1057.83935546875, + "completions/mean_terminated_length": 658.888916015625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 13.634402332361516, + "grad_norm": 0.14694525301456451, + "learning_rate": 1e-06, + "loss": -0.0687, + "num_tokens": 873018506.0, + "reward": 0.6707589626312256, + "reward_std": 0.13339193165302277, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1459 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2342.0, + "completions/mean_length": 1110.458740234375, + "completions/mean_terminated_length": 644.3290405273438, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 13.643731778425655, + "grad_norm": 0.16294415295124054, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 873612685.0, + "reward": 0.6830357313156128, + "reward_std": 0.13955634832382202, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1460 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3086.0, + "completions/mean_length": 1041.673095703125, + "completions/mean_terminated_length": 649.3035278320312, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 13.653061224489797, + "grad_norm": 0.1573956310749054, + "learning_rate": 1e-06, + "loss": -0.071, + "num_tokens": 874208744.0, + "reward": 0.7131696939468384, + "reward_std": 0.14301365613937378, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3341.0, + "completions/mean_length": 980.3527221679688, + "completions/mean_terminated_length": 593.3400268554688, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 13.662390670553936, + "grad_norm": 0.17042101919651031, + "learning_rate": 1e-06, + "loss": -0.0671, + "num_tokens": 874772964.0, + "reward": 0.7053571939468384, + "reward_std": 0.14042328298091888, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1462 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 1108.044677734375, + "completions/mean_terminated_length": 619.1065063476562, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 13.671720116618076, + "grad_norm": 0.1893538534641266, + "learning_rate": 1e-06, + "loss": -0.0987, + "num_tokens": 875337116.0, + "reward": 0.640625, + "reward_std": 0.16003653407096863, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 1463 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4095.0, + "completions/mean_length": 1091.2835693359375, + "completions/mean_terminated_length": 644.42822265625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 13.681049562682215, + "grad_norm": 0.1623772531747818, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 875924458.0, + "reward": 0.637276828289032, + "reward_std": 0.11272881925106049, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1464 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3126.0, + "completions/mean_length": 853.2600708007812, + "completions/mean_terminated_length": 603.8185424804688, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 13.690379008746355, + "grad_norm": 0.2092238962650299, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 876523275.0, + "reward": 0.7020089626312256, + "reward_std": 0.18216319382190704, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1465 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2462.0, + "completions/mean_length": 922.2467041015625, + "completions/mean_terminated_length": 576.59033203125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 13.699708454810496, + "grad_norm": 0.15257404744625092, + "learning_rate": 1e-06, + "loss": -0.0474, + "num_tokens": 877067528.0, + "reward": 0.746651828289032, + "reward_std": 0.11204595863819122, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3467.0, + "completions/mean_length": 960.7176513671875, + "completions/mean_terminated_length": 640.6334228515625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 13.709037900874636, + "grad_norm": 0.15851299464702606, + "learning_rate": 1e-06, + "loss": -0.0426, + "num_tokens": 877672347.0, + "reward": 0.7131696939468384, + "reward_std": 0.13000810146331787, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3065.0, + "completions/mean_length": 990.8058471679688, + "completions/mean_terminated_length": 591.9017333984375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.718367346938775, + "grad_norm": 0.17357619106769562, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 878243837.0, + "reward": 0.6595982313156128, + "reward_std": 0.12613599002361298, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1468 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3254.0, + "completions/mean_length": 1113.1785888671875, + "completions/mean_terminated_length": 691.40380859375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 13.727696793002915, + "grad_norm": 0.16974762082099915, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 878877197.0, + "reward": 0.6484375, + "reward_std": 0.14838209748268127, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1469 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2407.0, + "completions/mean_length": 1057.8192138671875, + "completions/mean_terminated_length": 614.9130249023438, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.737026239067056, + "grad_norm": 0.16159725189208984, + "learning_rate": 1e-06, + "loss": -0.0706, + "num_tokens": 879437771.0, + "reward": 0.7109375596046448, + "reward_std": 0.12685208022594452, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1470 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3467.0, + "completions/mean_length": 1014.474365234375, + "completions/mean_terminated_length": 614.2257080078125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 13.746355685131196, + "grad_norm": 0.19176022708415985, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 880013188.0, + "reward": 0.7198660969734192, + "reward_std": 0.16735707223415375, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1471 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2491.0, + "completions/mean_length": 1065.8817138671875, + "completions/mean_terminated_length": 624.15087890625, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 13.755685131195335, + "grad_norm": 0.15909433364868164, + "learning_rate": 1e-06, + "loss": -0.0818, + "num_tokens": 880600018.0, + "reward": 0.6875000596046448, + "reward_std": 0.15015266835689545, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1472 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3102.0, + "completions/mean_length": 1013.0558471679688, + "completions/mean_terminated_length": 647.4132080078125, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 13.765014577259475, + "grad_norm": 0.14070793986320496, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 881202588.0, + "reward": 0.6741071939468384, + "reward_std": 0.12464035302400589, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1473 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 1051.0201416015625, + "completions/mean_terminated_length": 642.4531860351562, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 13.774344023323614, + "grad_norm": 0.1834116131067276, + "learning_rate": 1e-06, + "loss": -0.0703, + "num_tokens": 881808030.0, + "reward": 0.65625, + "reward_std": 0.1377544105052948, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1474 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2916.0, + "completions/mean_length": 1013.9375610351562, + "completions/mean_terminated_length": 690.9099731445312, + "completions/min_length": 208.0, + "completions/min_terminated_length": 208.0, + "epoch": 13.783673469387756, + "grad_norm": 0.16175587475299835, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 882462102.0, + "reward": 0.6573660969734192, + "reward_std": 0.14248555898666382, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3226.0, + "completions/mean_length": 862.0982666015625, + "completions/mean_terminated_length": 562.3707275390625, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 13.793002915451895, + "grad_norm": 0.19891230762004852, + "learning_rate": 1e-06, + "loss": -0.078, + "num_tokens": 883019774.0, + "reward": 0.7299107313156128, + "reward_std": 0.164614737033844, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1476 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3850.0, + "completions/mean_length": 1164.677490234375, + "completions/mean_terminated_length": 680.5708618164062, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 13.802332361516035, + "grad_norm": 0.16575460135936737, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 883638445.0, + "reward": 0.6350446939468384, + "reward_std": 0.1350851058959961, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1477 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3474.0, + "completions/mean_length": 1026.149658203125, + "completions/mean_terminated_length": 631.785888671875, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 13.811661807580174, + "grad_norm": 0.15212048590183258, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 884218523.0, + "reward": 0.7354910969734192, + "reward_std": 0.11809726059436798, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 1478 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2888.0, + "completions/mean_length": 980.6250610351562, + "completions/mean_terminated_length": 645.5970458984375, + "completions/min_length": 198.0, + "completions/min_terminated_length": 198.0, + "epoch": 13.820991253644316, + "grad_norm": 0.18683385848999023, + "learning_rate": 1e-06, + "loss": -0.0314, + "num_tokens": 884828347.0, + "reward": 0.6439732313156128, + "reward_std": 0.1507924646139145, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1479 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3962.0, + "completions/mean_length": 968.4576416015625, + "completions/mean_terminated_length": 601.8878173828125, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 13.830320699708455, + "grad_norm": 0.1512003242969513, + "learning_rate": 1e-06, + "loss": -0.0325, + "num_tokens": 885395469.0, + "reward": 0.7477678656578064, + "reward_std": 0.10310779511928558, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.434536337852478, + "step": 1480 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4015.0, + "completions/mean_length": 929.6160888671875, + "completions/mean_terminated_length": 614.9202270507812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.839650145772595, + "grad_norm": 0.15268227458000183, + "learning_rate": 1e-06, + "loss": -0.0399, + "num_tokens": 885979421.0, + "reward": 0.7354910969734192, + "reward_std": 0.11974698305130005, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 1481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 1374.6082763671875, + "completions/mean_terminated_length": 746.5947875976562, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.848979591836734, + "grad_norm": 0.17460675537586212, + "learning_rate": 1e-06, + "loss": -0.0944, + "num_tokens": 886622726.0, + "reward": 0.582589328289032, + "reward_std": 0.1363222450017929, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 1482 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 988.5022583007812, + "completions/mean_terminated_length": 624.2817993164062, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 13.858309037900874, + "grad_norm": 0.19615179300308228, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 887216496.0, + "reward": 0.676339328289032, + "reward_std": 0.1649598926305771, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1483 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2445.0, + "completions/mean_length": 964.9576416015625, + "completions/mean_terminated_length": 602.333740234375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 13.867638483965015, + "grad_norm": 0.16442091763019562, + "learning_rate": 1e-06, + "loss": -0.0583, + "num_tokens": 887789738.0, + "reward": 0.6796875596046448, + "reward_std": 0.1258022040128708, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3406.0, + "completions/mean_length": 957.2288208007812, + "completions/mean_terminated_length": 584.9650268554688, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 13.876967930029155, + "grad_norm": 0.1913428008556366, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 888339695.0, + "reward": 0.691964328289032, + "reward_std": 0.1349346935749054, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2628.0, + "completions/mean_length": 1037.868408203125, + "completions/mean_terminated_length": 618.7334594726562, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 13.886297376093294, + "grad_norm": 0.14787276089191437, + "learning_rate": 1e-06, + "loss": -0.0296, + "num_tokens": 888914169.0, + "reward": 0.6484375, + "reward_std": 0.10502780973911285, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1486 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 969.2645263671875, + "completions/mean_terminated_length": 585.2794189453125, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 13.895626822157434, + "grad_norm": 0.17590682208538055, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 889470590.0, + "reward": 0.6986607313156128, + "reward_std": 0.16074123978614807, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1487 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4073.0, + "completions/mean_length": 902.0357666015625, + "completions/mean_terminated_length": 610.2606811523438, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 13.904956268221575, + "grad_norm": 0.1257428377866745, + "learning_rate": 1e-06, + "loss": -0.0487, + "num_tokens": 890049350.0, + "reward": 0.7444196939468384, + "reward_std": 0.09679364413022995, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1488 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2802.0, + "completions/mean_length": 1040.3148193359375, + "completions/mean_terminated_length": 590.3739013671875, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 13.914285714285715, + "grad_norm": 0.17225463688373566, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 890605344.0, + "reward": 0.6584821939468384, + "reward_std": 0.1392507553100586, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1489 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2248.0, + "completions/mean_length": 975.9922485351562, + "completions/mean_terminated_length": 588.4378662109375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 13.923615160349854, + "grad_norm": 0.19903920590877533, + "learning_rate": 1e-06, + "loss": -0.0693, + "num_tokens": 891162729.0, + "reward": 0.7053571939468384, + "reward_std": 0.146812304854393, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1490 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2746.0, + "completions/mean_length": 960.2701416015625, + "completions/mean_terminated_length": 635.8842163085938, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 13.932944606413994, + "grad_norm": 0.17317359149456024, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 891766315.0, + "reward": 0.6662946939468384, + "reward_std": 0.16029545664787292, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1491 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2907.0, + "completions/mean_length": 956.4397583007812, + "completions/mean_terminated_length": 601.5328979492188, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 13.942274052478133, + "grad_norm": 0.16615170240402222, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 892339717.0, + "reward": 0.645089328289032, + "reward_std": 0.1494365781545639, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1492 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1768.0, + "completions/mean_length": 826.8303833007812, + "completions/mean_terminated_length": 583.798583984375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.951603498542275, + "grad_norm": 0.17302149534225464, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 892913909.0, + "reward": 0.7276785969734192, + "reward_std": 0.13572493195533752, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1493 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2086.0, + "completions/mean_length": 1025.3695068359375, + "completions/mean_terminated_length": 630.905517578125, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 13.960932944606414, + "grad_norm": 0.1573459655046463, + "learning_rate": 1e-06, + "loss": -0.0859, + "num_tokens": 893511112.0, + "reward": 0.6964285969734192, + "reward_std": 0.14624707400798798, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1494 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3187.0, + "completions/mean_length": 1127.03466796875, + "completions/mean_terminated_length": 636.7113037109375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.970262390670554, + "grad_norm": 0.149027481675148, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 894081831.0, + "reward": 0.6149553656578064, + "reward_std": 0.13421930372714996, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1495 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3001.0, + "completions/mean_length": 891.9766235351562, + "completions/mean_terminated_length": 582.1627807617188, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 13.979591836734693, + "grad_norm": 0.16230741143226624, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 894642658.0, + "reward": 0.7332589626312256, + "reward_std": 0.10998578369617462, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 1496 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2666.0, + "completions/mean_length": 1073.5625, + "completions/mean_terminated_length": 680.9886474609375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 13.988921282798835, + "grad_norm": 0.18213573098182678, + "learning_rate": 1e-06, + "loss": -0.0481, + "num_tokens": 895275730.0, + "reward": 0.6183035969734192, + "reward_std": 0.17404988408088684, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 1497 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.11931818181818177, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2855.0, + "completions/mean_length": 1023.227294921875, + "completions/mean_terminated_length": 606.9161376953125, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 13.998250728862974, + "grad_norm": 0.17152397334575653, + "learning_rate": 1e-06, + "loss": -0.0619, + "num_tokens": 895843375.0, + "reward": 0.7109375596046448, + "reward_std": 0.1328292191028595, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3177.0, + "completions/mean_length": 1059.9765625, + "completions/mean_terminated_length": 599.5, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.00932944606414, + "grad_norm": 0.2022089809179306, + "learning_rate": 1e-06, + "loss": -0.0283, + "num_tokens": 896397698.0, + "reward": 0.645089328289032, + "reward_std": 0.10757722705602646, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1499 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2431.0, + "completions/mean_length": 847.0335083007812, + "completions/mean_terminated_length": 592.9025268554688, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 14.018658892128279, + "grad_norm": 0.16060009598731995, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 896986088.0, + "reward": 0.7667410969734192, + "reward_std": 0.12628935277462006, + "rewards/verify_math_reward/mean": 0.7667410969734192, + "rewards/verify_math_reward/std": 0.4231418967247009, + "step": 1500 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2513.0, + "completions/mean_length": 1096.0982666015625, + "completions/mean_terminated_length": 645.5353393554688, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 14.02798833819242, + "grad_norm": 0.17677782475948334, + "learning_rate": 1e-06, + "loss": -0.0658, + "num_tokens": 897587488.0, + "reward": 0.6227678656578064, + "reward_std": 0.16623122990131378, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 1501 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2656.0, + "completions/mean_length": 947.55810546875, + "completions/mean_terminated_length": 630.3931274414062, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 14.03731778425656, + "grad_norm": 0.1595323532819748, + "learning_rate": 1e-06, + "loss": -0.0549, + "num_tokens": 898181756.0, + "reward": 0.7187500596046448, + "reward_std": 0.13865482807159424, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3637.0, + "completions/mean_length": 913.8225708007812, + "completions/mean_terminated_length": 623.124267578125, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 14.0466472303207, + "grad_norm": 0.1728564351797104, + "learning_rate": 1e-06, + "loss": -0.0748, + "num_tokens": 898783421.0, + "reward": 0.7410714626312256, + "reward_std": 0.16751113533973694, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.4382909834384918, + "step": 1503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3101.0, + "completions/mean_length": 923.2734985351562, + "completions/mean_terminated_length": 616.4859619140625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 14.055976676384839, + "grad_norm": 0.17132966220378876, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 899372250.0, + "reward": 0.731026828289032, + "reward_std": 0.1465826779603958, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1504 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2858.0, + "completions/mean_length": 869.1194458007812, + "completions/mean_terminated_length": 608.3220825195312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 14.06530612244898, + "grad_norm": 0.14070063829421997, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 899963989.0, + "reward": 0.738839328289032, + "reward_std": 0.12328347563743591, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 922.3638916015625, + "completions/mean_terminated_length": 598.3640747070312, + "completions/min_length": 194.0, + "completions/min_terminated_length": 194.0, + "epoch": 14.07463556851312, + "grad_norm": 0.17546145617961884, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 900545227.0, + "reward": 0.6886160969734192, + "reward_std": 0.14410065114498138, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1506 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 1105.091552734375, + "completions/mean_terminated_length": 620.1842041015625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 14.08396501457726, + "grad_norm": 0.17474296689033508, + "learning_rate": 1e-06, + "loss": -0.0855, + "num_tokens": 901111565.0, + "reward": 0.6953125596046448, + "reward_std": 0.1361721009016037, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1507 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2709.0, + "completions/mean_length": 914.505615234375, + "completions/mean_terminated_length": 568.0061645507812, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 14.093294460641399, + "grad_norm": 0.15863321721553802, + "learning_rate": 1e-06, + "loss": -0.0654, + "num_tokens": 901658418.0, + "reward": 0.7477678656578064, + "reward_std": 0.12166838347911835, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1508 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2591.0, + "completions/mean_length": 965.6998291015625, + "completions/mean_terminated_length": 590.063720703125, + "completions/min_length": 205.0, + "completions/min_terminated_length": 205.0, + "epoch": 14.102623906705539, + "grad_norm": 0.15326561033725739, + "learning_rate": 1e-06, + "loss": -0.0873, + "num_tokens": 902221901.0, + "reward": 0.7176339626312256, + "reward_std": 0.13914267718791962, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1509 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2757.0, + "completions/mean_length": 839.755615234375, + "completions/mean_terminated_length": 576.5850219726562, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.11195335276968, + "grad_norm": 0.1533268392086029, + "learning_rate": 1e-06, + "loss": -0.056, + "num_tokens": 902782122.0, + "reward": 0.7823660969734192, + "reward_std": 0.12156030535697937, + "rewards/verify_math_reward/mean": 0.7823660969734192, + "rewards/verify_math_reward/std": 0.41286757588386536, + "step": 1510 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 1012.7500610351562, + "completions/mean_terminated_length": 647.0711669921875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 14.12128279883382, + "grad_norm": 0.13378357887268066, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 903383978.0, + "reward": 0.684151828289032, + "reward_std": 0.11129778623580933, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.46511244773864746, + "step": 1511 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3779.0, + "completions/mean_length": 741.2745971679688, + "completions/mean_terminated_length": 572.1617431640625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 14.130612244897959, + "grad_norm": 0.16233642399311066, + "learning_rate": 1e-06, + "loss": -0.0198, + "num_tokens": 903952352.0, + "reward": 0.7767857313156128, + "reward_std": 0.11791250854730606, + "rewards/verify_math_reward/mean": 0.7767857313156128, + "rewards/verify_math_reward/std": 0.41663339734077454, + "step": 1512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3494.0, + "completions/mean_length": 879.114990234375, + "completions/mean_terminated_length": 606.4976196289062, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 14.139941690962099, + "grad_norm": 0.1455606073141098, + "learning_rate": 1e-06, + "loss": -0.0504, + "num_tokens": 904547167.0, + "reward": 0.691964328289032, + "reward_std": 0.12426057457923889, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1513 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4087.0, + "completions/mean_length": 1049.782470703125, + "completions/mean_terminated_length": 627.8792724609375, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 14.14927113702624, + "grad_norm": 0.14874128997325897, + "learning_rate": 1e-06, + "loss": -0.0455, + "num_tokens": 905135476.0, + "reward": 0.6506696939468384, + "reward_std": 0.1328292191028595, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1514 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 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": 4096.0, + "completions/max_terminated_length": 3750.0, + "completions/mean_length": 1090.1763916015625, + "completions/mean_terminated_length": 691.1732177734375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 14.15860058309038, + "grad_norm": 0.18435132503509521, + "learning_rate": 1e-06, + "loss": -0.0497, + "num_tokens": 905772634.0, + "reward": 0.6540178656578064, + "reward_std": 0.14120353758335114, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1515 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3794.0, + "completions/mean_length": 912.9397583007812, + "completions/mean_terminated_length": 605.1530151367188, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.167930029154519, + "grad_norm": 0.13334131240844727, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 906356308.0, + "reward": 0.7678571939468384, + "reward_std": 0.09795550256967545, + "rewards/verify_math_reward/mean": 0.7678571343421936, + "rewards/verify_math_reward/std": 0.422435462474823, + "step": 1516 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2613.0, + "completions/mean_length": 858.5201416015625, + "completions/mean_terminated_length": 588.4038696289062, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 14.177259475218658, + "grad_norm": 0.19183936715126038, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 906931110.0, + "reward": 0.7276785969734192, + "reward_std": 0.16093555092811584, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1517 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2864.0, + "completions/mean_length": 969.8605346679688, + "completions/mean_terminated_length": 629.3898315429688, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 14.186588921282798, + "grad_norm": 0.16061101853847504, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 907525097.0, + "reward": 0.6986607313156128, + "reward_std": 0.13906529545783997, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2864.0, + "completions/mean_length": 1060.1060791015625, + "completions/mean_terminated_length": 567.9053344726562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.19591836734694, + "grad_norm": 0.15217699110507965, + "learning_rate": 1e-06, + "loss": -0.0704, + "num_tokens": 908055944.0, + "reward": 0.6975446939468384, + "reward_std": 0.11550647020339966, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1519 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3806.0, + "completions/mean_length": 798.654052734375, + "completions/mean_terminated_length": 553.527587890625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 14.205247813411079, + "grad_norm": 0.1444699615240097, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 908606898.0, + "reward": 0.754464328289032, + "reward_std": 0.1216326504945755, + "rewards/verify_math_reward/mean": 0.7544642686843872, + "rewards/verify_math_reward/std": 0.4306447505950928, + "step": 1520 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3694.0, + "completions/mean_length": 974.63623046875, + "completions/mean_terminated_length": 582.5050048828125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 14.214577259475218, + "grad_norm": 0.14740106463432312, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 909155908.0, + "reward": 0.6941964626312256, + "reward_std": 0.10904286801815033, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1521 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3105.0, + "completions/mean_length": 1033.37060546875, + "completions/mean_terminated_length": 609.1943969726562, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 14.223906705539358, + "grad_norm": 0.20302633941173553, + "learning_rate": 1e-06, + "loss": -0.0707, + "num_tokens": 909716880.0, + "reward": 0.7042410969734192, + "reward_std": 0.15597687661647797, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1522 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4069.0, + "completions/mean_length": 1000.3370971679688, + "completions/mean_terminated_length": 671.6617431640625, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 14.2332361516035, + "grad_norm": 0.161456897854805, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 910349222.0, + "reward": 0.6674107313156128, + "reward_std": 0.14943771064281464, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140389680862427, + "step": 1523 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3216.0, + "completions/mean_length": 958.6094360351562, + "completions/mean_terminated_length": 612.6022338867188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 14.242565597667639, + "grad_norm": 0.16442476212978363, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 910932048.0, + "reward": 0.7053571939468384, + "reward_std": 0.14327649772167206, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1524 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3293.0, + "completions/mean_length": 951.3092041015625, + "completions/mean_terminated_length": 600.1649780273438, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 14.251895043731778, + "grad_norm": 0.16809825599193573, + "learning_rate": 1e-06, + "loss": -0.0596, + "num_tokens": 911516661.0, + "reward": 0.6930803656578064, + "reward_std": 0.11404222995042801, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 1525 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 981.1641235351562, + "completions/mean_terminated_length": 598.6403198242188, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 14.261224489795918, + "grad_norm": 0.18891319632530212, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 912084616.0, + "reward": 0.6696428656578064, + "reward_std": 0.13485799729824066, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1526 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 972.2254638671875, + "completions/mean_terminated_length": 623.4168701171875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 14.270553935860057, + "grad_norm": 0.18301506340503693, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 912684906.0, + "reward": 0.6629464626312256, + "reward_std": 0.14751701056957245, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1527 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1540178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4075.0, + "completions/mean_length": 1189.430908203125, + "completions/mean_terminated_length": 660.2665405273438, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.279883381924199, + "grad_norm": 0.1512364149093628, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 913278100.0, + "reward": 0.5803571939468384, + "reward_std": 0.12249071896076202, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 1528 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2563.0, + "completions/mean_length": 1008.7210083007812, + "completions/mean_terminated_length": 659.7241821289062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 14.289212827988338, + "grad_norm": 0.153366357088089, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 913889898.0, + "reward": 0.6707589626312256, + "reward_std": 0.14289602637290955, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1529 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2060.0, + "completions/mean_length": 814.7567138671875, + "completions/mean_terminated_length": 558.10107421875, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 14.298542274052478, + "grad_norm": 0.14239616692066193, + "learning_rate": 1e-06, + "loss": -0.0463, + "num_tokens": 914426248.0, + "reward": 0.7812500596046448, + "reward_std": 0.10562442243099213, + "rewards/verify_math_reward/mean": 0.78125, + "rewards/verify_math_reward/std": 0.41362953186035156, + "step": 1530 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 916.3092041015625, + "completions/mean_terminated_length": 604.5747680664062, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 14.307871720116617, + "grad_norm": 0.17770318686962128, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 915007085.0, + "reward": 0.7187500596046448, + "reward_std": 0.14004097878932953, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1531 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2443.0, + "completions/mean_length": 903.86279296875, + "completions/mean_terminated_length": 624.9381103515625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 14.317201166180759, + "grad_norm": 0.15887369215488434, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 915615378.0, + "reward": 0.7053571939468384, + "reward_std": 0.1247912049293518, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1532 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3424.0, + "completions/mean_length": 858.7053833007812, + "completions/mean_terminated_length": 580.101806640625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 14.326530612244898, + "grad_norm": 0.15962766110897064, + "learning_rate": 1e-06, + "loss": -0.052, + "num_tokens": 916174026.0, + "reward": 0.707589328289032, + "reward_std": 0.110549196600914, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 1062.739990234375, + "completions/mean_terminated_length": 638.2379150390625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 14.335860058309038, + "grad_norm": 0.18271981179714203, + "learning_rate": 1e-06, + "loss": -0.0876, + "num_tokens": 916764025.0, + "reward": 0.7131696939468384, + "reward_std": 0.15713942050933838, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3734.0, + "completions/mean_length": 1076.1942138671875, + "completions/mean_terminated_length": 627.0949096679688, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 14.345189504373177, + "grad_norm": 0.19564831256866455, + "learning_rate": 1e-06, + "loss": -0.0496, + "num_tokens": 917344279.0, + "reward": 0.660714328289032, + "reward_std": 0.15161803364753723, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1535 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3894.0, + "completions/mean_length": 888.4732666015625, + "completions/mean_terminated_length": 603.9660034179688, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.354518950437317, + "grad_norm": 0.17494149506092072, + "learning_rate": 1e-06, + "loss": -0.0561, + "num_tokens": 917924279.0, + "reward": 0.7399553656578064, + "reward_std": 0.15706387162208557, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1536 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2869.0, + "completions/mean_length": 925.4063110351562, + "completions/mean_terminated_length": 644.1749877929688, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 14.363848396501458, + "grad_norm": 0.18259690701961517, + "learning_rate": 1e-06, + "loss": -0.0323, + "num_tokens": 918553667.0, + "reward": 0.6316964626312256, + "reward_std": 0.1428646296262741, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1537 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 924.8928833007812, + "completions/mean_terminated_length": 618.261962890625, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 14.373177842565598, + "grad_norm": 0.1602715402841568, + "learning_rate": 1e-06, + "loss": -0.055, + "num_tokens": 919139235.0, + "reward": 0.7087053656578064, + "reward_std": 0.13508623838424683, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1538 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3708.0, + "completions/mean_length": 924.3326416015625, + "completions/mean_terminated_length": 613.3848266601562, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 14.382507288629737, + "grad_norm": 0.17896686494350433, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 919724333.0, + "reward": 0.6886160969734192, + "reward_std": 0.14522719383239746, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2709.0, + "completions/mean_length": 830.2935791015625, + "completions/mean_terminated_length": 523.2612915039062, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.391836734693877, + "grad_norm": 0.1672126203775406, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 920235484.0, + "reward": 0.7566964626312256, + "reward_std": 0.11637409776449203, + "rewards/verify_math_reward/mean": 0.7566964030265808, + "rewards/verify_math_reward/std": 0.42931652069091797, + "step": 1540 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2532.0, + "completions/mean_length": 965.9029541015625, + "completions/mean_terminated_length": 667.4340209960938, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 14.401166180758018, + "grad_norm": 0.15895900130271912, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 920876397.0, + "reward": 0.668526828289032, + "reward_std": 0.15277989208698273, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1541 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3454.0, + "completions/mean_length": 918.87060546875, + "completions/mean_terminated_length": 632.8515625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 14.410495626822158, + "grad_norm": 0.16459904611110687, + "learning_rate": 1e-06, + "loss": -0.045, + "num_tokens": 921477281.0, + "reward": 0.7087053656578064, + "reward_std": 0.1353151649236679, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1542 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3966.0, + "completions/mean_length": 880.33154296875, + "completions/mean_terminated_length": 628.8050537109375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.419825072886297, + "grad_norm": 0.18135885894298553, + "learning_rate": 1e-06, + "loss": -0.0407, + "num_tokens": 922081482.0, + "reward": 0.7265625596046448, + "reward_std": 0.11201275140047073, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1543 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2717.0, + "completions/mean_length": 1003.5569458007812, + "completions/mean_terminated_length": 606.2908935546875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 14.429154518950437, + "grad_norm": 0.18828460574150085, + "learning_rate": 1e-06, + "loss": -0.0547, + "num_tokens": 922665005.0, + "reward": 0.668526828289032, + "reward_std": 0.1460929960012436, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1544 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3698.0, + "completions/mean_length": 1003.732177734375, + "completions/mean_terminated_length": 597.6767578125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.438483965014576, + "grad_norm": 0.1579473912715912, + "learning_rate": 1e-06, + "loss": -0.0672, + "num_tokens": 923227429.0, + "reward": 0.676339328289032, + "reward_std": 0.1159176379442215, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 953.3582763671875, + "completions/mean_terminated_length": 641.0220947265625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.447813411078718, + "grad_norm": 0.18425829708576202, + "learning_rate": 1e-06, + "loss": -0.0258, + "num_tokens": 923832742.0, + "reward": 0.7064732313156128, + "reward_std": 0.1443602740764618, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1546 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3102.0, + "completions/mean_length": 905.3460083007812, + "completions/mean_terminated_length": 592.5368041992188, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.457142857142857, + "grad_norm": 0.1750928908586502, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 924410324.0, + "reward": 0.6517857313156128, + "reward_std": 0.14286212623119354, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 1000.3381958007812, + "completions/mean_terminated_length": 633.187255859375, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.466472303206997, + "grad_norm": 0.1698337346315384, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 925019963.0, + "reward": 0.7098214626312256, + "reward_std": 0.12264407426118851, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1548 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 913.950927734375, + "completions/mean_terminated_length": 614.7838745117188, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 14.475801749271136, + "grad_norm": 0.1929028183221817, + "learning_rate": 1e-06, + "loss": -0.0634, + "num_tokens": 925604751.0, + "reward": 0.7098214626312256, + "reward_std": 0.16585004329681396, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1549 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 939.6797485351562, + "completions/mean_terminated_length": 591.5848999023438, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 14.485131195335278, + "grad_norm": 0.1784702092409134, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 926181240.0, + "reward": 0.6785714626312256, + "reward_std": 0.1397392898797989, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1550 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3929.0, + "completions/mean_length": 997.9241333007812, + "completions/mean_terminated_length": 591.1060791015625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 14.494460641399417, + "grad_norm": 0.18176329135894775, + "learning_rate": 1e-06, + "loss": -0.0498, + "num_tokens": 926734540.0, + "reward": 0.691964328289032, + "reward_std": 0.1472894698381424, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1551 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3565.0, + "completions/mean_length": 1041.3248291015625, + "completions/mean_terminated_length": 622.6636962890625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 14.503790087463557, + "grad_norm": 0.16912685334682465, + "learning_rate": 1e-06, + "loss": -0.0436, + "num_tokens": 927319215.0, + "reward": 0.6718750596046448, + "reward_std": 0.11437922716140747, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1552 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3266.0, + "completions/mean_length": 1041.8248291015625, + "completions/mean_terminated_length": 587.6141357421875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.513119533527696, + "grad_norm": 0.15282073616981506, + "learning_rate": 1e-06, + "loss": -0.0698, + "num_tokens": 927866218.0, + "reward": 0.7243303656578064, + "reward_std": 0.11204735189676285, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1553 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4087.0, + "completions/mean_length": 1037.977783203125, + "completions/mean_terminated_length": 618.8578491210938, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.522448979591836, + "grad_norm": 0.18081989884376526, + "learning_rate": 1e-06, + "loss": -0.1079, + "num_tokens": 928444342.0, + "reward": 0.6886160969734192, + "reward_std": 0.17641356587409973, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2874.0, + "completions/mean_length": 944.5614013671875, + "completions/mean_terminated_length": 631.3508911132812, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 14.531778425655977, + "grad_norm": 0.12914735078811646, + "learning_rate": 1e-06, + "loss": -0.0338, + "num_tokens": 929057421.0, + "reward": 0.699776828289032, + "reward_std": 0.09551232308149338, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1555 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2666.0, + "completions/mean_length": 1052.9051513671875, + "completions/mean_terminated_length": 644.5911254882812, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.541107871720117, + "grad_norm": 0.15479591488838196, + "learning_rate": 1e-06, + "loss": -0.0366, + "num_tokens": 929661992.0, + "reward": 0.6484375, + "reward_std": 0.1129884421825409, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1556 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2691.0, + "completions/mean_length": 888.8516235351562, + "completions/mean_terminated_length": 591.6036376953125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 14.550437317784256, + "grad_norm": 0.1862279772758484, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 930233555.0, + "reward": 0.715401828289032, + "reward_std": 0.17844417691230774, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1557 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2884.0, + "completions/mean_length": 902.0379638671875, + "completions/mean_terminated_length": 536.5596923828125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.559766763848396, + "grad_norm": 0.16136878728866577, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 930744357.0, + "reward": 0.7455357313156128, + "reward_std": 0.11783722788095474, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1558 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2471.0, + "completions/mean_length": 893.32373046875, + "completions/mean_terminated_length": 575.0208740234375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 14.569096209912537, + "grad_norm": 0.1651744842529297, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 931303359.0, + "reward": 0.7042410969734192, + "reward_std": 0.10832468420267105, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1559 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2199.0, + "completions/mean_length": 1137.6942138671875, + "completions/mean_terminated_length": 635.6318359375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 14.578425655976677, + "grad_norm": 0.1699349582195282, + "learning_rate": 1e-06, + "loss": -0.0893, + "num_tokens": 931878421.0, + "reward": 0.668526828289032, + "reward_std": 0.1383524239063263, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1560 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2864.0, + "completions/mean_length": 910.4241333007812, + "completions/mean_terminated_length": 559.1028442382812, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 14.587755102040816, + "grad_norm": 0.14631903171539307, + "learning_rate": 1e-06, + "loss": -0.08, + "num_tokens": 932416937.0, + "reward": 0.7377232313156128, + "reward_std": 0.12249141931533813, + "rewards/verify_math_reward/mean": 0.7377232313156128, + "rewards/verify_math_reward/std": 0.4401180148124695, + "step": 1561 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2742.0, + "completions/mean_length": 929.83154296875, + "completions/mean_terminated_length": 585.001220703125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "epoch": 14.597084548104956, + "grad_norm": 0.1636786311864853, + "learning_rate": 1e-06, + "loss": -0.0436, + "num_tokens": 932983594.0, + "reward": 0.7109375596046448, + "reward_std": 0.13192768394947052, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2659.0, + "completions/mean_length": 814.4219360351562, + "completions/mean_terminated_length": 595.6500244140625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.606413994169095, + "grad_norm": 0.16003815829753876, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 933569260.0, + "reward": 0.7488839626312256, + "reward_std": 0.1385364830493927, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 1563 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3954.0, + "completions/mean_length": 935.6172485351562, + "completions/mean_terminated_length": 646.909912109375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 14.615743440233237, + "grad_norm": 0.1716504544019699, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 934189829.0, + "reward": 0.6863839626312256, + "reward_std": 0.1140076145529747, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1564 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3102.0, + "completions/mean_length": 865.536865234375, + "completions/mean_terminated_length": 570.4275512695312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 14.625072886297376, + "grad_norm": 0.17306075990200043, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 934746710.0, + "reward": 0.762276828289032, + "reward_std": 0.1145307719707489, + "rewards/verify_math_reward/mean": 0.7622767686843872, + "rewards/verify_math_reward/std": 0.42592647671699524, + "step": 1565 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3584.0, + "completions/mean_length": 925.8114013671875, + "completions/mean_terminated_length": 602.16357421875, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 14.634402332361516, + "grad_norm": 0.15716712176799774, + "learning_rate": 1e-06, + "loss": -0.0183, + "num_tokens": 935321117.0, + "reward": 0.7064732313156128, + "reward_std": 0.11750134825706482, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2317.0, + "completions/mean_length": 798.2042846679688, + "completions/mean_terminated_length": 540.25390625, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 14.643731778425655, + "grad_norm": 0.18043197691440582, + "learning_rate": 1e-06, + "loss": -0.0121, + "num_tokens": 935866140.0, + "reward": 0.7511160969734192, + "reward_std": 0.11475831270217896, + "rewards/verify_math_reward/mean": 0.7511160969734192, + "rewards/verify_math_reward/std": 0.43260788917541504, + "step": 1567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3016.0, + "completions/mean_length": 963.6607666015625, + "completions/mean_terminated_length": 635.3637084960938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 14.653061224489797, + "grad_norm": 0.17410822212696075, + "learning_rate": 1e-06, + "loss": -0.076, + "num_tokens": 936464028.0, + "reward": 0.723214328289032, + "reward_std": 0.1356828510761261, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1568 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3863.0, + "completions/mean_length": 1080.49560546875, + "completions/mean_terminated_length": 640.8951416015625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.662390670553936, + "grad_norm": 0.20159533619880676, + "learning_rate": 1e-06, + "loss": -0.0985, + "num_tokens": 937066184.0, + "reward": 0.6551339626312256, + "reward_std": 0.14684370160102844, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 1569 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2792.0, + "completions/mean_length": 949.8013916015625, + "completions/mean_terminated_length": 628.6027221679688, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 14.671720116618076, + "grad_norm": 0.14958810806274414, + "learning_rate": 1e-06, + "loss": -0.0319, + "num_tokens": 937664318.0, + "reward": 0.6729910969734192, + "reward_std": 0.1137365996837616, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1570 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3328.0, + "completions/mean_length": 987.4542846679688, + "completions/mean_terminated_length": 610.0713500976562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.681049562682215, + "grad_norm": 0.20107974112033844, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 938244053.0, + "reward": 0.7008928656578064, + "reward_std": 0.13354460895061493, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1571 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 978.37060546875, + "completions/mean_terminated_length": 568.98486328125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.690379008746355, + "grad_norm": 0.14747817814350128, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 938792185.0, + "reward": 0.7064732313156128, + "reward_std": 0.11490915715694427, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1572 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4042.0, + "completions/mean_length": 1109.232177734375, + "completions/mean_terminated_length": 669.439208984375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 14.699708454810496, + "grad_norm": 0.2102978229522705, + "learning_rate": 1e-06, + "loss": -0.0525, + "num_tokens": 939406257.0, + "reward": 0.590401828289032, + "reward_std": 0.16176220774650574, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 1573 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2735.0, + "completions/mean_length": 874.357177734375, + "completions/mean_terminated_length": 580.0536499023438, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.709037900874636, + "grad_norm": 0.14909407496452332, + "learning_rate": 1e-06, + "loss": -0.0576, + "num_tokens": 939970217.0, + "reward": 0.7533482313156128, + "reward_std": 0.11550647765398026, + "rewards/verify_math_reward/mean": 0.7533482313156128, + "rewards/verify_math_reward/std": 0.4313030242919922, + "step": 1574 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2710.0, + "completions/mean_length": 914.7422485351562, + "completions/mean_terminated_length": 589.9642944335938, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 14.718367346938775, + "grad_norm": 0.18469861149787903, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 940531434.0, + "reward": 0.699776828289032, + "reward_std": 0.17528842389583588, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1575 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2730.0, + "completions/mean_length": 960.1094360351562, + "completions/mean_terminated_length": 622.8751220703125, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 14.727696793002915, + "grad_norm": 0.1627260446548462, + "learning_rate": 1e-06, + "loss": -0.06, + "num_tokens": 941142028.0, + "reward": 0.6718750596046448, + "reward_std": 0.15450304746627808, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1576 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 917.2779541015625, + "completions/mean_terminated_length": 575.4375610351562, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 14.737026239067056, + "grad_norm": 0.18065008521080017, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 941706885.0, + "reward": 0.6953125596046448, + "reward_std": 0.14774522185325623, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3381.0, + "completions/mean_length": 726.5313110351562, + "completions/mean_terminated_length": 560.8196411132812, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 14.746355685131196, + "grad_norm": 0.2056432068347931, + "learning_rate": 1e-06, + "loss": -0.0211, + "num_tokens": 942266377.0, + "reward": 0.7723214626312256, + "reward_std": 0.15575115382671356, + "rewards/verify_math_reward/mean": 0.7723214030265808, + "rewards/verify_math_reward/std": 0.4195682406425476, + "step": 1578 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2808.0, + "completions/mean_length": 820.7221069335938, + "completions/mean_terminated_length": 581.4503173828125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 14.755685131195335, + "grad_norm": 0.16682174801826477, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 942846976.0, + "reward": 0.7522321939468384, + "reward_std": 0.11701097339391708, + "rewards/verify_math_reward/mean": 0.7522321343421936, + "rewards/verify_math_reward/std": 0.4319573938846588, + "step": 1579 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2814.0, + "completions/mean_length": 884.51904296875, + "completions/mean_terminated_length": 578.2897338867188, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 14.765014577259475, + "grad_norm": 0.19107991456985474, + "learning_rate": 1e-06, + "loss": -0.0656, + "num_tokens": 943410809.0, + "reward": 0.7343750596046448, + "reward_std": 0.14368949830532074, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 1580 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3842.0, + "completions/mean_length": 1076.5201416015625, + "completions/mean_terminated_length": 692.9132080078125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 14.774344023323614, + "grad_norm": 0.16536971926689148, + "learning_rate": 1e-06, + "loss": -0.0372, + "num_tokens": 944063211.0, + "reward": 0.6149553656578064, + "reward_std": 0.15631456673145294, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1581 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3346.0, + "completions/mean_length": 1020.96435546875, + "completions/mean_terminated_length": 586.1503295898438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 14.783673469387756, + "grad_norm": 0.1759994477033615, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 944615707.0, + "reward": 0.6383928656578064, + "reward_std": 0.13395646214485168, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1582 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2325.0, + "completions/mean_length": 819.0904541015625, + "completions/mean_terminated_length": 549.9722290039062, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 14.793002915451895, + "grad_norm": 0.14643263816833496, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 945155924.0, + "reward": 0.7823660969734192, + "reward_std": 0.09307490289211273, + "rewards/verify_math_reward/mean": 0.7823660969734192, + "rewards/verify_math_reward/std": 0.41286760568618774, + "step": 1583 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2789.0, + "completions/mean_length": 979.1194458007812, + "completions/mean_terminated_length": 600.7246704101562, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.802332361516035, + "grad_norm": 0.16532468795776367, + "learning_rate": 1e-06, + "loss": -0.0345, + "num_tokens": 945721887.0, + "reward": 0.6696428656578064, + "reward_std": 0.14079055190086365, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2755.0, + "completions/mean_length": 1057.6685791015625, + "completions/mean_terminated_length": 645.6260986328125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 14.811661807580174, + "grad_norm": 0.1641041487455368, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 946327278.0, + "reward": 0.6819196939468384, + "reward_std": 0.12936870753765106, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1585 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3809.0, + "completions/mean_length": 1001.3203735351562, + "completions/mean_terminated_length": 599.3631591796875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 14.820991253644316, + "grad_norm": 0.18239086866378784, + "learning_rate": 1e-06, + "loss": -0.0599, + "num_tokens": 946895229.0, + "reward": 0.7131696939468384, + "reward_std": 0.11761124432086945, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1586 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2862.0, + "completions/mean_length": 1020.6038208007812, + "completions/mean_terminated_length": 634.2474975585938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 14.830320699708455, + "grad_norm": 0.18540169298648834, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 947485266.0, + "reward": 0.6595982313156128, + "reward_std": 0.1345965713262558, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1587 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2403.0, + "completions/mean_length": 885.0614013671875, + "completions/mean_terminated_length": 570.2634887695312, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 14.839650145772595, + "grad_norm": 0.20089595019817352, + "learning_rate": 1e-06, + "loss": -0.0562, + "num_tokens": 948043145.0, + "reward": 0.7299107313156128, + "reward_std": 0.1475936770439148, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1588 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3955.0, + "completions/mean_length": 1085.3382568359375, + "completions/mean_terminated_length": 672.7093505859375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 14.848979591836734, + "grad_norm": 0.1703149527311325, + "learning_rate": 1e-06, + "loss": -0.0736, + "num_tokens": 948662048.0, + "reward": 0.6674107313156128, + "reward_std": 0.15221534669399261, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1589 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2960.0, + "completions/mean_length": 1021.3426513671875, + "completions/mean_terminated_length": 617.5997314453125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.858309037900874, + "grad_norm": 0.17019563913345337, + "learning_rate": 1e-06, + "loss": -0.0259, + "num_tokens": 949243971.0, + "reward": 0.6729910969734192, + "reward_std": 0.10833607614040375, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1590 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2421.0, + "completions/mean_length": 964.60498046875, + "completions/mean_terminated_length": 580.047607421875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.867638483965015, + "grad_norm": 0.17162199318408966, + "learning_rate": 1e-06, + "loss": -0.0483, + "num_tokens": 949795065.0, + "reward": 0.7120535969734192, + "reward_std": 0.12140876054763794, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530588984489441, + "step": 1591 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2732.0, + "completions/mean_length": 887.66748046875, + "completions/mean_terminated_length": 538.2450561523438, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 14.876967930029155, + "grad_norm": 0.22743330895900726, + "learning_rate": 1e-06, + "loss": -0.0586, + "num_tokens": 950317799.0, + "reward": 0.7410714626312256, + "reward_std": 0.13835172355175018, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829092383384705, + "step": 1592 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3222.0, + "completions/mean_length": 1014.8839721679688, + "completions/mean_terminated_length": 619.072998046875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 14.886297376093294, + "grad_norm": 0.15725521743297577, + "learning_rate": 1e-06, + "loss": -0.0822, + "num_tokens": 950905575.0, + "reward": 0.7031250596046448, + "reward_std": 0.1409093141555786, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1593 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 1017.2109985351562, + "completions/mean_terminated_length": 572.89013671875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 14.895626822157434, + "grad_norm": 0.15078191459178925, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 951446428.0, + "reward": 0.6428571939468384, + "reward_std": 0.09995036572217941, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1594 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3209.0, + "completions/mean_length": 972.05029296875, + "completions/mean_terminated_length": 592.7972412109375, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 14.904956268221575, + "grad_norm": 0.160890594124794, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 952010705.0, + "reward": 0.6696428656578064, + "reward_std": 0.11163367331027985, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 938.4029541015625, + "completions/mean_terminated_length": 641.5348510742188, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.914285714285715, + "grad_norm": 0.178226038813591, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 952621818.0, + "reward": 0.7008928656578064, + "reward_std": 0.1528201550245285, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1596 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3119.0, + "completions/mean_length": 895.9263916015625, + "completions/mean_terminated_length": 560.5302124023438, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 14.923615160349854, + "grad_norm": 0.1800370216369629, + "learning_rate": 1e-06, + "loss": -0.0674, + "num_tokens": 953158088.0, + "reward": 0.7488839626312256, + "reward_std": 0.1208428293466568, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 1597 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3897.0, + "completions/mean_length": 967.5960083007812, + "completions/mean_terminated_length": 618.2704467773438, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 14.932944606413994, + "grad_norm": 0.15556591749191284, + "learning_rate": 1e-06, + "loss": -0.0523, + "num_tokens": 953740398.0, + "reward": 0.7299107313156128, + "reward_std": 0.11644826829433441, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1598 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3027.0, + "completions/mean_length": 945.122802734375, + "completions/mean_terminated_length": 593.287841796875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 14.942274052478133, + "grad_norm": 0.18332384526729584, + "learning_rate": 1e-06, + "loss": -0.0304, + "num_tokens": 954311060.0, + "reward": 0.6953125596046448, + "reward_std": 0.10983241349458694, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1599 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2373.0, + "completions/mean_length": 1155.8070068359375, + "completions/mean_terminated_length": 634.2220458984375, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 14.951603498542275, + "grad_norm": 0.17527787387371063, + "learning_rate": 1e-06, + "loss": -0.0843, + "num_tokens": 954894103.0, + "reward": 0.6037946939468384, + "reward_std": 0.14135649800300598, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 1600 + }, + { + "epoch": 14.951603498542275, + "step": 1600, + "total_flos": 0.0, + "train_loss": -0.03323389737907974, + "train_runtime": 206541.0301, + "train_samples_per_second": 6.941, + "train_steps_per_second": 0.008 + } + ], + "logging_steps": 1, + "max_steps": 1600, + "num_input_tokens_seen": 954894103, + "num_train_epochs": 15, + "save_steps": 80, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +}