diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,2967 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.5599626691553896,
+  "eval_steps": 500,
+  "global_step": 450,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "completion_length": 503.56141090393066,
+      "epoch": 0.002488722974023954,
+      "grad_norm": 0.0948848358664139,
+      "kl": 0.0,
+      "learning_rate": 7.142857142857142e-08,
+      "loss": -0.0,
+      "reward": 0.302455373108387,
+      "reward_std": 0.4470902141183615,
+      "rewards/equation_reward_func": 0.05691964563447982,
+      "rewards/format_reward_func": 0.24553572107106447,
+      "step": 2
+    },
+    {
+      "completion_length": 496.7712287902832,
+      "epoch": 0.004977445948047908,
+      "grad_norm": 0.11039430968253656,
+      "kl": 0.0003851652145385742,
+      "learning_rate": 1.4285714285714285e-07,
+      "loss": 0.0,
+      "reward": 0.313616088591516,
+      "reward_std": 0.4562107492238283,
+      "rewards/equation_reward_func": 0.051339288242161274,
+      "rewards/format_reward_func": 0.26227679569274187,
+      "step": 4
+    },
+    {
+      "completion_length": 484.37056159973145,
+      "epoch": 0.007466168922071862,
+      "grad_norm": 0.09700064038560771,
+      "kl": 0.0003807544708251953,
+      "learning_rate": 2.1428571428571426e-07,
+      "loss": 0.0,
+      "reward": 0.31026787031441927,
+      "reward_std": 0.4778581503778696,
+      "rewards/equation_reward_func": 0.04464285960420966,
+      "rewards/format_reward_func": 0.2656250102445483,
+      "step": 6
+    },
+    {
+      "completion_length": 491.7064971923828,
+      "epoch": 0.009954891896095816,
+      "grad_norm": 0.11222944095605243,
+      "kl": 0.0003857612609863281,
+      "learning_rate": 2.857142857142857e-07,
+      "loss": 0.0,
+      "reward": 0.3526785857975483,
+      "reward_std": 0.48208368569612503,
+      "rewards/equation_reward_func": 0.06138393119908869,
+      "rewards/format_reward_func": 0.29129465483129025,
+      "step": 8
+    },
+    {
+      "completion_length": 481.55136489868164,
+      "epoch": 0.01244361487011977,
+      "grad_norm": 0.10010029385278177,
+      "kl": 0.0004470348358154297,
+      "learning_rate": 3.5714285714285716e-07,
+      "loss": 0.0,
+      "reward": 0.34933037497103214,
+      "reward_std": 0.48176000639796257,
+      "rewards/equation_reward_func": 0.06473214644938707,
+      "rewards/format_reward_func": 0.28459822945296764,
+      "step": 10
+    },
+    {
+      "completion_length": 482.3136405944824,
+      "epoch": 0.014932337844143724,
+      "grad_norm": 0.10991646335937744,
+      "kl": 0.0005280971527099609,
+      "learning_rate": 4.285714285714285e-07,
+      "loss": 0.0,
+      "reward": 0.36160716600716114,
+      "reward_std": 0.4869873467832804,
+      "rewards/equation_reward_func": 0.05245535890571773,
+      "rewards/format_reward_func": 0.3091518050059676,
+      "step": 12
+    },
+    {
+      "completion_length": 474.49109268188477,
+      "epoch": 0.017421060818167678,
+      "grad_norm": 0.09629390166600237,
+      "kl": 0.0010502338409423828,
+      "learning_rate": 5e-07,
+      "loss": 0.0,
+      "reward": 0.483258955180645,
+      "reward_std": 0.5381666142493486,
+      "rewards/equation_reward_func": 0.06250000209547579,
+      "rewards/format_reward_func": 0.4207589514553547,
+      "step": 14
+    },
+    {
+      "completion_length": 461.3962230682373,
+      "epoch": 0.019909783792191632,
+      "grad_norm": 0.10141033715590766,
+      "kl": 0.0015168190002441406,
+      "learning_rate": 4.999740409224932e-07,
+      "loss": 0.0,
+      "reward": 0.5424107331782579,
+      "reward_std": 0.5326656997203827,
+      "rewards/equation_reward_func": 0.058035717345774174,
+      "rewards/format_reward_func": 0.4843750186264515,
+      "step": 16
+    },
+    {
+      "completion_length": 460.481050491333,
+      "epoch": 0.022398506766215586,
+      "grad_norm": 0.07526369827636718,
+      "kl": 0.005082130432128906,
+      "learning_rate": 4.998961690809627e-07,
+      "loss": 0.0,
+      "reward": 0.7901786044239998,
+      "reward_std": 0.5360442269593477,
+      "rewards/equation_reward_func": 0.08035714610014111,
+      "rewards/format_reward_func": 0.709821455180645,
+      "step": 18
+    },
+    {
+      "completion_length": 452.7879695892334,
+      "epoch": 0.02488722974023954,
+      "grad_norm": 0.07040025423308008,
+      "kl": 0.00662994384765625,
+      "learning_rate": 4.997664006472578e-07,
+      "loss": 0.0,
+      "reward": 0.8392857536673546,
+      "reward_std": 0.497211379930377,
+      "rewards/equation_reward_func": 0.08928571699652821,
+      "rewards/format_reward_func": 0.7500000447034836,
+      "step": 20
+    },
+    {
+      "completion_length": 457.1462287902832,
+      "epoch": 0.027375952714263494,
+      "grad_norm": 0.0706145552092105,
+      "kl": 0.0069255828857421875,
+      "learning_rate": 4.995847625707292e-07,
+      "loss": 0.0,
+      "reward": 0.8738839700818062,
+      "reward_std": 0.4637769144028425,
+      "rewards/equation_reward_func": 0.08928571839351207,
+      "rewards/format_reward_func": 0.7845982611179352,
+      "step": 22
+    },
+    {
+      "completion_length": 440.41743087768555,
+      "epoch": 0.029864675688287448,
+      "grad_norm": 0.06311767376346097,
+      "kl": 0.0074863433837890625,
+      "learning_rate": 4.993512925726318e-07,
+      "loss": 0.0,
+      "reward": 0.9408482499420643,
+      "reward_std": 0.423699701204896,
+      "rewards/equation_reward_func": 0.09486607497092336,
+      "rewards/format_reward_func": 0.8459821790456772,
+      "step": 24
+    },
+    {
+      "completion_length": 469.8939914703369,
+      "epoch": 0.0323533986623114,
+      "grad_norm": 0.062583919157641,
+      "kl": 0.008693695068359375,
+      "learning_rate": 4.990660391382923e-07,
+      "loss": 0.0,
+      "reward": 0.9553571902215481,
+      "reward_std": 0.43422539345920086,
+      "rewards/equation_reward_func": 0.10044643352739513,
+      "rewards/format_reward_func": 0.8549107573926449,
+      "step": 26
+    },
+    {
+      "completion_length": 443.16185188293457,
+      "epoch": 0.034842121636335356,
+      "grad_norm": 0.06052609584603851,
+      "kl": 0.0123748779296875,
+      "learning_rate": 4.987290615070384e-07,
+      "loss": 0.0,
+      "reward": 1.041294701397419,
+      "reward_std": 0.3484124131500721,
+      "rewards/equation_reward_func": 0.12053572025615722,
+      "rewards/format_reward_func": 0.9207589663565159,
+      "step": 28
+    },
+    {
+      "completion_length": 423.43640327453613,
+      "epoch": 0.03733084461035931,
+      "grad_norm": 0.06438768655706453,
+      "kl": 0.037322998046875,
+      "learning_rate": 4.983404296598978e-07,
+      "loss": 0.0,
+      "reward": 1.0691964775323868,
+      "reward_std": 0.3513711765408516,
+      "rewards/equation_reward_func": 0.137276794295758,
+      "rewards/format_reward_func": 0.9319196827709675,
+      "step": 30
+    },
+    {
+      "completion_length": 430.5067195892334,
+      "epoch": 0.039819567584383264,
+      "grad_norm": 0.058402210713399266,
+      "kl": 0.016407012939453125,
+      "learning_rate": 4.979002243050646e-07,
+      "loss": 0.0,
+      "reward": 1.0636161379516125,
+      "reward_std": 0.3065283698961139,
+      "rewards/equation_reward_func": 0.11718750465661287,
+      "rewards/format_reward_func": 0.9464286081492901,
+      "step": 32
+    },
+    {
+      "completion_length": 406.2924289703369,
+      "epoch": 0.042308290558407215,
+      "grad_norm": 0.05357769148482046,
+      "kl": 0.018497467041015625,
+      "learning_rate": 4.974085368611381e-07,
+      "loss": 0.0,
+      "reward": 1.072544701397419,
+      "reward_std": 0.26174191664904356,
+      "rewards/equation_reward_func": 0.10937500570435077,
+      "rewards/format_reward_func": 0.9631696678698063,
+      "step": 34
+    },
+    {
+      "completion_length": 396.38394927978516,
+      "epoch": 0.04479701353243117,
+      "grad_norm": 0.052051460097072955,
+      "kl": 0.018001556396484375,
+      "learning_rate": 4.968654694381379e-07,
+      "loss": 0.0,
+      "reward": 1.0792411267757416,
+      "reward_std": 0.264671778306365,
+      "rewards/equation_reward_func": 0.11495536263100803,
+      "rewards/format_reward_func": 0.9642857499420643,
+      "step": 36
+    },
+    {
+      "completion_length": 377.9855155944824,
+      "epoch": 0.04728573650645512,
+      "grad_norm": 0.04517660218693028,
+      "kl": 0.02022552490234375,
+      "learning_rate": 4.962711348162987e-07,
+      "loss": 0.0,
+      "reward": 1.09933041036129,
+      "reward_std": 0.24636342655867338,
+      "rewards/equation_reward_func": 0.12388393352739513,
+      "rewards/format_reward_func": 0.9754464626312256,
+      "step": 38
+    },
+    {
+      "completion_length": 379.65738105773926,
+      "epoch": 0.04977445948047908,
+      "grad_norm": 0.05744479093406108,
+      "kl": 0.02159881591796875,
+      "learning_rate": 4.956256564226487e-07,
+      "loss": 0.0,
+      "reward": 1.0781250521540642,
+      "reward_std": 0.25441541336476803,
+      "rewards/equation_reward_func": 0.10825893341097981,
+      "rewards/format_reward_func": 0.9698661044239998,
+      "step": 40
+    },
+    {
+      "completion_length": 374.02791595458984,
+      "epoch": 0.05226318245450303,
+      "grad_norm": 0.058753380813439324,
+      "kl": 0.0220794677734375,
+      "learning_rate": 4.949291683053768e-07,
+      "loss": 0.0,
+      "reward": 1.1138393431901932,
+      "reward_std": 0.2881670519709587,
+      "rewards/equation_reward_func": 0.14285714994184673,
+      "rewards/format_reward_func": 0.9709821678698063,
+      "step": 42
+    },
+    {
+      "completion_length": 376.6361789703369,
+      "epoch": 0.05475190542852699,
+      "grad_norm": 0.04253532297873144,
+      "kl": 0.02449798583984375,
+      "learning_rate": 4.941818151059955e-07,
+      "loss": 0.0,
+      "reward": 1.0915179252624512,
+      "reward_std": 0.2087155058979988,
+      "rewards/equation_reward_func": 0.11272321897558868,
+      "rewards/format_reward_func": 0.9787946715950966,
+      "step": 44
+    },
+    {
+      "completion_length": 372.0290336608887,
+      "epoch": 0.05724062840255094,
+      "grad_norm": 0.055294299722758004,
+      "kl": 0.02376556396484375,
+      "learning_rate": 4.933837520293017e-07,
+      "loss": 0.0,
+      "reward": 1.1004464849829674,
+      "reward_std": 0.2569141434505582,
+      "rewards/equation_reward_func": 0.1250000069849193,
+      "rewards/format_reward_func": 0.9754464700818062,
+      "step": 46
+    },
+    {
+      "completion_length": 357.9397449493408,
+      "epoch": 0.059729351376574896,
+      "grad_norm": 0.05485318549956521,
+      "kl": 0.02727508544921875,
+      "learning_rate": 4.925351448111454e-07,
+      "loss": 0.0,
+      "reward": 1.1361607685685158,
+      "reward_std": 0.26225551683455706,
+      "rewards/equation_reward_func": 0.15848215075675398,
+      "rewards/format_reward_func": 0.9776785969734192,
+      "step": 48
+    },
+    {
+      "completion_length": 344.12947845458984,
+      "epoch": 0.06221807435059885,
+      "grad_norm": 0.06195502205989122,
+      "kl": 0.02738189697265625,
+      "learning_rate": 4.91636169684011e-07,
+      "loss": 0.0,
+      "reward": 1.1763393431901932,
+      "reward_std": 0.25935907755047083,
+      "rewards/equation_reward_func": 0.18415179545991123,
+      "rewards/format_reward_func": 0.9921875223517418,
+      "step": 50
+    },
+    {
+      "completion_length": 364.3404197692871,
+      "epoch": 0.0647067973246228,
+      "grad_norm": 0.053362806131777296,
+      "kl": 0.0282440185546875,
+      "learning_rate": 4.906870133404186e-07,
+      "loss": 0.0,
+      "reward": 1.150669701397419,
+      "reward_std": 0.2615376366302371,
+      "rewards/equation_reward_func": 0.16852679406292737,
+      "rewards/format_reward_func": 0.9821428917348385,
+      "step": 52
+    },
+    {
+      "completion_length": 349.52233695983887,
+      "epoch": 0.06719552029864675,
+      "grad_norm": 0.05603847653858231,
+      "kl": 0.0309600830078125,
+      "learning_rate": 4.896878728941531e-07,
+      "loss": 0.0,
+      "reward": 1.1640625670552254,
+      "reward_std": 0.25859395135194063,
+      "rewards/equation_reward_func": 0.1752232222352177,
+      "rewards/format_reward_func": 0.9888393133878708,
+      "step": 54
+    },
+    {
+      "completion_length": 357.9721164703369,
+      "epoch": 0.06968424327267071,
+      "grad_norm": 0.05692788775474882,
+      "kl": 0.0357513427734375,
+      "learning_rate": 4.886389558393284e-07,
+      "loss": 0.0,
+      "reward": 1.1830357536673546,
+      "reward_std": 0.29094040393829346,
+      "rewards/equation_reward_func": 0.19866072107106447,
+      "rewards/format_reward_func": 0.9843750409781933,
+      "step": 56
+    },
+    {
+      "completion_length": 356.3794803619385,
+      "epoch": 0.07217296624669467,
+      "grad_norm": 0.05896567816690595,
+      "kl": 0.03997802734375,
+      "learning_rate": 4.875404800072976e-07,
+      "loss": 0.0,
+      "reward": 1.1875000521540642,
+      "reward_std": 0.2795195486396551,
+      "rewards/equation_reward_func": 0.20200893841683865,
+      "rewards/format_reward_func": 0.9854910969734192,
+      "step": 58
+    },
+    {
+      "completion_length": 347.1930961608887,
+      "epoch": 0.07466168922071861,
+      "grad_norm": 0.050230804578636344,
+      "kl": 0.039276123046875,
+      "learning_rate": 4.86392673521415e-07,
+      "loss": 0.0,
+      "reward": 1.1830357685685158,
+      "reward_std": 0.25950881000608206,
+      "rewards/equation_reward_func": 0.19531250791624188,
+      "rewards/format_reward_func": 0.9877232350409031,
+      "step": 60
+    },
+    {
+      "completion_length": 337.48215675354004,
+      "epoch": 0.07715041219474257,
+      "grad_norm": 0.052408896199406675,
+      "kl": 0.0442047119140625,
+      "learning_rate": 4.851957747496606e-07,
+      "loss": 0.0,
+      "reward": 1.1841518469154835,
+      "reward_std": 0.2687631882727146,
+      "rewards/equation_reward_func": 0.20200893853325397,
+      "rewards/format_reward_func": 0.9821428805589676,
+      "step": 62
+    },
+    {
+      "completion_length": 345.3080520629883,
+      "epoch": 0.07963913516876653,
+      "grad_norm": 0.06294945230632734,
+      "kl": 0.0493011474609375,
+      "learning_rate": 4.839500322551386e-07,
+      "loss": 0.0,
+      "reward": 1.1986607611179352,
+      "reward_std": 0.28346721874549985,
+      "rewards/equation_reward_func": 0.20647322351578623,
+      "rewards/format_reward_func": 0.9921875223517418,
+      "step": 64
+    },
+    {
+      "completion_length": 350.44755935668945,
+      "epoch": 0.08212785814279049,
+      "grad_norm": 0.06086093842006542,
+      "kl": 0.050140380859375,
+      "learning_rate": 4.826557047444563e-07,
+      "loss": 0.0001,
+      "reward": 1.2064732760190964,
+      "reward_std": 0.30727679654955864,
+      "rewards/equation_reward_func": 0.22209822852164507,
+      "rewards/format_reward_func": 0.9843750298023224,
+      "step": 66
+    },
+    {
+      "completion_length": 331.2901954650879,
+      "epoch": 0.08461658111681443,
+      "grad_norm": 0.06362960710704729,
+      "kl": 0.0567474365234375,
+      "learning_rate": 4.813130610139993e-07,
+      "loss": 0.0001,
+      "reward": 1.2455357611179352,
+      "reward_std": 0.28113956190645695,
+      "rewards/equation_reward_func": 0.2566964430734515,
+      "rewards/format_reward_func": 0.9888393133878708,
+      "step": 68
+    },
+    {
+      "completion_length": 351.97880935668945,
+      "epoch": 0.08710530409083839,
+      "grad_norm": 0.06267945194538582,
+      "kl": 0.05755615234375,
+      "learning_rate": 4.799223798941089e-07,
+      "loss": 0.0001,
+      "reward": 1.1930804252624512,
+      "reward_std": 0.2554375780746341,
+      "rewards/equation_reward_func": 0.20758929662406445,
+      "rewards/format_reward_func": 0.9854910895228386,
+      "step": 70
+    },
+    {
+      "completion_length": 335.7288064956665,
+      "epoch": 0.08959402706486234,
+      "grad_norm": 0.06583286191782552,
+      "kl": 0.0636749267578125,
+      "learning_rate": 4.78483950191177e-07,
+      "loss": 0.0001,
+      "reward": 1.2488839849829674,
+      "reward_std": 0.31368037685751915,
+      "rewards/equation_reward_func": 0.26562501210719347,
+      "rewards/format_reward_func": 0.9832589663565159,
+      "step": 72
+    },
+    {
+      "completion_length": 359.6205539703369,
+      "epoch": 0.0920827500388863,
+      "grad_norm": 0.05887884685941389,
+      "kl": 0.067291259765625,
+      "learning_rate": 4.769980706276687e-07,
+      "loss": 0.0001,
+      "reward": 1.2299107685685158,
+      "reward_std": 0.33461445197463036,
+      "rewards/equation_reward_func": 0.26116072945296764,
+      "rewards/format_reward_func": 0.9687500298023224,
+      "step": 74
+    },
+    {
+      "completion_length": 356.4163131713867,
+      "epoch": 0.09457147301291025,
+      "grad_norm": 0.06235340328160086,
+      "kl": 0.074432373046875,
+      "learning_rate": 4.7546504978008595e-07,
+      "loss": 0.0001,
+      "reward": 1.3080357611179352,
+      "reward_std": 0.3593567367643118,
+      "rewards/equation_reward_func": 0.3303571594879031,
+      "rewards/format_reward_func": 0.977678619325161,
+      "step": 76
+    },
+    {
+      "completion_length": 375.3147506713867,
+      "epoch": 0.0970601959869342,
+      "grad_norm": 0.056324331796943045,
+      "kl": 0.07568359375,
+      "learning_rate": 4.738852060148848e-07,
+      "loss": 0.0001,
+      "reward": 1.2433036267757416,
+      "reward_std": 0.310510310344398,
+      "rewards/equation_reward_func": 0.26897322572767735,
+      "rewards/format_reward_func": 0.9743303880095482,
+      "step": 78
+    },
+    {
+      "completion_length": 378.6038131713867,
+      "epoch": 0.09954891896095816,
+      "grad_norm": 0.06544056788087135,
+      "kl": 0.079559326171875,
+      "learning_rate": 4.722588674223593e-07,
+      "loss": 0.0001,
+      "reward": 1.241071492433548,
+      "reward_std": 0.3315843231976032,
+      "rewards/equation_reward_func": 0.2767857280559838,
+      "rewards/format_reward_func": 0.9642857424914837,
+      "step": 80
+    },
+    {
+      "completion_length": 371.5558223724365,
+      "epoch": 0.10203764193498212,
+      "grad_norm": 0.06000338506612649,
+      "kl": 0.09228515625,
+      "learning_rate": 4.70586371748506e-07,
+      "loss": 0.0001,
+      "reward": 1.2734375521540642,
+      "reward_std": 0.32439365796744823,
+      "rewards/equation_reward_func": 0.3113839365541935,
+      "rewards/format_reward_func": 0.9620535932481289,
+      "step": 82
+    },
+    {
+      "completion_length": 360.88952445983887,
+      "epoch": 0.10452636490900606,
+      "grad_norm": 0.0662228531842243,
+      "kl": 0.095611572265625,
+      "learning_rate": 4.6886806632488363e-07,
+      "loss": 0.0001,
+      "reward": 1.306919701397419,
+      "reward_std": 0.3095084084197879,
+      "rewards/equation_reward_func": 0.3314732317812741,
+      "rewards/format_reward_func": 0.9754464700818062,
+      "step": 84
+    },
+    {
+      "completion_length": 379.3147487640381,
+      "epoch": 0.10701508788303002,
+      "grad_norm": 0.060174589846167865,
+      "kl": 0.093414306640625,
+      "learning_rate": 4.6710430799648143e-07,
+      "loss": 0.0001,
+      "reward": 1.2667411416769028,
+      "reward_std": 0.3022028226405382,
+      "rewards/equation_reward_func": 0.3035714430734515,
+      "rewards/format_reward_func": 0.9631696790456772,
+      "step": 86
+    },
+    {
+      "completion_length": 371.4241237640381,
+      "epoch": 0.10950381085705398,
+      "grad_norm": 0.058997581427704456,
+      "kl": 0.1055908203125,
+      "learning_rate": 4.652954630476127e-07,
+      "loss": 0.0001,
+      "reward": 1.3046875596046448,
+      "reward_std": 0.32442515436559916,
+      "rewards/equation_reward_func": 0.34040180407464504,
+      "rewards/format_reward_func": 0.9642857387661934,
+      "step": 88
+    },
+    {
+      "completion_length": 399.9319362640381,
+      "epoch": 0.11199253383107793,
+      "grad_norm": 0.06053909398263049,
+      "kl": 0.101837158203125,
+      "learning_rate": 4.6344190712584713e-07,
+      "loss": 0.0001,
+      "reward": 1.3102679252624512,
+      "reward_std": 0.3171618916094303,
+      "rewards/equation_reward_func": 0.34709823597222567,
+      "rewards/format_reward_func": 0.9631696753203869,
+      "step": 90
+    },
+    {
+      "completion_length": 408.24667167663574,
+      "epoch": 0.11448125680510188,
+      "grad_norm": 0.05672443722539237,
+      "kl": 0.103118896484375,
+      "learning_rate": 4.615440251639995e-07,
+      "loss": 0.0001,
+      "reward": 1.284598283469677,
+      "reward_std": 0.361779160797596,
+      "rewards/equation_reward_func": 0.34486608672887087,
+      "rewards/format_reward_func": 0.9397321753203869,
+      "step": 92
+    },
+    {
+      "completion_length": 415.2946605682373,
+      "epoch": 0.11696997977912583,
+      "grad_norm": 0.05701659561324134,
+      "kl": 0.10345458984375,
+      "learning_rate": 4.596022113001894e-07,
+      "loss": 0.0001,
+      "reward": 1.2979911342263222,
+      "reward_std": 0.32484395429491997,
+      "rewards/equation_reward_func": 0.3582589440047741,
+      "rewards/format_reward_func": 0.9397321753203869,
+      "step": 94
+    },
+    {
+      "completion_length": 399.2433166503906,
+      "epoch": 0.11945870275314979,
+      "grad_norm": 0.06305880755834348,
+      "kl": 0.118804931640625,
+      "learning_rate": 4.576168687959895e-07,
+      "loss": 0.0001,
+      "reward": 1.3872768357396126,
+      "reward_std": 0.3481318447738886,
+      "rewards/equation_reward_func": 0.4397321678698063,
+      "rewards/format_reward_func": 0.9475446678698063,
+      "step": 96
+    },
+    {
+      "completion_length": 426.3482360839844,
+      "epoch": 0.12194742572717375,
+      "grad_norm": 0.047999431553062445,
+      "kl": 0.10955810546875,
+      "learning_rate": 4.555884099526793e-07,
+      "loss": 0.0001,
+      "reward": 1.3370536267757416,
+      "reward_std": 0.29146868363022804,
+      "rewards/equation_reward_func": 0.3917410857975483,
+      "rewards/format_reward_func": 0.9453125335276127,
+      "step": 98
+    },
+    {
+      "completion_length": 422.6651945114136,
+      "epoch": 0.1244361487011977,
+      "grad_norm": 0.050797654511719595,
+      "kl": 0.111175537109375,
+      "learning_rate": 4.5351725602562174e-07,
+      "loss": 0.0001,
+      "reward": 1.3337054178118706,
+      "reward_std": 0.2792348740622401,
+      "rewards/equation_reward_func": 0.3816964514553547,
+      "rewards/format_reward_func": 0.9520089626312256,
+      "step": 100
+    },
+    {
+      "completion_length": 452.9520320892334,
+      "epoch": 0.12692487167522165,
+      "grad_norm": 0.05486761712388945,
+      "kl": 0.114166259765625,
+      "learning_rate": 4.514038371367791e-07,
+      "loss": 0.0001,
+      "reward": 1.2745536230504513,
+      "reward_std": 0.3251532930880785,
+      "rewards/equation_reward_func": 0.35602680617012084,
+      "rewards/format_reward_func": 0.9185268245637417,
+      "step": 102
+    },
+    {
+      "completion_length": 412.76118087768555,
+      "epoch": 0.1294135946492456,
+      "grad_norm": 0.05179336901472077,
+      "kl": 0.12225341796875,
+      "learning_rate": 4.4924859218538936e-07,
+      "loss": 0.0001,
+      "reward": 1.3984375521540642,
+      "reward_std": 0.2861457858234644,
+      "rewards/equation_reward_func": 0.4520089542493224,
+      "rewards/format_reward_func": 0.9464286118745804,
+      "step": 104
+    },
+    {
+      "completion_length": 404.86943435668945,
+      "epoch": 0.13190231762326957,
+      "grad_norm": 0.053536495931847755,
+      "kl": 0.123687744140625,
+      "learning_rate": 4.470519687568185e-07,
+      "loss": 0.0001,
+      "reward": 1.381696492433548,
+      "reward_std": 0.2666415721178055,
+      "rewards/equation_reward_func": 0.4274553805589676,
+      "rewards/format_reward_func": 0.9542411118745804,
+      "step": 106
+    },
+    {
+      "completion_length": 407.1082763671875,
+      "epoch": 0.1343910405972935,
+      "grad_norm": 0.05215816107862381,
+      "kl": 0.13446044921875,
+      "learning_rate": 4.4481442302960923e-07,
+      "loss": 0.0001,
+      "reward": 1.3805804252624512,
+      "reward_std": 0.3346856739372015,
+      "rewards/equation_reward_func": 0.43750002048909664,
+      "rewards/format_reward_func": 0.9430803917348385,
+      "step": 108
+    },
+    {
+      "completion_length": 414.9018039703369,
+      "epoch": 0.13687976357131748,
+      "grad_norm": 0.06646824899059828,
+      "kl": 0.138336181640625,
+      "learning_rate": 4.4253641968074505e-07,
+      "loss": 0.0001,
+      "reward": 1.3604911416769028,
+      "reward_std": 0.25776811968535185,
+      "rewards/equation_reward_func": 0.4051339514553547,
+      "rewards/format_reward_func": 0.9553571715950966,
+      "step": 110
+    },
+    {
+      "completion_length": 445.246675491333,
+      "epoch": 0.13936848654534142,
+      "grad_norm": 0.05260455171982223,
+      "kl": 0.124481201171875,
+      "learning_rate": 4.402184317891501e-07,
+      "loss": 0.0001,
+      "reward": 1.2991072088479996,
+      "reward_std": 0.2543983329087496,
+      "rewards/equation_reward_func": 0.3404018022119999,
+      "rewards/format_reward_func": 0.9587054029107094,
+      "step": 112
+    },
+    {
+      "completion_length": 386.75336170196533,
+      "epoch": 0.14185720951936537,
+      "grad_norm": 0.06276933769672523,
+      "kl": 0.12603759765625,
+      "learning_rate": 4.37860940737443e-07,
+      "loss": 0.0001,
+      "reward": 1.4441965073347092,
+      "reward_std": 0.27539225295186043,
+      "rewards/equation_reward_func": 0.4843750214204192,
+      "rewards/format_reward_func": 0.9598214626312256,
+      "step": 114
+    },
+    {
+      "completion_length": 409.517879486084,
+      "epoch": 0.14434593249338934,
+      "grad_norm": 0.046210489105058865,
+      "kl": 0.13018798828125,
+      "learning_rate": 4.354644361119671e-07,
+      "loss": 0.0001,
+      "reward": 1.3660714998841286,
+      "reward_std": 0.2605076776817441,
+      "rewards/equation_reward_func": 0.4185268050059676,
+      "rewards/format_reward_func": 0.9475446715950966,
+      "step": 116
+    },
+    {
+      "completion_length": 381.48215675354004,
+      "epoch": 0.14683465546741328,
+      "grad_norm": 0.04626076501226659,
+      "kl": 0.13671875,
+      "learning_rate": 4.3302941560111716e-07,
+      "loss": 0.0001,
+      "reward": 1.448660783469677,
+      "reward_std": 0.24480629619210958,
+      "rewards/equation_reward_func": 0.48437502048909664,
+      "rewards/format_reward_func": 0.9642857387661934,
+      "step": 118
+    },
+    {
+      "completion_length": 375.8002405166626,
+      "epoch": 0.14932337844143723,
+      "grad_norm": 0.064347259929319,
+      "kl": 0.132171630859375,
+      "learning_rate": 4.3055638489198236e-07,
+      "loss": 0.0001,
+      "reward": 1.4508929252624512,
+      "reward_std": 0.31021546106785536,
+      "rewards/equation_reward_func": 0.48660717345774174,
+      "rewards/format_reward_func": 0.964285746216774,
+      "step": 120
+    },
+    {
+      "completion_length": 395.19197845458984,
+      "epoch": 0.1518121014154612,
+      "grad_norm": 0.05844318115286761,
+      "kl": 0.141693115234375,
+      "learning_rate": 4.280458575653296e-07,
+      "loss": 0.0001,
+      "reward": 1.4040179029107094,
+      "reward_std": 0.2540714908391237,
+      "rewards/equation_reward_func": 0.4363839477300644,
+      "rewards/format_reward_func": 0.9676339514553547,
+      "step": 122
+    },
+    {
+      "completion_length": 358.8236789703369,
+      "epoch": 0.15430082438948514,
+      "grad_norm": 0.052114516892564085,
+      "kl": 0.14666748046875,
+      "learning_rate": 4.2549835498894665e-07,
+      "loss": 0.0001,
+      "reward": 1.504464365541935,
+      "reward_std": 0.2538653062656522,
+      "rewards/equation_reward_func": 0.5290178880095482,
+      "rewards/format_reward_func": 0.9754464663565159,
+      "step": 124
+    },
+    {
+      "completion_length": 393.2901954650879,
+      "epoch": 0.1567895473635091,
+      "grad_norm": 0.05823650722611721,
+      "kl": 0.162750244140625,
+      "learning_rate": 4.229144062093679e-07,
+      "loss": 0.0002,
+      "reward": 1.4051339849829674,
+      "reward_std": 0.2523822980001569,
+      "rewards/equation_reward_func": 0.4375000260770321,
+      "rewards/format_reward_func": 0.9676339589059353,
+      "step": 126
+    },
+    {
+      "completion_length": 405.56251525878906,
+      "epoch": 0.15927827033753306,
+      "grad_norm": 0.06351569565773,
+      "kl": 0.142913818359375,
+      "learning_rate": 4.2029454784200675e-07,
+      "loss": 0.0001,
+      "reward": 1.369419701397419,
+      "reward_std": 0.23465343983843923,
+      "rewards/equation_reward_func": 0.40736608766019344,
+      "rewards/format_reward_func": 0.9620536044239998,
+      "step": 128
+    },
+    {
+      "completion_length": 374.26675510406494,
+      "epoch": 0.161766993311557,
+      "grad_norm": 0.058600891047310276,
+      "kl": 0.141693115234375,
+      "learning_rate": 4.1763932395971433e-07,
+      "loss": 0.0001,
+      "reward": 1.452008992433548,
+      "reward_std": 0.2408639071509242,
+      "rewards/equation_reward_func": 0.4765625260770321,
+      "rewards/format_reward_func": 0.9754464589059353,
+      "step": 130
+    },
+    {
+      "completion_length": 358.2567148208618,
+      "epoch": 0.16425571628558097,
+      "grad_norm": 0.05496069479334609,
+      "kl": 0.1519775390625,
+      "learning_rate": 4.1494928597979117e-07,
+      "loss": 0.0002,
+      "reward": 1.500000074505806,
+      "reward_std": 0.21514854487031698,
+      "rewards/equation_reward_func": 0.5200893096625805,
+      "rewards/format_reward_func": 0.9799107536673546,
+      "step": 132
+    },
+    {
+      "completion_length": 352.5524682998657,
+      "epoch": 0.16674443925960492,
+      "grad_norm": 0.05248231582575076,
+      "kl": 0.16265869140625,
+      "learning_rate": 4.122249925494726e-07,
+      "loss": 0.0002,
+      "reward": 1.4966518506407738,
+      "reward_std": 0.215032197535038,
+      "rewards/equation_reward_func": 0.5089285932481289,
+      "rewards/format_reward_func": 0.9877232536673546,
+      "step": 134
+    },
+    {
+      "completion_length": 366.6317138671875,
+      "epoch": 0.16923316223362886,
+      "grad_norm": 0.05679988919891403,
+      "kl": 0.142303466796875,
+      "learning_rate": 4.094670094299131e-07,
+      "loss": 0.0001,
+      "reward": 1.444196492433548,
+      "reward_std": 0.2186202765442431,
+      "rewards/equation_reward_func": 0.4732143059372902,
+      "rewards/format_reward_func": 0.9709821753203869,
+      "step": 136
+    },
+    {
+      "completion_length": 359.1785888671875,
+      "epoch": 0.17172188520765283,
+      "grad_norm": 0.0670324226853738,
+      "kl": 0.156280517578125,
+      "learning_rate": 4.066759093786931e-07,
+      "loss": 0.0002,
+      "reward": 1.4252232685685158,
+      "reward_std": 0.1616944009438157,
+      "rewards/equation_reward_func": 0.43750002095475793,
+      "rewards/format_reward_func": 0.9877232499420643,
+      "step": 138
+    },
+    {
+      "completion_length": 383.5435447692871,
+      "epoch": 0.17421060818167677,
+      "grad_norm": 0.04649149273407711,
+      "kl": 0.14532470703125,
+      "learning_rate": 4.038522720308732e-07,
+      "loss": 0.0001,
+      "reward": 1.3526786342263222,
+      "reward_std": 0.16552409389987588,
+      "rewards/equation_reward_func": 0.36607144365552813,
+      "rewards/format_reward_func": 0.9866071566939354,
+      "step": 140
+    },
+    {
+      "completion_length": 372.57144927978516,
+      "epoch": 0.17669933115570072,
+      "grad_norm": 0.05097142037201984,
+      "kl": 0.22589111328125,
+      "learning_rate": 4.009966837786194e-07,
+      "loss": 0.0002,
+      "reward": 1.3928572311997414,
+      "reward_std": 0.20025141583755612,
+      "rewards/equation_reward_func": 0.40513395331799984,
+      "rewards/format_reward_func": 0.9877232536673546,
+      "step": 142
+    },
+    {
+      "completion_length": 356.07032585144043,
+      "epoch": 0.1791880541297247,
+      "grad_norm": 0.05501640500795023,
+      "kl": 0.153076171875,
+      "learning_rate": 3.981097376494259e-07,
+      "loss": 0.0002,
+      "reward": 1.4654018506407738,
+      "reward_std": 0.2332756402902305,
+      "rewards/equation_reward_func": 0.4787946678698063,
+      "rewards/format_reward_func": 0.9866071753203869,
+      "step": 144
+    },
+    {
+      "completion_length": 359.5089416503906,
+      "epoch": 0.18167677710374863,
+      "grad_norm": 0.03960504940392292,
+      "kl": 0.1627197265625,
+      "learning_rate": 3.951920331829592e-07,
+      "loss": 0.0002,
+      "reward": 1.3950893431901932,
+      "reward_std": 0.20504061924293637,
+      "rewards/equation_reward_func": 0.4140625176951289,
+      "rewards/format_reward_func": 0.9810268096625805,
+      "step": 146
+    },
+    {
+      "completion_length": 340.59376525878906,
+      "epoch": 0.1841655000777726,
+      "grad_norm": 0.0522427513812835,
+      "kl": 0.16888427734375,
+      "learning_rate": 3.922441763065506e-07,
+      "loss": 0.0002,
+      "reward": 1.4464286416769028,
+      "reward_std": 0.16819261107593775,
+      "rewards/equation_reward_func": 0.45982144959270954,
+      "rewards/format_reward_func": 0.9866071790456772,
+      "step": 148
+    },
+    {
+      "completion_length": 355.54354095458984,
+      "epoch": 0.18665422305179655,
+      "grad_norm": 0.05897953255292171,
+      "kl": 0.16357421875,
+      "learning_rate": 3.8926677920936093e-07,
+      "loss": 0.0002,
+      "reward": 1.453125074505806,
+      "reward_std": 0.22553037758916616,
+      "rewards/equation_reward_func": 0.47433038242161274,
+      "rewards/format_reward_func": 0.9787946678698063,
+      "step": 150
+    },
+    {
+      "completion_length": 330.8404178619385,
+      "epoch": 0.1891429460258205,
+      "grad_norm": 0.05627495445361396,
+      "kl": 0.17462158203125,
+      "learning_rate": 3.862604602152464e-07,
+      "loss": 0.0002,
+      "reward": 1.4352679178118706,
+      "reward_std": 0.19060763390734792,
+      "rewards/equation_reward_func": 0.45982144959270954,
+      "rewards/format_reward_func": 0.9754464700818062,
+      "step": 152
+    },
+    {
+      "completion_length": 338.2745656967163,
+      "epoch": 0.19163166899984446,
+      "grad_norm": 0.04910199366203511,
+      "kl": 0.16571044921875,
+      "learning_rate": 3.8322584365434934e-07,
+      "loss": 0.0002,
+      "reward": 1.4654018580913544,
+      "reward_std": 0.2223955481313169,
+      "rewards/equation_reward_func": 0.48214288242161274,
+      "rewards/format_reward_func": 0.9832589626312256,
+      "step": 154
+    },
+    {
+      "completion_length": 352.1216697692871,
+      "epoch": 0.1941203919738684,
+      "grad_norm": 0.05356327173771425,
+      "kl": 0.16558837890625,
+      "learning_rate": 3.8016355973344173e-07,
+      "loss": 0.0002,
+      "reward": 1.4308036416769028,
+      "reward_std": 0.18950919900089502,
+      "rewards/equation_reward_func": 0.45089287869632244,
+      "rewards/format_reward_func": 0.979910746216774,
+      "step": 156
+    },
+    {
+      "completion_length": 351.04577255249023,
+      "epoch": 0.19660911494789235,
+      "grad_norm": 0.06141645509073623,
+      "kl": 0.17547607421875,
+      "learning_rate": 3.7707424440504863e-07,
+      "loss": 0.0002,
+      "reward": 1.4363839998841286,
+      "reward_std": 0.21332781156525016,
+      "rewards/equation_reward_func": 0.45535716600716114,
+      "rewards/format_reward_func": 0.9810268171131611,
+      "step": 158
+    },
+    {
+      "completion_length": 320.54688930511475,
+      "epoch": 0.19909783792191632,
+      "grad_norm": 0.05781227739788321,
+      "kl": 0.18389892578125,
+      "learning_rate": 3.739585392353787e-07,
+      "loss": 0.0002,
+      "reward": 1.4341518580913544,
+      "reward_std": 0.20938572334125638,
+      "rewards/equation_reward_func": 0.4520089542493224,
+      "rewards/format_reward_func": 0.9821428991854191,
+      "step": 160
+    },
+    {
+      "completion_length": 340.25336360931396,
+      "epoch": 0.20158656089594026,
+      "grad_norm": 0.06477984340724512,
+      "kl": 0.18438720703125,
+      "learning_rate": 3.7081709127108767e-07,
+      "loss": 0.0002,
+      "reward": 1.3917411416769028,
+      "reward_std": 0.24160758964717388,
+      "rewards/equation_reward_func": 0.4319196632131934,
+      "rewards/format_reward_func": 0.9598214700818062,
+      "step": 162
+    },
+    {
+      "completion_length": 347.57479190826416,
+      "epoch": 0.20407528386996424,
+      "grad_norm": 0.07120213043946341,
+      "kl": 0.186279296875,
+      "learning_rate": 3.6765055290490513e-07,
+      "loss": 0.0002,
+      "reward": 1.373883992433548,
+      "reward_std": 0.21425227960571647,
+      "rewards/equation_reward_func": 0.4084821604192257,
+      "rewards/format_reward_func": 0.9654018245637417,
+      "step": 164
+    },
+    {
+      "completion_length": 342.25671195983887,
+      "epoch": 0.20656400684398818,
+      "grad_norm": 0.06700342546108162,
+      "kl": 0.1981201171875,
+      "learning_rate": 3.644595817401501e-07,
+      "loss": 0.0002,
+      "reward": 1.4140625521540642,
+      "reward_std": 0.24193571787327528,
+      "rewards/equation_reward_func": 0.45200894959270954,
+      "rewards/format_reward_func": 0.9620536081492901,
+      "step": 166
+    },
+    {
+      "completion_length": 340.312518119812,
+      "epoch": 0.20905272981801212,
+      "grad_norm": 0.08762287914289317,
+      "kl": 0.2086181640625,
+      "learning_rate": 3.6124484045416483e-07,
+      "loss": 0.0002,
+      "reward": 1.400669701397419,
+      "reward_std": 0.24582227924838662,
+      "rewards/equation_reward_func": 0.4453125214204192,
+      "rewards/format_reward_func": 0.9553571827709675,
+      "step": 168
+    },
+    {
+      "completion_length": 354.2154188156128,
+      "epoch": 0.2115414527920361,
+      "grad_norm": 0.05256734039133035,
+      "kl": 0.20733642578125,
+      "learning_rate": 3.580069966606949e-07,
+      "loss": 0.0002,
+      "reward": 1.3705357760190964,
+      "reward_std": 0.22536195535212755,
+      "rewards/equation_reward_func": 0.4185268005821854,
+      "rewards/format_reward_func": 0.9520089626312256,
+      "step": 170
+    },
+    {
+      "completion_length": 335.11161613464355,
+      "epoch": 0.21403017576606004,
+      "grad_norm": 0.06944613130624071,
+      "kl": 0.19964599609375,
+      "learning_rate": 3.547467227712444e-07,
+      "loss": 0.0002,
+      "reward": 1.4285714775323868,
+      "reward_std": 0.27147548273205757,
+      "rewards/equation_reward_func": 0.47991073317825794,
+      "rewards/format_reward_func": 0.9486607424914837,
+      "step": 172
+    },
+    {
+      "completion_length": 311.1305913925171,
+      "epoch": 0.21651889874008398,
+      "grad_norm": 0.08671053778094895,
+      "kl": 0.2025146484375,
+      "learning_rate": 3.5146469585543386e-07,
+      "loss": 0.0002,
+      "reward": 1.484375074505806,
+      "reward_std": 0.26195389963686466,
+      "rewards/equation_reward_func": 0.5368303805589676,
+      "rewards/format_reward_func": 0.9475446790456772,
+      "step": 174
+    },
+    {
+      "completion_length": 350.00894355773926,
+      "epoch": 0.21900762171410795,
+      "grad_norm": 0.06615215528616859,
+      "kl": 0.2117919921875,
+      "learning_rate": 3.481615975003922e-07,
+      "loss": 0.0002,
+      "reward": 1.3694197088479996,
+      "reward_std": 0.27513020671904087,
+      "rewards/equation_reward_func": 0.43191966973245144,
+      "rewards/format_reward_func": 0.9375000335276127,
+      "step": 176
+    },
+    {
+      "completion_length": 338.06139945983887,
+      "epoch": 0.2214963446881319,
+      "grad_norm": 0.054349578220732504,
+      "kl": 0.21484375,
+      "learning_rate": 3.448381136692089e-07,
+      "loss": 0.0002,
+      "reward": 1.429687574505806,
+      "reward_std": 0.2586075942963362,
+      "rewards/equation_reward_func": 0.48995537497103214,
+      "rewards/format_reward_func": 0.9397321790456772,
+      "step": 178
+    },
+    {
+      "completion_length": 359.5156421661377,
+      "epoch": 0.22398506766215587,
+      "grad_norm": 0.05374122628526808,
+      "kl": 0.20916748046875,
+      "learning_rate": 3.4149493455847897e-07,
+      "loss": 0.0002,
+      "reward": 1.3493304252624512,
+      "reward_std": 0.23989787977188826,
+      "rewards/equation_reward_func": 0.4095982275903225,
+      "rewards/format_reward_func": 0.9397321827709675,
+      "step": 180
+    },
+    {
+      "completion_length": 342.4587211608887,
+      "epoch": 0.2264737906361798,
+      "grad_norm": 0.06811808864322311,
+      "kl": 0.23675537109375,
+      "learning_rate": 3.3813275445496766e-07,
+      "loss": 0.0002,
+      "reward": 1.38839291036129,
+      "reward_std": 0.24801104608923197,
+      "rewards/equation_reward_func": 0.45089287497103214,
+      "rewards/format_reward_func": 0.937500037252903,
+      "step": 182
+    },
+    {
+      "completion_length": 345.0446605682373,
+      "epoch": 0.22896251361020376,
+      "grad_norm": 0.06987324492754955,
+      "kl": 0.22637939453125,
+      "learning_rate": 3.347522715914262e-07,
+      "loss": 0.0002,
+      "reward": 1.3917411267757416,
+      "reward_std": 0.2438207296654582,
+      "rewards/equation_reward_func": 0.45312501722946763,
+      "rewards/format_reward_func": 0.9386161118745804,
+      "step": 184
+    },
+    {
+      "completion_length": 307.60269260406494,
+      "epoch": 0.23145123658422773,
+      "grad_norm": 0.08559276579960963,
+      "kl": 0.22283935546875,
+      "learning_rate": 3.313541880015877e-07,
+      "loss": 0.0002,
+      "reward": 1.4631697162985802,
+      "reward_std": 0.27263635816052556,
+      "rewards/equation_reward_func": 0.5200893082655966,
+      "rewards/format_reward_func": 0.9430803954601288,
+      "step": 186
+    },
+    {
+      "completion_length": 333.93305015563965,
+      "epoch": 0.23393995955825167,
+      "grad_norm": 0.06886906420831151,
+      "kl": 0.21905517578125,
+      "learning_rate": 3.279392093743747e-07,
+      "loss": 0.0002,
+      "reward": 1.4029018431901932,
+      "reward_std": 0.26216446328908205,
+      "rewards/equation_reward_func": 0.47321430779993534,
+      "rewards/format_reward_func": 0.929687537252903,
+      "step": 188
+    },
+    {
+      "completion_length": 334.85827350616455,
+      "epoch": 0.2364286825322756,
+      "grad_norm": 0.054404156906439176,
+      "kl": 0.22149658203125,
+      "learning_rate": 3.245080449073459e-07,
+      "loss": 0.0002,
+      "reward": 1.400669701397419,
+      "reward_std": 0.22941731940954924,
+      "rewards/equation_reward_func": 0.45200895238667727,
+      "rewards/format_reward_func": 0.948660746216774,
+      "step": 190
+    },
+    {
+      "completion_length": 318.9564838409424,
+      "epoch": 0.23891740550629958,
+      "grad_norm": 0.07343897062684883,
+      "kl": 0.228271484375,
+      "learning_rate": 3.210614071594162e-07,
+      "loss": 0.0002,
+      "reward": 1.439732201397419,
+      "reward_std": 0.2519109146669507,
+      "rewards/equation_reward_func": 0.49330359511077404,
+      "rewards/format_reward_func": 0.9464286044239998,
+      "step": 192
+    },
+    {
+      "completion_length": 338.8672037124634,
+      "epoch": 0.24140612848032353,
+      "grad_norm": 0.061350768494073026,
+      "kl": 0.2132568359375,
+      "learning_rate": 3.1760001190287695e-07,
+      "loss": 0.0002,
+      "reward": 1.3816964887082577,
+      "reward_std": 0.23298237565904856,
+      "rewards/equation_reward_func": 0.4363839430734515,
+      "rewards/format_reward_func": 0.945312537252903,
+      "step": 194
+    },
+    {
+      "completion_length": 345.0937671661377,
+      "epoch": 0.2438948514543475,
+      "grad_norm": 0.06872187414027332,
+      "kl": 0.213134765625,
+      "learning_rate": 3.141245779747502e-07,
+      "loss": 0.0002,
+      "reward": 1.3638393431901932,
+      "reward_std": 0.239340515807271,
+      "rewards/equation_reward_func": 0.41294644493609667,
+      "rewards/format_reward_func": 0.9508928954601288,
+      "step": 196
+    },
+    {
+      "completion_length": 310.77680015563965,
+      "epoch": 0.24638357442837144,
+      "grad_norm": 5.443595942007072,
+      "kl": 7.49005126953125,
+      "learning_rate": 3.106358271275056e-07,
+      "loss": 0.0075,
+      "reward": 1.4464286342263222,
+      "reward_std": 0.2085065320134163,
+      "rewards/equation_reward_func": 0.48995538614690304,
+      "rewards/format_reward_func": 0.9564732499420643,
+      "step": 198
+    },
+    {
+      "completion_length": 312.64845085144043,
+      "epoch": 0.2488722974023954,
+      "grad_norm": 0.06598180406690658,
+      "kl": 0.22442626953125,
+      "learning_rate": 3.0713448387917227e-07,
+      "loss": 0.0002,
+      "reward": 1.4129464961588383,
+      "reward_std": 0.23078837245702744,
+      "rewards/equation_reward_func": 0.45535716926679015,
+      "rewards/format_reward_func": 0.9575893171131611,
+      "step": 200
+    },
+    {
+      "completion_length": 312.81363010406494,
+      "epoch": 0.25136102037641933,
+      "grad_norm": 0.4754247351711236,
+      "kl": 0.2811279296875,
+      "learning_rate": 3.0362127536287636e-07,
+      "loss": 0.0003,
+      "reward": 1.4542411416769028,
+      "reward_std": 0.23245403449982405,
+      "rewards/equation_reward_func": 0.5022321697324514,
+      "rewards/format_reward_func": 0.9520089626312256,
+      "step": 202
+    },
+    {
+      "completion_length": 314.4732246398926,
+      "epoch": 0.2538497433504433,
+      "grad_norm": 0.07156843273118213,
+      "kl": 0.25048828125,
+      "learning_rate": 3.0009693117583523e-07,
+      "loss": 0.0003,
+      "reward": 1.4062500596046448,
+      "reward_std": 0.20789815485477448,
+      "rewards/equation_reward_func": 0.45089287497103214,
+      "rewards/format_reward_func": 0.9553571864962578,
+      "step": 204
+    },
+    {
+      "completion_length": 326.76229190826416,
+      "epoch": 0.2563384663244673,
+      "grad_norm": 0.06704094931913333,
+      "kl": 0.20635986328125,
+      "learning_rate": 2.965621832278401e-07,
+      "loss": 0.0002,
+      "reward": 1.4084822088479996,
+      "reward_std": 0.24688747199252248,
+      "rewards/equation_reward_func": 0.46428573690354824,
+      "rewards/format_reward_func": 0.9441964663565159,
+      "step": 206
+    },
+    {
+      "completion_length": 319.1551504135132,
+      "epoch": 0.2588271892984912,
+      "grad_norm": 0.06458932073742,
+      "kl": 0.21240234375,
+      "learning_rate": 2.9301776558925875e-07,
+      "loss": 0.0002,
+      "reward": 1.4196429327130318,
+      "reward_std": 0.24886887054890394,
+      "rewards/equation_reward_func": 0.4698660932481289,
+      "rewards/format_reward_func": 0.9497768171131611,
+      "step": 208
+    },
+    {
+      "completion_length": 330.76676177978516,
+      "epoch": 0.26131591227251516,
+      "grad_norm": 0.06508155230407775,
+      "kl": 0.21636962890625,
+      "learning_rate": 2.894644143385885e-07,
+      "loss": 0.0002,
+      "reward": 1.3638393506407738,
+      "reward_std": 0.26955581549555063,
+      "rewards/equation_reward_func": 0.42075895611196756,
+      "rewards/format_reward_func": 0.9430803917348385,
+      "step": 210
+    },
+    {
+      "completion_length": 291.4464406967163,
+      "epoch": 0.26380463524653913,
+      "grad_norm": 0.0833854899576959,
+      "kl": 0.226318359375,
+      "learning_rate": 2.859028674095937e-07,
+      "loss": 0.0002,
+      "reward": 1.4408482760190964,
+      "reward_std": 0.22698505548760295,
+      "rewards/equation_reward_func": 0.49665181897580624,
+      "rewards/format_reward_func": 0.9441964626312256,
+      "step": 212
+    },
+    {
+      "completion_length": 309.15961360931396,
+      "epoch": 0.26629335822056305,
+      "grad_norm": 0.05103001730987287,
+      "kl": 0.21319580078125,
+      "learning_rate": 2.823338644380566e-07,
+      "loss": 0.0002,
+      "reward": 1.4430804327130318,
+      "reward_std": 0.2395549127832055,
+      "rewards/equation_reward_func": 0.497767879627645,
+      "rewards/format_reward_func": 0.945312537252903,
+      "step": 214
+    },
+    {
+      "completion_length": 310.4118423461914,
+      "epoch": 0.268782081194587,
+      "grad_norm": 0.07403319863620571,
+      "kl": 0.22412109375,
+      "learning_rate": 2.7875814660817504e-07,
+      "loss": 0.0002,
+      "reward": 1.386160783469677,
+      "reward_std": 0.24659961089491844,
+      "rewards/equation_reward_func": 0.43973216135054827,
+      "rewards/format_reward_func": 0.9464286044239998,
+      "step": 216
+    },
+    {
+      "completion_length": 305.60938930511475,
+      "epoch": 0.271270804168611,
+      "grad_norm": 0.05773673792340038,
+      "kl": 0.213623046875,
+      "learning_rate": 2.751764564986396e-07,
+      "loss": 0.0002,
+      "reward": 1.3872768506407738,
+      "reward_std": 0.19503082614392042,
+      "rewards/equation_reward_func": 0.4308035932481289,
+      "rewards/format_reward_func": 0.9564732499420643,
+      "step": 218
+    },
+    {
+      "completion_length": 290.44420623779297,
+      "epoch": 0.27375952714263496,
+      "grad_norm": 0.05392804528737292,
+      "kl": 0.2159423828125,
+      "learning_rate": 2.715895379284194e-07,
+      "loss": 0.0002,
+      "reward": 1.4475447162985802,
+      "reward_std": 0.20527297724038363,
+      "rewards/equation_reward_func": 0.49776787776499987,
+      "rewards/format_reward_func": 0.9497768208384514,
+      "step": 220
+    },
+    {
+      "completion_length": 305.9788074493408,
+      "epoch": 0.2762482501166589,
+      "grad_norm": 0.05980760400922627,
+      "kl": 0.216796875,
+      "learning_rate": 2.6799813580229174e-07,
+      "loss": 0.0002,
+      "reward": 1.4207590073347092,
+      "reward_std": 0.22403404489159584,
+      "rewards/equation_reward_func": 0.4687500260770321,
+      "rewards/format_reward_func": 0.9520089663565159,
+      "step": 222
+    },
+    {
+      "completion_length": 272.02791595458984,
+      "epoch": 0.27873697309068285,
+      "grad_norm": 0.08303222816911335,
+      "kl": 0.24041748046875,
+      "learning_rate": 2.6440299595614606e-07,
+      "loss": 0.0002,
+      "reward": 1.457589365541935,
+      "reward_std": 0.23546006111428142,
+      "rewards/equation_reward_func": 0.5044643115252256,
+      "rewards/format_reward_func": 0.953125037252903,
+      "step": 224
+    },
+    {
+      "completion_length": 280.05023670196533,
+      "epoch": 0.2812256960647068,
+      "grad_norm": 0.05404811960392472,
+      "kl": 0.2314453125,
+      "learning_rate": 2.6080486500209347e-07,
+      "loss": 0.0002,
+      "reward": 1.5022322237491608,
+      "reward_std": 0.22998644690960646,
+      "rewards/equation_reward_func": 0.5479910960420966,
+      "rewards/format_reward_func": 0.9542410969734192,
+      "step": 226
+    },
+    {
+      "completion_length": 304.1093921661377,
+      "epoch": 0.28371441903873074,
+      "grad_norm": 0.06300739916900569,
+      "kl": 0.241943359375,
+      "learning_rate": 2.572044901734166e-07,
+      "loss": 0.0002,
+      "reward": 1.3872768469154835,
+      "reward_std": 0.2272398821078241,
+      "rewards/equation_reward_func": 0.4408482362050563,
+      "rewards/format_reward_func": 0.9464286118745804,
+      "step": 228
+    },
+    {
+      "completion_length": 264.1930923461914,
+      "epoch": 0.2862031420127547,
+      "grad_norm": 0.06856109499172447,
+      "kl": 0.22003173828125,
+      "learning_rate": 2.536026191693893e-07,
+      "loss": 0.0002,
+      "reward": 1.5446429327130318,
+      "reward_std": 0.23560786852613091,
+      "rewards/equation_reward_func": 0.5837053805589676,
+      "rewards/format_reward_func": 0.9609375298023224,
+      "step": 230
+    },
+    {
+      "completion_length": 280.08818531036377,
+      "epoch": 0.2886918649867787,
+      "grad_norm": 0.06460311962128547,
+      "kl": 0.22894287109375,
+      "learning_rate": 2.5e-07,
+      "loss": 0.0002,
+      "reward": 1.4776786416769028,
+      "reward_std": 0.25147050712257624,
+      "rewards/equation_reward_func": 0.5334821678698063,
+      "rewards/format_reward_func": 0.9441964663565159,
+      "step": 232
+    },
+    {
+      "completion_length": 291.1105031967163,
+      "epoch": 0.2911805879608026,
+      "grad_norm": 0.06575064642151073,
+      "kl": 0.2237548828125,
+      "learning_rate": 2.4639738083061073e-07,
+      "loss": 0.0002,
+      "reward": 1.3683036118745804,
+      "reward_std": 0.23962750937789679,
+      "rewards/equation_reward_func": 0.4263393087312579,
+      "rewards/format_reward_func": 0.9419643133878708,
+      "step": 234
+    },
+    {
+      "completion_length": 289.46653175354004,
+      "epoch": 0.29366931093482657,
+      "grad_norm": 0.06296182617357601,
+      "kl": 0.21917724609375,
+      "learning_rate": 2.4279550982658345e-07,
+      "loss": 0.0002,
+      "reward": 1.421875074505806,
+      "reward_std": 0.23982096277177334,
+      "rewards/equation_reward_func": 0.4698660895228386,
+      "rewards/format_reward_func": 0.9520089663565159,
+      "step": 236
+    },
+    {
+      "completion_length": 277.73662281036377,
+      "epoch": 0.29615803390885054,
+      "grad_norm": 0.06887502467667868,
+      "kl": 0.21881103515625,
+      "learning_rate": 2.3919513499790646e-07,
+      "loss": 0.0002,
+      "reward": 1.4430804252624512,
+      "reward_std": 0.23861057311296463,
+      "rewards/equation_reward_func": 0.47991072945296764,
+      "rewards/format_reward_func": 0.9631696715950966,
+      "step": 238
+    },
+    {
+      "completion_length": 298.07479190826416,
+      "epoch": 0.29864675688287445,
+      "grad_norm": 0.06290785295542843,
+      "kl": 0.22564697265625,
+      "learning_rate": 2.3559700404385394e-07,
+      "loss": 0.0002,
+      "reward": 1.3906250596046448,
+      "reward_std": 0.27012846898287535,
+      "rewards/equation_reward_func": 0.4520089514553547,
+      "rewards/format_reward_func": 0.9386161081492901,
+      "step": 240
+    },
+    {
+      "completion_length": 267.7533597946167,
+      "epoch": 0.3011354798568984,
+      "grad_norm": 0.059074395935352374,
+      "kl": 0.25567626953125,
+      "learning_rate": 2.3200186419770823e-07,
+      "loss": 0.0003,
+      "reward": 1.48214291036129,
+      "reward_std": 0.19563588872551918,
+      "rewards/equation_reward_func": 0.5178571660071611,
+      "rewards/format_reward_func": 0.9642857424914837,
+      "step": 242
+    },
+    {
+      "completion_length": 270.7366199493408,
+      "epoch": 0.3036242028309224,
+      "grad_norm": 0.0709982715482787,
+      "kl": 0.23052978515625,
+      "learning_rate": 2.284104620715807e-07,
+      "loss": 0.0002,
+      "reward": 1.4497768431901932,
+      "reward_std": 0.1896849451586604,
+      "rewards/equation_reward_func": 0.48325894959270954,
+      "rewards/format_reward_func": 0.9665178917348385,
+      "step": 244
+    },
+    {
+      "completion_length": 264.1417541503906,
+      "epoch": 0.3061129258049463,
+      "grad_norm": 0.06526960275231504,
+      "kl": 0.2496337890625,
+      "learning_rate": 2.2482354350136043e-07,
+      "loss": 0.0002,
+      "reward": 1.4386161267757416,
+      "reward_std": 0.1712889587506652,
+      "rewards/equation_reward_func": 0.47767859511077404,
+      "rewards/format_reward_func": 0.9609375260770321,
+      "step": 246
+    },
+    {
+      "completion_length": 297.9285840988159,
+      "epoch": 0.3086016487789703,
+      "grad_norm": 0.0739927662660479,
+      "kl": 0.22344970703125,
+      "learning_rate": 2.2124185339182496e-07,
+      "loss": 0.0002,
+      "reward": 1.3694197088479996,
+      "reward_std": 0.2540249917656183,
+      "rewards/equation_reward_func": 0.4285714498255402,
+      "rewards/format_reward_func": 0.9408482424914837,
+      "step": 248
+    },
+    {
+      "completion_length": 263.3125104904175,
+      "epoch": 0.31109037175299425,
+      "grad_norm": 0.07715164326379821,
+      "kl": 0.22210693359375,
+      "learning_rate": 2.1766613556194344e-07,
+      "loss": 0.0002,
+      "reward": 1.4218750670552254,
+      "reward_std": 0.2036459231749177,
+      "rewards/equation_reward_func": 0.45424109511077404,
+      "rewards/format_reward_func": 0.967633955180645,
+      "step": 250
+    },
+    {
+      "completion_length": 256.2232255935669,
+      "epoch": 0.3135790947270182,
+      "grad_norm": 0.061273497621395105,
+      "kl": 0.24237060546875,
+      "learning_rate": 2.1409713259040628e-07,
+      "loss": 0.0002,
+      "reward": 1.4832589849829674,
+      "reward_std": 0.19279880495741963,
+      "rewards/equation_reward_func": 0.5111607387661934,
+      "rewards/format_reward_func": 0.9720982424914837,
+      "step": 252
+    },
+    {
+      "completion_length": 251.2701005935669,
+      "epoch": 0.31606781770104214,
+      "grad_norm": 0.06269540097192026,
+      "kl": 0.22998046875,
+      "learning_rate": 2.105355856614115e-07,
+      "loss": 0.0002,
+      "reward": 1.5000000596046448,
+      "reward_std": 0.181473882868886,
+      "rewards/equation_reward_func": 0.530133955180645,
+      "rewards/format_reward_func": 0.9698661006987095,
+      "step": 254
+    },
+    {
+      "completion_length": 270.8817081451416,
+      "epoch": 0.3185565406750661,
+      "grad_norm": 0.05361492662421592,
+      "kl": 0.212646484375,
+      "learning_rate": 2.069822344107413e-07,
+      "loss": 0.0002,
+      "reward": 1.4854911491274834,
+      "reward_std": 0.22021894436329603,
+      "rewards/equation_reward_func": 0.5234375316649675,
+      "rewards/format_reward_func": 0.9620536081492901,
+      "step": 256
+    },
+    {
+      "completion_length": 261.3270215988159,
+      "epoch": 0.3210452636490901,
+      "grad_norm": 0.0533864967760835,
+      "kl": 0.22186279296875,
+      "learning_rate": 2.034378167721599e-07,
+      "loss": 0.0002,
+      "reward": 1.4832590073347092,
+      "reward_std": 0.16471941815689206,
+      "rewards/equation_reward_func": 0.5000000186264515,
+      "rewards/format_reward_func": 0.9832589514553547,
+      "step": 258
+    },
+    {
+      "completion_length": 258.474347114563,
+      "epoch": 0.323533986623114,
+      "grad_norm": 0.06040063699395629,
+      "kl": 0.2305908203125,
+      "learning_rate": 1.9990306882416485e-07,
+      "loss": 0.0002,
+      "reward": 1.452008992433548,
+      "reward_std": 0.1812770888209343,
+      "rewards/equation_reward_func": 0.4877232313156128,
+      "rewards/format_reward_func": 0.964285746216774,
+      "step": 260
+    },
+    {
+      "completion_length": 294.9553689956665,
+      "epoch": 0.32602270959713797,
+      "grad_norm": 0.06320814691259159,
+      "kl": 0.21978759765625,
+      "learning_rate": 1.9637872463712362e-07,
+      "loss": 0.0002,
+      "reward": 1.4263393506407738,
+      "reward_std": 0.2837945078499615,
+      "rewards/equation_reward_func": 0.49107145611196756,
+      "rewards/format_reward_func": 0.9352678917348385,
+      "step": 262
+    },
+    {
+      "completion_length": 264.8705463409424,
+      "epoch": 0.32851143257116194,
+      "grad_norm": 0.058570436981231384,
+      "kl": 0.22821044921875,
+      "learning_rate": 1.9286551612082773e-07,
+      "loss": 0.0002,
+      "reward": 1.41964291036129,
+      "reward_std": 0.20483079506084323,
+      "rewards/equation_reward_func": 0.4631696669384837,
+      "rewards/format_reward_func": 0.9564732499420643,
+      "step": 264
+    },
+    {
+      "completion_length": 260.60046005249023,
+      "epoch": 0.33100015554518586,
+      "grad_norm": 0.06631119975940275,
+      "kl": 0.2242431640625,
+      "learning_rate": 1.8936417287249446e-07,
+      "loss": 0.0002,
+      "reward": 1.4799107909202576,
+      "reward_std": 0.18420447735115886,
+      "rewards/equation_reward_func": 0.5212053805589676,
+      "rewards/format_reward_func": 0.9587053917348385,
+      "step": 266
+    },
+    {
+      "completion_length": 274.5435400009155,
+      "epoch": 0.33348887851920983,
+      "grad_norm": 0.05960228260776588,
+      "kl": 0.21954345703125,
+      "learning_rate": 1.8587542202524985e-07,
+      "loss": 0.0002,
+      "reward": 1.3984375596046448,
+      "reward_std": 0.19745240919291973,
+      "rewards/equation_reward_func": 0.4375000218860805,
+      "rewards/format_reward_func": 0.9609375223517418,
+      "step": 268
+    },
+    {
+      "completion_length": 299.59822940826416,
+      "epoch": 0.3359776014932338,
+      "grad_norm": 0.06150696597576238,
+      "kl": 0.22930908203125,
+      "learning_rate": 1.82399988097123e-07,
+      "loss": 0.0002,
+      "reward": 1.3482143431901932,
+      "reward_std": 0.23468840029090643,
+      "rewards/equation_reward_func": 0.40178573317825794,
+      "rewards/format_reward_func": 0.9464286044239998,
+      "step": 270
+    },
+    {
+      "completion_length": 236.88170623779297,
+      "epoch": 0.3384663244672577,
+      "grad_norm": 0.061559128134109485,
+      "kl": 0.22576904296875,
+      "learning_rate": 1.7893859284058378e-07,
+      "loss": 0.0002,
+      "reward": 1.5546875670552254,
+      "reward_std": 0.17562847631052136,
+      "rewards/equation_reward_func": 0.5837053880095482,
+      "rewards/format_reward_func": 0.9709821864962578,
+      "step": 272
+    },
+    {
+      "completion_length": 286.8727807998657,
+      "epoch": 0.3409550474412817,
+      "grad_norm": 0.06418627638571622,
+      "kl": 0.20733642578125,
+      "learning_rate": 1.7549195509265407e-07,
+      "loss": 0.0002,
+      "reward": 1.373883992433548,
+      "reward_std": 0.24791530705988407,
+      "rewards/equation_reward_func": 0.42410716507583857,
+      "rewards/format_reward_func": 0.9497768171131611,
+      "step": 274
+    },
+    {
+      "completion_length": 266.1897430419922,
+      "epoch": 0.34344377041530566,
+      "grad_norm": 0.7673239333850856,
+      "kl": 0.2283935546875,
+      "learning_rate": 1.7206079062562536e-07,
+      "loss": 0.0002,
+      "reward": 1.45089291036129,
+      "reward_std": 0.173080294393003,
+      "rewards/equation_reward_func": 0.4899553805589676,
+      "rewards/format_reward_func": 0.9609375298023224,
+      "step": 276
+    },
+    {
+      "completion_length": 282.8348331451416,
+      "epoch": 0.3459324933893296,
+      "grad_norm": 0.0632244397075646,
+      "kl": 0.21356201171875,
+      "learning_rate": 1.6864581199841226e-07,
+      "loss": 0.0002,
+      "reward": 1.4174107685685158,
+      "reward_std": 0.20238550938665867,
+      "rewards/equation_reward_func": 0.470982164144516,
+      "rewards/format_reward_func": 0.9464286155998707,
+      "step": 278
+    },
+    {
+      "completion_length": 283.8951015472412,
+      "epoch": 0.34842121636335355,
+      "grad_norm": 0.05348187564500839,
+      "kl": 0.2279052734375,
+      "learning_rate": 1.6524772840857388e-07,
+      "loss": 0.0002,
+      "reward": 1.4040179252624512,
+      "reward_std": 0.21677289670333266,
+      "rewards/equation_reward_func": 0.450892873108387,
+      "rewards/format_reward_func": 0.9531250409781933,
+      "step": 280
+    },
+    {
+      "completion_length": 265.0759038925171,
+      "epoch": 0.3509099393373775,
+      "grad_norm": 0.07679910791742155,
+      "kl": 0.22210693359375,
+      "learning_rate": 1.6186724554503237e-07,
+      "loss": 0.0002,
+      "reward": 1.4330357611179352,
+      "reward_std": 0.20752421114593744,
+      "rewards/equation_reward_func": 0.4687500223517418,
+      "rewards/format_reward_func": 0.9642857499420643,
+      "step": 282
+    },
+    {
+      "completion_length": 284.56363010406494,
+      "epoch": 0.35339866231140143,
+      "grad_norm": 0.05732289464788312,
+      "kl": 0.22857666015625,
+      "learning_rate": 1.5850506544152103e-07,
+      "loss": 0.0002,
+      "reward": 1.4430804178118706,
+      "reward_std": 0.17447989620268345,
+      "rewards/equation_reward_func": 0.47321430779993534,
+      "rewards/format_reward_func": 0.9698661006987095,
+      "step": 284
+    },
+    {
+      "completion_length": 264.10938453674316,
+      "epoch": 0.3558873852854254,
+      "grad_norm": 0.04971277203023392,
+      "kl": 0.22442626953125,
+      "learning_rate": 1.5516188633079107e-07,
+      "loss": 0.0002,
+      "reward": 1.4821429252624512,
+      "reward_std": 0.19169320166110992,
+      "rewards/equation_reward_func": 0.5156250260770321,
+      "rewards/format_reward_func": 0.9665178954601288,
+      "step": 286
+    },
+    {
+      "completion_length": 276.925235748291,
+      "epoch": 0.3583761082594494,
+      "grad_norm": 0.05844528789981532,
+      "kl": 0.21966552734375,
+      "learning_rate": 1.5183840249960784e-07,
+      "loss": 0.0002,
+      "reward": 1.433035783469677,
+      "reward_std": 0.2321711997501552,
+      "rewards/equation_reward_func": 0.47991073690354824,
+      "rewards/format_reward_func": 0.9531250447034836,
+      "step": 288
+    },
+    {
+      "completion_length": 249.3024673461914,
+      "epoch": 0.36086483123347335,
+      "grad_norm": 0.04774574092660203,
+      "kl": 0.225341796875,
+      "learning_rate": 1.4853530414456612e-07,
+      "loss": 0.0002,
+      "reward": 1.5066964998841286,
+      "reward_std": 0.16551457298919559,
+      "rewards/equation_reward_func": 0.5357143115252256,
+      "rewards/format_reward_func": 0.9709821715950966,
+      "step": 290
+    },
+    {
+      "completion_length": 260.86608123779297,
+      "epoch": 0.36335355420749726,
+      "grad_norm": 0.050668244719322277,
+      "kl": 0.212890625,
+      "learning_rate": 1.4525327722875568e-07,
+      "loss": 0.0002,
+      "reward": 1.4709822088479996,
+      "reward_std": 0.1885156724601984,
+      "rewards/equation_reward_func": 0.5000000251457095,
+      "rewards/format_reward_func": 0.9709821678698063,
+      "step": 292
+    },
+    {
+      "completion_length": 285.5926465988159,
+      "epoch": 0.36584227718152124,
+      "grad_norm": 0.05273095674745574,
+      "kl": 0.22235107421875,
+      "learning_rate": 1.4199300333930515e-07,
+      "loss": 0.0002,
+      "reward": 1.4587054178118706,
+      "reward_std": 0.20572060579434037,
+      "rewards/equation_reward_func": 0.510044664144516,
+      "rewards/format_reward_func": 0.9486607499420643,
+      "step": 294
+    },
+    {
+      "completion_length": 301.3437662124634,
+      "epoch": 0.3683310001555452,
+      "grad_norm": 0.05580536848172712,
+      "kl": 0.22265625,
+      "learning_rate": 1.3875515954583523e-07,
+      "loss": 0.0002,
+      "reward": 1.3404018431901932,
+      "reward_std": 0.18478884501382709,
+      "rewards/equation_reward_func": 0.3839285899884999,
+      "rewards/format_reward_func": 0.9564732424914837,
+      "step": 296
+    },
+    {
+      "completion_length": 283.79131031036377,
+      "epoch": 0.3708197231295691,
+      "grad_norm": 0.06174432474758232,
+      "kl": 0.22003173828125,
+      "learning_rate": 1.3554041825985e-07,
+      "loss": 0.0002,
+      "reward": 1.4375000521540642,
+      "reward_std": 0.20120970671996474,
+      "rewards/equation_reward_func": 0.47991073224693537,
+      "rewards/format_reward_func": 0.957589328289032,
+      "step": 298
+    },
+    {
+      "completion_length": 289.1685390472412,
+      "epoch": 0.3733084461035931,
+      "grad_norm": 0.0549328940573922,
+      "kl": 0.23248291015625,
+      "learning_rate": 1.323494470950949e-07,
+      "loss": 0.0002,
+      "reward": 1.36495541036129,
+      "reward_std": 0.21155750146135688,
+      "rewards/equation_reward_func": 0.4185268050059676,
+      "rewards/format_reward_func": 0.9464286118745804,
+      "step": 300
+    },
+    {
+      "completion_length": 267.3917512893677,
+      "epoch": 0.37579716907761707,
+      "grad_norm": 0.05147542755960161,
+      "kl": 0.2188720703125,
+      "learning_rate": 1.2918290872891236e-07,
+      "loss": 0.0002,
+      "reward": 1.4988840073347092,
+      "reward_std": 0.17496886011213064,
+      "rewards/equation_reward_func": 0.5368303880095482,
+      "rewards/format_reward_func": 0.9620536081492901,
+      "step": 302
+    },
+    {
+      "completion_length": 276.30582094192505,
+      "epoch": 0.378285892051641,
+      "grad_norm": 0.05300792192239273,
+      "kl": 0.22711181640625,
+      "learning_rate": 1.260414607646213e-07,
+      "loss": 0.0002,
+      "reward": 1.4486607685685158,
+      "reward_std": 0.18876845808699727,
+      "rewards/equation_reward_func": 0.4921875186264515,
+      "rewards/format_reward_func": 0.9564732387661934,
+      "step": 304
+    },
+    {
+      "completion_length": 266.9196557998657,
+      "epoch": 0.38077461502566495,
+      "grad_norm": 0.06514244269360671,
+      "kl": 0.28143310546875,
+      "learning_rate": 1.2292575559495143e-07,
+      "loss": 0.0003,
+      "reward": 1.4754464775323868,
+      "reward_std": 0.20895800599828362,
+      "rewards/equation_reward_func": 0.5245536006987095,
+      "rewards/format_reward_func": 0.9508928880095482,
+      "step": 306
+    },
+    {
+      "completion_length": 263.41630840301514,
+      "epoch": 0.3832633379996889,
+      "grad_norm": 0.0520463127094474,
+      "kl": 0.2337646484375,
+      "learning_rate": 1.1983644026655835e-07,
+      "loss": 0.0002,
+      "reward": 1.488839365541935,
+      "reward_std": 0.2131196865811944,
+      "rewards/equation_reward_func": 0.5234375260770321,
+      "rewards/format_reward_func": 0.9654018171131611,
+      "step": 308
+    },
+    {
+      "completion_length": 266.73103857040405,
+      "epoch": 0.38575206097371284,
+      "grad_norm": 0.052587830849904815,
+      "kl": 0.22607421875,
+      "learning_rate": 1.1677415634565066e-07,
+      "loss": 0.0002,
+      "reward": 1.4631697088479996,
+      "reward_std": 0.17361521907150745,
+      "rewards/equation_reward_func": 0.4944196594879031,
+      "rewards/format_reward_func": 0.9687500298023224,
+      "step": 310
+    },
+    {
+      "completion_length": 260.9263496398926,
+      "epoch": 0.3882407839477368,
+      "grad_norm": 0.05132512982351145,
+      "kl": 0.2310791015625,
+      "learning_rate": 1.1373953978475353e-07,
+      "loss": 0.0002,
+      "reward": 1.4665179178118706,
+      "reward_std": 0.19973930763080716,
+      "rewards/equation_reward_func": 0.5066964514553547,
+      "rewards/format_reward_func": 0.9598214663565159,
+      "step": 312
+    },
+    {
+      "completion_length": 292.9397449493408,
+      "epoch": 0.3907295069217608,
+      "grad_norm": 0.04204105376528588,
+      "kl": 0.2447509765625,
+      "learning_rate": 1.1073322079063913e-07,
+      "loss": 0.0002,
+      "reward": 1.405133992433548,
+      "reward_std": 0.1639846209436655,
+      "rewards/equation_reward_func": 0.4453125223517418,
+      "rewards/format_reward_func": 0.9598214626312256,
+      "step": 314
+    },
+    {
+      "completion_length": 281.0435390472412,
+      "epoch": 0.3932182298957847,
+      "grad_norm": 0.05904069968992011,
+      "kl": 0.21405029296875,
+      "learning_rate": 1.0775582369344946e-07,
+      "loss": 0.0002,
+      "reward": 1.4185268506407738,
+      "reward_std": 0.20928293373435736,
+      "rewards/equation_reward_func": 0.4575893133878708,
+      "rewards/format_reward_func": 0.9609375298023224,
+      "step": 316
+    },
+    {
+      "completion_length": 282.73438835144043,
+      "epoch": 0.39570695286980867,
+      "grad_norm": 0.05546896578155631,
+      "kl": 0.21636962890625,
+      "learning_rate": 1.0480796681704077e-07,
+      "loss": 0.0002,
+      "reward": 1.4453125596046448,
+      "reward_std": 0.21966069657355547,
+      "rewards/equation_reward_func": 0.4910714477300644,
+      "rewards/format_reward_func": 0.9542411081492901,
+      "step": 318
+    },
+    {
+      "completion_length": 290.85715770721436,
+      "epoch": 0.39819567584383264,
+      "grad_norm": 0.051125363447751745,
+      "kl": 0.217041015625,
+      "learning_rate": 1.018902623505741e-07,
+      "loss": 0.0002,
+      "reward": 1.3772322088479996,
+      "reward_std": 0.17292431741952896,
+      "rewards/equation_reward_func": 0.4241071678698063,
+      "rewards/format_reward_func": 0.953125037252903,
+      "step": 320
+    },
+    {
+      "completion_length": 271.14286708831787,
+      "epoch": 0.4006843988178566,
+      "grad_norm": 0.05569159591979702,
+      "kl": 0.23260498046875,
+      "learning_rate": 9.900331622138063e-08,
+      "loss": 0.0002,
+      "reward": 1.4575893580913544,
+      "reward_std": 0.19401329522952437,
+      "rewards/equation_reward_func": 0.49888395611196756,
+      "rewards/format_reward_func": 0.9587053842842579,
+      "step": 322
+    },
+    {
+      "completion_length": 266.0435380935669,
+      "epoch": 0.40317312179188053,
+      "grad_norm": 0.0643989995982714,
+      "kl": 0.2467041015625,
+      "learning_rate": 9.614772796912681e-08,
+      "loss": 0.0002,
+      "reward": 1.4330357760190964,
+      "reward_std": 0.1952164708636701,
+      "rewards/equation_reward_func": 0.47433038242161274,
+      "rewards/format_reward_func": 0.9587053917348385,
+      "step": 324
+    },
+    {
+      "completion_length": 296.36162281036377,
+      "epoch": 0.4056618447659045,
+      "grad_norm": 0.06592474865214347,
+      "kl": 0.216064453125,
+      "learning_rate": 9.332409062130686e-08,
+      "loss": 0.0002,
+      "reward": 1.4029018506407738,
+      "reward_std": 0.23388848965987563,
+      "rewards/equation_reward_func": 0.4464285932481289,
+      "rewards/format_reward_func": 0.956473246216774,
+      "step": 326
+    },
+    {
+      "completion_length": 265.6294755935669,
+      "epoch": 0.40815056773992847,
+      "grad_norm": 0.040286899772741865,
+      "kl": 0.2259521484375,
+      "learning_rate": 9.053299057008699e-08,
+      "loss": 0.0002,
+      "reward": 1.4832589775323868,
+      "reward_std": 0.1687624640762806,
+      "rewards/equation_reward_func": 0.5156250260770321,
+      "rewards/format_reward_func": 0.9676339663565159,
+      "step": 328
+    },
+    {
+      "completion_length": 247.71653175354004,
+      "epoch": 0.4106392907139524,
+      "grad_norm": 0.08330283837230541,
+      "kl": 0.2242431640625,
+      "learning_rate": 8.777500745052743e-08,
+      "loss": 0.0002,
+      "reward": 1.4609375670552254,
+      "reward_std": 0.1921851597726345,
+      "rewards/equation_reward_func": 0.49553573690354824,
+      "rewards/format_reward_func": 0.9654018171131611,
+      "step": 330
+    },
+    {
+      "completion_length": 258.3069314956665,
+      "epoch": 0.41312801368797636,
+      "grad_norm": 0.047241469109823984,
+      "kl": 0.2232666015625,
+      "learning_rate": 8.505071402020892e-08,
+      "loss": 0.0002,
+      "reward": 1.4441964998841286,
+      "reward_std": 0.15772674046456814,
+      "rewards/equation_reward_func": 0.47991072945296764,
+      "rewards/format_reward_func": 0.964285746216774,
+      "step": 332
+    },
+    {
+      "completion_length": 257.712064743042,
+      "epoch": 0.41561673666200033,
+      "grad_norm": 0.05922356243283228,
+      "kl": 0.21270751953125,
+      "learning_rate": 8.236067604028562e-08,
+      "loss": 0.0002,
+      "reward": 1.4776786491274834,
+      "reward_std": 0.1794830970466137,
+      "rewards/equation_reward_func": 0.505580386146903,
+      "rewards/format_reward_func": 0.9720982536673546,
+      "step": 334
+    },
+    {
+      "completion_length": 261.08929920196533,
+      "epoch": 0.41810545963602425,
+      "grad_norm": 0.05512439495800802,
+      "kl": 0.27520751953125,
+      "learning_rate": 7.970545215799327e-08,
+      "loss": 0.0003,
+      "reward": 1.453125074505806,
+      "reward_std": 0.17044777097180486,
+      "rewards/equation_reward_func": 0.482142879627645,
+      "rewards/format_reward_func": 0.9709821753203869,
+      "step": 336
+    },
+    {
+      "completion_length": 253.13282299041748,
+      "epoch": 0.4205941826100482,
+      "grad_norm": 0.0546878420888323,
+      "kl": 0.23211669921875,
+      "learning_rate": 7.708559379063204e-08,
+      "loss": 0.0002,
+      "reward": 1.4520089998841286,
+      "reward_std": 0.160722223110497,
+      "rewards/equation_reward_func": 0.47767859511077404,
+      "rewards/format_reward_func": 0.9743303954601288,
+      "step": 338
+    },
+    {
+      "completion_length": 268.42523288726807,
+      "epoch": 0.4230829055840722,
+      "grad_norm": 0.0643372128173253,
+      "kl": 0.24578857421875,
+      "learning_rate": 7.45016450110534e-08,
+      "loss": 0.0002,
+      "reward": 1.392857201397419,
+      "reward_std": 0.15963875455781817,
+      "rewards/equation_reward_func": 0.4263393054716289,
+      "rewards/format_reward_func": 0.9665178842842579,
+      "step": 340
+    },
+    {
+      "completion_length": 289.94197940826416,
+      "epoch": 0.4255716285580961,
+      "grad_norm": 0.12614189846940307,
+      "kl": 0.35626220703125,
+      "learning_rate": 7.195414243467029e-08,
+      "loss": 0.0004,
+      "reward": 1.3549107909202576,
+      "reward_std": 0.1969244135543704,
+      "rewards/equation_reward_func": 0.3962053820723668,
+      "rewards/format_reward_func": 0.9587053880095482,
+      "step": 342
+    },
+    {
+      "completion_length": 259.41407203674316,
+      "epoch": 0.4280603515321201,
+      "grad_norm": 0.054571007367345555,
+      "kl": 0.21270751953125,
+      "learning_rate": 6.944361510801763e-08,
+      "loss": 0.0002,
+      "reward": 1.4241072088479996,
+      "reward_std": 0.1726059396751225,
+      "rewards/equation_reward_func": 0.4620535969734192,
+      "rewards/format_reward_func": 0.9620536118745804,
+      "step": 344
+    },
+    {
+      "completion_length": 255.23326969146729,
+      "epoch": 0.43054907450614405,
+      "grad_norm": 0.07334212596865244,
+      "kl": 0.220703125,
+      "learning_rate": 6.697058439888283e-08,
+      "loss": 0.0002,
+      "reward": 1.4832589849829674,
+      "reward_std": 0.17638651933521032,
+      "rewards/equation_reward_func": 0.5167411025613546,
+      "rewards/format_reward_func": 0.9665178842842579,
+      "step": 346
+    },
+    {
+      "completion_length": 272.0904130935669,
+      "epoch": 0.43303779748016796,
+      "grad_norm": 0.0639218845007823,
+      "kl": 0.216064453125,
+      "learning_rate": 6.453556388803288e-08,
+      "loss": 0.0002,
+      "reward": 1.4542411491274834,
+      "reward_std": 0.21514991857111454,
+      "rewards/equation_reward_func": 0.4888393050059676,
+      "rewards/format_reward_func": 0.9654018208384514,
+      "step": 348
+    },
+    {
+      "completion_length": 246.30581283569336,
+      "epoch": 0.43552652045419193,
+      "grad_norm": 0.0840915601049537,
+      "kl": 0.2225341796875,
+      "learning_rate": 6.213905926255697e-08,
+      "loss": 0.0002,
+      "reward": 1.5279018580913544,
+      "reward_std": 0.18220655526965857,
+      "rewards/equation_reward_func": 0.5647321715950966,
+      "rewards/format_reward_func": 0.9631696790456772,
+      "step": 350
+    },
+    {
+      "completion_length": 258.50113010406494,
+      "epoch": 0.4380152434282159,
+      "grad_norm": 0.05057611458509387,
+      "kl": 0.2137451171875,
+      "learning_rate": 5.978156821084987e-08,
+      "loss": 0.0002,
+      "reward": 1.4654018357396126,
+      "reward_std": 0.1536577446386218,
+      "rewards/equation_reward_func": 0.49888394959270954,
+      "rewards/format_reward_func": 0.9665178917348385,
+      "step": 352
+    },
+    {
+      "completion_length": 246.66184329986572,
+      "epoch": 0.4405039664022399,
+      "grad_norm": 0.055435192319033814,
+      "kl": 0.2176513671875,
+      "learning_rate": 5.7463580319254853e-08,
+      "loss": 0.0002,
+      "reward": 1.4687500670552254,
+      "reward_std": 0.17097870633006096,
+      "rewards/equation_reward_func": 0.49553573597222567,
+      "rewards/format_reward_func": 0.9732143245637417,
+      "step": 354
+    },
+    {
+      "completion_length": 250.23550128936768,
+      "epoch": 0.4429926893762638,
+      "grad_norm": 0.04991979968995625,
+      "kl": 0.21051025390625,
+      "learning_rate": 5.518557697039081e-08,
+      "loss": 0.0002,
+      "reward": 1.4944197088479996,
+      "reward_std": 0.18009533546864986,
+      "rewards/equation_reward_func": 0.520089304074645,
+      "rewards/format_reward_func": 0.9743303768336773,
+      "step": 356
+    },
+    {
+      "completion_length": 255.1149673461914,
+      "epoch": 0.44548141235028776,
+      "grad_norm": 0.05353950405443061,
+      "kl": 0.20758056640625,
+      "learning_rate": 5.294803124318145e-08,
+      "loss": 0.0002,
+      "reward": 1.468750074505806,
+      "reward_std": 0.18782880948856473,
+      "rewards/equation_reward_func": 0.5044643077999353,
+      "rewards/format_reward_func": 0.9642857573926449,
+      "step": 358
+    },
+    {
+      "completion_length": 255.90067958831787,
+      "epoch": 0.44797013532431174,
+      "grad_norm": 0.05725118559311897,
+      "kl": 0.21295166015625,
+      "learning_rate": 5.07514078146106e-08,
+      "loss": 0.0002,
+      "reward": 1.4464286267757416,
+      "reward_std": 0.18124625319615006,
+      "rewards/equation_reward_func": 0.4810268124565482,
+      "rewards/format_reward_func": 0.9654018320143223,
+      "step": 360
+    },
+    {
+      "completion_length": 262.5714416503906,
+      "epoch": 0.45045885829833565,
+      "grad_norm": 0.06839221438705215,
+      "kl": 0.22894287109375,
+      "learning_rate": 4.859616286322094e-08,
+      "loss": 0.0002,
+      "reward": 1.4352679252624512,
+      "reward_std": 0.19061506912112236,
+      "rewards/equation_reward_func": 0.4776785938302055,
+      "rewards/format_reward_func": 0.9575893208384514,
+      "step": 362
+    },
+    {
+      "completion_length": 270.0591630935669,
+      "epoch": 0.4529475812723596,
+      "grad_norm": 0.06771361462731067,
+      "kl": 0.23846435546875,
+      "learning_rate": 4.648274397437829e-08,
+      "loss": 0.0002,
+      "reward": 1.4241072088479996,
+      "reward_std": 0.1951315077021718,
+      "rewards/equation_reward_func": 0.46428574062883854,
+      "rewards/format_reward_func": 0.9598214626312256,
+      "step": 364
+    },
+    {
+      "completion_length": 244.5680913925171,
+      "epoch": 0.4554363042463836,
+      "grad_norm": 0.06570494423275842,
+      "kl": 0.22015380859375,
+      "learning_rate": 4.4411590047320617e-08,
+      "loss": 0.0002,
+      "reward": 1.5491072162985802,
+      "reward_std": 0.17699437169358134,
+      "rewards/equation_reward_func": 0.5714285969734192,
+      "rewards/format_reward_func": 0.9776785932481289,
+      "step": 366
+    },
+    {
+      "completion_length": 258.7991189956665,
+      "epoch": 0.4579250272204075,
+      "grad_norm": 0.0720114335252616,
+      "kl": 0.2042236328125,
+      "learning_rate": 4.2383131204010494e-08,
+      "loss": 0.0002,
+      "reward": 1.4709822088479996,
+      "reward_std": 0.19236048124730587,
+      "rewards/equation_reward_func": 0.5066964570432901,
+      "rewards/format_reward_func": 0.9642857387661934,
+      "step": 368
+    },
+    {
+      "completion_length": 256.76228618621826,
+      "epoch": 0.4604137501944315,
+      "grad_norm": 0.051725385509834826,
+      "kl": 0.22265625,
+      "learning_rate": 4.039778869981064e-08,
+      "loss": 0.0002,
+      "reward": 1.4776786416769028,
+      "reward_std": 0.18397637410089374,
+      "rewards/equation_reward_func": 0.5022321697324514,
+      "rewards/format_reward_func": 0.975446455180645,
+      "step": 370
+    },
+    {
+      "completion_length": 266.7600574493408,
+      "epoch": 0.46290247316845545,
+      "grad_norm": 0.059635161912602624,
+      "kl": 0.21502685546875,
+      "learning_rate": 3.845597483600049e-08,
+      "loss": 0.0002,
+      "reward": 1.402901865541935,
+      "reward_std": 0.18843654869124293,
+      "rewards/equation_reward_func": 0.4441964514553547,
+      "rewards/format_reward_func": 0.9587053880095482,
+      "step": 372
+    },
+    {
+      "completion_length": 269.5904130935669,
+      "epoch": 0.46539119614247937,
+      "grad_norm": 0.052058607348838166,
+      "kl": 0.239990234375,
+      "learning_rate": 3.655809287415284e-08,
+      "loss": 0.0002,
+      "reward": 1.4475447162985802,
+      "reward_std": 0.18611292028799653,
+      "rewards/equation_reward_func": 0.47879466973245144,
+      "rewards/format_reward_func": 0.9687500335276127,
+      "step": 374
+    },
+    {
+      "completion_length": 261.80804920196533,
+      "epoch": 0.46787991911650334,
+      "grad_norm": 0.05654912472263546,
+      "kl": 0.21923828125,
+      "learning_rate": 3.4704536952387285e-08,
+      "loss": 0.0002,
+      "reward": 1.4296875596046448,
+      "reward_std": 0.19141265889629722,
+      "rewards/equation_reward_func": 0.4676339477300644,
+      "rewards/format_reward_func": 0.9620536044239998,
+      "step": 376
+    },
+    {
+      "completion_length": 257.9151906967163,
+      "epoch": 0.4703686420905273,
+      "grad_norm": 0.05359040211565284,
+      "kl": 0.22039794921875,
+      "learning_rate": 3.2895692003518575e-08,
+      "loss": 0.0002,
+      "reward": 1.4352679178118706,
+      "reward_std": 0.2002840624190867,
+      "rewards/equation_reward_func": 0.4787946715950966,
+      "rewards/format_reward_func": 0.9564732499420643,
+      "step": 378
+    },
+    {
+      "completion_length": 252.555814743042,
+      "epoch": 0.4728573650645512,
+      "grad_norm": 0.05665983443995847,
+      "kl": 0.23614501953125,
+      "learning_rate": 3.113193367511635e-08,
+      "loss": 0.0002,
+      "reward": 1.4587054252624512,
+      "reward_std": 0.18507013376802206,
+      "rewards/equation_reward_func": 0.4955357387661934,
+      "rewards/format_reward_func": 0.9631696715950966,
+      "step": 380
+    },
+    {
+      "completion_length": 246.92746829986572,
+      "epoch": 0.4753460880385752,
+      "grad_norm": 0.05986371207599387,
+      "kl": 0.2442626953125,
+      "learning_rate": 2.9413628251493934e-08,
+      "loss": 0.0002,
+      "reward": 1.487723283469677,
+      "reward_std": 0.15949935605749488,
+      "rewards/equation_reward_func": 0.5234375260770321,
+      "rewards/format_reward_func": 0.9642857499420643,
+      "step": 382
+    },
+    {
+      "completion_length": 264.3817071914673,
+      "epoch": 0.47783481101259917,
+      "grad_norm": 0.05453811145707849,
+      "kl": 0.2188720703125,
+      "learning_rate": 2.774113257764066e-08,
+      "loss": 0.0002,
+      "reward": 1.434151865541935,
+      "reward_std": 0.18349733110517263,
+      "rewards/equation_reward_func": 0.47656252048909664,
+      "rewards/format_reward_func": 0.9575893171131611,
+      "step": 384
+    },
+    {
+      "completion_length": 245.20536708831787,
+      "epoch": 0.48032353398662314,
+      "grad_norm": 0.05552843745329939,
+      "kl": 0.22064208984375,
+      "learning_rate": 2.611479398511518e-08,
+      "loss": 0.0002,
+      "reward": 1.4899554327130318,
+      "reward_std": 0.1519217761233449,
+      "rewards/equation_reward_func": 0.5145089514553547,
+      "rewards/format_reward_func": 0.9754464626312256,
+      "step": 386
+    },
+    {
+      "completion_length": 269.5904140472412,
+      "epoch": 0.48281225696064706,
+      "grad_norm": 0.05743635058931217,
+      "kl": 0.2388916015625,
+      "learning_rate": 2.4534950219914057e-08,
+      "loss": 0.0002,
+      "reward": 1.4185268357396126,
+      "reward_std": 0.19744081050157547,
+      "rewards/equation_reward_func": 0.44419645238667727,
+      "rewards/format_reward_func": 0.9743303842842579,
+      "step": 388
+    },
+    {
+      "completion_length": 233.9810390472412,
+      "epoch": 0.48530097993467103,
+      "grad_norm": 0.060693747477844985,
+      "kl": 0.2197265625,
+      "learning_rate": 2.300192937233128e-08,
+      "loss": 0.0002,
+      "reward": 1.5803572162985802,
+      "reward_std": 0.16287201223894954,
+      "rewards/equation_reward_func": 0.6093750288709998,
+      "rewards/format_reward_func": 0.970982164144516,
+      "step": 390
+    },
+    {
+      "completion_length": 257.4933156967163,
+      "epoch": 0.487789702908695,
+      "grad_norm": 0.061473806975209856,
+      "kl": 0.2171630859375,
+      "learning_rate": 2.1516049808822935e-08,
+      "loss": 0.0002,
+      "reward": 1.4687500596046448,
+      "reward_std": 0.18804124649614096,
+      "rewards/equation_reward_func": 0.505580386146903,
+      "rewards/format_reward_func": 0.9631696790456772,
+      "step": 392
+    },
+    {
+      "completion_length": 266.17189025878906,
+      "epoch": 0.4902784258827189,
+      "grad_norm": 0.057251931052709155,
+      "kl": 0.2237548828125,
+      "learning_rate": 2.007762010589098e-08,
+      "loss": 0.0002,
+      "reward": 1.4062500596046448,
+      "reward_std": 0.16503014508634806,
+      "rewards/equation_reward_func": 0.43638394586741924,
+      "rewards/format_reward_func": 0.9698661044239998,
+      "step": 394
+    },
+    {
+      "completion_length": 288.45872020721436,
+      "epoch": 0.4927671488567429,
+      "grad_norm": 0.05683738580050447,
+      "kl": 0.22113037109375,
+      "learning_rate": 1.8686938986000627e-08,
+      "loss": 0.0002,
+      "reward": 1.3292411267757416,
+      "reward_std": 0.17618125723674893,
+      "rewards/equation_reward_func": 0.3705357308499515,
+      "rewards/format_reward_func": 0.9587053880095482,
+      "step": 396
+    },
+    {
+      "completion_length": 274.2555932998657,
+      "epoch": 0.49525587183076686,
+      "grad_norm": 0.05631696950465828,
+      "kl": 0.2208251953125,
+      "learning_rate": 1.734429525554365e-08,
+      "loss": 0.0002,
+      "reward": 1.4140625670552254,
+      "reward_std": 0.1975519866682589,
+      "rewards/equation_reward_func": 0.4520089467987418,
+      "rewards/format_reward_func": 0.9620536006987095,
+      "step": 398
+    },
+    {
+      "completion_length": 245.83148384094238,
+      "epoch": 0.4977445948047908,
+      "grad_norm": 0.05014202042525302,
+      "kl": 0.2230224609375,
+      "learning_rate": 1.604996774486145e-08,
+      "loss": 0.0002,
+      "reward": 1.4866072088479996,
+      "reward_std": 0.16337728220969439,
+      "rewards/equation_reward_func": 0.5212053786963224,
+      "rewards/format_reward_func": 0.9654018171131611,
+      "step": 400
+    },
+    {
+      "completion_length": 253.30804443359375,
+      "epoch": 0.5002333177788147,
+      "grad_norm": 0.060614389948920294,
+      "kl": 0.22021484375,
+      "learning_rate": 1.4804225250339281e-08,
+      "loss": 0.0002,
+      "reward": 1.4687500670552254,
+      "reward_std": 0.1655928883701563,
+      "rewards/equation_reward_func": 0.502232164144516,
+      "rewards/format_reward_func": 0.9665178805589676,
+      "step": 402
+    },
+    {
+      "completion_length": 268.2712173461914,
+      "epoch": 0.5027220407528387,
+      "grad_norm": 0.06043605047844046,
+      "kl": 0.21038818359375,
+      "learning_rate": 1.360732647858498e-08,
+      "loss": 0.0002,
+      "reward": 1.4419643506407738,
+      "reward_std": 0.18442421313375235,
+      "rewards/equation_reward_func": 0.486607164144516,
+      "rewards/format_reward_func": 0.9553571753203869,
+      "step": 404
+    },
+    {
+      "completion_length": 256.667423248291,
+      "epoch": 0.5052107637268627,
+      "grad_norm": 0.03748084182110105,
+      "kl": 0.21240234375,
+      "learning_rate": 1.2459519992702311e-08,
+      "loss": 0.0002,
+      "reward": 1.4732143580913544,
+      "reward_std": 0.11906880373135209,
+      "rewards/equation_reward_func": 0.4910714514553547,
+      "rewards/format_reward_func": 0.9821428842842579,
+      "step": 406
+    },
+    {
+      "completion_length": 236.2924222946167,
+      "epoch": 0.5076994867008866,
+      "grad_norm": 0.0629615161229278,
+      "kl": 0.22723388671875,
+      "learning_rate": 1.1361044160671629e-08,
+      "loss": 0.0002,
+      "reward": 1.5323661491274834,
+      "reward_std": 0.17991887778043747,
+      "rewards/equation_reward_func": 0.5647321660071611,
+      "rewards/format_reward_func": 0.9676339589059353,
+      "step": 408
+    },
+    {
+      "completion_length": 267.2890729904175,
+      "epoch": 0.5101882096749105,
+      "grad_norm": 0.08454099786217265,
+      "kl": 0.21246337890625,
+      "learning_rate": 1.0312127105846947e-08,
+      "loss": 0.0002,
+      "reward": 1.3917411416769028,
+      "reward_std": 0.200666856020689,
+      "rewards/equation_reward_func": 0.42857144866138697,
+      "rewards/format_reward_func": 0.9631696864962578,
+      "step": 410
+    },
+    {
+      "completion_length": 241.93639659881592,
+      "epoch": 0.5126769326489345,
+      "grad_norm": 0.05678911579880298,
+      "kl": 0.2169189453125,
+      "learning_rate": 9.312986659581301e-09,
+      "loss": 0.0002,
+      "reward": 1.5145090073347092,
+      "reward_std": 0.18534906907007098,
+      "rewards/equation_reward_func": 0.5435268152505159,
+      "rewards/format_reward_func": 0.970982164144516,
+      "step": 412
+    },
+    {
+      "completion_length": 262.33148670196533,
+      "epoch": 0.5151656556229585,
+      "grad_norm": 0.07237569221975489,
+      "kl": 0.2265625,
+      "learning_rate": 8.363830315988945e-09,
+      "loss": 0.0002,
+      "reward": 1.4308036416769028,
+      "reward_std": 0.18102896073833108,
+      "rewards/equation_reward_func": 0.4799107415601611,
+      "rewards/format_reward_func": 0.9508928880095482,
+      "step": 414
+    },
+    {
+      "completion_length": 290.8247900009155,
+      "epoch": 0.5176543785969824,
+      "grad_norm": 0.06291574216632526,
+      "kl": 0.211181640625,
+      "learning_rate": 7.46485518885462e-09,
+      "loss": 0.0002,
+      "reward": 1.3325893431901932,
+      "reward_std": 0.2226958447135985,
+      "rewards/equation_reward_func": 0.3828125229338184,
+      "rewards/format_reward_func": 0.9497768133878708,
+      "step": 416
+    },
+    {
+      "completion_length": 250.63059425354004,
+      "epoch": 0.5201431015710064,
+      "grad_norm": 0.06285425713534079,
+      "kl": 0.2586669921875,
+      "learning_rate": 6.616247970698319e-09,
+      "loss": 0.0003,
+      "reward": 1.4854911491274834,
+      "reward_std": 0.17498166300356388,
+      "rewards/equation_reward_func": 0.5156250260770321,
+      "rewards/format_reward_func": 0.9698660969734192,
+      "step": 418
+    },
+    {
+      "completion_length": 256.33260345458984,
+      "epoch": 0.5226318245450303,
+      "grad_norm": 0.05299668721681115,
+      "kl": 0.21307373046875,
+      "learning_rate": 5.8181848940044855e-09,
+      "loss": 0.0002,
+      "reward": 1.4241072088479996,
+      "reward_std": 0.2053914200514555,
+      "rewards/equation_reward_func": 0.4631696604192257,
+      "rewards/format_reward_func": 0.9609375223517418,
+      "step": 420
+    },
+    {
+      "completion_length": 251.9442081451416,
+      "epoch": 0.5251205475190542,
+      "grad_norm": 0.03956351552241899,
+      "kl": 0.26885986328125,
+      "learning_rate": 5.070831694623135e-09,
+      "loss": 0.0003,
+      "reward": 1.4899554327130318,
+      "reward_std": 0.13848364166915417,
+      "rewards/equation_reward_func": 0.5200893077999353,
+      "rewards/format_reward_func": 0.9698661006987095,
+      "step": 422
+    },
+    {
+      "completion_length": 256.5011262893677,
+      "epoch": 0.5276092704930783,
+      "grad_norm": 0.0943308246377347,
+      "kl": 0.21270751953125,
+      "learning_rate": 4.374343577351336e-09,
+      "loss": 0.0002,
+      "reward": 1.4866072162985802,
+      "reward_std": 0.22044320311397314,
+      "rewards/equation_reward_func": 0.5212053824216127,
+      "rewards/format_reward_func": 0.9654018171131611,
+      "step": 424
+    },
+    {
+      "completion_length": 263.80693340301514,
+      "epoch": 0.5300979934671022,
+      "grad_norm": 0.06623549471839515,
+      "kl": 0.2095947265625,
+      "learning_rate": 3.7288651837012745e-09,
+      "loss": 0.0002,
+      "reward": 1.4475447311997414,
+      "reward_std": 0.18339544022455812,
+      "rewards/equation_reward_func": 0.47879466228187084,
+      "rewards/format_reward_func": 0.968750037252903,
+      "step": 426
+    },
+    {
+      "completion_length": 268.9073781967163,
+      "epoch": 0.5325867164411261,
+      "grad_norm": 0.052449792813201314,
+      "kl": 0.2080078125,
+      "learning_rate": 3.134530561862081e-09,
+      "loss": 0.0002,
+      "reward": 1.392857201397419,
+      "reward_std": 0.19783430639654398,
+      "rewards/equation_reward_func": 0.43415180779993534,
+      "rewards/format_reward_func": 0.9587053880095482,
+      "step": 428
+    },
+    {
+      "completion_length": 280.67747020721436,
+      "epoch": 0.5350754394151501,
+      "grad_norm": 0.08827545780059727,
+      "kl": 0.21624755859375,
+      "learning_rate": 2.5914631388619103e-09,
+      "loss": 0.0002,
+      "reward": 1.3750000447034836,
+      "reward_std": 0.18710578652098775,
+      "rewards/equation_reward_func": 0.4140625229338184,
+      "rewards/format_reward_func": 0.960937537252903,
+      "step": 430
+    },
+    {
+      "completion_length": 264.04130840301514,
+      "epoch": 0.537564162389174,
+      "grad_norm": 0.053078250209355325,
+      "kl": 0.21514892578125,
+      "learning_rate": 2.0997756949353297e-09,
+      "loss": 0.0002,
+      "reward": 1.4620536267757416,
+      "reward_std": 0.16093172412365675,
+      "rewards/equation_reward_func": 0.4977678833529353,
+      "rewards/format_reward_func": 0.9642857499420643,
+      "step": 432
+    },
+    {
+      "completion_length": 278.2064847946167,
+      "epoch": 0.540052885363198,
+      "grad_norm": 0.07760000928814792,
+      "kl": 0.21234130859375,
+      "learning_rate": 1.6595703401020844e-09,
+      "loss": 0.0002,
+      "reward": 1.3883929252624512,
+      "reward_std": 0.2151358723640442,
+      "rewards/equation_reward_func": 0.4341518059372902,
+      "rewards/format_reward_func": 0.9542410969734192,
+      "step": 434
+    },
+    {
+      "completion_length": 264.3091688156128,
+      "epoch": 0.542541608337222,
+      "grad_norm": 0.05086715788627145,
+      "kl": 0.21771240234375,
+      "learning_rate": 1.2709384929615596e-09,
+      "loss": 0.0002,
+      "reward": 1.437500074505806,
+      "reward_std": 0.1830471744760871,
+      "rewards/equation_reward_func": 0.47098216507583857,
+      "rewards/format_reward_func": 0.9665178842842579,
+      "step": 436
+    },
+    {
+      "completion_length": 254.76452159881592,
+      "epoch": 0.5450303313112459,
+      "grad_norm": 0.06208573799074421,
+      "kl": 0.22509765625,
+      "learning_rate": 9.339608617077165e-10,
+      "loss": 0.0002,
+      "reward": 1.4921875819563866,
+      "reward_std": 0.19756911788135767,
+      "rewards/equation_reward_func": 0.5323660969734192,
+      "rewards/format_reward_func": 0.9598214700818062,
+      "step": 438
+    },
+    {
+      "completion_length": 268.79241943359375,
+      "epoch": 0.5475190542852699,
+      "grad_norm": 0.04727521654296364,
+      "kl": 0.23455810546875,
+      "learning_rate": 6.487074273681114e-10,
+      "loss": 0.0002,
+      "reward": 1.4497768580913544,
+      "reward_std": 0.22120195254683495,
+      "rewards/equation_reward_func": 0.4877232424914837,
+      "rewards/format_reward_func": 0.9620535969734192,
+      "step": 440
+    },
+    {
+      "completion_length": 258.01563453674316,
+      "epoch": 0.5500077772592938,
+      "grad_norm": 0.05943481158821016,
+      "kl": 0.23046875,
+      "learning_rate": 4.152374292708538e-10,
+      "loss": 0.0002,
+      "reward": 1.4107143580913544,
+      "reward_std": 0.1552983969449997,
+      "rewards/equation_reward_func": 0.44754466600716114,
+      "rewards/format_reward_func": 0.9631696790456772,
+      "step": 442
+    },
+    {
+      "completion_length": 259.3951025009155,
+      "epoch": 0.5524965002333178,
+      "grad_norm": 0.0527539277018196,
+      "kl": 0.2109375,
+      "learning_rate": 2.3359935274214204e-10,
+      "loss": 0.0002,
+      "reward": 1.4654018431901932,
+      "reward_std": 0.18068964406847954,
+      "rewards/equation_reward_func": 0.49776788614690304,
+      "rewards/format_reward_func": 0.9676339626312256,
+      "step": 444
+    },
+    {
+      "completion_length": 252.39510250091553,
+      "epoch": 0.5549852232073418,
+      "grad_norm": 0.05940159867969958,
+      "kl": 0.21868896484375,
+      "learning_rate": 1.0383091903720665e-10,
+      "loss": 0.0002,
+      "reward": 1.4497768506407738,
+      "reward_std": 0.18734711268916726,
+      "rewards/equation_reward_func": 0.4743303842842579,
+      "rewards/format_reward_func": 0.9754464626312256,
+      "step": 446
+    },
+    {
+      "completion_length": 253.82255363464355,
+      "epoch": 0.5574739461813657,
+      "grad_norm": 0.060085523030670523,
+      "kl": 0.220458984375,
+      "learning_rate": 2.595907750671533e-11,
+      "loss": 0.0002,
+      "reward": 1.4587054401636124,
+      "reward_std": 0.1627471437677741,
+      "rewards/equation_reward_func": 0.48214288242161274,
+      "rewards/format_reward_func": 0.976562537252903,
+      "step": 448
+    },
+    {
+      "completion_length": 261.6841678619385,
+      "epoch": 0.5599626691553896,
+      "grad_norm": 0.06873722690901284,
+      "kl": 0.2308349609375,
+      "learning_rate": 0.0,
+      "loss": 0.0002,
+      "reward": 1.4330357909202576,
+      "reward_std": 0.2197583164088428,
+      "rewards/equation_reward_func": 0.4743303870782256,
+      "rewards/format_reward_func": 0.9587053917348385,
+      "step": 450
+    },
+    {
+      "epoch": 0.5599626691553896,
+      "step": 450,
+      "total_flos": 0.0,
+      "train_loss": 0.00020639907060823268,
+      "train_runtime": 22455.4825,
+      "train_samples_per_second": 1.122,
+      "train_steps_per_second": 0.02
+    }
+  ],
+  "logging_steps": 2,
+  "max_steps": 450,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 25,
+  "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": 1,
+  "trial_name": null,
+  "trial_params": null
+}