{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6113092205807438,
  "eval_steps": 500,
  "global_step": 750,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.007859776524128392,
      "clip_ratio/high_mean": 0.0020467473841563333,
      "clip_ratio/low_mean": 0.00029163302679080516,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023383804109471384,
      "entropy": 0.41038384661078453,
      "epoch": 0.000815078960774325,
      "grad_norm": 12.521499633789062,
      "learning_rate": 7.993480032599837e-07,
      "loss": -1.7572,
      "reward": 3.2622119585673013,
      "reward_std": 1.1541353166103363,
      "rewards/reasoning_density_reward/mean": 0.4151560266812642,
      "rewards/reasoning_density_reward/std": 0.09147774428129196,
      "rewards/success_reward/mean": 2.3125,
      "rewards/success_reward/std": 0.9487520456314087,
      "rewards/trajectory_similarity_reward/mean": 0.5345559815565745,
      "rewards/trajectory_similarity_reward/std": 0.33653077483177185,
      "step": 1
    },
    {
      "clip_ratio/high_max": 0.008624728885479271,
      "clip_ratio/high_mean": 0.0015955669296090491,
      "clip_ratio/low_mean": 0.0006949721719138324,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022905391379026696,
      "entropy": 0.38022736832499504,
      "epoch": 0.00163015792154865,
      "grad_norm": 13.362013816833496,
      "learning_rate": 7.986960065199674e-07,
      "loss": -3.8417,
      "reward": 1.9967880249023438,
      "reward_std": 1.5318949222564697,
      "rewards/reasoning_density_reward/mean": 0.3804687559604645,
      "rewards/reasoning_density_reward/std": 0.09584804624319077,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.3038192093372345,
      "rewards/trajectory_similarity_reward/std": 0.391468346118927,
      "step": 2
    },
    {
      "clip_ratio/high_max": 0.008515521243680269,
      "clip_ratio/high_mean": 0.002141209479304962,
      "clip_ratio/low_mean": 0.0007361595562542789,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002877369028283283,
      "entropy": 0.3913104347884655,
      "epoch": 0.002445236882322975,
      "grad_norm": 17.39444923400879,
      "learning_rate": 7.980440097799511e-07,
      "loss": 7.5485,
      "reward": 3.78806471824646,
      "reward_std": 0.687467634677887,
      "rewards/reasoning_density_reward/mean": 0.4416666626930237,
      "rewards/reasoning_density_reward/std": 0.0898558497428894,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.6276480555534363,
      "rewards/trajectory_similarity_reward/std": 0.32661935687065125,
      "step": 3
    },
    {
      "clip_ratio/high_max": 0.01465495937736705,
      "clip_ratio/high_mean": 0.003104845825873781,
      "clip_ratio/low_mean": 0.0005387703713495284,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036436161535675637,
      "entropy": 0.3811783604323864,
      "epoch": 0.0032603158430973,
      "grad_norm": 12.952384948730469,
      "learning_rate": 7.973920130399348e-07,
      "loss": 4.8579,
      "reward": 3.6348241567611694,
      "reward_std": 0.7719940543174744,
      "rewards/reasoning_density_reward/mean": 0.4181423485279083,
      "rewards/reasoning_density_reward/std": 0.0870109274983406,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.7330788373947144,
      "rewards/trajectory_similarity_reward/mean": 0.5916817486286163,
      "rewards/trajectory_similarity_reward/std": 0.34440699219703674,
      "step": 4
    },
    {
      "clip_ratio/high_max": 0.010123982443474233,
      "clip_ratio/high_mean": 0.002097139171382878,
      "clip_ratio/low_mean": 0.00026786292437464,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002365002110309433,
      "entropy": 0.3790912292897701,
      "epoch": 0.004075394803871625,
      "grad_norm": 13.61019515991211,
      "learning_rate": 7.967400162999185e-07,
      "loss": -5.6803,
      "reward": 3.4796690940856934,
      "reward_std": 1.6300475597381592,
      "rewards/reasoning_density_reward/mean": 0.3843005895614624,
      "rewards/reasoning_density_reward/std": 0.09743615984916687,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6578683853149414,
      "rewards/trajectory_similarity_reward/std": 0.35712963342666626,
      "step": 5
    },
    {
      "clip_ratio/high_max": 0.003762884938623756,
      "clip_ratio/high_mean": 0.001275071277632378,
      "clip_ratio/low_mean": 0.0005715068837162107,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018465782013663556,
      "entropy": 0.4013095982372761,
      "epoch": 0.00489047376464595,
      "grad_norm": 14.584135055541992,
      "learning_rate": 7.960880195599022e-07,
      "loss": -0.9102,
      "reward": 1.8393546342849731,
      "reward_std": 1.4517347812652588,
      "rewards/reasoning_density_reward/mean": 0.3333333432674408,
      "rewards/reasoning_density_reward/std": 0.06992058455944061,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.19352132081985474,
      "rewards/trajectory_similarity_reward/std": 0.3487545847892761,
      "step": 6
    },
    {
      "clip_ratio/high_max": 0.009444680355954915,
      "clip_ratio/high_mean": 0.002164443016226869,
      "clip_ratio/low_mean": 0.0007736272236797959,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002938070279924432,
      "entropy": 0.3809679038822651,
      "epoch": 0.005705552725420275,
      "grad_norm": 12.568232536315918,
      "learning_rate": 7.954360228198859e-07,
      "loss": 4.0611,
      "reward": 3.2083773612976074,
      "reward_std": 1.7187258005142212,
      "rewards/reasoning_density_reward/mean": 0.3939732313156128,
      "rewards/reasoning_density_reward/std": 0.12620505690574646,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5644038915634155,
      "rewards/trajectory_similarity_reward/std": 0.3941784203052521,
      "step": 7
    },
    {
      "clip_ratio/high_max": 0.012981796055100858,
      "clip_ratio/high_mean": 0.002625926717882976,
      "clip_ratio/low_mean": 0.00021594019199255854,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002841866938979365,
      "entropy": 0.3893021456897259,
      "epoch": 0.0065206316861946,
      "grad_norm": 11.039835929870605,
      "learning_rate": 7.947840260798696e-07,
      "loss": -1.3529,
      "reward": 3.2191858291625977,
      "reward_std": 1.7478536367416382,
      "rewards/reasoning_density_reward/mean": 0.4227430820465088,
      "rewards/reasoning_density_reward/std": 0.0888771265745163,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.6401925086975098,
      "rewards/trajectory_similarity_reward/std": 0.45292434096336365,
      "step": 8
    },
    {
      "clip_ratio/high_max": 0.01910276967100799,
      "clip_ratio/high_mean": 0.0032809551630634815,
      "clip_ratio/low_mean": 0.00024579951423220336,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035267546481918544,
      "entropy": 0.39846257492899895,
      "epoch": 0.007335710646968925,
      "grad_norm": 12.049652099609375,
      "learning_rate": 7.941320293398533e-07,
      "loss": 10.9283,
      "reward": 2.4605400562286377,
      "reward_std": 1.4281959533691406,
      "rewards/reasoning_density_reward/mean": 0.35877978801727295,
      "rewards/reasoning_density_reward/std": 0.09032869338989258,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.2512494325637817,
      "rewards/trajectory_similarity_reward/mean": 0.3205103278160095,
      "rewards/trajectory_similarity_reward/std": 0.3986041843891144,
      "step": 9
    },
    {
      "clip_ratio/high_max": 0.00945054937619716,
      "clip_ratio/high_mean": 0.0017620270809857175,
      "clip_ratio/low_mean": 0.0008624666152172722,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026244937325827777,
      "entropy": 0.3519035205245018,
      "epoch": 0.00815078960774325,
      "grad_norm": 10.72104549407959,
      "learning_rate": 7.93480032599837e-07,
      "loss": 10.884,
      "reward": 1.9378045797348022,
      "reward_std": 1.5494487285614014,
      "rewards/reasoning_density_reward/mean": 0.36108630895614624,
      "rewards/reasoning_density_reward/std": 0.08137461543083191,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.264218270778656,
      "rewards/trajectory_similarity_reward/std": 0.31887319684028625,
      "step": 10
    },
    {
      "clip_ratio/high_max": 0.008095223049167544,
      "clip_ratio/high_mean": 0.0021544756018556654,
      "clip_ratio/low_mean": 0.0007612036060891114,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002915679171564989,
      "entropy": 0.42530952021479607,
      "epoch": 0.008965868568517575,
      "grad_norm": 12.492753982543945,
      "learning_rate": 7.928280358598207e-07,
      "loss": -7.0991,
      "reward": 3.0185335874557495,
      "reward_std": 0.9666886627674103,
      "rewards/reasoning_density_reward/mean": 0.4312500059604645,
      "rewards/reasoning_density_reward/std": 0.09595610946416855,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.4393417239189148,
      "rewards/trajectory_similarity_reward/mean": 0.6185336410999298,
      "rewards/trajectory_similarity_reward/std": 0.4545029401779175,
      "step": 11
    },
    {
      "clip_ratio/high_max": 0.00813884031958878,
      "clip_ratio/high_mean": 0.0017036209610523656,
      "clip_ratio/low_mean": 0.0001487227709731087,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018523437320254743,
      "entropy": 0.4255274683237076,
      "epoch": 0.0097809475292919,
      "grad_norm": 12.413070678710938,
      "learning_rate": 7.921760391198044e-07,
      "loss": -0.776,
      "reward": 2.646998643875122,
      "reward_std": 1.501129150390625,
      "rewards/reasoning_density_reward/mean": 0.37049853801727295,
      "rewards/reasoning_density_reward/std": 0.0925363227725029,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.1618950366973877,
      "rewards/trajectory_similarity_reward/mean": 0.4015001654624939,
      "rewards/trajectory_similarity_reward/std": 0.4715266227722168,
      "step": 12
    },
    {
      "clip_ratio/high_max": 0.01149896765127778,
      "clip_ratio/high_mean": 0.002402821060968563,
      "clip_ratio/low_mean": 0.00027169084205524996,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026745119248516858,
      "entropy": 0.38611266016960144,
      "epoch": 0.010596026490066225,
      "grad_norm": 13.47108268737793,
      "learning_rate": 7.915240423797881e-07,
      "loss": -6.6575,
      "reward": 1.84372079372406,
      "reward_std": 1.9108951091766357,
      "rewards/reasoning_density_reward/mean": 0.3489583432674408,
      "rewards/reasoning_density_reward/std": 0.10173836350440979,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.2760124206542969,
      "rewards/trajectory_similarity_reward/std": 0.3772246539592743,
      "step": 13
    },
    {
      "clip_ratio/high_max": 0.010418549936730415,
      "clip_ratio/high_mean": 0.0018016824760707095,
      "clip_ratio/low_mean": 0.00039847465086495504,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002200157134211622,
      "entropy": 0.38171733915805817,
      "epoch": 0.01141110545084055,
      "grad_norm": 11.886825561523438,
      "learning_rate": 7.908720456397718e-07,
      "loss": -12.203,
      "reward": 2.094133198261261,
      "reward_std": 1.4255470633506775,
      "rewards/reasoning_density_reward/mean": 0.3463541716337204,
      "rewards/reasoning_density_reward/std": 0.09789494052529335,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.4010097980499268,
      "rewards/trajectory_similarity_reward/mean": 0.29465413093566895,
      "rewards/trajectory_similarity_reward/std": 0.371559277176857,
      "step": 14
    },
    {
      "clip_ratio/high_max": 0.012846630183048546,
      "clip_ratio/high_mean": 0.0025962256477214396,
      "clip_ratio/low_mean": 0.0003824857012659777,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029787112871417776,
      "entropy": 0.3693549111485481,
      "epoch": 0.012226184411614875,
      "grad_norm": 12.709497451782227,
      "learning_rate": 7.902200488997555e-07,
      "loss": 11.5254,
      "reward": 3.2679052352905273,
      "reward_std": 1.8606152534484863,
      "rewards/reasoning_density_reward/mean": 0.43169641494750977,
      "rewards/reasoning_density_reward/std": 0.10895198583602905,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5862088799476624,
      "rewards/trajectory_similarity_reward/std": 0.41185083985328674,
      "step": 15
    },
    {
      "clip_ratio/high_max": 0.007397249573841691,
      "clip_ratio/high_mean": 0.0014249961532186717,
      "clip_ratio/low_mean": 0.000572818229557015,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019978144628112204,
      "entropy": 0.37656884267926216,
      "epoch": 0.0130412633723892,
      "grad_norm": 16.078630447387695,
      "learning_rate": 7.895680521597392e-07,
      "loss": 8.9224,
      "reward": 2.938059091567993,
      "reward_std": 0.734161913394928,
      "rewards/reasoning_density_reward/mean": 0.40037204325199127,
      "rewards/reasoning_density_reward/std": 0.08752777799963951,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 0.5809475183486938,
      "rewards/trajectory_similarity_reward/mean": 0.47518689930438995,
      "rewards/trajectory_similarity_reward/std": 0.27149156481027603,
      "step": 16
    },
    {
      "clip_ratio/high_max": 0.01289211370749399,
      "clip_ratio/high_mean": 0.0020305066063883714,
      "clip_ratio/low_mean": 0.0010653849312802777,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030958915303926915,
      "entropy": 0.40612415596842766,
      "epoch": 0.013856342333163525,
      "grad_norm": 13.875617027282715,
      "learning_rate": 7.889160554197228e-07,
      "loss": -0.4427,
      "reward": 3.209893226623535,
      "reward_std": 1.6665213108062744,
      "rewards/reasoning_density_reward/mean": 0.4048878252506256,
      "rewards/reasoning_density_reward/std": 0.11086837947368622,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5550055503845215,
      "rewards/trajectory_similarity_reward/std": 0.38807398080825806,
      "step": 17
    },
    {
      "clip_ratio/high_max": 0.008637343591544777,
      "clip_ratio/high_mean": 0.0016370482262573205,
      "clip_ratio/low_mean": 0.0001482785155531019,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00178532674908638,
      "entropy": 0.388773862272501,
      "epoch": 0.01467142129393785,
      "grad_norm": 12.720061302185059,
      "learning_rate": 7.882640586797066e-07,
      "loss": 10.1798,
      "reward": 3.069821357727051,
      "reward_std": 1.988619327545166,
      "rewards/reasoning_density_reward/mean": 0.4164930582046509,
      "rewards/reasoning_density_reward/std": 0.09179546684026718,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5908280611038208,
      "rewards/trajectory_similarity_reward/std": 0.43369945883750916,
      "step": 18
    },
    {
      "clip_ratio/high_max": 0.01083475089399144,
      "clip_ratio/high_mean": 0.0022815284901298583,
      "clip_ratio/low_mean": 0.0008069094219536055,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030884378647897393,
      "entropy": 0.38309188559651375,
      "epoch": 0.015486500254712175,
      "grad_norm": 11.791181564331055,
      "learning_rate": 7.876120619396902e-07,
      "loss": 5.9776,
      "reward": 2.587360143661499,
      "reward_std": 1.841870903968811,
      "rewards/reasoning_density_reward/mean": 0.35486114025115967,
      "rewards/reasoning_density_reward/std": 0.07132995128631592,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4512489438056946,
      "rewards/trajectory_similarity_reward/std": 0.4415634274482727,
      "step": 19
    },
    {
      "clip_ratio/high_max": 0.00882295856717974,
      "clip_ratio/high_mean": 0.0013191869074944407,
      "clip_ratio/low_mean": 0.0006177019640745129,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001936888846103102,
      "entropy": 0.3883170895278454,
      "epoch": 0.0163015792154865,
      "grad_norm": 11.376957893371582,
      "learning_rate": 7.86960065199674e-07,
      "loss": 12.208,
      "reward": 1.8601847887039185,
      "reward_std": 1.250142216682434,
      "rewards/reasoning_density_reward/mean": 0.3020833432674408,
      "rewards/reasoning_density_reward/std": 0.0858018770813942,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.0201103687286377,
      "rewards/trajectory_similarity_reward/mean": 0.15185154974460602,
      "rewards/trajectory_similarity_reward/std": 0.3367954194545746,
      "step": 20
    },
    {
      "clip_ratio/high_max": 0.01059138891287148,
      "clip_ratio/high_mean": 0.001953897430212237,
      "clip_ratio/low_mean": 4.189008177490905e-05,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001995787511987146,
      "entropy": 0.4000932015478611,
      "epoch": 0.017116658176260825,
      "grad_norm": 17.00652503967285,
      "learning_rate": 7.863080684596576e-07,
      "loss": -4.7472,
      "step": 21
    },
    {
      "clip_ratio/high_max": 0.012072266486939043,
      "clip_ratio/high_mean": 0.002198079600930214,
      "clip_ratio/low_mean": 0.0001544327096780762,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023525123251602054,
      "entropy": 0.40602242574095726,
      "epoch": 0.01793173713703515,
      "grad_norm": 13.365681648254395,
      "learning_rate": 7.856560717196414e-07,
      "loss": 2.6745,
      "reward": 3.9301629066467285,
      "reward_std": 0.7464324831962585,
      "rewards/reasoning_density_reward/mean": 0.4089657664299011,
      "rewards/reasoning_density_reward/std": 0.06682140380144119,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.7086969614028931,
      "rewards/trajectory_similarity_reward/std": 0.3320273160934448,
      "step": 22
    },
    {
      "clip_ratio/high_max": 0.01033702792483382,
      "clip_ratio/high_mean": 0.002311987609573407,
      "clip_ratio/low_mean": 0.0007027549945632927,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003014742673258297,
      "entropy": 0.40249646827578545,
      "epoch": 0.018746816097809475,
      "grad_norm": 11.500093460083008,
      "learning_rate": 7.85004074979625e-07,
      "loss": -0.0164,
      "reward": 2.1643571853637695,
      "reward_std": 0.33569949865341187,
      "rewards/reasoning_density_reward/mean": 0.3882812559604645,
      "rewards/reasoning_density_reward/std": 0.10180339217185974,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.18232598900794983,
      "rewards/trajectory_similarity_reward/std": 0.19711516797542572,
      "step": 23
    },
    {
      "clip_ratio/high_max": 0.007841259706765413,
      "clip_ratio/high_mean": 0.0012821430136682466,
      "clip_ratio/low_mean": 0.0007857897435314953,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002067932713543996,
      "entropy": 0.3761267885565758,
      "epoch": 0.0195618950585838,
      "grad_norm": 13.018792152404785,
      "learning_rate": 7.843520782396088e-07,
      "loss": -0.5297,
      "reward": 1.9293029308319092,
      "reward_std": 1.4434521198272705,
      "rewards/reasoning_density_reward/mean": 0.33348214626312256,
      "rewards/reasoning_density_reward/std": 0.06141730025410652,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.2808688879013062,
      "rewards/trajectory_similarity_reward/mean": 0.18957073986530304,
      "rewards/trajectory_similarity_reward/std": 0.3052114248275757,
      "step": 24
    },
    {
      "clip_ratio/high_max": 0.008468169951811433,
      "clip_ratio/high_mean": 0.0017096884839702398,
      "clip_ratio/low_mean": 0.000630402319075074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002340090803045314,
      "entropy": 0.3720705918967724,
      "epoch": 0.020376974019358125,
      "grad_norm": 12.41434383392334,
      "learning_rate": 7.837000814995924e-07,
      "loss": -2.139,
      "reward": 2.76846981048584,
      "reward_std": 1.4001634120941162,
      "rewards/reasoning_density_reward/mean": 0.3743055462837219,
      "rewards/reasoning_density_reward/std": 0.09293732047080994,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.3316643238067627,
      "rewards/trajectory_similarity_reward/std": 0.4065845012664795,
      "step": 25
    },
    {
      "clip_ratio/high_max": 0.006848955526947975,
      "clip_ratio/high_mean": 0.0009821274870773777,
      "clip_ratio/low_mean": 0.0005009113119740505,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014830387881374918,
      "entropy": 0.36622409150004387,
      "epoch": 0.02119205298013245,
      "grad_norm": 13.468979835510254,
      "learning_rate": 7.830480847595762e-07,
      "loss": -0.593,
      "reward": 1.8094419240951538,
      "reward_std": 1.1419930458068848,
      "rewards/reasoning_density_reward/mean": 0.3560554087162018,
      "rewards/reasoning_density_reward/std": 0.07812166959047318,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.23463642597198486,
      "rewards/trajectory_similarity_reward/std": 0.3242618143558502,
      "step": 26
    },
    {
      "clip_ratio/high_max": 0.014098172105150297,
      "clip_ratio/high_mean": 0.0025445934625167865,
      "clip_ratio/low_mean": 0.00045481944107450545,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029994129145052284,
      "entropy": 0.38702986389398575,
      "epoch": 0.022007131940906775,
      "grad_norm": 10.025175094604492,
      "learning_rate": 7.823960880195598e-07,
      "loss": -12.1301,
      "step": 27
    },
    {
      "clip_ratio/high_max": 0.008215140929678455,
      "clip_ratio/high_mean": 0.001683556201896863,
      "clip_ratio/low_mean": 6.537656736327335e-05,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017489327692601364,
      "entropy": 0.36855505034327507,
      "epoch": 0.0228222109016811,
      "grad_norm": 15.635762214660645,
      "learning_rate": 7.817440912795436e-07,
      "loss": 2.4332,
      "reward": 2.4929356575012207,
      "reward_std": 1.825430989265442,
      "rewards/reasoning_density_reward/mean": 0.3868551552295685,
      "rewards/reasoning_density_reward/std": 0.09320401400327682,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.32483047246932983,
      "rewards/trajectory_similarity_reward/std": 0.3603823781013489,
      "step": 28
    },
    {
      "clip_ratio/high_max": 0.012339716835413128,
      "clip_ratio/high_mean": 0.002353316383960191,
      "clip_ratio/low_mean": 0.0005386052871472202,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002891921700211242,
      "entropy": 0.40603727474808693,
      "epoch": 0.023637289862455425,
      "grad_norm": 11.24703311920166,
      "learning_rate": 7.810920945395272e-07,
      "loss": 7.5223,
      "reward": 2.8720836639404297,
      "reward_std": 1.5154728889465332,
      "rewards/reasoning_density_reward/mean": 0.3824442923069,
      "rewards/reasoning_density_reward/std": 0.10641347616910934,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.3333892226219177,
      "rewards/trajectory_similarity_reward/std": 0.3254243731498718,
      "step": 29
    },
    {
      "clip_ratio/high_max": 0.0035803994978778064,
      "clip_ratio/high_mean": 0.0006763349811080843,
      "clip_ratio/low_mean": 0.000665383275190834,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013417182562989183,
      "entropy": 0.37067585438489914,
      "epoch": 0.02445236882322975,
      "grad_norm": 12.155799865722656,
      "learning_rate": 7.80440097799511e-07,
      "loss": 11.4955,
      "reward": 1.9905742406845093,
      "reward_std": 1.4096477031707764,
      "rewards/reasoning_density_reward/mean": 0.35803571343421936,
      "rewards/reasoning_density_reward/std": 0.09042134881019592,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.13253852725028992,
      "rewards/trajectory_similarity_reward/std": 0.24700158834457397,
      "step": 30
    },
    {
      "clip_ratio/high_max": 0.010327437688829377,
      "clip_ratio/high_mean": 0.001750317318510497,
      "clip_ratio/low_mean": 0.000592503696680069,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002342820997000672,
      "entropy": 0.3968063220381737,
      "epoch": 0.025267447784004075,
      "grad_norm": 18.070541381835938,
      "learning_rate": 7.797881010594947e-07,
      "loss": 1.2454,
      "reward": 1.600077509880066,
      "reward_std": 1.6434695720672607,
      "rewards/reasoning_density_reward/mean": 0.3266369104385376,
      "rewards/reasoning_density_reward/std": 0.04349801689386368,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.24219061434268951,
      "rewards/trajectory_similarity_reward/std": 0.3727206587791443,
      "step": 31
    },
    {
      "clip_ratio/high_max": 0.010534149303566664,
      "clip_ratio/high_mean": 0.001544669496070128,
      "clip_ratio/low_mean": 0.0002654226482263766,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018100921588484198,
      "entropy": 0.367525827139616,
      "epoch": 0.0260825267447784,
      "grad_norm": 14.136274337768555,
      "learning_rate": 7.791361043194784e-07,
      "loss": 1.2516,
      "reward": 0.5252794027328491,
      "reward_std": 0.6371865272521973,
      "rewards/reasoning_density_reward/mean": 0.3020833432674408,
      "rewards/reasoning_density_reward/std": 0.008333333767950535,
      "rewards/success_reward/mean": 0.1875,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.0356961190700531,
      "rewards/trajectory_similarity_reward/std": 0.1427844613790512,
      "step": 32
    },
    {
      "clip_ratio/high_max": 0.011171146237757057,
      "clip_ratio/high_mean": 0.0016880335024325177,
      "clip_ratio/low_mean": 0.0005863933911314234,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022744269081158563,
      "entropy": 0.3936876095831394,
      "epoch": 0.026897605705552725,
      "grad_norm": 8.294105529785156,
      "learning_rate": 7.784841075794621e-07,
      "loss": -1.1423,
      "reward": 3.445535659790039,
      "reward_std": 1.123462200164795,
      "rewards/reasoning_density_reward/mean": 0.4334697425365448,
      "rewards/reasoning_density_reward/std": 0.10052753984928131,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.5745657682418823,
      "rewards/trajectory_similarity_reward/std": 0.3378354609012604,
      "step": 33
    },
    {
      "clip_ratio/high_max": 0.008225160650908947,
      "clip_ratio/high_mean": 0.0016372287063859403,
      "clip_ratio/low_mean": 0.0004229458609188441,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020601745854946785,
      "entropy": 0.37133844569325447,
      "epoch": 0.02771268466632705,
      "grad_norm": 12.370161056518555,
      "learning_rate": 7.778321108394458e-07,
      "loss": 11.573,
      "reward": 3.399420738220215,
      "reward_std": 1.5146331787109375,
      "rewards/reasoning_density_reward/mean": 0.43095237016677856,
      "rewards/reasoning_density_reward/std": 0.08495696634054184,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.5309681296348572,
      "rewards/trajectory_similarity_reward/std": 0.3819196820259094,
      "step": 34
    },
    {
      "clip_ratio/high_max": 0.014310821366962045,
      "clip_ratio/high_mean": 0.0026903004618361592,
      "clip_ratio/low_mean": 5.841121674166061e-05,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002748711674939841,
      "entropy": 0.41421492770314217,
      "epoch": 0.028527763627101375,
      "grad_norm": 11.206388473510742,
      "learning_rate": 7.771801140994295e-07,
      "loss": -7.6214,
      "reward": 2.5901639461517334,
      "reward_std": 1.6561309099197388,
      "rewards/reasoning_density_reward/mean": 0.3921875059604645,
      "rewards/reasoning_density_reward/std": 0.09988272935152054,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.32297641038894653,
      "rewards/trajectory_similarity_reward/std": 0.4291108250617981,
      "step": 35
    },
    {
      "clip_ratio/high_max": 0.015765533258672804,
      "clip_ratio/high_mean": 0.003418349224375561,
      "clip_ratio/low_mean": 0.0013017970231885556,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004720146331237629,
      "entropy": 0.40097955241799355,
      "epoch": 0.0293428425878757,
      "grad_norm": 13.552125930786133,
      "learning_rate": 7.765281173594132e-07,
      "loss": -7.8976,
      "reward": 1.870129108428955,
      "reward_std": 1.3373193740844727,
      "rewards/reasoning_density_reward/mean": 0.33271104097366333,
      "rewards/reasoning_density_reward/std": 0.07297610491514206,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.157853603363037,
      "rewards/trajectory_similarity_reward/mean": 0.13116803765296936,
      "rewards/trajectory_similarity_reward/std": 0.2511459290981293,
      "step": 36
    },
    {
      "clip_ratio/high_max": 0.010484583268407732,
      "clip_ratio/high_mean": 0.0025313855512649752,
      "clip_ratio/low_mean": 0.0005432693578768522,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003074654938245658,
      "entropy": 0.41323524713516235,
      "epoch": 0.030157921548650025,
      "grad_norm": 13.042089462280273,
      "learning_rate": 7.758761206193969e-07,
      "loss": 1.6761,
      "reward": 3.282583236694336,
      "reward_std": 1.5460436344146729,
      "rewards/reasoning_density_reward/mean": 0.39887940883636475,
      "rewards/reasoning_density_reward/std": 0.08201368153095245,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.5399540066719055,
      "rewards/trajectory_similarity_reward/std": 0.38468894362449646,
      "step": 37
    },
    {
      "clip_ratio/high_max": 0.009271231421735138,
      "clip_ratio/high_mean": 0.0017663503822404891,
      "clip_ratio/low_mean": 0.00048291530765709467,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022492657008115202,
      "entropy": 0.4282389096915722,
      "epoch": 0.03097300050942435,
      "grad_norm": 13.600410461425781,
      "learning_rate": 7.752241238793806e-07,
      "loss": 1.6546,
      "reward": 2.449902296066284,
      "reward_std": 1.4987552165985107,
      "rewards/reasoning_density_reward/mean": 0.3674479126930237,
      "rewards/reasoning_density_reward/std": 0.0758015587925911,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.20745432376861572,
      "rewards/trajectory_similarity_reward/std": 0.3380158245563507,
      "step": 38
    },
    {
      "clip_ratio/high_max": 0.007575604482553899,
      "clip_ratio/high_mean": 0.0013556054764194414,
      "clip_ratio/low_mean": 0.000784342541010119,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021399480174295604,
      "entropy": 0.3956512324512005,
      "epoch": 0.031788079470198675,
      "grad_norm": 13.522518157958984,
      "learning_rate": 7.745721271393643e-07,
      "loss": -5.893,
      "reward": 2.4626083374023438,
      "reward_std": 0.7695376873016357,
      "rewards/reasoning_density_reward/mean": 0.37174874544143677,
      "rewards/reasoning_density_reward/std": 0.07838289439678192,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.403359591960907,
      "rewards/trajectory_similarity_reward/std": 0.3868091106414795,
      "step": 39
    },
    {
      "clip_ratio/high_max": 0.009787353337742388,
      "clip_ratio/high_mean": 0.00217960421286989,
      "clip_ratio/low_mean": 0.0007625155631103553,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029421198123600334,
      "entropy": 0.40306348353624344,
      "epoch": 0.032603158430973,
      "grad_norm": 13.620566368103027,
      "learning_rate": 7.73920130399348e-07,
      "loss": -0.2347,
      "reward": 3.704893112182617,
      "reward_std": 0.9524960517883301,
      "rewards/reasoning_density_reward/mean": 0.4292658567428589,
      "rewards/reasoning_density_reward/std": 0.08231482654809952,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.5568775534629822,
      "rewards/trajectory_similarity_reward/std": 0.35786157846450806,
      "step": 40
    },
    {
      "clip_ratio/high_max": 0.009275955671910197,
      "clip_ratio/high_mean": 0.0018230690475320444,
      "clip_ratio/low_mean": 0.00019518865155987442,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020182576990919188,
      "entropy": 0.3331269100308418,
      "epoch": 0.033418237391747325,
      "grad_norm": 12.218531608581543,
      "learning_rate": 7.732681336593317e-07,
      "loss": -3.0496,
      "reward": 4.055032253265381,
      "reward_std": 0.69251948595047,
      "rewards/reasoning_density_reward/mean": 0.4703125059604645,
      "rewards/reasoning_density_reward/std": 0.05602320656180382,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.7722196578979492,
      "rewards/trajectory_similarity_reward/std": 0.31565535068511963,
      "step": 41
    },
    {
      "clip_ratio/high_max": 0.005086818244308233,
      "clip_ratio/high_mean": 0.0007639260438736528,
      "clip_ratio/low_mean": 0.0007853527058614418,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015492787497350946,
      "entropy": 0.3765343241393566,
      "epoch": 0.03423331635252165,
      "grad_norm": 11.417027473449707,
      "learning_rate": 7.726161369193154e-07,
      "loss": 15.4077,
      "reward": 1.8246607780456543,
      "reward_std": 2.0303993225097656,
      "rewards/reasoning_density_reward/mean": 0.3567708730697632,
      "rewards/reasoning_density_reward/std": 0.08762391656637192,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.34289002418518066,
      "rewards/trajectory_similarity_reward/std": 0.4584181308746338,
      "step": 42
    },
    {
      "clip_ratio/high_max": 0.00596946181030944,
      "clip_ratio/high_mean": 0.0010919155101873912,
      "clip_ratio/low_mean": 0.0004942744890286122,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001586189988302067,
      "entropy": 0.37504440173506737,
      "epoch": 0.035048395313295976,
      "grad_norm": 11.808359146118164,
      "learning_rate": 7.719641401792991e-07,
      "loss": -2.7218,
      "reward": 1.608153223991394,
      "reward_std": 0.9072791934013367,
      "rewards/reasoning_density_reward/mean": 0.34732142090797424,
      "rewards/reasoning_density_reward/std": 0.07625927031040192,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.2295818030834198,
      "rewards/trajectory_similarity_reward/std": 0.3543436825275421,
      "step": 43
    },
    {
      "clip_ratio/high_max": 0.00774767150869593,
      "clip_ratio/high_mean": 0.001423423338565044,
      "clip_ratio/low_mean": 0.00047843762149568647,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019018609309569001,
      "entropy": 0.3837415762245655,
      "epoch": 0.0358634742740703,
      "grad_norm": 11.643194198608398,
      "learning_rate": 7.713121434392828e-07,
      "loss": 2.6115,
      "reward": 2.2679197788238525,
      "reward_std": 1.626136064529419,
      "rewards/reasoning_density_reward/mean": 0.3673611283302307,
      "rewards/reasoning_density_reward/std": 0.10089873522520065,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.30680862069129944,
      "rewards/trajectory_similarity_reward/std": 0.3932991325855255,
      "step": 44
    },
    {
      "clip_ratio/high_max": 0.004595555452397093,
      "clip_ratio/high_mean": 0.0008941302185121458,
      "clip_ratio/low_mean": 0.0002996013790834695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001193731604871573,
      "entropy": 0.36634326353669167,
      "epoch": 0.036678553234844626,
      "grad_norm": 12.331775665283203,
      "learning_rate": 7.706601466992665e-07,
      "loss": 5.9725,
      "reward": 2.5841710567474365,
      "reward_std": 1.8813647031784058,
      "rewards/reasoning_density_reward/mean": 0.37118056416511536,
      "rewards/reasoning_density_reward/std": 0.08488599210977554,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.431740403175354,
      "rewards/trajectory_similarity_reward/std": 0.4570370018482208,
      "step": 45
    },
    {
      "clip_ratio/high_max": 0.008579818881116807,
      "clip_ratio/high_mean": 0.0016704537702025846,
      "clip_ratio/low_mean": 0.0002803377101372462,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019507914475980215,
      "entropy": 0.42372840642929077,
      "epoch": 0.03749363219561895,
      "grad_norm": 12.786086082458496,
      "learning_rate": 7.700081499592503e-07,
      "loss": 2.8538,
      "reward": 2.664923667907715,
      "reward_std": 1.7742314338684082,
      "rewards/reasoning_density_reward/mean": 0.363167405128479,
      "rewards/reasoning_density_reward/std": 0.08593014627695084,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.4267563223838806,
      "rewards/trajectory_similarity_reward/std": 0.4531887173652649,
      "step": 46
    },
    {
      "clip_ratio/high_max": 0.011128685844596475,
      "clip_ratio/high_mean": 0.0020314836656325497,
      "clip_ratio/low_mean": 0.0002699101187317865,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023013938043732196,
      "entropy": 0.37759581953287125,
      "epoch": 0.038308711156393276,
      "grad_norm": 11.821990966796875,
      "learning_rate": 7.693561532192339e-07,
      "loss": -0.6891,
      "reward": 2.528106927871704,
      "reward_std": 1.1531633138656616,
      "rewards/reasoning_density_reward/mean": 0.3743179440498352,
      "rewards/reasoning_density_reward/std": 0.10550232976675034,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.37253886461257935,
      "rewards/trajectory_similarity_reward/std": 0.3894842267036438,
      "step": 47
    },
    {
      "clip_ratio/high_max": 0.008950168848969042,
      "clip_ratio/high_mean": 0.001980014509172179,
      "clip_ratio/low_mean": 0.0005675486609106883,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002547563206462655,
      "entropy": 0.39086730405688286,
      "epoch": 0.0391237901171676,
      "grad_norm": 17.126033782958984,
      "learning_rate": 7.687041564792176e-07,
      "loss": -7.5243,
      "reward": 3.0080385208129883,
      "reward_std": 1.4505555629730225,
      "rewards/reasoning_density_reward/mean": 0.38452380895614624,
      "rewards/reasoning_density_reward/std": 0.10689388960599899,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.3735148012638092,
      "rewards/trajectory_similarity_reward/std": 0.3396790325641632,
      "step": 48
    },
    {
      "clip_ratio/high_max": 0.009602706792065874,
      "clip_ratio/high_mean": 0.002211471823102329,
      "clip_ratio/low_mean": 0.0005230822498560883,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027345540729584172,
      "entropy": 0.39226942881941795,
      "epoch": 0.039938869077941926,
      "grad_norm": 15.912209510803223,
      "learning_rate": 7.680521597392013e-07,
      "loss": -1.0174,
      "reward": 2.819272756576538,
      "reward_std": 0.9264771342277527,
      "rewards/reasoning_density_reward/mean": 0.40069445967674255,
      "rewards/reasoning_density_reward/std": 0.0907360091805458,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5435782074928284,
      "rewards/trajectory_similarity_reward/std": 0.4595146179199219,
      "step": 49
    },
    {
      "clip_ratio/high_max": 0.013075993047095835,
      "clip_ratio/high_mean": 0.0025862200709525496,
      "clip_ratio/low_mean": 0.0005601635275525041,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003146383533021435,
      "entropy": 0.4194720983505249,
      "epoch": 0.04075394803871625,
      "grad_norm": 16.060813903808594,
      "learning_rate": 7.67400162999185e-07,
      "loss": -6.2949,
      "reward": 2.23777174949646,
      "reward_std": 1.3939740657806396,
      "rewards/reasoning_density_reward/mean": 0.3682291805744171,
      "rewards/reasoning_density_reward/std": 0.07947854697704315,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.3695424497127533,
      "rewards/trajectory_similarity_reward/std": 0.38748934864997864,
      "step": 50
    },
    {
      "clip_ratio/high_max": 0.01398004055954516,
      "clip_ratio/high_mean": 0.0026551064802333713,
      "clip_ratio/low_mean": 0.0005711726553272456,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032262791937682778,
      "entropy": 0.39375386387109756,
      "epoch": 0.041569026999490576,
      "grad_norm": 9.149628639221191,
      "learning_rate": 7.667481662591686e-07,
      "loss": -4.0981,
      "reward": 2.560924530029297,
      "reward_std": 1.5052707195281982,
      "rewards/reasoning_density_reward/mean": 0.35524338483810425,
      "rewards/reasoning_density_reward/std": 0.05022336170077324,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.3306809663772583,
      "rewards/trajectory_similarity_reward/std": 0.34178388118743896,
      "step": 51
    },
    {
      "clip_ratio/high_max": 0.014428187336307019,
      "clip_ratio/high_mean": 0.0029552413689089008,
      "clip_ratio/low_mean": 0.00035374950675759465,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033089908756664954,
      "entropy": 0.3927930220961571,
      "epoch": 0.0423841059602649,
      "grad_norm": 15.01794719696045,
      "learning_rate": 7.660961695191524e-07,
      "loss": 0.1048,
      "reward": 2.5342326164245605,
      "reward_std": 1.5871641635894775,
      "rewards/reasoning_density_reward/mean": 0.3747361898422241,
      "rewards/reasoning_density_reward/std": 0.08767645806074142,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.28449636697769165,
      "rewards/trajectory_similarity_reward/std": 0.31795307993888855,
      "step": 52
    },
    {
      "clip_ratio/high_max": 0.00996944890357554,
      "clip_ratio/high_mean": 0.0021074789110571146,
      "clip_ratio/low_mean": 0.00021669361740350723,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023241725211846642,
      "entropy": 0.388361606746912,
      "epoch": 0.043199184921039226,
      "grad_norm": 13.605493545532227,
      "learning_rate": 7.65444172779136e-07,
      "loss": 0.7871,
      "reward": 2.6952104568481445,
      "reward_std": 1.5655244588851929,
      "rewards/reasoning_density_reward/mean": 0.40234375,
      "rewards/reasoning_density_reward/std": 0.09706299006938934,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.41786643862724304,
      "rewards/trajectory_similarity_reward/std": 0.3985233008861542,
      "step": 53
    },
    {
      "clip_ratio/high_max": 0.00785830058157444,
      "clip_ratio/high_mean": 0.0017957580494112335,
      "clip_ratio/low_mean": 0.0007073937449604273,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025031518234754913,
      "entropy": 0.40704749897122383,
      "epoch": 0.04401426388181355,
      "grad_norm": 13.771414756774902,
      "learning_rate": 7.647921760391198e-07,
      "loss": 9.2311,
      "reward": 2.99031400680542,
      "reward_std": 1.994948148727417,
      "rewards/reasoning_density_reward/mean": 0.40677082538604736,
      "rewards/reasoning_density_reward/std": 0.09809381514787674,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.6147930026054382,
      "rewards/trajectory_similarity_reward/std": 0.4934585392475128,
      "step": 54
    },
    {
      "clip_ratio/high_max": 0.010499781812541187,
      "clip_ratio/high_mean": 0.0028489500837167725,
      "clip_ratio/low_mean": 0.0005846838939760346,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003433633944950998,
      "entropy": 0.4001888781785965,
      "epoch": 0.044829342842587876,
      "grad_norm": 13.307802200317383,
      "learning_rate": 7.641401792991034e-07,
      "loss": 3.6608,
      "reward": 3.7499741315841675,
      "reward_std": 1.1213011741638184,
      "rewards/reasoning_density_reward/mean": 0.44999998807907104,
      "rewards/reasoning_density_reward/std": 0.08818264305591583,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.6613698303699493,
      "rewards/trajectory_similarity_reward/mean": 0.6749741435050964,
      "rewards/trajectory_similarity_reward/std": 0.4104381650686264,
      "step": 55
    },
    {
      "clip_ratio/high_max": 0.009336592047475278,
      "clip_ratio/high_mean": 0.0016059127083281055,
      "clip_ratio/low_mean": 0.0007502674052375369,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002356180084461812,
      "entropy": 0.417436346411705,
      "epoch": 0.0456444218033622,
      "grad_norm": 12.433035850524902,
      "learning_rate": 7.634881825590872e-07,
      "loss": 8.7999,
      "reward": 1.3389360904693604,
      "reward_std": 1.7033448219299316,
      "rewards/reasoning_density_reward/mean": 0.3074979782104492,
      "rewards/reasoning_density_reward/std": 0.12252127379179001,
      "rewards/success_reward/mean": 0.84375,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.18768811225891113,
      "rewards/trajectory_similarity_reward/std": 0.34366127848625183,
      "step": 56
    },
    {
      "clip_ratio/high_max": 0.004143277590628713,
      "clip_ratio/high_mean": 0.0007454764636349864,
      "clip_ratio/low_mean": 0.0009602927311789244,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017057692166417837,
      "entropy": 0.38056887313723564,
      "epoch": 0.046459500764136526,
      "grad_norm": 11.68018627166748,
      "learning_rate": 7.628361858190708e-07,
      "loss": 7.0803,
      "reward": 2.950498104095459,
      "reward_std": 1.4176007509231567,
      "rewards/reasoning_density_reward/mean": 0.40442711114883423,
      "rewards/reasoning_density_reward/std": 0.08222179114818573,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.29607093334198,
      "rewards/trajectory_similarity_reward/std": 0.22726063430309296,
      "step": 57
    },
    {
      "clip_ratio/high_max": 0.01211570220766589,
      "clip_ratio/high_mean": 0.002394643939624075,
      "clip_ratio/low_mean": 0.0004616901023837272,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028563340165419504,
      "entropy": 0.37350625172257423,
      "epoch": 0.04727457972491085,
      "grad_norm": 12.497527122497559,
      "learning_rate": 7.621841890790546e-07,
      "loss": -7.9406,
      "reward": 2.2789666652679443,
      "reward_std": 1.1253360509872437,
      "rewards/reasoning_density_reward/mean": 0.34715908765792847,
      "rewards/reasoning_density_reward/std": 0.08497082442045212,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 0.8158584833145142,
      "rewards/trajectory_similarity_reward/mean": 0.15055763721466064,
      "rewards/trajectory_similarity_reward/std": 0.2809356451034546,
      "step": 58
    },
    {
      "clip_ratio/high_max": 0.010915369144640863,
      "clip_ratio/high_mean": 0.0016982024681055918,
      "clip_ratio/low_mean": 0.00116630466436618,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002864507187041454,
      "entropy": 0.36868639290332794,
      "epoch": 0.048089658685685176,
      "grad_norm": 15.073084831237793,
      "learning_rate": 7.615321923390382e-07,
      "loss": 7.7038,
      "reward": 3.9502639770507812,
      "reward_std": 0.8137335181236267,
      "rewards/reasoning_density_reward/mean": 0.4520833492279053,
      "rewards/reasoning_density_reward/std": 0.07790663838386536,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8158584833145142,
      "rewards/trajectory_similarity_reward/mean": 0.7794308662414551,
      "rewards/trajectory_similarity_reward/std": 0.3159634470939636,
      "step": 59
    },
    {
      "clip_ratio/high_max": 0.010881754162255675,
      "clip_ratio/high_mean": 0.002530091645894572,
      "clip_ratio/low_mean": 0.0006667685993306804,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031968602997949347,
      "entropy": 0.3731381483376026,
      "epoch": 0.0489047376464595,
      "grad_norm": 12.888964653015137,
      "learning_rate": 7.60880195599022e-07,
      "loss": 10.1272,
      "reward": 2.872649669647217,
      "reward_std": 1.0361273288726807,
      "rewards/reasoning_density_reward/mean": 0.3818204402923584,
      "rewards/reasoning_density_reward/std": 0.09364991635084152,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 0.9437293410301208,
      "rewards/trajectory_similarity_reward/mean": 0.3345792293548584,
      "rewards/trajectory_similarity_reward/std": 0.3774014413356781,
      "step": 60
    },
    {
      "clip_ratio/high_max": 0.00836676312610507,
      "clip_ratio/high_mean": 0.0022389110235963017,
      "clip_ratio/low_mean": 0.0007510695941164158,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00298998061043676,
      "entropy": 0.38083334639668465,
      "epoch": 0.049719816607233826,
      "grad_norm": 12.694096565246582,
      "learning_rate": 7.602281988590057e-07,
      "loss": -3.0015,
      "step": 61
    },
    {
      "clip_ratio/high_max": 0.003610127721913159,
      "clip_ratio/high_mean": 0.000492601942823967,
      "clip_ratio/low_mean": 0.0006639954954152927,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0011565974637051113,
      "entropy": 0.3874463737010956,
      "epoch": 0.05053489556800815,
      "grad_norm": 14.215960502624512,
      "learning_rate": 7.595762021189894e-07,
      "loss": -7.0807,
      "reward": 1.4311048984527588,
      "reward_std": 1.812423586845398,
      "rewards/reasoning_density_reward/mean": 0.3360164165496826,
      "rewards/reasoning_density_reward/std": 0.06632942333817482,
      "rewards/success_reward/mean": 0.84375,
      "rewards/success_reward/std": 1.3546693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.2513384073972702,
      "rewards/trajectory_similarity_reward/std": 0.40264730155467987,
      "step": 62
    },
    {
      "clip_ratio/high_max": 0.00636323617072776,
      "clip_ratio/high_mean": 0.0011081686097895727,
      "clip_ratio/low_mean": 0.001232019865710754,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023401884973281994,
      "entropy": 0.43159306794404984,
      "epoch": 0.051349974528782476,
      "grad_norm": 25.600139617919922,
      "learning_rate": 7.589242053789731e-07,
      "loss": 3.2884,
      "reward": 2.971198081970215,
      "reward_std": 1.126588225364685,
      "rewards/reasoning_density_reward/mean": 0.3828125,
      "rewards/reasoning_density_reward/std": 0.11500678956508636,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 0.9437293410301208,
      "rewards/trajectory_similarity_reward/mean": 0.43213552236557007,
      "rewards/trajectory_similarity_reward/std": 0.45934852957725525,
      "step": 63
    },
    {
      "clip_ratio/high_max": 0.008690592076163739,
      "clip_ratio/high_mean": 0.0017426507620257325,
      "clip_ratio/low_mean": 0.0006316464423434809,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023742971825413406,
      "entropy": 0.40060946345329285,
      "epoch": 0.0521650534895568,
      "grad_norm": 11.760992050170898,
      "learning_rate": 7.582722086389568e-07,
      "loss": -3.9461,
      "reward": 4.395369529724121,
      "reward_std": 0.08993902802467346,
      "rewards/reasoning_density_reward/mean": 0.4546875059604645,
      "rewards/reasoning_density_reward/std": 0.0607033409178257,
      "rewards/success_reward/mean": 3.0,
      "rewards/success_reward/std": 0.0,
      "rewards/trajectory_similarity_reward/mean": 0.940682590007782,
      "rewards/trajectory_similarity_reward/std": 0.08018424361944199,
      "step": 64
    },
    {
      "clip_ratio/high_max": 0.014463506668107584,
      "clip_ratio/high_mean": 0.002992144425661536,
      "clip_ratio/low_mean": 0.00039146686322055757,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033836112888820935,
      "entropy": 0.38492605835199356,
      "epoch": 0.052980132450331126,
      "grad_norm": 13.80599594116211,
      "learning_rate": 7.576202118989405e-07,
      "loss": -4.3406,
      "reward": 3.665117621421814,
      "reward_std": 1.3457306623458862,
      "rewards/reasoning_density_reward/mean": 0.4258192330598831,
      "rewards/reasoning_density_reward/std": 0.07829459384083748,
      "rewards/success_reward/mean": 2.578125,
      "rewards/success_reward/std": 0.947025716304779,
      "rewards/trajectory_similarity_reward/mean": 0.6611732542514801,
      "rewards/trajectory_similarity_reward/std": 0.37063828110694885,
      "step": 65
    },
    {
      "clip_ratio/high_max": 0.0184517502784729,
      "clip_ratio/high_mean": 0.003712194797117263,
      "clip_ratio/low_mean": 0.00047715257096569985,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004189347368082963,
      "entropy": 0.39641857519745827,
      "epoch": 0.05379521141110545,
      "grad_norm": 12.762567520141602,
      "learning_rate": 7.569682151589242e-07,
      "loss": -3.8246,
      "reward": 2.8577425479888916,
      "reward_std": 1.7262508869171143,
      "rewards/reasoning_density_reward/mean": 0.3756212592124939,
      "rewards/reasoning_density_reward/std": 0.07759054005146027,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.4196212589740753,
      "rewards/trajectory_similarity_reward/std": 0.38321223855018616,
      "step": 66
    },
    {
      "clip_ratio/high_max": 0.012834838591516018,
      "clip_ratio/high_mean": 0.002887967108108569,
      "clip_ratio/low_mean": 0.000285057642031461,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031730247428640723,
      "entropy": 0.3853634148836136,
      "epoch": 0.054610290371879776,
      "grad_norm": 11.58923625946045,
      "learning_rate": 7.563162184189079e-07,
      "loss": -6.9117,
      "reward": 1.993050456047058,
      "reward_std": 1.2376649379730225,
      "rewards/reasoning_density_reward/mean": 0.32187503576278687,
      "rewards/reasoning_density_reward/std": 0.13287682831287384,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.17117543518543243,
      "rewards/trajectory_similarity_reward/std": 0.26233983039855957,
      "step": 67
    },
    {
      "clip_ratio/high_max": 0.012676209909841418,
      "clip_ratio/high_mean": 0.0028458970846259035,
      "clip_ratio/low_mean": 0.0005760463282058481,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003421943460125476,
      "entropy": 0.40147677436470985,
      "epoch": 0.0554253693326541,
      "grad_norm": 12.954485893249512,
      "learning_rate": 7.556642216788916e-07,
      "loss": -8.3021,
      "reward": 3.0348451137542725,
      "reward_std": 1.3245049715042114,
      "rewards/reasoning_density_reward/mean": 0.3583333492279053,
      "rewards/reasoning_density_reward/std": 0.08563487976789474,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 0.7745966911315918,
      "rewards/trajectory_similarity_reward/mean": 0.42651188373565674,
      "rewards/trajectory_similarity_reward/std": 0.4511069655418396,
      "step": 68
    },
    {
      "clip_ratio/high_max": 0.009144828538410366,
      "clip_ratio/high_mean": 0.0023918721126392484,
      "clip_ratio/low_mean": 0.0006949734233785421,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003086845521465875,
      "entropy": 0.39229365438222885,
      "epoch": 0.056240448293428426,
      "grad_norm": 12.73189640045166,
      "learning_rate": 7.550122249388753e-07,
      "loss": -10.357,
      "reward": 3.0065780878067017,
      "reward_std": 0.9408829808235168,
      "rewards/reasoning_density_reward/mean": 0.3507874608039856,
      "rewards/reasoning_density_reward/std": 0.11689263209700584,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 0.839354395866394,
      "rewards/trajectory_similarity_reward/mean": 0.4057907462120056,
      "rewards/trajectory_similarity_reward/std": 0.43393878638744354,
      "step": 69
    },
    {
      "clip_ratio/high_max": 0.008590885641751811,
      "clip_ratio/high_mean": 0.0015785393661644775,
      "clip_ratio/low_mean": 0.0003228592104278505,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001901398558402434,
      "entropy": 0.38441379740834236,
      "epoch": 0.05705552725420275,
      "grad_norm": 14.538728713989258,
      "learning_rate": 7.54360228198859e-07,
      "loss": -0.1875,
      "reward": 3.230435371398926,
      "reward_std": 1.7468980550765991,
      "rewards/reasoning_density_reward/mean": 0.4268229156732559,
      "rewards/reasoning_density_reward/std": 0.09354004263877869,
      "rewards/success_reward/mean": 2.203125,
      "rewards/success_reward/std": 1.3285260796546936,
      "rewards/trajectory_similarity_reward/mean": 0.6004873514175415,
      "rewards/trajectory_similarity_reward/std": 0.3958117961883545,
      "step": 70
    },
    {
      "clip_ratio/high_max": 0.010424280830193311,
      "clip_ratio/high_mean": 0.0017261643006349914,
      "clip_ratio/low_mean": 0.00012478265489335172,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018509469446144067,
      "entropy": 0.3884718082845211,
      "epoch": 0.057870606214977076,
      "grad_norm": 13.074050903320312,
      "learning_rate": 7.537082314588427e-07,
      "loss": 6.3994,
      "reward": 3.190948724746704,
      "reward_std": 1.2352278232574463,
      "rewards/reasoning_density_reward/mean": 0.421875,
      "rewards/reasoning_density_reward/std": 0.11397185176610947,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5190736651420593,
      "rewards/trajectory_similarity_reward/std": 0.35613352060317993,
      "step": 71
    },
    {
      "clip_ratio/high_max": 0.01048097643069923,
      "clip_ratio/high_mean": 0.001989768701605499,
      "clip_ratio/low_mean": 0.0009479770051257219,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002937745735835051,
      "entropy": 0.40269992128014565,
      "epoch": 0.0586856851757514,
      "grad_norm": 14.169676780700684,
      "learning_rate": 7.530562347188264e-07,
      "loss": 14.6861,
      "reward": 2.4058523178100586,
      "reward_std": 1.9238004684448242,
      "rewards/reasoning_density_reward/mean": 0.328125,
      "rewards/reasoning_density_reward/std": 0.06451147049665451,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.39022713899612427,
      "rewards/trajectory_similarity_reward/std": 0.43470141291618347,
      "step": 72
    },
    {
      "clip_ratio/high_max": 0.00988182210130617,
      "clip_ratio/high_mean": 0.0015655708411941305,
      "clip_ratio/low_mean": 0.0005688439487130381,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002134414840838872,
      "entropy": 0.37859955057501793,
      "epoch": 0.059500764136525726,
      "grad_norm": 13.9193696975708,
      "learning_rate": 7.524042379788101e-07,
      "loss": 7.2691,
      "reward": 3.9818503856658936,
      "reward_std": 0.5083592534065247,
      "rewards/reasoning_density_reward/mean": 0.4444769024848938,
      "rewards/reasoning_density_reward/std": 0.07208810746669769,
      "rewards/success_reward/mean": 2.90625,
      "rewards/success_reward/std": 0.375,
      "rewards/trajectory_similarity_reward/mean": 0.6311233043670654,
      "rewards/trajectory_similarity_reward/std": 0.23543983697891235,
      "step": 73
    },
    {
      "clip_ratio/high_max": 0.01117798761697486,
      "clip_ratio/high_mean": 0.0023023093308438547,
      "clip_ratio/low_mean": 0.00027138817677041516,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00257369750761427,
      "entropy": 0.39074845239520073,
      "epoch": 0.06031584309730005,
      "grad_norm": 10.823390007019043,
      "learning_rate": 7.517522412387939e-07,
      "loss": 4.8812,
      "reward": 2.0759811401367188,
      "reward_std": 0.4306200444698334,
      "rewards/reasoning_density_reward/mean": 0.33385416865348816,
      "rewards/reasoning_density_reward/std": 0.10057559609413147,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.33587706089019775,
      "rewards/trajectory_similarity_reward/std": 0.4032735228538513,
      "step": 74
    },
    {
      "clip_ratio/high_max": 0.011692053900333121,
      "clip_ratio/high_mean": 0.0023486262616643216,
      "clip_ratio/low_mean": 0.00027654807490762323,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026251743220200296,
      "entropy": 0.4042360857129097,
      "epoch": 0.061130922058074376,
      "grad_norm": 8.58728313446045,
      "learning_rate": 7.511002444987775e-07,
      "loss": -2.3601,
      "step": 75
    },
    {
      "clip_ratio/high_max": 0.010118345060618594,
      "clip_ratio/high_mean": 0.0019494446787575725,
      "clip_ratio/low_mean": 0.0010543222051637713,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030037668475415558,
      "entropy": 0.4147445373237133,
      "epoch": 0.0619460010188487,
      "grad_norm": 10.073269844055176,
      "learning_rate": 7.504482477587613e-07,
      "loss": 3.9379,
      "reward": 2.139889657497406,
      "reward_std": 1.3298376202583313,
      "rewards/reasoning_density_reward/mean": 0.32760417461395264,
      "rewards/reasoning_density_reward/std": 0.11367396265268326,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.4665062427520752,
      "rewards/trajectory_similarity_reward/mean": 0.3591604083776474,
      "rewards/trajectory_similarity_reward/std": 0.42159049212932587,
      "step": 76
    },
    {
      "clip_ratio/high_max": 0.009173763857688755,
      "clip_ratio/high_mean": 0.002080853286315687,
      "clip_ratio/low_mean": 0.00025539923808537424,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002336252582608722,
      "entropy": 0.4410474933683872,
      "epoch": 0.06276107997962303,
      "grad_norm": 12.545037269592285,
      "learning_rate": 7.497962510187449e-07,
      "loss": -1.7975,
      "reward": 1.8489015102386475,
      "reward_std": 0.6966423392295837,
      "rewards/reasoning_density_reward/mean": 0.32499998807907104,
      "rewards/reasoning_density_reward/std": 0.08316649496555328,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.30515146255493164,
      "rewards/trajectory_similarity_reward/std": 0.46766749024391174,
      "step": 77
    },
    {
      "clip_ratio/high_max": 0.00876466406043619,
      "clip_ratio/high_mean": 0.0017985984013648704,
      "clip_ratio/low_mean": 6.793477950850502e-05,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018665331808733754,
      "entropy": 0.41209088638424873,
      "epoch": 0.06357615894039735,
      "grad_norm": 12.532767295837402,
      "learning_rate": 7.491442542787287e-07,
      "loss": -8.5343,
      "reward": 2.886443614959717,
      "reward_std": 1.6251652240753174,
      "rewards/reasoning_density_reward/mean": 0.36380207538604736,
      "rewards/reasoning_density_reward/std": 0.09066624566912651,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3888907432556152,
      "rewards/trajectory_similarity_reward/mean": 0.5538914799690247,
      "rewards/trajectory_similarity_reward/std": 0.44304654002189636,
      "step": 78
    },
    {
      "clip_ratio/high_max": 0.013052655209321529,
      "clip_ratio/high_mean": 0.003234988143958617,
      "clip_ratio/low_mean": 0.00011394392276997678,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003348932084918488,
      "entropy": 0.36943015828728676,
      "epoch": 0.06439123790117168,
      "grad_norm": 12.222606658935547,
      "learning_rate": 7.484922575387123e-07,
      "loss": 10.9382,
      "reward": 2.971004009246826,
      "reward_std": 1.7194501161575317,
      "rewards/reasoning_density_reward/mean": 0.3570188581943512,
      "rewards/reasoning_density_reward/std": 0.10506927222013474,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5514851212501526,
      "rewards/trajectory_similarity_reward/std": 0.4218115508556366,
      "step": 79
    },
    {
      "clip_ratio/high_max": 0.010106658563017845,
      "clip_ratio/high_mean": 0.0018305744597455487,
      "clip_ratio/low_mean": 0.0006813563668401912,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025119308120338246,
      "entropy": 0.3999580852687359,
      "epoch": 0.065206316861946,
      "grad_norm": 11.34425163269043,
      "learning_rate": 7.47840260798696e-07,
      "loss": 14.3197,
      "reward": 2.6055893898010254,
      "reward_std": 1.7489244937896729,
      "rewards/reasoning_density_reward/mean": 0.30416667461395264,
      "rewards/reasoning_density_reward/std": 0.11512095481157303,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.42642247676849365,
      "rewards/trajectory_similarity_reward/std": 0.455314964056015,
      "step": 80
    },
    {
      "clip_ratio/high_max": 0.009134847670793533,
      "clip_ratio/high_mean": 0.002084262319840491,
      "clip_ratio/low_mean": 0.0013481453497661278,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003432407625950873,
      "entropy": 0.4311261475086212,
      "epoch": 0.06602139582272033,
      "grad_norm": 12.698168754577637,
      "learning_rate": 7.471882640586797e-07,
      "loss": 8.4744,
      "reward": 3.257112979888916,
      "reward_std": 1.4439314603805542,
      "rewards/reasoning_density_reward/mean": 0.4035714268684387,
      "rewards/reasoning_density_reward/std": 0.08200528472661972,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5097914934158325,
      "rewards/trajectory_similarity_reward/std": 0.3711496889591217,
      "step": 81
    },
    {
      "clip_ratio/high_max": 0.009873063943814486,
      "clip_ratio/high_mean": 0.002101016347296536,
      "clip_ratio/low_mean": 0.0007624792960996274,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028634956543100998,
      "entropy": 0.44826582819223404,
      "epoch": 0.06683647478349465,
      "grad_norm": 13.439857482910156,
      "learning_rate": 7.465362673186633e-07,
      "loss": -1.0762,
      "reward": 1.807541847229004,
      "reward_std": 1.403374433517456,
      "rewards/reasoning_density_reward/mean": 0.3161458373069763,
      "rewards/reasoning_density_reward/std": 0.030348092317581177,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.17889603972434998,
      "rewards/trajectory_similarity_reward/std": 0.3225190043449402,
      "step": 82
    },
    {
      "clip_ratio/high_max": 0.005772447970230132,
      "clip_ratio/high_mean": 0.0010407836780359503,
      "clip_ratio/low_mean": 0.0009366043741465546,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019773880412685685,
      "entropy": 0.42903079092502594,
      "epoch": 0.06765155374426898,
      "grad_norm": 10.118024826049805,
      "learning_rate": 7.45884270578647e-07,
      "loss": 0.0278,
      "reward": 2.2298383712768555,
      "reward_std": 0.02137605845928192,
      "rewards/reasoning_density_reward/mean": 0.4000000059604645,
      "rewards/reasoning_density_reward/std": 0.10327955335378647,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.3298383951187134,
      "rewards/trajectory_similarity_reward/std": 0.34190526604652405,
      "step": 83
    },
    {
      "clip_ratio/high_max": 0.008928747905883938,
      "clip_ratio/high_mean": 0.0015060031437315047,
      "clip_ratio/low_mean": 0.001005987978714984,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002511991115170531,
      "entropy": 0.3896900974214077,
      "epoch": 0.0684666327050433,
      "grad_norm": 13.2094087600708,
      "learning_rate": 7.452322738386307e-07,
      "loss": 12.4082,
      "reward": 3.228626251220703,
      "reward_std": 1.1493841409683228,
      "rewards/reasoning_density_reward/mean": 0.3911706507205963,
      "rewards/reasoning_density_reward/std": 0.10694057494401932,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.587455689907074,
      "rewards/trajectory_similarity_reward/std": 0.3648388385772705,
      "step": 84
    },
    {
      "clip_ratio/high_max": 0.010124671505764127,
      "clip_ratio/high_mean": 0.0019426351645961404,
      "clip_ratio/low_mean": 0.0006592905701836571,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026019257347797975,
      "entropy": 0.39912351593375206,
      "epoch": 0.06928171166581763,
      "grad_norm": 12.524714469909668,
      "learning_rate": 7.445802770986144e-07,
      "loss": -6.4934,
      "reward": 2.0483317375183105,
      "reward_std": 1.2638603448867798,
      "rewards/reasoning_density_reward/mean": 0.3447916805744171,
      "rewards/reasoning_density_reward/std": 0.08226396888494492,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.20354022085666656,
      "rewards/trajectory_similarity_reward/std": 0.36749422550201416,
      "step": 85
    },
    {
      "clip_ratio/high_max": 0.00992424855940044,
      "clip_ratio/high_mean": 0.001905419550894294,
      "clip_ratio/low_mean": 0.0005976069369353354,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002503026458725799,
      "entropy": 0.3912870064377785,
      "epoch": 0.07009679062659195,
      "grad_norm": 12.827573776245117,
      "learning_rate": 7.439282803585981e-07,
      "loss": 2.2371,
      "reward": 2.482821464538574,
      "reward_std": 1.3732976913452148,
      "rewards/reasoning_density_reward/mean": 0.3347267508506775,
      "rewards/reasoning_density_reward/std": 0.1049395352602005,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 0.866025447845459,
      "rewards/trajectory_similarity_reward/mean": 0.27309471368789673,
      "rewards/trajectory_similarity_reward/std": 0.42067328095436096,
      "step": 86
    },
    {
      "clip_ratio/high_max": 0.004116729076486081,
      "clip_ratio/high_mean": 0.000614943441178184,
      "clip_ratio/low_mean": 0.0006552189879585057,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0012701624291366898,
      "entropy": 0.4100978560745716,
      "epoch": 0.07091186958736628,
      "grad_norm": 13.063579559326172,
      "learning_rate": 7.432762836185818e-07,
      "loss": 1.508,
      "reward": 1.8687537908554077,
      "reward_std": 1.6106373071670532,
      "rewards/reasoning_density_reward/mean": 0.3322916626930237,
      "rewards/reasoning_density_reward/std": 0.07922512292861938,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.31771203875541687,
      "rewards/trajectory_similarity_reward/std": 0.4343677759170532,
      "step": 87
    },
    {
      "clip_ratio/high_max": 0.01232010597595945,
      "clip_ratio/high_mean": 0.001843068159359973,
      "clip_ratio/low_mean": 0.0005758274746767711,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024188955649151467,
      "entropy": 0.4022018276154995,
      "epoch": 0.0717269485481406,
      "grad_norm": 14.1510009765625,
      "learning_rate": 7.426242868785655e-07,
      "loss": 1.9689,
      "reward": 3.028843879699707,
      "reward_std": 1.6514050960540771,
      "rewards/reasoning_density_reward/mean": 0.37552085518836975,
      "rewards/reasoning_density_reward/std": 0.13795693218708038,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.49707287549972534,
      "rewards/trajectory_similarity_reward/std": 0.438860148191452,
      "step": 88
    },
    {
      "clip_ratio/high_max": 0.01160558417905122,
      "clip_ratio/high_mean": 0.002190460276324302,
      "clip_ratio/low_mean": 0.00033463917861809023,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002525099480408244,
      "entropy": 0.3943169601261616,
      "epoch": 0.07254202750891493,
      "grad_norm": 12.859684944152832,
      "learning_rate": 7.419722901385493e-07,
      "loss": -5.9661,
      "reward": 2.8071448802948,
      "reward_std": 2.0417107343673706,
      "rewards/reasoning_density_reward/mean": 0.4043402820825577,
      "rewards/reasoning_density_reward/std": 0.0972960814833641,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.4456666707992554,
      "rewards/trajectory_similarity_reward/mean": 0.5278044939041138,
      "rewards/trajectory_similarity_reward/std": 0.46718375384807587,
      "step": 89
    },
    {
      "clip_ratio/high_max": 0.008113264542771503,
      "clip_ratio/high_mean": 0.001903011212561978,
      "clip_ratio/low_mean": 0.0002978700213134289,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022008812084095553,
      "entropy": 0.4149659425020218,
      "epoch": 0.07335710646968925,
      "grad_norm": 12.730341911315918,
      "learning_rate": 7.413202933985329e-07,
      "loss": -10.7816,
      "reward": 2.2670230865478516,
      "reward_std": 1.5808131694793701,
      "rewards/reasoning_density_reward/mean": 0.33020833134651184,
      "rewards/reasoning_density_reward/std": 0.05716083198785782,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.24931496381759644,
      "rewards/trajectory_similarity_reward/std": 0.35256466269493103,
      "step": 90
    },
    {
      "clip_ratio/high_max": 0.010234113549813628,
      "clip_ratio/high_mean": 0.0017315566656179726,
      "clip_ratio/low_mean": 0.00038159036193974316,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021131470566615462,
      "entropy": 0.4260617308318615,
      "epoch": 0.07417218543046358,
      "grad_norm": 13.068678855895996,
      "learning_rate": 7.406682966585167e-07,
      "loss": 9.1235,
      "reward": 2.7331905364990234,
      "reward_std": 1.5474060773849487,
      "rewards/reasoning_density_reward/mean": 0.32884424924850464,
      "rewards/reasoning_density_reward/std": 0.1253787726163864,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.3418463468551636,
      "rewards/trajectory_similarity_reward/std": 0.36024823784828186,
      "step": 91
    },
    {
      "clip_ratio/high_max": 0.007482225453713909,
      "clip_ratio/high_mean": 0.0016892352396098431,
      "clip_ratio/low_mean": 0.0002648817535373382,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019541169713193085,
      "entropy": 0.42399443686008453,
      "epoch": 0.0749872643912379,
      "grad_norm": 13.654683113098145,
      "learning_rate": 7.400162999185003e-07,
      "loss": 1.8828,
      "reward": 2.9242210388183594,
      "reward_std": 1.5566799640655518,
      "rewards/reasoning_density_reward/mean": 0.3773561716079712,
      "rewards/reasoning_density_reward/std": 0.08353006094694138,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.3906148672103882,
      "rewards/trajectory_similarity_reward/std": 0.36342746019363403,
      "step": 92
    },
    {
      "clip_ratio/high_max": 0.007931840606033802,
      "clip_ratio/high_mean": 0.001543864214909263,
      "clip_ratio/low_mean": 0.0010062727633339819,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025501370073470753,
      "entropy": 0.4307224750518799,
      "epoch": 0.07580234335201223,
      "grad_norm": 12.592246055603027,
      "learning_rate": 7.393643031784841e-07,
      "loss": 6.8596,
      "reward": 2.3604636192321777,
      "reward_std": 2.1365127563476562,
      "rewards/reasoning_density_reward/mean": 0.4000000059604645,
      "rewards/reasoning_density_reward/std": 0.10327955335378647,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.4604635536670685,
      "rewards/trajectory_similarity_reward/std": 0.477953165769577,
      "step": 93
    },
    {
      "clip_ratio/high_max": 0.009837876714300364,
      "clip_ratio/high_mean": 0.0018338983427383937,
      "clip_ratio/low_mean": 0.00042407297587487847,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002257971304061357,
      "entropy": 0.4045519344508648,
      "epoch": 0.07661742231278655,
      "grad_norm": 14.129582405090332,
      "learning_rate": 7.387123064384677e-07,
      "loss": -6.5446,
      "reward": 2.9387670755386353,
      "reward_std": 1.7567320466041565,
      "rewards/reasoning_density_reward/mean": 0.39010417461395264,
      "rewards/reasoning_density_reward/std": 0.10634449496865273,
      "rewards/success_reward/mean": 2.015625,
      "rewards/success_reward/std": 1.348175585269928,
      "rewards/trajectory_similarity_reward/mean": 0.5330378711223602,
      "rewards/trajectory_similarity_reward/std": 0.4452061206102371,
      "step": 94
    },
    {
      "clip_ratio/high_max": 0.006522592215333134,
      "clip_ratio/high_mean": 0.0014180327852955088,
      "clip_ratio/low_mean": 0.0005762850487371907,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019943178485846147,
      "entropy": 0.39954593777656555,
      "epoch": 0.07743250127356088,
      "grad_norm": 14.21096134185791,
      "learning_rate": 7.380603096984515e-07,
      "loss": 7.4343,
      "reward": 1.9616262912750244,
      "reward_std": 1.6865642070770264,
      "rewards/reasoning_density_reward/mean": 0.36250001192092896,
      "rewards/reasoning_density_reward/std": 0.09574270993471146,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.2808688879013062,
      "rewards/trajectory_similarity_reward/mean": 0.19287623465061188,
      "rewards/trajectory_similarity_reward/std": 0.30871346592903137,
      "step": 95
    },
    {
      "clip_ratio/high_max": 0.013795634149573743,
      "clip_ratio/high_mean": 0.0025708769680932164,
      "clip_ratio/low_mean": 0.0005770203060819767,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003147897296003066,
      "entropy": 0.41927606612443924,
      "epoch": 0.0782475802343352,
      "grad_norm": 15.971545219421387,
      "learning_rate": 7.374083129584351e-07,
      "loss": -13.6791,
      "reward": 1.965996503829956,
      "reward_std": 1.2552995681762695,
      "rewards/reasoning_density_reward/mean": 0.3427083492279053,
      "rewards/reasoning_density_reward/std": 0.1057545617222786,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.21703821420669556,
      "rewards/trajectory_similarity_reward/std": 0.31833288073539734,
      "step": 96
    },
    {
      "clip_ratio/high_max": 0.01315725891618058,
      "clip_ratio/high_mean": 0.002665467451151926,
      "clip_ratio/low_mean": 0.0006315340579021722,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032970014653983526,
      "entropy": 0.38862523436546326,
      "epoch": 0.07906265919510953,
      "grad_norm": 15.3977632522583,
      "learning_rate": 7.367563162184189e-07,
      "loss": -2.5389,
      "reward": 4.0067458152771,
      "reward_std": 0.5177525468170643,
      "rewards/reasoning_density_reward/mean": 0.4410460740327835,
      "rewards/reasoning_density_reward/std": 0.05462482571601868,
      "rewards/success_reward/mean": 2.765625,
      "rewards/success_reward/std": 0.5281157493591309,
      "rewards/trajectory_similarity_reward/mean": 0.8000745177268982,
      "rewards/trajectory_similarity_reward/std": 0.20378848165273666,
      "step": 97
    },
    {
      "clip_ratio/high_max": 0.012190619658213109,
      "clip_ratio/high_mean": 0.0025064930377993733,
      "clip_ratio/low_mean": 0.000522415735758841,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030289088244899176,
      "entropy": 0.38256434723734856,
      "epoch": 0.07987773815588385,
      "grad_norm": 12.290925979614258,
      "learning_rate": 7.361043194784025e-07,
      "loss": -7.346,
      "reward": 3.308230400085449,
      "reward_std": 1.1283650398254395,
      "rewards/reasoning_density_reward/mean": 0.3872023820877075,
      "rewards/reasoning_density_reward/std": 0.13591289520263672,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6710280776023865,
      "rewards/trajectory_similarity_reward/std": 0.41945797204971313,
      "step": 98
    },
    {
      "clip_ratio/high_max": 0.012369863165076822,
      "clip_ratio/high_mean": 0.003046818426810205,
      "clip_ratio/low_mean": 0.0009327930310973898,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039796114433556795,
      "entropy": 0.3798425532877445,
      "epoch": 0.08069281711665818,
      "grad_norm": 13.359941482543945,
      "learning_rate": 7.354523227383863e-07,
      "loss": 1.3789,
      "reward": 3.746951937675476,
      "reward_std": 1.1107196807861328,
      "rewards/reasoning_density_reward/mean": 0.4294270724058151,
      "rewards/reasoning_density_reward/std": 0.10018717125058174,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.7433394491672516,
      "rewards/trajectory_similarity_reward/mean": 0.6456498503684998,
      "rewards/trajectory_similarity_reward/std": 0.32606610655784607,
      "step": 99
    },
    {
      "clip_ratio/high_max": 0.008228461723774672,
      "clip_ratio/high_mean": 0.001863207624410279,
      "clip_ratio/low_mean": 0.0005252495866443496,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023884571783128195,
      "entropy": 0.4063984677195549,
      "epoch": 0.0815078960774325,
      "grad_norm": 12.278055191040039,
      "learning_rate": 7.348003259983699e-07,
      "loss": 5.5366,
      "reward": 1.666050910949707,
      "reward_std": 1.866332769393921,
      "rewards/reasoning_density_reward/mean": 0.3239583373069763,
      "rewards/reasoning_density_reward/std": 0.12094821780920029,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.3108425736427307,
      "rewards/trajectory_similarity_reward/std": 0.47618967294692993,
      "step": 100
    },
    {
      "clip_ratio/high_max": 0.005038036179030314,
      "clip_ratio/high_mean": 0.001091449576051673,
      "clip_ratio/low_mean": 0.00035544499405659735,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014468945882981643,
      "entropy": 0.42624181509017944,
      "epoch": 0.08232297503820683,
      "grad_norm": 13.971607208251953,
      "learning_rate": 7.341483292583537e-07,
      "loss": -9.9352,
      "reward": 2.060678482055664,
      "reward_std": 1.4147560596466064,
      "rewards/reasoning_density_reward/mean": 0.28680557012557983,
      "rewards/reasoning_density_reward/std": 0.07477331906557083,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.2738730311393738,
      "rewards/trajectory_similarity_reward/std": 0.3952414095401764,
      "step": 101
    },
    {
      "clip_ratio/high_max": 0.007314139453228563,
      "clip_ratio/high_mean": 0.0012204287995700724,
      "clip_ratio/low_mean": 0.0004430055014381651,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016634343082841951,
      "entropy": 0.4115152508020401,
      "epoch": 0.08313805399898115,
      "grad_norm": 13.421957969665527,
      "learning_rate": 7.334963325183373e-07,
      "loss": -3.4275,
      "reward": 2.711223840713501,
      "reward_std": 1.185329258441925,
      "rewards/reasoning_density_reward/mean": 0.3335193544626236,
      "rewards/reasoning_density_reward/std": 0.11304861307144165,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.1924119889736176,
      "rewards/trajectory_similarity_reward/mean": 0.4089545011520386,
      "rewards/trajectory_similarity_reward/std": 0.42651331424713135,
      "step": 102
    },
    {
      "clip_ratio/high_max": 0.01538753934437409,
      "clip_ratio/high_mean": 0.003329378378111869,
      "clip_ratio/low_mean": 0.0005931131599936634,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003922491552657448,
      "entropy": 0.4208606258034706,
      "epoch": 0.08395313295975548,
      "grad_norm": 13.691299438476562,
      "learning_rate": 7.328443357783211e-07,
      "loss": 8.6717,
      "reward": 2.232236623764038,
      "reward_std": 1.3128751516342163,
      "rewards/reasoning_density_reward/mean": 0.34315475821495056,
      "rewards/reasoning_density_reward/std": 0.08292587101459503,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.2015819251537323,
      "rewards/trajectory_similarity_reward/std": 0.3179975748062134,
      "step": 103
    },
    {
      "clip_ratio/high_max": 0.009795513295102865,
      "clip_ratio/high_mean": 0.0021753816108684987,
      "clip_ratio/low_mean": 0.000535449380549835,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027108310096082278,
      "entropy": 0.41022003814578056,
      "epoch": 0.0847682119205298,
      "grad_norm": 12.230087280273438,
      "learning_rate": 7.321923390383048e-07,
      "loss": 0.8742,
      "step": 104
    },
    {
      "clip_ratio/high_max": 0.006724592676619068,
      "clip_ratio/high_mean": 0.0011830015027953777,
      "clip_ratio/low_mean": 0.0003476714337011799,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015306729328585789,
      "entropy": 0.3867288827896118,
      "epoch": 0.08558329088130413,
      "grad_norm": 11.559967994689941,
      "learning_rate": 7.315403422982885e-07,
      "loss": 4.5466,
      "reward": 1.927454948425293,
      "reward_std": 2.0553712844848633,
      "rewards/reasoning_density_reward/mean": 0.24969953298568726,
      "rewards/reasoning_density_reward/std": 0.17441707849502563,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.36525553464889526,
      "rewards/trajectory_similarity_reward/std": 0.4373611807823181,
      "step": 105
    },
    {
      "clip_ratio/high_max": 0.004884231952019036,
      "clip_ratio/high_mean": 0.0012027644843328744,
      "clip_ratio/low_mean": 0.000757791094656568,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001960555622645188,
      "entropy": 0.3799471892416477,
      "epoch": 0.08639836984207845,
      "grad_norm": 11.993608474731445,
      "learning_rate": 7.308883455582722e-07,
      "loss": -2.5413,
      "reward": 2.575575351715088,
      "reward_std": 1.9435489177703857,
      "rewards/reasoning_density_reward/mean": 0.3917613625526428,
      "rewards/reasoning_density_reward/std": 0.10784922540187836,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4025641679763794,
      "rewards/trajectory_similarity_reward/std": 0.38692423701286316,
      "step": 106
    },
    {
      "clip_ratio/high_max": 0.013824931578710675,
      "clip_ratio/high_mean": 0.003812712777289562,
      "clip_ratio/low_mean": 0.0006164055357658071,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004429118300322443,
      "entropy": 0.4325575605034828,
      "epoch": 0.08721344880285278,
      "grad_norm": 14.137404441833496,
      "learning_rate": 7.302363488182559e-07,
      "loss": 2.8813,
      "reward": 3.115222454071045,
      "reward_std": 1.7825796008110046,
      "rewards/reasoning_density_reward/mean": 0.42656250298023224,
      "rewards/reasoning_density_reward/std": 0.10771246999502182,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.420820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.6261599957942963,
      "rewards/trajectory_similarity_reward/std": 0.4373759776353836,
      "step": 107
    },
    {
      "clip_ratio/high_max": 0.010367794777266681,
      "clip_ratio/high_mean": 0.002466555670253001,
      "clip_ratio/low_mean": 0.00045943107397761196,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029259867151267827,
      "entropy": 0.3788311704993248,
      "epoch": 0.0880285277636271,
      "grad_norm": 14.470721244812012,
      "learning_rate": 7.295843520782396e-07,
      "loss": -8.9323,
      "reward": 1.8157618641853333,
      "reward_std": 1.577126532793045,
      "rewards/reasoning_density_reward/mean": 0.36796876788139343,
      "rewards/reasoning_density_reward/std": 0.07917395047843456,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.3731906414031982,
      "rewards/trajectory_similarity_reward/mean": 0.32279306650161743,
      "rewards/trajectory_similarity_reward/std": 0.3740277886390686,
      "step": 108
    },
    {
      "clip_ratio/high_max": 0.011547476897248998,
      "clip_ratio/high_mean": 0.002293549885507673,
      "clip_ratio/low_mean": 0.00028525794550660066,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002578807805548422,
      "entropy": 0.4210098423063755,
      "epoch": 0.08884360672440143,
      "grad_norm": 21.079328536987305,
      "learning_rate": 7.289323553382233e-07,
      "loss": 9.6629,
      "reward": 3.7690805196762085,
      "reward_std": 0.7633872032165527,
      "rewards/reasoning_density_reward/mean": 0.41492265462875366,
      "rewards/reasoning_density_reward/std": 0.10393134504556656,
      "rewards/success_reward/mean": 2.765625,
      "rewards/success_reward/std": 0.5585084557533264,
      "rewards/trajectory_similarity_reward/mean": 0.5885328352451324,
      "rewards/trajectory_similarity_reward/std": 0.32027123868465424,
      "step": 109
    },
    {
      "clip_ratio/high_max": 0.014525482198223472,
      "clip_ratio/high_mean": 0.0040031706157606095,
      "clip_ratio/low_mean": 0.00048221998440567404,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004485390614718199,
      "entropy": 0.4100766032934189,
      "epoch": 0.08965868568517575,
      "grad_norm": 17.04603385925293,
      "learning_rate": 7.28280358598207e-07,
      "loss": -20.8577,
      "reward": 3.373192310333252,
      "reward_std": 1.699789047241211,
      "rewards/reasoning_density_reward/mean": 0.40072113275527954,
      "rewards/reasoning_density_reward/std": 0.08883017301559448,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.6287210583686829,
      "rewards/trajectory_similarity_reward/std": 0.40198129415512085,
      "step": 110
    },
    {
      "clip_ratio/high_max": 0.01266779235447757,
      "clip_ratio/high_mean": 0.0028360130927467253,
      "clip_ratio/low_mean": 0.0006503511031041853,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034863641230913345,
      "entropy": 0.40329694747924805,
      "epoch": 0.09047376464595008,
      "grad_norm": 11.329288482666016,
      "learning_rate": 7.276283618581907e-07,
      "loss": 11.767,
      "reward": 2.485553026199341,
      "reward_std": 1.5701525211334229,
      "rewards/reasoning_density_reward/mean": 0.3374255895614624,
      "rewards/reasoning_density_reward/std": 0.1494952291250229,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.4606275260448456,
      "rewards/trajectory_similarity_reward/std": 0.4590940773487091,
      "step": 111
    },
    {
      "clip_ratio/high_max": 0.013761651702225208,
      "clip_ratio/high_mean": 0.0030711421568412334,
      "clip_ratio/low_mean": 0.0005521217281057034,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003623263954068534,
      "entropy": 0.4187561497092247,
      "epoch": 0.0912888436067244,
      "grad_norm": 11.682222366333008,
      "learning_rate": 7.269763651181743e-07,
      "loss": 2.2843,
      "reward": 2.785010814666748,
      "reward_std": 1.8236756324768066,
      "rewards/reasoning_density_reward/mean": 0.3481026887893677,
      "rewards/reasoning_density_reward/std": 0.11424379795789719,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.4681580662727356,
      "rewards/trajectory_similarity_reward/std": 0.45206016302108765,
      "step": 112
    },
    {
      "clip_ratio/high_max": 0.015193283266853541,
      "clip_ratio/high_mean": 0.0037191537267062813,
      "clip_ratio/low_mean": 0.0009977530389733147,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004716906667454168,
      "entropy": 0.39803583920001984,
      "epoch": 0.09210392256749873,
      "grad_norm": 14.415155410766602,
      "learning_rate": 7.263243683781581e-07,
      "loss": 9.0478,
      "reward": 3.497145175933838,
      "reward_std": 1.62192702293396,
      "rewards/reasoning_density_reward/mean": 0.3857072591781616,
      "rewards/reasoning_density_reward/std": 0.13648837804794312,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6739378571510315,
      "rewards/trajectory_similarity_reward/std": 0.36391502618789673,
      "step": 113
    },
    {
      "clip_ratio/high_max": 0.012792276655090973,
      "clip_ratio/high_mean": 0.003132829715468688,
      "clip_ratio/low_mean": 0.00025129239656962454,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033841221120383125,
      "entropy": 0.39321062713861465,
      "epoch": 0.09291900152827305,
      "grad_norm": 11.726900100708008,
      "learning_rate": 7.256723716381417e-07,
      "loss": -5.2935,
      "reward": 2.2716190814971924,
      "reward_std": 2.048884391784668,
      "rewards/reasoning_density_reward/mean": 0.3586929440498352,
      "rewards/reasoning_density_reward/std": 0.08061307668685913,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.4129261374473572,
      "rewards/trajectory_similarity_reward/std": 0.4454958736896515,
      "step": 114
    },
    {
      "clip_ratio/high_max": 0.01500163262244314,
      "clip_ratio/high_mean": 0.002678214557818137,
      "clip_ratio/low_mean": 0.0004654806980397552,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031436952558578923,
      "entropy": 0.41402987018227577,
      "epoch": 0.09373408048904738,
      "grad_norm": 12.227312088012695,
      "learning_rate": 7.250203748981255e-07,
      "loss": 1.2717,
      "reward": 3.3105993270874023,
      "reward_std": 1.4664783477783203,
      "rewards/reasoning_density_reward/mean": 0.41961807012557983,
      "rewards/reasoning_density_reward/std": 0.08787892013788223,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.547231137752533,
      "rewards/trajectory_similarity_reward/std": 0.3792457580566406,
      "step": 115
    },
    {
      "clip_ratio/high_max": 0.010535035835346207,
      "clip_ratio/high_mean": 0.002102305992593756,
      "clip_ratio/low_mean": 0.00022484230430563912,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023271482750715222,
      "entropy": 0.3886296935379505,
      "epoch": 0.0945491594498217,
      "grad_norm": 12.787339210510254,
      "learning_rate": 7.243683781581091e-07,
      "loss": -4.3924,
      "reward": 0.7639956474304199,
      "reward_std": 0.9499197602272034,
      "rewards/reasoning_density_reward/mean": 0.29215317964553833,
      "rewards/reasoning_density_reward/std": 0.08645332604646683,
      "rewards/success_reward/mean": 0.375,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.09684248268604279,
      "rewards/trajectory_similarity_reward/std": 0.27711281180381775,
      "step": 116
    },
    {
      "clip_ratio/high_max": 0.005701223009964451,
      "clip_ratio/high_mean": 0.0011757573629438411,
      "clip_ratio/low_mean": 0.00048766809049993753,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001663425409788033,
      "entropy": 0.41133878752589226,
      "epoch": 0.09536423841059603,
      "grad_norm": 15.606783866882324,
      "learning_rate": 7.237163814180928e-07,
      "loss": -7.0901,
      "reward": 2.2044951915740967,
      "reward_std": 1.3918193578720093,
      "rewards/reasoning_density_reward/mean": 0.33191967010498047,
      "rewards/reasoning_density_reward/std": 0.08626281470060349,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.185075581073761,
      "rewards/trajectory_similarity_reward/std": 0.29314130544662476,
      "step": 117
    },
    {
      "clip_ratio/high_max": 0.007828497386071831,
      "clip_ratio/high_mean": 0.0013543933600885794,
      "clip_ratio/low_mean": 0.0005582532903645188,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019126466504530981,
      "entropy": 0.4146486967802048,
      "epoch": 0.09617931737137035,
      "grad_norm": 13.513701438903809,
      "learning_rate": 7.230643846780765e-07,
      "loss": 10.475,
      "reward": 0.746760368347168,
      "reward_std": 0.9029960632324219,
      "rewards/reasoning_density_reward/mean": 0.29107144474983215,
      "rewards/reasoning_density_reward/std": 0.07552537322044373,
      "rewards/success_reward/mean": 0.375,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.08068893849849701,
      "rewards/trajectory_similarity_reward/std": 0.22561952471733093,
      "step": 118
    },
    {
      "clip_ratio/high_max": 0.013782548514427617,
      "clip_ratio/high_mean": 0.002387008604273433,
      "clip_ratio/low_mean": 0.0005862681646249257,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002973276819830062,
      "entropy": 0.4298718310892582,
      "epoch": 0.09699439633214468,
      "grad_norm": 13.38536548614502,
      "learning_rate": 7.224123879380603e-07,
      "loss": 13.5919,
      "reward": 3.6598434448242188,
      "reward_std": 1.1985636949539185,
      "rewards/reasoning_density_reward/mean": 0.4106026887893677,
      "rewards/reasoning_density_reward/std": 0.1428905874490738,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.7179907560348511,
      "rewards/trajectory_similarity_reward/std": 0.4089668095111847,
      "step": 119
    },
    {
      "clip_ratio/high_max": 0.006778395734727383,
      "clip_ratio/high_mean": 0.0017845475158537738,
      "clip_ratio/low_mean": 0.00092219334692345,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002706740880967118,
      "entropy": 0.4273788742721081,
      "epoch": 0.097809475292919,
      "grad_norm": 14.043580055236816,
      "learning_rate": 7.217603911980439e-07,
      "loss": 8.8442,
      "reward": 2.7496860027313232,
      "reward_std": 1.445129156112671,
      "rewards/reasoning_density_reward/mean": 0.3396780490875244,
      "rewards/reasoning_density_reward/std": 0.08540934324264526,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.4412580728530884,
      "rewards/trajectory_similarity_reward/std": 0.4326867163181305,
      "step": 120
    },
    {
      "clip_ratio/high_max": 0.006658323167357594,
      "clip_ratio/high_mean": 0.0012371543198241852,
      "clip_ratio/low_mean": 0.0011839130384032615,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024210673582274467,
      "entropy": 0.4129839949309826,
      "epoch": 0.09862455425369333,
      "grad_norm": 16.41416358947754,
      "learning_rate": 7.211083944580277e-07,
      "loss": -3.6336,
      "reward": 2.170681953430176,
      "reward_std": 1.4745197296142578,
      "rewards/reasoning_density_reward/mean": 0.35287702083587646,
      "rewards/reasoning_density_reward/std": 0.09128795564174652,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.157853603363037,
      "rewards/trajectory_similarity_reward/mean": 0.22405506670475006,
      "rewards/trajectory_similarity_reward/std": 0.3906078636646271,
      "step": 121
    },
    {
      "clip_ratio/high_max": 0.00660419708583504,
      "clip_ratio/high_mean": 0.001199063473904971,
      "clip_ratio/low_mean": 0.0005195435333007481,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017186070253956132,
      "entropy": 0.3643096648156643,
      "epoch": 0.09943963321446765,
      "grad_norm": 13.112727165222168,
      "learning_rate": 7.204563977180113e-07,
      "loss": 4.0295,
      "reward": 3.5608458518981934,
      "reward_std": 1.6424708366394043,
      "rewards/reasoning_density_reward/mean": 0.42544642090797424,
      "rewards/reasoning_density_reward/std": 0.0799977257847786,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.697899341583252,
      "rewards/trajectory_similarity_reward/std": 0.35566219687461853,
      "step": 122
    },
    {
      "clip_ratio/high_max": 0.014093314646743238,
      "clip_ratio/high_mean": 0.003458755265455693,
      "clip_ratio/low_mean": 0.0002929664551629685,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037517217569984496,
      "entropy": 0.38356638327240944,
      "epoch": 0.10025471217524198,
      "grad_norm": 12.855864524841309,
      "learning_rate": 7.198044009779951e-07,
      "loss": 1.0111,
      "reward": 3.236358404159546,
      "reward_std": 1.638232946395874,
      "rewards/reasoning_density_reward/mean": 0.3973958492279053,
      "rewards/reasoning_density_reward/std": 0.11473473906517029,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.5889624357223511,
      "rewards/trajectory_similarity_reward/std": 0.43947267532348633,
      "step": 123
    },
    {
      "clip_ratio/high_max": 0.012005313066765666,
      "clip_ratio/high_mean": 0.003348989979713224,
      "clip_ratio/low_mean": 0.00022524141968460754,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035742314212257043,
      "entropy": 0.39064013957977295,
      "epoch": 0.1010697911360163,
      "grad_norm": 13.388768196105957,
      "learning_rate": 7.191524042379787e-07,
      "loss": 2.3684,
      "reward": 3.8572402000427246,
      "reward_std": 1.4400691986083984,
      "rewards/reasoning_density_reward/mean": 0.43214285373687744,
      "rewards/reasoning_density_reward/std": 0.09511896967887878,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.8000973463058472,
      "rewards/trajectory_similarity_reward/std": 0.31746625900268555,
      "step": 124
    },
    {
      "clip_ratio/high_max": 0.009565273823682219,
      "clip_ratio/high_mean": 0.0030449857003986835,
      "clip_ratio/low_mean": 0.00014388083218364045,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003188866510754451,
      "entropy": 0.3992129936814308,
      "epoch": 0.10188487009679063,
      "grad_norm": 14.60083293914795,
      "learning_rate": 7.185004074979625e-07,
      "loss": 14.8752,
      "reward": 3.5469837188720703,
      "reward_std": 1.4289809465408325,
      "rewards/reasoning_density_reward/mean": 0.4087797701358795,
      "rewards/reasoning_density_reward/std": 0.0857972502708435,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.1831679940223694,
      "rewards/trajectory_similarity_reward/mean": 0.7007037401199341,
      "rewards/trajectory_similarity_reward/std": 0.37968575954437256,
      "step": 125
    },
    {
      "clip_ratio/high_max": 0.010577268985798582,
      "clip_ratio/high_mean": 0.0017693242079985794,
      "clip_ratio/low_mean": 0.0009742327638377901,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027435569863882847,
      "entropy": 0.3790353536605835,
      "epoch": 0.10269994905756495,
      "grad_norm": 14.075222969055176,
      "learning_rate": 7.178484107579461e-07,
      "loss": 10.8541,
      "reward": 3.7598021030426025,
      "reward_std": 0.9493705034255981,
      "rewards/reasoning_density_reward/mean": 0.4572916626930237,
      "rewards/reasoning_density_reward/std": 0.08073017001152039,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.7712604403495789,
      "rewards/trajectory_similarity_reward/std": 0.39286062121391296,
      "step": 126
    },
    {
      "clip_ratio/high_max": 0.010027483745943755,
      "clip_ratio/high_mean": 0.002285152528202161,
      "clip_ratio/low_mean": 0.0004609682619047817,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002746120808296837,
      "entropy": 0.38836725801229477,
      "epoch": 0.10351502801833928,
      "grad_norm": 11.48889446258545,
      "learning_rate": 7.171964140179299e-07,
      "loss": 0.9623,
      "reward": 3.7459875345230103,
      "reward_std": 1.0976549685001373,
      "rewards/reasoning_density_reward/mean": 0.41093750298023224,
      "rewards/reasoning_density_reward/std": 0.12173208594322205,
      "rewards/success_reward/mean": 2.578125,
      "rewards/success_reward/std": 0.8666295409202576,
      "rewards/trajectory_similarity_reward/mean": 0.7569250166416168,
      "rewards/trajectory_similarity_reward/std": 0.38206858932971954,
      "step": 127
    },
    {
      "clip_ratio/high_max": 0.009956729569239542,
      "clip_ratio/high_mean": 0.00218489502731245,
      "clip_ratio/low_mean": 0.0007897432296886109,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029746381915174425,
      "entropy": 0.41538476571440697,
      "epoch": 0.1043301069791136,
      "grad_norm": 14.688623428344727,
      "learning_rate": 7.165444172779135e-07,
      "loss": 0.9793,
      "reward": 2.0047168731689453,
      "reward_std": 1.9073963165283203,
      "rewards/reasoning_density_reward/mean": 0.36294645071029663,
      "rewards/reasoning_density_reward/std": 0.08740384131669998,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.235520139336586,
      "rewards/trajectory_similarity_reward/std": 0.3067294657230377,
      "step": 128
    },
    {
      "clip_ratio/high_max": 0.009259733778890222,
      "clip_ratio/high_mean": 0.0016919046829571016,
      "clip_ratio/low_mean": 0.0006386148015735671,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002330519499082584,
      "entropy": 0.42109450325369835,
      "epoch": 0.10514518593988793,
      "grad_norm": 12.568429946899414,
      "learning_rate": 7.158924205378973e-07,
      "loss": -0.8104,
      "reward": 2.5073468685150146,
      "reward_std": 1.7497374415397644,
      "rewards/reasoning_density_reward/mean": 0.3499069958925247,
      "rewards/reasoning_density_reward/std": 0.10036593303084373,
      "rewards/success_reward/mean": 1.734375,
      "rewards/success_reward/std": 1.4375925064086914,
      "rewards/trajectory_similarity_reward/mean": 0.4230649322271347,
      "rewards/trajectory_similarity_reward/std": 0.3837912529706955,
      "step": 129
    },
    {
      "clip_ratio/high_max": 0.01154330198187381,
      "clip_ratio/high_mean": 0.002905304507294204,
      "clip_ratio/low_mean": 0.00099042682631989,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038957313518039882,
      "entropy": 0.42996587604284286,
      "epoch": 0.10596026490066225,
      "grad_norm": 32.98195266723633,
      "learning_rate": 7.152404237978809e-07,
      "loss": -5.9684,
      "reward": 1.6808441877365112,
      "reward_std": 0.9954308271408081,
      "rewards/reasoning_density_reward/mean": 0.30418696999549866,
      "rewards/reasoning_density_reward/std": 0.08974776417016983,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.15790721774101257,
      "rewards/trajectory_similarity_reward/std": 0.2875151038169861,
      "step": 130
    },
    {
      "clip_ratio/high_max": 0.0086608060519211,
      "clip_ratio/high_mean": 0.0017645798870944418,
      "clip_ratio/low_mean": 0.0008994864110718481,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00266406629816629,
      "entropy": 0.41765541583299637,
      "epoch": 0.10677534386143658,
      "grad_norm": 13.455292701721191,
      "learning_rate": 7.145884270578647e-07,
      "loss": 1.1026,
      "reward": 2.813427448272705,
      "reward_std": 1.8351097106933594,
      "rewards/reasoning_density_reward/mean": 0.3683035671710968,
      "rewards/reasoning_density_reward/std": 0.11922991275787354,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.4763736128807068,
      "rewards/trajectory_similarity_reward/std": 0.3950500190258026,
      "step": 131
    },
    {
      "clip_ratio/high_max": 0.01490003673825413,
      "clip_ratio/high_mean": 0.003348680096678436,
      "clip_ratio/low_mean": 0.00019899968538084067,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003547679807525128,
      "entropy": 0.40691235661506653,
      "epoch": 0.1075904228222109,
      "grad_norm": 10.605008125305176,
      "learning_rate": 7.139364303178483e-07,
      "loss": 1.8035,
      "reward": 2.1520142555236816,
      "reward_std": 0.5006915330886841,
      "rewards/reasoning_density_reward/mean": 0.33750003576278687,
      "rewards/reasoning_density_reward/std": 0.11798093467950821,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.4082643389701843,
      "rewards/trajectory_similarity_reward/std": 0.4795984625816345,
      "step": 132
    },
    {
      "clip_ratio/high_max": 0.01464744855184108,
      "clip_ratio/high_mean": 0.002909335831645876,
      "clip_ratio/low_mean": 0.0010758147036540322,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003985150455264375,
      "entropy": 0.3927535302937031,
      "epoch": 0.10840550178298523,
      "grad_norm": 12.775403022766113,
      "learning_rate": 7.132844335778321e-07,
      "loss": 3.5679,
      "reward": 3.149346351623535,
      "reward_std": 1.716435432434082,
      "rewards/reasoning_density_reward/mean": 0.35804811120033264,
      "rewards/reasoning_density_reward/std": 0.11159070581197739,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5412980914115906,
      "rewards/trajectory_similarity_reward/std": 0.36030513048171997,
      "step": 133
    },
    {
      "clip_ratio/high_max": 0.010470565466675907,
      "clip_ratio/high_mean": 0.0025513654545648023,
      "clip_ratio/low_mean": 0.00036477501271292567,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029161404381738976,
      "entropy": 0.38942592591047287,
      "epoch": 0.10922058074375955,
      "grad_norm": 12.385831832885742,
      "learning_rate": 7.126324368378158e-07,
      "loss": 2.0249,
      "reward": 2.2305495738983154,
      "reward_std": 1.5520544052124023,
      "rewards/reasoning_density_reward/mean": 0.33645835518836975,
      "rewards/reasoning_density_reward/std": 0.12855590879917145,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.3003412187099457,
      "rewards/trajectory_similarity_reward/std": 0.3686550557613373,
      "step": 134
    },
    {
      "clip_ratio/high_max": 0.012338117405306548,
      "clip_ratio/high_mean": 0.0025608250580262393,
      "clip_ratio/low_mean": 0.0006963220657780766,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032571471529081464,
      "entropy": 0.3963875472545624,
      "epoch": 0.11003565970453388,
      "grad_norm": 12.89198112487793,
      "learning_rate": 7.119804400977995e-07,
      "loss": 8.0831,
      "reward": 2.772329568862915,
      "reward_std": 1.2836329936981201,
      "rewards/reasoning_density_reward/mean": 0.3887152671813965,
      "rewards/reasoning_density_reward/std": 0.09680316597223282,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.189800500869751,
      "rewards/trajectory_similarity_reward/mean": 0.4148642122745514,
      "rewards/trajectory_similarity_reward/std": 0.4433470070362091,
      "step": 135
    },
    {
      "clip_ratio/high_max": 0.015368734661024064,
      "clip_ratio/high_mean": 0.002956614138383884,
      "clip_ratio/low_mean": 0.0003167252798448317,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032733394546085037,
      "entropy": 0.38504334166646004,
      "epoch": 0.1108507386653082,
      "grad_norm": 21.773099899291992,
      "learning_rate": 7.113284433577832e-07,
      "loss": -15.6011,
      "reward": 2.739035129547119,
      "reward_std": 0.7265345454216003,
      "rewards/reasoning_density_reward/mean": 0.3619791865348816,
      "rewards/reasoning_density_reward/std": 0.08073554933071136,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.5020560622215271,
      "rewards/trajectory_similarity_reward/std": 0.4704548120498657,
      "step": 136
    },
    {
      "clip_ratio/high_max": 0.019019578758161515,
      "clip_ratio/high_mean": 0.0033786182902986184,
      "clip_ratio/low_mean": 0.001005371887004003,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004383990206406452,
      "entropy": 0.3862803094089031,
      "epoch": 0.11166581762608253,
      "grad_norm": 11.028555870056152,
      "learning_rate": 7.106764466177669e-07,
      "loss": 10.0026,
      "reward": 4.035966873168945,
      "reward_std": 0.9145456552505493,
      "rewards/reasoning_density_reward/mean": 0.4281250238418579,
      "rewards/reasoning_density_reward/std": 0.08638174086809158,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.5123475790023804,
      "rewards/trajectory_similarity_reward/mean": 0.7953417301177979,
      "rewards/trajectory_similarity_reward/std": 0.33445703983306885,
      "step": 137
    },
    {
      "clip_ratio/high_max": 0.008130995032843202,
      "clip_ratio/high_mean": 0.001419855652784463,
      "clip_ratio/low_mean": 0.0004339627121225931,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018538183721830137,
      "entropy": 0.42531975731253624,
      "epoch": 0.11248089658685685,
      "grad_norm": 16.375030517578125,
      "learning_rate": 7.100244498777506e-07,
      "loss": 6.0329,
      "reward": 2.4678995609283447,
      "reward_std": 1.4341795444488525,
      "rewards/reasoning_density_reward/mean": 0.3895833492279053,
      "rewards/reasoning_density_reward/std": 0.10089873522520065,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.3908163011074066,
      "rewards/trajectory_similarity_reward/std": 0.41544046998023987,
      "step": 138
    },
    {
      "clip_ratio/high_max": 0.006433037808164954,
      "clip_ratio/high_mean": 0.001035063323797658,
      "clip_ratio/low_mean": 0.0006042288514436223,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016392921752412803,
      "entropy": 0.42558496072888374,
      "epoch": 0.11329597554763118,
      "grad_norm": 15.316346168518066,
      "learning_rate": 7.093724531377343e-07,
      "loss": -5.6402,
      "reward": 2.31181001663208,
      "reward_std": 1.3169147968292236,
      "rewards/reasoning_density_reward/mean": 0.3821428418159485,
      "rewards/reasoning_density_reward/std": 0.10823255032300949,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.42966699600219727,
      "rewards/trajectory_similarity_reward/std": 0.46758270263671875,
      "step": 139
    },
    {
      "clip_ratio/high_max": 0.006272852246183902,
      "clip_ratio/high_mean": 0.0012149137328378856,
      "clip_ratio/low_mean": 0.00041006561878020875,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001624979377083946,
      "entropy": 0.40650906041264534,
      "epoch": 0.1141110545084055,
      "grad_norm": 11.267071723937988,
      "learning_rate": 7.08720456397718e-07,
      "loss": -5.3677,
      "reward": 2.2851513028144836,
      "reward_std": 0.8895585536956787,
      "rewards/reasoning_density_reward/mean": 0.3668898940086365,
      "rewards/reasoning_density_reward/std": 0.07418611645698547,
      "rewards/success_reward/mean": 1.546875,
      "rewards/success_reward/std": 1.2163992524147034,
      "rewards/trajectory_similarity_reward/mean": 0.3713865205645561,
      "rewards/trajectory_similarity_reward/std": 0.3097982481122017,
      "step": 140
    },
    {
      "clip_ratio/high_max": 0.013700049894396216,
      "clip_ratio/high_mean": 0.002960452540719416,
      "clip_ratio/low_mean": 0.00104591078707017,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004006363349617459,
      "entropy": 0.3859316073358059,
      "epoch": 0.11492613346917983,
      "grad_norm": 34.727359771728516,
      "learning_rate": 7.080684596577017e-07,
      "loss": -6.4376,
      "reward": 3.9740967750549316,
      "reward_std": 0.3341333121061325,
      "rewards/reasoning_density_reward/mean": 0.4497023820877075,
      "rewards/reasoning_density_reward/std": 0.08173554763197899,
      "rewards/success_reward/mean": 2.953125,
      "rewards/success_reward/std": 0.1875,
      "rewards/trajectory_similarity_reward/mean": 0.5712693631649017,
      "rewards/trajectory_similarity_reward/std": 0.3466876596212387,
      "step": 141
    },
    {
      "clip_ratio/high_max": 0.019000602304004133,
      "clip_ratio/high_mean": 0.003392241371329874,
      "clip_ratio/low_mean": 0.0008305845731229056,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004222826049954165,
      "entropy": 0.39660463109612465,
      "epoch": 0.11574121242995415,
      "grad_norm": 11.149782180786133,
      "learning_rate": 7.074164629176854e-07,
      "loss": 7.2192,
      "reward": 2.325479745864868,
      "reward_std": 1.3713749647140503,
      "rewards/reasoning_density_reward/mean": 0.3673611283302307,
      "rewards/reasoning_density_reward/std": 0.1249999925494194,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.4581184983253479,
      "rewards/trajectory_similarity_reward/std": 0.494777649641037,
      "step": 142
    },
    {
      "clip_ratio/high_max": 0.010096678975969553,
      "clip_ratio/high_mean": 0.001727395698253531,
      "clip_ratio/low_mean": 0.0004539915025816299,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021813871426275,
      "entropy": 0.398218110203743,
      "epoch": 0.11655629139072848,
      "grad_norm": 12.448598861694336,
      "learning_rate": 7.067644661776691e-07,
      "loss": 10.4884,
      "reward": 2.7593088150024414,
      "reward_std": 0.8319442272186279,
      "rewards/reasoning_density_reward/mean": 0.37998175621032715,
      "rewards/reasoning_density_reward/std": 0.08500321209430695,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.4105769991874695,
      "rewards/trajectory_similarity_reward/std": 0.4123113453388214,
      "step": 143
    },
    {
      "clip_ratio/high_max": 0.008165470731910318,
      "clip_ratio/high_mean": 0.0017048229201463982,
      "clip_ratio/low_mean": 0.0005614021138171665,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022662249975837767,
      "entropy": 0.38561052456498146,
      "epoch": 0.1173713703515028,
      "grad_norm": 14.366609573364258,
      "learning_rate": 7.061124694376528e-07,
      "loss": -4.0329,
      "reward": 2.5205217599868774,
      "reward_std": 1.679360270500183,
      "rewards/reasoning_density_reward/mean": 0.3673921227455139,
      "rewards/reasoning_density_reward/std": 0.09335056319832802,
      "rewards/success_reward/mean": 1.828125,
      "rewards/success_reward/std": 1.3129705786705017,
      "rewards/trajectory_similarity_reward/mean": 0.32500453293323517,
      "rewards/trajectory_similarity_reward/std": 0.3819954842329025,
      "step": 144
    },
    {
      "clip_ratio/high_max": 0.009190324577502906,
      "clip_ratio/high_mean": 0.002210744882177096,
      "clip_ratio/low_mean": 0.0009887909982353449,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031995358440326527,
      "entropy": 0.40411194041371346,
      "epoch": 0.11818644931227713,
      "grad_norm": 11.761322021484375,
      "learning_rate": 7.054604726976365e-07,
      "loss": 2.9629,
      "reward": 2.746427297592163,
      "reward_std": 1.8174784183502197,
      "rewards/reasoning_density_reward/mean": 0.35739850997924805,
      "rewards/reasoning_density_reward/std": 0.09475551545619965,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.42027872800827026,
      "rewards/trajectory_similarity_reward/std": 0.35441187024116516,
      "step": 145
    },
    {
      "clip_ratio/high_max": 0.009257589699700475,
      "clip_ratio/high_mean": 0.00222003401722759,
      "clip_ratio/low_mean": 0.0008200490337912925,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030400830146390945,
      "entropy": 0.4005694203078747,
      "epoch": 0.11900152827305145,
      "grad_norm": 16.49759864807129,
      "learning_rate": 7.048084759576201e-07,
      "loss": 5.4304,
      "reward": 2.587724208831787,
      "reward_std": 1.3760197162628174,
      "rewards/reasoning_density_reward/mean": 0.33125001192092896,
      "rewards/reasoning_density_reward/std": 0.15798206627368927,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4752241373062134,
      "rewards/trajectory_similarity_reward/std": 0.4444490671157837,
      "step": 146
    },
    {
      "clip_ratio/high_max": 0.0069012478343211114,
      "clip_ratio/high_mean": 0.001304795005125925,
      "clip_ratio/low_mean": 0.0007185118447523564,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002023306835326366,
      "entropy": 0.4040742255747318,
      "epoch": 0.11981660723382578,
      "grad_norm": 12.841859817504883,
      "learning_rate": 7.041564792176039e-07,
      "loss": 0.3757,
      "reward": 3.0619869232177734,
      "reward_std": 1.8234646320343018,
      "rewards/reasoning_density_reward/mean": 0.3770066797733307,
      "rewards/reasoning_density_reward/std": 0.12291164696216583,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.5287305116653442,
      "rewards/trajectory_similarity_reward/std": 0.39171135425567627,
      "step": 147
    },
    {
      "clip_ratio/high_max": 0.010737568169133738,
      "clip_ratio/high_mean": 0.002178132901462959,
      "clip_ratio/low_mean": 0.00019677678756124806,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023749096872052178,
      "entropy": 0.39571545645594597,
      "epoch": 0.1206316861946001,
      "grad_norm": 14.199832916259766,
      "learning_rate": 7.035044824775875e-07,
      "loss": -3.1535,
      "reward": 2.7866201400756836,
      "reward_std": 1.6529579162597656,
      "rewards/reasoning_density_reward/mean": 0.34036457538604736,
      "rewards/reasoning_density_reward/std": 0.10994230955839157,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.3837556540966034,
      "rewards/trajectory_similarity_reward/std": 0.37400731444358826,
      "step": 148
    },
    {
      "clip_ratio/high_max": 0.013719257724005729,
      "clip_ratio/high_mean": 0.002154515270376578,
      "clip_ratio/low_mean": 5.351027357392013e-05,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002208025543950498,
      "entropy": 0.3895655609667301,
      "epoch": 0.12144676515537443,
      "grad_norm": 10.755170822143555,
      "learning_rate": 7.028524857375713e-07,
      "loss": -3.1393,
      "reward": 1.5501967668533325,
      "reward_std": 0.8626880049705505,
      "rewards/reasoning_density_reward/mean": 0.3265625238418579,
      "rewards/reasoning_density_reward/std": 0.055127691477537155,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.19238433241844177,
      "rewards/trajectory_similarity_reward/std": 0.30521160364151,
      "step": 149
    },
    {
      "clip_ratio/high_max": 0.006714242103043944,
      "clip_ratio/high_mean": 0.0010764557737275027,
      "clip_ratio/low_mean": 0.00025283676586695947,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013292925141286105,
      "entropy": 0.3737179934978485,
      "epoch": 0.12226184411614875,
      "grad_norm": 11.765026092529297,
      "learning_rate": 7.022004889975549e-07,
      "loss": -0.7364,
      "reward": 2.3648887276649475,
      "reward_std": 1.4015721380710602,
      "rewards/reasoning_density_reward/mean": 0.3773599714040756,
      "rewards/reasoning_density_reward/std": 0.07578011229634285,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3888907432556152,
      "rewards/trajectory_similarity_reward/mean": 0.3937787786126137,
      "rewards/trajectory_similarity_reward/std": 0.3534849286079407,
      "step": 150
    },
    {
      "clip_ratio/high_max": 0.012794107780791819,
      "clip_ratio/high_mean": 0.0034131778083974496,
      "clip_ratio/low_mean": 0.00033527310370118357,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037484508720808662,
      "entropy": 0.38559572398662567,
      "epoch": 0.12307692307692308,
      "grad_norm": 11.864717483520508,
      "learning_rate": 7.015484922575387e-07,
      "loss": 0.536,
      "reward": 2.8137240409851074,
      "reward_std": 1.0960065126419067,
      "rewards/reasoning_density_reward/mean": 0.3861810266971588,
      "rewards/reasoning_density_reward/std": 0.09848460555076599,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5525431036949158,
      "rewards/trajectory_similarity_reward/std": 0.462913453578949,
      "step": 151
    },
    {
      "clip_ratio/high_max": 0.009433104540221393,
      "clip_ratio/high_mean": 0.001736468679155223,
      "clip_ratio/low_mean": 0.00043508496310096234,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021715536422561854,
      "entropy": 0.3827439062297344,
      "epoch": 0.1238920020376974,
      "grad_norm": 16.466890335083008,
      "learning_rate": 7.008964955175223e-07,
      "loss": 2.8891,
      "reward": 2.5773134231567383,
      "reward_std": 2.005769729614258,
      "rewards/reasoning_density_reward/mean": 0.3373652398586273,
      "rewards/reasoning_density_reward/std": 0.12231392413377762,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4586978554725647,
      "rewards/trajectory_similarity_reward/std": 0.4340452253818512,
      "step": 152
    },
    {
      "clip_ratio/high_max": 0.010968711983878165,
      "clip_ratio/high_mean": 0.002021770029386971,
      "clip_ratio/low_mean": 0.00033814075504778884,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002359910780796781,
      "entropy": 0.3706898204982281,
      "epoch": 0.12470708099847173,
      "grad_norm": 10.561605453491211,
      "learning_rate": 7.002444987775061e-07,
      "loss": -12.8761,
      "reward": 2.6176185607910156,
      "reward_std": 1.5123642086982727,
      "rewards/reasoning_density_reward/mean": 0.38142360746860504,
      "rewards/reasoning_density_reward/std": 0.1245909035205841,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.458299458026886,
      "rewards/trajectory_similarity_reward/mean": 0.45494501292705536,
      "rewards/trajectory_similarity_reward/std": 0.41190703213214874,
      "step": 153
    },
    {
      "clip_ratio/high_max": 0.01231579773593694,
      "clip_ratio/high_mean": 0.0019934272131649777,
      "clip_ratio/low_mean": 0.0009144480245595332,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029078752850182354,
      "entropy": 0.38287464156746864,
      "epoch": 0.12552215995924607,
      "grad_norm": 13.475027084350586,
      "learning_rate": 6.995925020374897e-07,
      "loss": -13.5999,
      "reward": 1.4877517223358154,
      "reward_std": 1.680891513824463,
      "rewards/reasoning_density_reward/mean": 0.3492187559604645,
      "rewards/reasoning_density_reward/std": 0.10413020104169846,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.20103278756141663,
      "rewards/trajectory_similarity_reward/std": 0.3243806064128876,
      "step": 154
    },
    {
      "clip_ratio/high_max": 0.0033275985915679485,
      "clip_ratio/high_mean": 0.0005980840032862034,
      "clip_ratio/low_mean": 0.0004532984130491968,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0010513824090594426,
      "entropy": 0.4294942617416382,
      "epoch": 0.12633723892002038,
      "grad_norm": 11.345733642578125,
      "learning_rate": 6.989405052974735e-07,
      "loss": 11.8605,
      "reward": 1.4660570621490479,
      "reward_std": 1.7926106452941895,
      "rewards/reasoning_density_reward/mean": 0.2941964268684387,
      "rewards/reasoning_density_reward/std": 0.0997842326760292,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.23436054587364197,
      "rewards/trajectory_similarity_reward/std": 0.42003318667411804,
      "step": 155
    },
    {
      "clip_ratio/high_max": 0.010387293848907575,
      "clip_ratio/high_mean": 0.0016946650684985798,
      "clip_ratio/low_mean": 0.000788085690146545,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024827507659210823,
      "entropy": 0.4087708964943886,
      "epoch": 0.1271523178807947,
      "grad_norm": 11.320816993713379,
      "learning_rate": 6.982885085574571e-07,
      "loss": 6.4965,
      "reward": 1.4296177625656128,
      "reward_std": 1.0370242595672607,
      "rewards/reasoning_density_reward/mean": 0.3500000238418579,
      "rewards/reasoning_density_reward/std": 0.08944271504878998,
      "rewards/success_reward/mean": 0.84375,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.2358677238225937,
      "rewards/trajectory_similarity_reward/std": 0.423627644777298,
      "step": 156
    },
    {
      "clip_ratio/high_max": 0.009847345703747123,
      "clip_ratio/high_mean": 0.0023103043640730903,
      "clip_ratio/low_mean": 0.0004770772757183295,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027873817016370595,
      "entropy": 0.41480574384331703,
      "epoch": 0.12796739684156902,
      "grad_norm": 12.9498929977417,
      "learning_rate": 6.976365118174409e-07,
      "loss": -2.0669,
      "reward": 3.033770203590393,
      "reward_std": 1.0101042091846466,
      "rewards/reasoning_density_reward/mean": 0.4260416775941849,
      "rewards/reasoning_density_reward/std": 0.09804466366767883,
      "rewards/success_reward/mean": 2.015625,
      "rewards/success_reward/std": 1.4279338717460632,
      "rewards/trajectory_similarity_reward/mean": 0.5921034812927246,
      "rewards/trajectory_similarity_reward/std": 0.44566042721271515,
      "step": 157
    },
    {
      "clip_ratio/high_max": 0.009682834497652948,
      "clip_ratio/high_mean": 0.001969913311768323,
      "clip_ratio/low_mean": 0.0015060009027365595,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034759142581606284,
      "entropy": 0.3781626336276531,
      "epoch": 0.12878247580234337,
      "grad_norm": 14.914186477661133,
      "learning_rate": 6.969845150774245e-07,
      "loss": 5.2746,
      "reward": 2.3577993512153625,
      "reward_std": 1.4898010194301605,
      "rewards/reasoning_density_reward/mean": 0.3317336440086365,
      "rewards/reasoning_density_reward/std": 0.11661557108163834,
      "rewards/success_reward/mean": 1.640625,
      "rewards/success_reward/std": 1.4783848524093628,
      "rewards/trajectory_similarity_reward/mean": 0.3854407072067261,
      "rewards/trajectory_similarity_reward/std": 0.38184550404548645,
      "step": 158
    },
    {
      "clip_ratio/high_max": 0.012849576596636325,
      "clip_ratio/high_mean": 0.003091376660449896,
      "clip_ratio/low_mean": 0.0005852367539773695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003676613479910884,
      "entropy": 0.3823826424777508,
      "epoch": 0.12959755476311768,
      "grad_norm": 16.987085342407227,
      "learning_rate": 6.963325183374083e-07,
      "loss": -2.7187,
      "reward": 3.5957531929016113,
      "reward_std": 1.178663194179535,
      "rewards/reasoning_density_reward/mean": 0.4401041567325592,
      "rewards/reasoning_density_reward/std": 0.08775114268064499,
      "rewards/success_reward/mean": 2.484375,
      "rewards/success_reward/std": 1.0434684753417969,
      "rewards/trajectory_similarity_reward/mean": 0.6712740361690521,
      "rewards/trajectory_similarity_reward/std": 0.3902585506439209,
      "step": 159
    },
    {
      "clip_ratio/high_max": 0.009328960033599287,
      "clip_ratio/high_mean": 0.002285638875036966,
      "clip_ratio/low_mean": 0.0010936411999864504,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033792800895753317,
      "entropy": 0.4125041738152504,
      "epoch": 0.130412633723892,
      "grad_norm": 14.033353805541992,
      "learning_rate": 6.956805215973919e-07,
      "loss": -5.7764,
      "reward": 2.5765857696533203,
      "reward_std": 1.9957246780395508,
      "rewards/reasoning_density_reward/mean": 0.36651334166526794,
      "rewards/reasoning_density_reward/std": 0.09272938966751099,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.42882227897644043,
      "rewards/trajectory_similarity_reward/std": 0.41966935992240906,
      "step": 160
    },
    {
      "clip_ratio/high_max": 0.014741258579306304,
      "clip_ratio/high_mean": 0.002675855503184721,
      "clip_ratio/low_mean": 0.0009373363791382872,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003613191918702796,
      "entropy": 0.40620074421167374,
      "epoch": 0.13122771268466632,
      "grad_norm": 13.349939346313477,
      "learning_rate": 6.950285248573757e-07,
      "loss": 2.6474,
      "reward": 2.7327282428741455,
      "reward_std": 1.775402545928955,
      "rewards/reasoning_density_reward/mean": 0.3646329641342163,
      "rewards/reasoning_density_reward/std": 0.1268080621957779,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.399345338344574,
      "rewards/trajectory_similarity_reward/std": 0.36612585186958313,
      "step": 161
    },
    {
      "clip_ratio/high_max": 0.012593241932336241,
      "clip_ratio/high_mean": 0.0021494696484296583,
      "clip_ratio/low_mean": 0.000763728457968682,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029131980882084463,
      "entropy": 0.37562988698482513,
      "epoch": 0.13204279164544067,
      "grad_norm": 13.839911460876465,
      "learning_rate": 6.943765281173593e-07,
      "loss": 14.8912,
      "reward": 3.3566133975982666,
      "reward_std": 1.0987136363983154,
      "rewards/reasoning_density_reward/mean": 0.40154534578323364,
      "rewards/reasoning_density_reward/std": 0.08451814204454422,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.9031196236610413,
      "rewards/trajectory_similarity_reward/mean": 0.4238179922103882,
      "rewards/trajectory_similarity_reward/std": 0.3739141523838043,
      "step": 162
    },
    {
      "clip_ratio/high_max": 0.009499966108705848,
      "clip_ratio/high_mean": 0.0023754550420562737,
      "clip_ratio/low_mean": 0.0002098247641697526,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002585279777122196,
      "entropy": 0.41071898862719536,
      "epoch": 0.13285787060621498,
      "grad_norm": 15.83351993560791,
      "learning_rate": 6.937245313773431e-07,
      "loss": 5.5001,
      "reward": 2.3140273094177246,
      "reward_std": 1.6421265602111816,
      "rewards/reasoning_density_reward/mean": 0.38229167461395264,
      "rewards/reasoning_density_reward/std": 0.09840745478868484,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.33798548579216003,
      "rewards/trajectory_similarity_reward/std": 0.4157569110393524,
      "step": 163
    },
    {
      "clip_ratio/high_max": 0.010159045225009322,
      "clip_ratio/high_mean": 0.0025646166322985664,
      "clip_ratio/low_mean": 0.0007257847173605114,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003290401364210993,
      "entropy": 0.4127842038869858,
      "epoch": 0.1336729495669893,
      "grad_norm": 13.404502868652344,
      "learning_rate": 6.930725346373268e-07,
      "loss": -1.6234,
      "reward": 2.1652655601501465,
      "reward_std": 1.7501590251922607,
      "rewards/reasoning_density_reward/mean": 0.3625496029853821,
      "rewards/reasoning_density_reward/std": 0.0731247216463089,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.3027160167694092,
      "rewards/trajectory_similarity_reward/std": 0.3289486765861511,
      "step": 164
    },
    {
      "clip_ratio/high_max": 0.009634578600525856,
      "clip_ratio/high_mean": 0.0019164202603860758,
      "clip_ratio/low_mean": 0.0003698933869600296,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022863136473461054,
      "entropy": 0.3919391520321369,
      "epoch": 0.13448802852776362,
      "grad_norm": 12.944774627685547,
      "learning_rate": 6.924205378973105e-07,
      "loss": 2.7782,
      "reward": 3.4794445037841797,
      "reward_std": 1.4218971729278564,
      "rewards/reasoning_density_reward/mean": 0.4452967047691345,
      "rewards/reasoning_density_reward/std": 0.08195393532514572,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.5028976202011108,
      "rewards/trajectory_similarity_reward/std": 0.31194913387298584,
      "step": 165
    },
    {
      "clip_ratio/high_max": 0.015124135999940336,
      "clip_ratio/high_mean": 0.0034159182832809165,
      "clip_ratio/low_mean": 0.00029876634653192014,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003714684586157091,
      "entropy": 0.38718075677752495,
      "epoch": 0.13530310748853797,
      "grad_norm": 15.592130661010742,
      "learning_rate": 6.917685411572942e-07,
      "loss": 10.6284,
      "reward": 3.997786521911621,
      "reward_std": 0.5522137880325317,
      "rewards/reasoning_density_reward/mean": 0.45773810148239136,
      "rewards/reasoning_density_reward/std": 0.0683254525065422,
      "rewards/success_reward/mean": 2.90625,
      "rewards/success_reward/std": 0.375,
      "rewards/trajectory_similarity_reward/mean": 0.633798360824585,
      "rewards/trajectory_similarity_reward/std": 0.2955058813095093,
      "step": 166
    },
    {
      "clip_ratio/high_max": 0.009276503405999392,
      "clip_ratio/high_mean": 0.0022113450104370713,
      "clip_ratio/low_mean": 0.0004929564202029724,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027043014124501497,
      "entropy": 0.40631843358278275,
      "epoch": 0.13611818644931228,
      "grad_norm": 15.167108535766602,
      "learning_rate": 6.911165444172779e-07,
      "loss": -1.9137,
      "reward": 2.613046646118164,
      "reward_std": 1.0563425123691559,
      "rewards/reasoning_density_reward/mean": 0.3826389014720917,
      "rewards/reasoning_density_reward/std": 0.08368143439292908,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 0.7979557514190674,
      "rewards/trajectory_similarity_reward/mean": 0.44915781915187836,
      "rewards/trajectory_similarity_reward/std": 0.3307528495788574,
      "step": 167
    },
    {
      "clip_ratio/high_max": 0.006010768993292004,
      "clip_ratio/high_mean": 0.0012935197373735718,
      "clip_ratio/low_mean": 0.0009553099080221727,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022488296672236174,
      "entropy": 0.37188999727368355,
      "epoch": 0.1369332654100866,
      "grad_norm": 12.864849090576172,
      "learning_rate": 6.904645476772616e-07,
      "loss": 5.606,
      "reward": 2.485769748687744,
      "reward_std": 1.7809841632843018,
      "rewards/reasoning_density_reward/mean": 0.3921875059604645,
      "rewards/reasoning_density_reward/std": 0.09988273680210114,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.4998322129249573,
      "rewards/trajectory_similarity_reward/std": 0.5162249207496643,
      "step": 168
    },
    {
      "clip_ratio/high_max": 0.005780037201475352,
      "clip_ratio/high_mean": 0.0012113103512092493,
      "clip_ratio/low_mean": 0.0005950320664851461,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018063424686260987,
      "entropy": 0.388936422765255,
      "epoch": 0.13774834437086092,
      "grad_norm": 13.979738235473633,
      "learning_rate": 6.898125509372453e-07,
      "loss": 5.7232,
      "reward": 2.6641623973846436,
      "reward_std": 1.5470315217971802,
      "rewards/reasoning_density_reward/mean": 0.35863998532295227,
      "rewards/reasoning_density_reward/std": 0.08378536999225616,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.3367723524570465,
      "rewards/trajectory_similarity_reward/std": 0.41622278094291687,
      "step": 169
    },
    {
      "clip_ratio/high_max": 0.01056658144807443,
      "clip_ratio/high_mean": 0.002264893162646331,
      "clip_ratio/low_mean": 0.0010327456329832785,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032976387738017365,
      "entropy": 0.3802441768348217,
      "epoch": 0.13856342333163527,
      "grad_norm": 12.52430248260498,
      "learning_rate": 6.89160554197229e-07,
      "loss": -11.3781,
      "reward": 2.324349880218506,
      "reward_std": 1.6687215566635132,
      "rewards/reasoning_density_reward/mean": 0.4000000059604645,
      "rewards/reasoning_density_reward/std": 0.10327955335378647,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.4243498593568802,
      "rewards/trajectory_similarity_reward/std": 0.4417187571525574,
      "step": 170
    },
    {
      "clip_ratio/high_max": 0.01180473854765296,
      "clip_ratio/high_mean": 0.0019539905770216137,
      "clip_ratio/low_mean": 0.0004306242408347316,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023846147960284725,
      "entropy": 0.4070592038333416,
      "epoch": 0.13937850229240958,
      "grad_norm": 15.775177955627441,
      "learning_rate": 6.885085574572127e-07,
      "loss": 4.6488,
      "reward": 3.399030089378357,
      "reward_std": 1.2371079325675964,
      "rewards/reasoning_density_reward/mean": 0.43549107015132904,
      "rewards/reasoning_density_reward/std": 0.08966847509145737,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.0183849930763245,
      "rewards/trajectory_similarity_reward/mean": 0.5260390788316727,
      "rewards/trajectory_similarity_reward/std": 0.33433592319488525,
      "step": 171
    },
    {
      "clip_ratio/high_max": 0.007762773282593116,
      "clip_ratio/high_mean": 0.0020541283156489953,
      "clip_ratio/low_mean": 0.0010064469242934138,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003060575334529858,
      "entropy": 0.3879299499094486,
      "epoch": 0.1401935812531839,
      "grad_norm": 15.775835037231445,
      "learning_rate": 6.878565607171964e-07,
      "loss": -1.2833,
      "reward": 3.249419331550598,
      "reward_std": 1.2527936697006226,
      "rewards/reasoning_density_reward/mean": 0.37291668355464935,
      "rewards/reasoning_density_reward/std": 0.14159830659627914,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2028923630714417,
      "rewards/trajectory_similarity_reward/mean": 0.62650266289711,
      "rewards/trajectory_similarity_reward/std": 0.45159852504730225,
      "step": 172
    },
    {
      "clip_ratio/high_max": 0.013393625733442605,
      "clip_ratio/high_mean": 0.0037880946474615484,
      "clip_ratio/low_mean": 0.0004790396196767688,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004267134208930656,
      "entropy": 0.4023267552256584,
      "epoch": 0.14100866021395822,
      "grad_norm": 13.806035041809082,
      "learning_rate": 6.872045639771801e-07,
      "loss": 5.1311,
      "reward": 3.211730480194092,
      "reward_std": 1.4363245964050293,
      "rewards/reasoning_density_reward/mean": 0.4052083492279053,
      "rewards/reasoning_density_reward/std": 0.11733439564704895,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.4627723693847656,
      "rewards/trajectory_similarity_reward/std": 0.343506783246994,
      "step": 173
    },
    {
      "clip_ratio/high_max": 0.007000158453593031,
      "clip_ratio/high_mean": 0.0013559540166170336,
      "clip_ratio/low_mean": 0.0004776358837261796,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018335899076191708,
      "entropy": 0.3974648006260395,
      "epoch": 0.14182373917473257,
      "grad_norm": 13.064712524414062,
      "learning_rate": 6.865525672371638e-07,
      "loss": -0.7884,
      "reward": 2.9248586893081665,
      "reward_std": 1.014636218547821,
      "rewards/reasoning_density_reward/mean": 0.35104165971279144,
      "rewards/reasoning_density_reward/std": 0.13621064275503159,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.2328757047653198,
      "rewards/trajectory_similarity_reward/mean": 0.511317104101181,
      "rewards/trajectory_similarity_reward/std": 0.4263229966163635,
      "step": 174
    },
    {
      "clip_ratio/high_max": 0.009250518312910572,
      "clip_ratio/high_mean": 0.0015027716326585505,
      "clip_ratio/low_mean": 0.0008876134306774475,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023903850378701463,
      "entropy": 0.36534805223345757,
      "epoch": 0.14263881813550688,
      "grad_norm": 12.878594398498535,
      "learning_rate": 6.859005704971475e-07,
      "loss": -7.3843,
      "reward": 2.4985930919647217,
      "reward_std": 1.3826785683631897,
      "rewards/reasoning_density_reward/mean": 0.3854662775993347,
      "rewards/reasoning_density_reward/std": 0.0915164202451706,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5245966911315918,
      "rewards/trajectory_similarity_reward/mean": 0.425626739859581,
      "rewards/trajectory_similarity_reward/std": 0.4066130071878433,
      "step": 175
    },
    {
      "clip_ratio/high_max": 0.013614132825750858,
      "clip_ratio/high_mean": 0.003309984182124026,
      "clip_ratio/low_mean": 0.0011600832222029567,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0044700675789499655,
      "entropy": 0.3998383358120918,
      "epoch": 0.1434538970962812,
      "grad_norm": 12.499394416809082,
      "learning_rate": 6.852485737571312e-07,
      "loss": 5.9419,
      "reward": 3.5042641162872314,
      "reward_std": 1.6375395059585571,
      "rewards/reasoning_density_reward/mean": 0.39306318759918213,
      "rewards/reasoning_density_reward/std": 0.1253402680158615,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6737009286880493,
      "rewards/trajectory_similarity_reward/std": 0.3824829161167145,
      "step": 176
    },
    {
      "clip_ratio/high_max": 0.013611948350444436,
      "clip_ratio/high_mean": 0.003048770420718938,
      "clip_ratio/low_mean": 0.00045723127186647616,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003506001696223393,
      "entropy": 0.384543813765049,
      "epoch": 0.14426897605705552,
      "grad_norm": 12.86613941192627,
      "learning_rate": 6.845965770171149e-07,
      "loss": 15.4498,
      "reward": 1.8390939235687256,
      "reward_std": 0.9684025049209595,
      "rewards/reasoning_density_reward/mean": 0.36183035373687744,
      "rewards/reasoning_density_reward/std": 0.09937490522861481,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.35226359963417053,
      "rewards/trajectory_similarity_reward/std": 0.4745177924633026,
      "step": 177
    },
    {
      "clip_ratio/high_max": 0.014815297559835017,
      "clip_ratio/high_mean": 0.0030111288360785693,
      "clip_ratio/low_mean": 0.00036487704346654937,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033760059013729915,
      "entropy": 0.3899564631283283,
      "epoch": 0.14508405501782987,
      "grad_norm": 13.294541358947754,
      "learning_rate": 6.839445802770985e-07,
      "loss": 6.6921,
      "reward": 2.802819609642029,
      "reward_std": 1.5882542133331299,
      "rewards/reasoning_density_reward/mean": 0.40601326525211334,
      "rewards/reasoning_density_reward/std": 0.09444880113005638,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.4454171061515808,
      "rewards/trajectory_similarity_reward/mean": 0.5218062996864319,
      "rewards/trajectory_similarity_reward/std": 0.4225466251373291,
      "step": 178
    },
    {
      "clip_ratio/high_max": 0.013600087142549455,
      "clip_ratio/high_mean": 0.0029924339178251103,
      "clip_ratio/low_mean": 0.0006515312634292059,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036439651885302737,
      "entropy": 0.3678429052233696,
      "epoch": 0.14589913397860418,
      "grad_norm": 12.505427360534668,
      "learning_rate": 6.832925835370823e-07,
      "loss": 4.8898,
      "reward": 2.8258447647094727,
      "reward_std": 1.9658246040344238,
      "rewards/reasoning_density_reward/mean": 0.36666667461395264,
      "rewards/reasoning_density_reward/std": 0.13416408002376556,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.49042800068855286,
      "rewards/trajectory_similarity_reward/std": 0.42679131031036377,
      "step": 179
    },
    {
      "clip_ratio/high_max": 0.010001414513681084,
      "clip_ratio/high_mean": 0.001884029210486915,
      "clip_ratio/low_mean": 0.0009898165735648945,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002873845733120106,
      "entropy": 0.3822389654815197,
      "epoch": 0.1467142129393785,
      "grad_norm": 11.523100852966309,
      "learning_rate": 6.826405867970659e-07,
      "loss": 8.3086,
      "reward": 2.1484715938568115,
      "reward_std": 1.5533933639526367,
      "rewards/reasoning_density_reward/mean": 0.33794641494750977,
      "rewards/reasoning_density_reward/std": 0.06902965158224106,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.3105250597000122,
      "rewards/trajectory_similarity_reward/std": 0.4211239814758301,
      "step": 180
    },
    {
      "clip_ratio/high_max": 0.01236625635647215,
      "clip_ratio/high_mean": 0.0020485016466409434,
      "clip_ratio/low_mean": 0.0008079896033450495,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028564912718138658,
      "entropy": 0.37500206753611565,
      "epoch": 0.14752929190015282,
      "grad_norm": 14.639012336730957,
      "learning_rate": 6.819885900570497e-07,
      "loss": 3.7165,
      "reward": 3.14108943939209,
      "reward_std": 1.0902849435806274,
      "rewards/reasoning_density_reward/mean": 0.38229167461395264,
      "rewards/reasoning_density_reward/std": 0.12386054545640945,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5087976455688477,
      "rewards/trajectory_similarity_reward/std": 0.3270317018032074,
      "step": 181
    },
    {
      "clip_ratio/high_max": 0.011576706019695848,
      "clip_ratio/high_mean": 0.0022260078039835207,
      "clip_ratio/low_mean": 0.0011049320310121402,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033309398204437457,
      "entropy": 0.37277961894869804,
      "epoch": 0.14834437086092717,
      "grad_norm": 15.377359390258789,
      "learning_rate": 6.813365933170333e-07,
      "loss": 1.6947,
      "reward": 3.143963098526001,
      "reward_std": 1.342300295829773,
      "rewards/reasoning_density_reward/mean": 0.39084821939468384,
      "rewards/reasoning_density_reward/std": 0.10783803835511208,
      "rewards/success_reward/mean": 2.296875,
      "rewards/success_reward/std": 0.8591630160808563,
      "rewards/trajectory_similarity_reward/mean": 0.4562399983406067,
      "rewards/trajectory_similarity_reward/std": 0.42167051136493683,
      "step": 182
    },
    {
      "clip_ratio/high_max": 0.010570888291113079,
      "clip_ratio/high_mean": 0.0024891560606192797,
      "clip_ratio/low_mean": 0.0007855419426050503,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032746979704825208,
      "entropy": 0.3982829414308071,
      "epoch": 0.14915944982170148,
      "grad_norm": 16.302087783813477,
      "learning_rate": 6.806845965770171e-07,
      "loss": 7.4585,
      "reward": 2.5030031204223633,
      "reward_std": 2.0753729343414307,
      "rewards/reasoning_density_reward/mean": 0.21770833432674408,
      "rewards/reasoning_density_reward/std": 0.1450251340866089,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.504044771194458,
      "rewards/trajectory_similarity_reward/std": 0.46401017904281616,
      "step": 183
    },
    {
      "clip_ratio/high_max": 0.009027595166116953,
      "clip_ratio/high_mean": 0.0020342562565929256,
      "clip_ratio/low_mean": 0.000600591203692602,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026348474857513793,
      "entropy": 0.41968884319067,
      "epoch": 0.1499745287824758,
      "grad_norm": 13.646903991699219,
      "learning_rate": 6.800325998370007e-07,
      "loss": 4.2425,
      "reward": 2.5822978019714355,
      "reward_std": 1.9313883781433105,
      "rewards/reasoning_density_reward/mean": 0.3739583492279053,
      "rewards/reasoning_density_reward/std": 0.1113833487033844,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.4270893335342407,
      "rewards/trajectory_similarity_reward/std": 0.44680118560791016,
      "step": 184
    },
    {
      "clip_ratio/high_max": 0.010095233068568632,
      "clip_ratio/high_mean": 0.0021438007352116983,
      "clip_ratio/low_mean": 0.0009301425161538646,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030739432622794993,
      "entropy": 0.39448364078998566,
      "epoch": 0.15078960774325012,
      "grad_norm": 12.589824676513672,
      "learning_rate": 6.793806030969845e-07,
      "loss": 8.5741,
      "reward": 3.3422040939331055,
      "reward_std": 1.52301025390625,
      "rewards/reasoning_density_reward/mean": 0.34337514638900757,
      "rewards/reasoning_density_reward/std": 0.15204022824764252,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.6550790071487427,
      "rewards/trajectory_similarity_reward/std": 0.39999982714653015,
      "step": 185
    },
    {
      "clip_ratio/high_max": 0.014869879116304219,
      "clip_ratio/high_mean": 0.0028013469782308675,
      "clip_ratio/low_mean": 0.000328998539771419,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031303454670705833,
      "entropy": 0.4117763489484787,
      "epoch": 0.15160468670402447,
      "grad_norm": 13.851509094238281,
      "learning_rate": 6.787286063569681e-07,
      "loss": 12.7446,
      "reward": 2.3865814208984375,
      "reward_std": 1.5967965126037598,
      "rewards/reasoning_density_reward/mean": 0.22801339626312256,
      "rewards/reasoning_density_reward/std": 0.15853244066238403,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.37731799483299255,
      "rewards/trajectory_similarity_reward/std": 0.43885600566864014,
      "step": 186
    },
    {
      "clip_ratio/high_max": 0.009288652567192912,
      "clip_ratio/high_mean": 0.0021040772990090773,
      "clip_ratio/low_mean": 0.0006196065514814109,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002723683817748679,
      "entropy": 0.40828223526477814,
      "epoch": 0.15241976566479878,
      "grad_norm": 14.511368751525879,
      "learning_rate": 6.780766096169519e-07,
      "loss": -5.8662,
      "reward": 1.6908168196678162,
      "reward_std": 1.1773008108139038,
      "rewards/reasoning_density_reward/mean": 0.3283854275941849,
      "rewards/reasoning_density_reward/std": 0.07353372871875763,
      "rewards/success_reward/mean": 1.265625,
      "rewards/success_reward/std": 1.2894206047058105,
      "rewards/trajectory_similarity_reward/mean": 0.09680644422769547,
      "rewards/trajectory_similarity_reward/std": 0.20303834974765778,
      "step": 187
    },
    {
      "clip_ratio/high_max": 0.011157759552588686,
      "clip_ratio/high_mean": 0.002038913491560379,
      "clip_ratio/low_mean": 0.0010250098348478787,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003063923359150067,
      "entropy": 0.3810150511562824,
      "epoch": 0.1532348446255731,
      "grad_norm": 13.379528045654297,
      "learning_rate": 6.774246128769355e-07,
      "loss": -2.3817,
      "reward": 2.278507649898529,
      "reward_std": 1.1344427168369293,
      "rewards/reasoning_density_reward/mean": 0.3843750059604645,
      "rewards/reasoning_density_reward/std": 0.10977387428283691,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.459863543510437,
      "rewards/trajectory_similarity_reward/mean": 0.4410076141357422,
      "rewards/trajectory_similarity_reward/std": 0.46693648397922516,
      "step": 188
    },
    {
      "clip_ratio/high_max": 0.0068621930840890855,
      "clip_ratio/high_mean": 0.0013586800159828272,
      "clip_ratio/low_mean": 0.0009883219900075346,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002347002016904298,
      "entropy": 0.3643890880048275,
      "epoch": 0.15404992358634742,
      "grad_norm": 14.030029296875,
      "learning_rate": 6.767726161369193e-07,
      "loss": 4.0265,
      "reward": 2.4079737663269043,
      "reward_std": 1.8823875188827515,
      "rewards/reasoning_density_reward/mean": 0.3671875,
      "rewards/reasoning_density_reward/std": 0.11965315788984299,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.4162823557853699,
      "rewards/trajectory_similarity_reward/mean": 0.44703616201877594,
      "rewards/trajectory_similarity_reward/std": 0.48242008686065674,
      "step": 189
    },
    {
      "clip_ratio/high_max": 0.008569017227273434,
      "clip_ratio/high_mean": 0.0021528387878788635,
      "clip_ratio/low_mean": 0.0009518638853478478,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031047026568558067,
      "entropy": 0.42652832716703415,
      "epoch": 0.15486500254712177,
      "grad_norm": 18.739139556884766,
      "learning_rate": 6.761206193969029e-07,
      "loss": 1.1878,
      "reward": 0.4922678768634796,
      "reward_std": 0.7202215790748596,
      "rewards/reasoning_density_reward/mean": 0.2552083432674408,
      "rewards/reasoning_density_reward/std": 0.09183697402477264,
      "rewards/success_reward/mean": 0.1875,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.04955955222249031,
      "rewards/trajectory_similarity_reward/std": 0.19823820888996124,
      "step": 190
    },
    {
      "clip_ratio/high_max": 0.009199935419019312,
      "clip_ratio/high_mean": 0.0022119562127045356,
      "clip_ratio/low_mean": 0.0005152832309249789,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027272394218016416,
      "entropy": 0.39669498801231384,
      "epoch": 0.15568008150789608,
      "grad_norm": 11.621088027954102,
      "learning_rate": 6.754686226568867e-07,
      "loss": -15.9562,
      "reward": 2.417794942855835,
      "reward_std": 1.3816693425178528,
      "rewards/reasoning_density_reward/mean": 0.3541666865348816,
      "rewards/reasoning_density_reward/std": 0.11718615517020226,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.2754897475242615,
      "rewards/trajectory_similarity_reward/mean": 0.4698782339692116,
      "rewards/trajectory_similarity_reward/std": 0.3513573259115219,
      "step": 191
    },
    {
      "clip_ratio/high_max": 0.010900053341174498,
      "clip_ratio/high_mean": 0.0019278493309684563,
      "clip_ratio/low_mean": 0.0003904369077645242,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023182862423709594,
      "entropy": 0.434473667293787,
      "epoch": 0.1564951604686704,
      "grad_norm": 13.661466598510742,
      "learning_rate": 6.748166259168703e-07,
      "loss": -3.5178,
      "reward": 1.4379825592041016,
      "reward_std": 1.7500710487365723,
      "rewards/reasoning_density_reward/mean": 0.31979167461395264,
      "rewards/reasoning_density_reward/std": 0.08101639896631241,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.1806909441947937,
      "rewards/trajectory_similarity_reward/std": 0.3369207978248596,
      "step": 192
    },
    {
      "clip_ratio/high_max": 0.007669130776776001,
      "clip_ratio/high_mean": 0.0017506778822280467,
      "clip_ratio/low_mean": 0.0006028050956956577,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023534829742857255,
      "entropy": 0.4205329902470112,
      "epoch": 0.15731023942944472,
      "grad_norm": 15.355896949768066,
      "learning_rate": 6.741646291768541e-07,
      "loss": -11.4802,
      "reward": 1.5706437230110168,
      "reward_std": 1.7430545091629028,
      "rewards/reasoning_density_reward/mean": 0.32816222310066223,
      "rewards/reasoning_density_reward/std": 0.07590590789914131,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.3462588787078857,
      "rewards/trajectory_similarity_reward/mean": 0.2112315446138382,
      "rewards/trajectory_similarity_reward/std": 0.3376599997282028,
      "step": 193
    },
    {
      "clip_ratio/high_max": 0.008992098330054432,
      "clip_ratio/high_mean": 0.0021627193709718995,
      "clip_ratio/low_mean": 0.0014992763244663365,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036619955935748294,
      "entropy": 0.4154961258172989,
      "epoch": 0.15812531839021907,
      "grad_norm": 14.557741165161133,
      "learning_rate": 6.735126324368378e-07,
      "loss": 1.657,
      "reward": 2.854020595550537,
      "reward_std": 1.476022720336914,
      "rewards/reasoning_density_reward/mean": 0.41354167461395264,
      "rewards/reasoning_density_reward/std": 0.09568829089403152,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.4717288613319397,
      "rewards/trajectory_similarity_reward/std": 0.39272671937942505,
      "step": 194
    },
    {
      "clip_ratio/high_max": 0.014254284120397642,
      "clip_ratio/high_mean": 0.0030918466109142173,
      "clip_ratio/low_mean": 0.0008295432890008669,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003921390001778491,
      "entropy": 0.3727341443300247,
      "epoch": 0.15894039735099338,
      "grad_norm": 13.068658828735352,
      "learning_rate": 6.728606356968215e-07,
      "loss": 1.8364,
      "reward": 1.7375197410583496,
      "reward_std": 2.0067298412323,
      "rewards/reasoning_density_reward/mean": 0.2906250059604645,
      "rewards/reasoning_density_reward/std": 0.1301417052745819,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.32189473509788513,
      "rewards/trajectory_similarity_reward/std": 0.4410015046596527,
      "step": 195
    },
    {
      "clip_ratio/high_max": 0.010070839314721525,
      "clip_ratio/high_mean": 0.0024960446899058297,
      "clip_ratio/low_mean": 0.0010423208004795015,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003538365592248738,
      "entropy": 0.4251212887465954,
      "epoch": 0.1597554763117677,
      "grad_norm": 15.201254844665527,
      "learning_rate": 6.722086389568052e-07,
      "loss": 20.1151,
      "reward": 3.3027689456939697,
      "reward_std": 1.5566644668579102,
      "rewards/reasoning_density_reward/mean": 0.34973958134651184,
      "rewards/reasoning_density_reward/std": 0.17838560789823532,
      "rewards/success_reward/mean": 2.390625,
      "rewards/success_reward/std": 1.1404882073402405,
      "rewards/trajectory_similarity_reward/mean": 0.5624043643474579,
      "rewards/trajectory_similarity_reward/std": 0.36137746274471283,
      "step": 196
    },
    {
      "clip_ratio/high_max": 0.012321467162109911,
      "clip_ratio/high_mean": 0.003165755930240266,
      "clip_ratio/low_mean": 0.0011522746426635422,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004318030652939342,
      "entropy": 0.4274259954690933,
      "epoch": 0.16057055527254202,
      "grad_norm": 13.487480163574219,
      "learning_rate": 6.715566422167889e-07,
      "loss": -2.4186,
      "reward": 1.880356788635254,
      "reward_std": 1.7862226963043213,
      "rewards/reasoning_density_reward/mean": 0.2864583432674408,
      "rewards/reasoning_density_reward/std": 0.17333534359931946,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.2813985049724579,
      "rewards/trajectory_similarity_reward/std": 0.3525673747062683,
      "step": 197
    },
    {
      "clip_ratio/high_max": 0.012482275604270399,
      "clip_ratio/high_mean": 0.002306920141563751,
      "clip_ratio/low_mean": 0.00047588204324711114,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027828021920868196,
      "entropy": 0.42776860296726227,
      "epoch": 0.16138563423331637,
      "grad_norm": 14.212177276611328,
      "learning_rate": 6.709046454767726e-07,
      "loss": 4.2407,
      "reward": 1.8437749147415161,
      "reward_std": 1.5584648251533508,
      "rewards/reasoning_density_reward/mean": 0.31011904776096344,
      "rewards/reasoning_density_reward/std": 0.10546908900141716,
      "rewards/success_reward/mean": 1.265625,
      "rewards/success_reward/std": 1.4295060634613037,
      "rewards/trajectory_similarity_reward/mean": 0.2680308669805527,
      "rewards/trajectory_similarity_reward/std": 0.37211769819259644,
      "step": 198
    },
    {
      "clip_ratio/high_max": 0.009112466184888035,
      "clip_ratio/high_mean": 0.0018908846832346171,
      "clip_ratio/low_mean": 0.00026603183141560294,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021569165292021353,
      "entropy": 0.40134814381599426,
      "epoch": 0.16220071319409068,
      "grad_norm": 11.406978607177734,
      "learning_rate": 6.702526487367563e-07,
      "loss": -10.8719,
      "step": 199
    },
    {
      "clip_ratio/high_max": 0.010839493275852874,
      "clip_ratio/high_mean": 0.0017960587902052794,
      "clip_ratio/low_mean": 0.0009875511095742695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002783609787002206,
      "entropy": 0.3971281312406063,
      "epoch": 0.163015792154865,
      "grad_norm": 20.610210418701172,
      "learning_rate": 6.6960065199674e-07,
      "loss": -6.6534,
      "reward": 1.9562755823135376,
      "reward_std": 1.8585047721862793,
      "rewards/reasoning_density_reward/mean": 0.32965031266212463,
      "rewards/reasoning_density_reward/std": 0.09633892774581909,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.22037525475025177,
      "rewards/trajectory_similarity_reward/std": 0.26803797483444214,
      "step": 200
    },
    {
      "clip_ratio/high_max": 0.006816069129854441,
      "clip_ratio/high_mean": 0.0012123807973694056,
      "clip_ratio/low_mean": 0.0008283125207526609,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020406933035701513,
      "entropy": 0.3884081318974495,
      "epoch": 0.16383087111563932,
      "grad_norm": 15.171896934509277,
      "learning_rate": 6.689486552567237e-07,
      "loss": 3.3492,
      "reward": 1.9186077117919922,
      "reward_std": 1.9475679397583008,
      "rewards/reasoning_density_reward/mean": 0.35624998807907104,
      "rewards/reasoning_density_reward/std": 0.08139410614967346,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.34360766410827637,
      "rewards/trajectory_similarity_reward/std": 0.4612559378147125,
      "step": 201
    },
    {
      "clip_ratio/high_max": 0.007867972948588431,
      "clip_ratio/high_mean": 0.0015086614221218042,
      "clip_ratio/low_mean": 0.0008110008857329376,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002319662264198996,
      "entropy": 0.3867642357945442,
      "epoch": 0.16464595007641367,
      "grad_norm": 15.056714057922363,
      "learning_rate": 6.682966585167074e-07,
      "loss": -2.9098,
      "reward": 2.7078007459640503,
      "reward_std": 1.9011127948760986,
      "rewards/reasoning_density_reward/mean": 0.3773437589406967,
      "rewards/reasoning_density_reward/std": 0.09074325114488602,
      "rewards/success_reward/mean": 1.828125,
      "rewards/success_reward/std": 1.36735200881958,
      "rewards/trajectory_similarity_reward/mean": 0.5023320317268372,
      "rewards/trajectory_similarity_reward/std": 0.43835969269275665,
      "step": 202
    },
    {
      "clip_ratio/high_max": 0.012861779250670224,
      "clip_ratio/high_mean": 0.0026174297454417683,
      "clip_ratio/low_mean": 0.00024397996821790002,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00286140973912552,
      "entropy": 0.40878527238965034,
      "epoch": 0.16546102903718798,
      "grad_norm": 12.076717376708984,
      "learning_rate": 6.676446617766911e-07,
      "loss": 3.9818,
      "reward": 0.27431008219718933,
      "reward_std": 0.042603202164173126,
      "rewards/reasoning_density_reward/mean": 0.27431008219718933,
      "rewards/reasoning_density_reward/std": 0.06396899372339249,
      "rewards/success_reward/mean": 0.0,
      "rewards/success_reward/std": 0.0,
      "rewards/trajectory_similarity_reward/mean": 0.0,
      "rewards/trajectory_similarity_reward/std": 0.0,
      "step": 203
    },
    {
      "clip_ratio/high_max": 0.00900199101306498,
      "clip_ratio/high_mean": 0.0020477658326854,
      "clip_ratio/low_mean": 0.0007906591381470207,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028384249308146536,
      "entropy": 0.40866315364837646,
      "epoch": 0.1662761079979623,
      "grad_norm": 12.474067687988281,
      "learning_rate": 6.669926650366748e-07,
      "loss": 9.6577,
      "reward": 3.5568830966949463,
      "reward_std": 1.1547958850860596,
      "rewards/reasoning_density_reward/mean": 0.36901043355464935,
      "rewards/reasoning_density_reward/std": 0.16355398297309875,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.1831679940223694,
      "rewards/trajectory_similarity_reward/mean": 0.7503727078437805,
      "rewards/trajectory_similarity_reward/std": 0.3705642819404602,
      "step": 204
    },
    {
      "clip_ratio/high_max": 0.010936850274447352,
      "clip_ratio/high_mean": 0.0022643383854301646,
      "clip_ratio/low_mean": 0.000541205929039279,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002805544252623804,
      "entropy": 0.378336526453495,
      "epoch": 0.16709118695873662,
      "grad_norm": 12.135371208190918,
      "learning_rate": 6.663406682966585e-07,
      "loss": -2.1487,
      "reward": 1.5982475876808167,
      "reward_std": 1.7663294076919556,
      "rewards/reasoning_density_reward/mean": 0.30260416865348816,
      "rewards/reasoning_density_reward/std": 0.14299274235963821,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.4680703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.2643933668732643,
      "rewards/trajectory_similarity_reward/std": 0.38810352981090546,
      "step": 205
    },
    {
      "clip_ratio/high_max": 0.008184676116798073,
      "clip_ratio/high_mean": 0.0012601193520822562,
      "clip_ratio/low_mean": 0.000944146464462392,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022042658310965635,
      "entropy": 0.4213388003408909,
      "epoch": 0.16790626591951097,
      "grad_norm": 23.72439956665039,
      "learning_rate": 6.656886715566422e-07,
      "loss": 7.6453,
      "reward": 1.9366759061813354,
      "reward_std": 1.4883842468261719,
      "rewards/reasoning_density_reward/mean": 0.32083335518836975,
      "rewards/reasoning_density_reward/std": 0.12594090402126312,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.30334246158599854,
      "rewards/trajectory_similarity_reward/std": 0.4138539731502533,
      "step": 206
    },
    {
      "clip_ratio/high_max": 0.012980960891582072,
      "clip_ratio/high_mean": 0.0028289632609812543,
      "clip_ratio/low_mean": 0.00040114209332386963,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032301053288392723,
      "entropy": 0.4311026744544506,
      "epoch": 0.16872134488028528,
      "grad_norm": 12.558481216430664,
      "learning_rate": 6.65036674816626e-07,
      "loss": -4.7631,
      "reward": 2.530846357345581,
      "reward_std": 1.7872974276542664,
      "rewards/reasoning_density_reward/mean": 0.3229166865348816,
      "rewards/reasoning_density_reward/std": 0.1445029005408287,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.3888907432556152,
      "rewards/trajectory_similarity_reward/mean": 0.42667968571186066,
      "rewards/trajectory_similarity_reward/std": 0.42734138667583466,
      "step": 207
    },
    {
      "clip_ratio/high_max": 0.009528428694466129,
      "clip_ratio/high_mean": 0.0014722289961355273,
      "clip_ratio/low_mean": 0.0010320379624317866,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025042670058610383,
      "entropy": 0.37615643814206123,
      "epoch": 0.1695364238410596,
      "grad_norm": 16.30551528930664,
      "learning_rate": 6.643846780766096e-07,
      "loss": 9.8301,
      "reward": 2.408722400665283,
      "reward_std": 1.5987976789474487,
      "rewards/reasoning_density_reward/mean": 0.3421875238418579,
      "rewards/reasoning_density_reward/std": 0.16271620988845825,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.37903499603271484,
      "rewards/trajectory_similarity_reward/std": 0.44682982563972473,
      "step": 208
    },
    {
      "clip_ratio/high_max": 0.00760275122593157,
      "clip_ratio/high_mean": 0.0017718409981171135,
      "clip_ratio/low_mean": 0.0007367488069576211,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025085898087127134,
      "entropy": 0.3847537785768509,
      "epoch": 0.17035150280183392,
      "grad_norm": 18.22587013244629,
      "learning_rate": 6.637326813365933e-07,
      "loss": -4.2789,
      "reward": 2.5141361951828003,
      "reward_std": 1.2665776014328003,
      "rewards/reasoning_density_reward/mean": 0.36480656266212463,
      "rewards/reasoning_density_reward/std": 0.13447962701320648,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.3304427862167358,
      "rewards/trajectory_similarity_reward/mean": 0.36807963252067566,
      "rewards/trajectory_similarity_reward/std": 0.37272436916828156,
      "step": 209
    },
    {
      "clip_ratio/high_max": 0.009760737360920757,
      "clip_ratio/high_mean": 0.0023823629526305012,
      "clip_ratio/low_mean": 0.0004422203310241457,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028245832872926258,
      "entropy": 0.40344857424497604,
      "epoch": 0.17116658176260827,
      "grad_norm": 15.076071739196777,
      "learning_rate": 6.63080684596577e-07,
      "loss": 5.2962,
      "reward": 3.477755546569824,
      "reward_std": 1.3262181282043457,
      "rewards/reasoning_density_reward/mean": 0.3577924817800522,
      "rewards/reasoning_density_reward/std": 0.1371065266430378,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.9742441177368164,
      "rewards/trajectory_similarity_reward/mean": 0.682463139295578,
      "rewards/trajectory_similarity_reward/std": 0.3820239007472992,
      "step": 210
    },
    {
      "clip_ratio/high_max": 0.009618124779080972,
      "clip_ratio/high_mean": 0.0021814734318468254,
      "clip_ratio/low_mean": 0.0006440041761379689,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002825477633450646,
      "entropy": 0.3794734850525856,
      "epoch": 0.17198166072338258,
      "grad_norm": 13.205920219421387,
      "learning_rate": 6.624286878565607e-07,
      "loss": 8.249,
      "reward": 4.438899993896484,
      "reward_std": 0.0713089182972908,
      "rewards/reasoning_density_reward/mean": 0.4671874940395355,
      "rewards/reasoning_density_reward/std": 0.062395744025707245,
      "rewards/success_reward/mean": 3.0,
      "rewards/success_reward/std": 0.0,
      "rewards/trajectory_similarity_reward/mean": 0.971712589263916,
      "rewards/trajectory_similarity_reward/std": 0.05802811682224274,
      "step": 211
    },
    {
      "clip_ratio/high_max": 0.00796051969518885,
      "clip_ratio/high_mean": 0.0016046721502789296,
      "clip_ratio/low_mean": 0.0004257549990143161,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002030427163845161,
      "entropy": 0.3504056930541992,
      "epoch": 0.1727967396841569,
      "grad_norm": 16.32909393310547,
      "learning_rate": 6.617766911165443e-07,
      "loss": -5.1335,
      "reward": 3.3396964073181152,
      "reward_std": 1.3214404582977295,
      "rewards/reasoning_density_reward/mean": 0.38147321343421936,
      "rewards/reasoning_density_reward/std": 0.1282055862247944,
      "rewards/success_reward/mean": 2.390625,
      "rewards/success_reward/std": 1.1228033304214478,
      "rewards/trajectory_similarity_reward/mean": 0.567597970366478,
      "rewards/trajectory_similarity_reward/std": 0.33839768171310425,
      "step": 212
    },
    {
      "clip_ratio/high_max": 0.0056570504384581,
      "clip_ratio/high_mean": 0.0010466304556757677,
      "clip_ratio/low_mean": 0.000890917144715786,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019375475822016597,
      "entropy": 0.4404926188290119,
      "epoch": 0.17361181864493122,
      "grad_norm": 17.940732955932617,
      "learning_rate": 6.611246943765281e-07,
      "loss": -2.2164,
      "reward": 1.9693751335144043,
      "reward_std": 1.8547470569610596,
      "rewards/reasoning_density_reward/mean": 0.23497024178504944,
      "rewards/reasoning_density_reward/std": 0.144023597240448,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.32815495133399963,
      "rewards/trajectory_similarity_reward/std": 0.39503344893455505,
      "step": 213
    },
    {
      "clip_ratio/high_max": 0.010513077024370432,
      "clip_ratio/high_mean": 0.00268908885482233,
      "clip_ratio/low_mean": 0.00024612996639916673,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029352188139455393,
      "entropy": 0.43870822340250015,
      "epoch": 0.17442689760570557,
      "grad_norm": 15.071953773498535,
      "learning_rate": 6.604726976365117e-07,
      "loss": -14.1188,
      "reward": 2.327479124069214,
      "reward_std": 1.7214199304580688,
      "rewards/reasoning_density_reward/mean": 0.28595781326293945,
      "rewards/reasoning_density_reward/std": 0.07513612508773804,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.3540213406085968,
      "rewards/trajectory_similarity_reward/std": 0.4027087092399597,
      "step": 214
    },
    {
      "clip_ratio/high_max": 0.014480903162620962,
      "clip_ratio/high_mean": 0.003456480466411449,
      "clip_ratio/low_mean": 0.00034053130184474867,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037970117264194414,
      "entropy": 0.38384050130844116,
      "epoch": 0.17524197656647988,
      "grad_norm": 18.001808166503906,
      "learning_rate": 6.598207008964955e-07,
      "loss": -4.6191,
      "reward": 2.830114006996155,
      "reward_std": 1.9118841886520386,
      "rewards/reasoning_density_reward/mean": 0.3810019791126251,
      "rewards/reasoning_density_reward/std": 0.11651447042822838,
      "rewards/success_reward/mean": 1.921875,
      "rewards/success_reward/std": 1.4538734555244446,
      "rewards/trajectory_similarity_reward/mean": 0.5272370874881744,
      "rewards/trajectory_similarity_reward/std": 0.44798336923122406,
      "step": 215
    },
    {
      "clip_ratio/high_max": 0.01097109541296959,
      "clip_ratio/high_mean": 0.0020959830580977723,
      "clip_ratio/low_mean": 0.0012277930582058616,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003323776094475761,
      "entropy": 0.3768860511481762,
      "epoch": 0.1760570555272542,
      "grad_norm": 12.076434135437012,
      "learning_rate": 6.591687041564791e-07,
      "loss": 3.9824,
      "reward": 2.0548949241638184,
      "reward_std": 1.8046026229858398,
      "rewards/reasoning_density_reward/mean": 0.34375,
      "rewards/reasoning_density_reward/std": 0.10267261415719986,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.2808688879013062,
      "rewards/trajectory_similarity_reward/mean": 0.3048948645591736,
      "rewards/trajectory_similarity_reward/std": 0.4677925705909729,
      "step": 216
    },
    {
      "clip_ratio/high_max": 0.014790251618251204,
      "clip_ratio/high_mean": 0.0027898005209863186,
      "clip_ratio/low_mean": 0.0009368196515424643,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037266201979946345,
      "entropy": 0.4185255765914917,
      "epoch": 0.17687213448802852,
      "grad_norm": 12.390375137329102,
      "learning_rate": 6.585167074164629e-07,
      "loss": 10.2602,
      "reward": 1.9842045307159424,
      "reward_std": 1.5292539596557617,
      "rewards/reasoning_density_reward/mean": 0.36250001192092896,
      "rewards/reasoning_density_reward/std": 0.08850612491369247,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.3092043995857239,
      "rewards/trajectory_similarity_reward/std": 0.41456523537635803,
      "step": 217
    },
    {
      "clip_ratio/high_max": 0.00621009522001259,
      "clip_ratio/high_mean": 0.001403634698363021,
      "clip_ratio/low_mean": 0.0005572693717112998,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019609040282375645,
      "entropy": 0.4258679263293743,
      "epoch": 0.17768721344880287,
      "grad_norm": 14.392851829528809,
      "learning_rate": 6.578647106764465e-07,
      "loss": -5.2028,
      "reward": 1.1439907550811768,
      "reward_std": 1.4061234593391418,
      "rewards/reasoning_density_reward/mean": 0.2637992277741432,
      "rewards/reasoning_density_reward/std": 0.1049049086868763,
      "rewards/success_reward/mean": 0.75,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.13019154965877533,
      "rewards/trajectory_similarity_reward/std": 0.2936646044254303,
      "step": 218
    },
    {
      "clip_ratio/high_max": 0.01169257564470172,
      "clip_ratio/high_mean": 0.002018424973357469,
      "clip_ratio/low_mean": 0.0010583546136331279,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003076779641560279,
      "entropy": 0.39171723276376724,
      "epoch": 0.17850229240957718,
      "grad_norm": 12.634140968322754,
      "learning_rate": 6.572127139364303e-07,
      "loss": -3.2267,
      "reward": 2.9371116161346436,
      "reward_std": 1.4628205299377441,
      "rewards/reasoning_density_reward/mean": 0.2604166865348816,
      "rewards/reasoning_density_reward/std": 0.17913904786109924,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.4266948997974396,
      "rewards/trajectory_similarity_reward/std": 0.37329739332199097,
      "step": 219
    },
    {
      "clip_ratio/high_max": 0.00967009732266888,
      "clip_ratio/high_mean": 0.0024016626994125545,
      "clip_ratio/low_mean": 0.001150601870904211,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035522645921446383,
      "entropy": 0.4111770763993263,
      "epoch": 0.1793173713703515,
      "grad_norm": 12.125751495361328,
      "learning_rate": 6.565607171964139e-07,
      "loss": 10.3898,
      "reward": 2.47548770904541,
      "reward_std": 1.3995863199234009,
      "rewards/reasoning_density_reward/mean": 0.3671875,
      "rewards/reasoning_density_reward/std": 0.10665228962898254,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.32705017924308777,
      "rewards/trajectory_similarity_reward/std": 0.34830451011657715,
      "step": 220
    },
    {
      "clip_ratio/high_max": 0.01672524760942906,
      "clip_ratio/high_mean": 0.003728142735781148,
      "clip_ratio/low_mean": 0.00036021503183292225,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004088357731234282,
      "entropy": 0.3930057920515537,
      "epoch": 0.18013245033112582,
      "grad_norm": 15.652580261230469,
      "learning_rate": 6.559087204563977e-07,
      "loss": 0.9199,
      "reward": 3.1900382041931152,
      "reward_std": 1.5990674495697021,
      "rewards/reasoning_density_reward/mean": 0.4084325432777405,
      "rewards/reasoning_density_reward/std": 0.08287572115659714,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.53160560131073,
      "rewards/trajectory_similarity_reward/std": 0.4103636145591736,
      "step": 221
    },
    {
      "clip_ratio/high_max": 0.012222428747918457,
      "clip_ratio/high_mean": 0.0030585447602788918,
      "clip_ratio/low_mean": 0.00022864471975481138,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032871894509298727,
      "entropy": 0.37092048674821854,
      "epoch": 0.18094752929190017,
      "grad_norm": 14.081640243530273,
      "learning_rate": 6.552567237163814e-07,
      "loss": 2.6478,
      "reward": 1.3879172801971436,
      "reward_std": 1.4824714660644531,
      "rewards/reasoning_density_reward/mean": 0.3073412775993347,
      "rewards/reasoning_density_reward/std": 0.10118284821510315,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.1430760622024536,
      "rewards/trajectory_similarity_reward/std": 0.3141404688358307,
      "step": 222
    },
    {
      "clip_ratio/high_max": 0.006371705327183008,
      "clip_ratio/high_mean": 0.0014425399058382027,
      "clip_ratio/low_mean": 0.0012727065077342559,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002715246402658522,
      "entropy": 0.38143138960003853,
      "epoch": 0.18176260825267448,
      "grad_norm": 14.291335105895996,
      "learning_rate": 6.546047269763651e-07,
      "loss": -0.0215,
      "reward": 2.8531885147094727,
      "reward_std": 1.887416124343872,
      "rewards/reasoning_density_reward/mean": 0.3281994163990021,
      "rewards/reasoning_density_reward/std": 0.13967469334602356,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.5562390685081482,
      "rewards/trajectory_similarity_reward/std": 0.45142126083374023,
      "step": 223
    },
    {
      "clip_ratio/high_max": 0.012094191915821284,
      "clip_ratio/high_mean": 0.0024704370443942025,
      "clip_ratio/low_mean": 0.0006965759494050872,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031670129828853533,
      "entropy": 0.39264532923698425,
      "epoch": 0.1825776872134488,
      "grad_norm": 12.152828216552734,
      "learning_rate": 6.539527302363488e-07,
      "loss": 1.924,
      "reward": 2.8079452514648438,
      "reward_std": 1.6690988540649414,
      "rewards/reasoning_density_reward/mean": 0.3211681544780731,
      "rewards/reasoning_density_reward/std": 0.13742293417453766,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.42427709698677063,
      "rewards/trajectory_similarity_reward/std": 0.39813002943992615,
      "step": 224
    },
    {
      "clip_ratio/high_max": 0.010098339582327753,
      "clip_ratio/high_mean": 0.001653487779549323,
      "clip_ratio/low_mean": 0.0004273486920283176,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020808365115954075,
      "entropy": 0.36625609919428825,
      "epoch": 0.18339276617422312,
      "grad_norm": 17.476276397705078,
      "learning_rate": 6.533007334963325e-07,
      "loss": -0.0036,
      "reward": 2.7217838764190674,
      "reward_std": 1.61286461353302,
      "rewards/reasoning_density_reward/mean": 0.3460937440395355,
      "rewards/reasoning_density_reward/std": 0.15971118211746216,
      "rewards/success_reward/mean": 1.921875,
      "rewards/success_reward/std": 1.3521251678466797,
      "rewards/trajectory_similarity_reward/mean": 0.45381516218185425,
      "rewards/trajectory_similarity_reward/std": 0.42445193231105804,
      "step": 225
    },
    {
      "clip_ratio/high_max": 0.009167157026240602,
      "clip_ratio/high_mean": 0.002038442195043899,
      "clip_ratio/low_mean": 0.0005591613044089172,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00259760348853888,
      "entropy": 0.39873121306300163,
      "epoch": 0.18420784513499747,
      "grad_norm": 14.546492576599121,
      "learning_rate": 6.526487367563162e-07,
      "loss": 3.8922,
      "reward": 1.994354248046875,
      "reward_std": 1.7774159908294678,
      "rewards/reasoning_density_reward/mean": 0.24071969091892242,
      "rewards/reasoning_density_reward/std": 0.13764844834804535,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.34738460183143616,
      "rewards/trajectory_similarity_reward/std": 0.4651830792427063,
      "step": 226
    },
    {
      "clip_ratio/high_max": 0.00999785796739161,
      "clip_ratio/high_mean": 0.0019937734468840063,
      "clip_ratio/low_mean": 0.0007984886215126608,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027922620938625187,
      "entropy": 0.4121651351451874,
      "epoch": 0.18502292409577178,
      "grad_norm": 9.397478103637695,
      "learning_rate": 6.519967400162999e-07,
      "loss": 23.1138,
      "step": 227
    },
    {
      "clip_ratio/high_max": 0.013505902548786253,
      "clip_ratio/high_mean": 0.0030738007844774984,
      "clip_ratio/low_mean": 0.0009611972564016469,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004034997982671484,
      "entropy": 0.4063638709485531,
      "epoch": 0.1858380030565461,
      "grad_norm": 13.610528945922852,
      "learning_rate": 6.513447432762836e-07,
      "loss": -6.0208,
      "reward": 2.9703140258789062,
      "reward_std": 1.4704012274742126,
      "rewards/reasoning_density_reward/mean": 0.37890625,
      "rewards/reasoning_density_reward/std": 0.10974327847361565,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.086819052696228,
      "rewards/trajectory_similarity_reward/mean": 0.43515780568122864,
      "rewards/trajectory_similarity_reward/std": 0.406582772731781,
      "step": 228
    },
    {
      "clip_ratio/high_max": 0.007635615882463753,
      "clip_ratio/high_mean": 0.0015121089236345142,
      "clip_ratio/low_mean": 0.0011669589293887839,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002679067889403086,
      "entropy": 0.40294381231069565,
      "epoch": 0.18665308201732042,
      "grad_norm": 16.683692932128906,
      "learning_rate": 6.506927465362673e-07,
      "loss": -3.0978,
      "reward": 1.7794148921966553,
      "reward_std": 1.9448235034942627,
      "rewards/reasoning_density_reward/mean": 0.24270835518836975,
      "rewards/reasoning_density_reward/std": 0.17543266713619232,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.31795650720596313,
      "rewards/trajectory_similarity_reward/std": 0.4307112693786621,
      "step": 229
    },
    {
      "clip_ratio/high_max": 0.0051728085381910205,
      "clip_ratio/high_mean": 0.0009047365310834721,
      "clip_ratio/low_mean": 0.000672769430821063,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015775059473526198,
      "entropy": 0.38489867374300957,
      "epoch": 0.18746816097809477,
      "grad_norm": 9.60844612121582,
      "learning_rate": 6.50040749796251e-07,
      "loss": 8.2065,
      "reward": 1.3076897859573364,
      "reward_std": 1.1600582599639893,
      "rewards/reasoning_density_reward/mean": 0.26884472370147705,
      "rewards/reasoning_density_reward/std": 0.0835762694478035,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.10134509205818176,
      "rewards/trajectory_similarity_reward/std": 0.2854185104370117,
      "step": 230
    },
    {
      "clip_ratio/high_max": 0.009395514847710729,
      "clip_ratio/high_mean": 0.0018210964917670935,
      "clip_ratio/low_mean": 0.0008296370069729164,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002650733571499586,
      "entropy": 0.38490836694836617,
      "epoch": 0.18828323993886908,
      "grad_norm": 14.869723320007324,
      "learning_rate": 6.493887530562347e-07,
      "loss": 10.332,
      "reward": 3.2803359031677246,
      "reward_std": 1.7180216312408447,
      "rewards/reasoning_density_reward/mean": 0.4080356955528259,
      "rewards/reasoning_density_reward/std": 0.11061439663171768,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6223001480102539,
      "rewards/trajectory_similarity_reward/std": 0.4115840196609497,
      "step": 231
    },
    {
      "clip_ratio/high_max": 0.010358628700487316,
      "clip_ratio/high_mean": 0.002906698442529887,
      "clip_ratio/low_mean": 0.00017182259034598246,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030785209964960814,
      "entropy": 0.4104515612125397,
      "epoch": 0.1890983188996434,
      "grad_norm": 12.61266803741455,
      "learning_rate": 6.487367563162184e-07,
      "loss": -1.9584,
      "reward": 3.2918968200683594,
      "reward_std": 1.6508336067199707,
      "rewards/reasoning_density_reward/mean": 0.3350198566913605,
      "rewards/reasoning_density_reward/std": 0.10487443208694458,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.6131269931793213,
      "rewards/trajectory_similarity_reward/std": 0.3983398973941803,
      "step": 232
    },
    {
      "clip_ratio/high_max": 0.00693597667850554,
      "clip_ratio/high_mean": 0.001628362078918144,
      "clip_ratio/low_mean": 0.0005245455613476224,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002152907669369597,
      "entropy": 0.3938688822090626,
      "epoch": 0.18991339786041772,
      "grad_norm": 14.865079879760742,
      "learning_rate": 6.480847595762021e-07,
      "loss": 7.9866,
      "reward": 2.9155752658843994,
      "reward_std": 1.8709650039672852,
      "rewards/reasoning_density_reward/mean": 0.40625,
      "rewards/reasoning_density_reward/std": 0.09141378849744797,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.44682517647743225,
      "rewards/trajectory_similarity_reward/std": 0.3650040626525879,
      "step": 233
    },
    {
      "clip_ratio/high_max": 0.011368985287845135,
      "clip_ratio/high_mean": 0.0035768255620496348,
      "clip_ratio/low_mean": 0.0009256395205738954,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004502465162659064,
      "entropy": 0.4352513924241066,
      "epoch": 0.19072847682119207,
      "grad_norm": 13.13321304321289,
      "learning_rate": 6.474327628361858e-07,
      "loss": -11.2337,
      "reward": 3.5938820838928223,
      "reward_std": 1.2085644006729126,
      "rewards/reasoning_density_reward/mean": 0.36536458134651184,
      "rewards/reasoning_density_reward/std": 0.14298179745674133,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.866025447845459,
      "rewards/trajectory_similarity_reward/mean": 0.603517472743988,
      "rewards/trajectory_similarity_reward/std": 0.3656633794307709,
      "step": 234
    },
    {
      "clip_ratio/high_max": 0.013846476242179051,
      "clip_ratio/high_mean": 0.003398128264961997,
      "clip_ratio/low_mean": 0.0006310669232334476,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004029195173643529,
      "entropy": 0.4285498522222042,
      "epoch": 0.19154355578196638,
      "grad_norm": 15.383270263671875,
      "learning_rate": 6.467807660961695e-07,
      "loss": 5.7166,
      "reward": 2.781540811061859,
      "reward_std": 1.5517498254776,
      "rewards/reasoning_density_reward/mean": 0.36015626788139343,
      "rewards/reasoning_density_reward/std": 0.11944442242383957,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.0968086123466492,
      "rewards/trajectory_similarity_reward/mean": 0.4526344835758209,
      "rewards/trajectory_similarity_reward/std": 0.39843443036079407,
      "step": 235
    },
    {
      "clip_ratio/high_max": 0.01194451143965125,
      "clip_ratio/high_mean": 0.002139831609383691,
      "clip_ratio/low_mean": 0.0006768727798771579,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028167043637949973,
      "entropy": 0.4164457209408283,
      "epoch": 0.1923586347427407,
      "grad_norm": 13.479515075683594,
      "learning_rate": 6.461287693561532e-07,
      "loss": 8.6293,
      "reward": 3.7240207195281982,
      "reward_std": 0.6454169750213623,
      "rewards/reasoning_density_reward/mean": 0.3619047701358795,
      "rewards/reasoning_density_reward/std": 0.1382777988910675,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.6433659195899963,
      "rewards/trajectory_similarity_reward/std": 0.346780925989151,
      "step": 236
    },
    {
      "clip_ratio/high_max": 0.0173221692093648,
      "clip_ratio/high_mean": 0.003448900140938349,
      "clip_ratio/low_mean": 0.00011623475438682362,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035651349026011303,
      "entropy": 0.4177265986800194,
      "epoch": 0.19317371370351502,
      "grad_norm": 12.927679061889648,
      "learning_rate": 6.45476772616137e-07,
      "loss": 1.0068,
      "reward": 2.3518459796905518,
      "reward_std": 1.7939839363098145,
      "rewards/reasoning_density_reward/mean": 0.2972284406423569,
      "rewards/reasoning_density_reward/std": 0.1320994272828102,
      "rewards/success_reward/mean": 1.640625,
      "rewards/success_reward/std": 1.3285260796546936,
      "rewards/trajectory_similarity_reward/mean": 0.4139925539493561,
      "rewards/trajectory_similarity_reward/std": 0.38472503423690796,
      "step": 237
    },
    {
      "clip_ratio/high_max": 0.0052727198344655335,
      "clip_ratio/high_mean": 0.0010315318431821652,
      "clip_ratio/low_mean": 0.0008490989202982746,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018806307634804398,
      "entropy": 0.42455385997891426,
      "epoch": 0.19398879266428937,
      "grad_norm": 13.2701416015625,
      "learning_rate": 6.448247758761206e-07,
      "loss": 9.1177,
      "reward": 1.0100371837615967,
      "reward_std": 1.659386396408081,
      "rewards/reasoning_density_reward/mean": 0.26770836114883423,
      "rewards/reasoning_density_reward/std": 0.11012934893369675,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.17982882261276245,
      "rewards/trajectory_similarity_reward/std": 0.3874852955341339,
      "step": 238
    },
    {
      "clip_ratio/high_max": 0.009380632604006678,
      "clip_ratio/high_mean": 0.002181718358770013,
      "clip_ratio/low_mean": 0.0009069379884749651,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030886563326930627,
      "entropy": 0.4176117479801178,
      "epoch": 0.19480387162506368,
      "grad_norm": 10.863207817077637,
      "learning_rate": 6.441727791361044e-07,
      "loss": 0.9643,
      "reward": 1.596393346786499,
      "reward_std": 0.9265512824058533,
      "rewards/reasoning_density_reward/mean": 0.21547618508338928,
      "rewards/reasoning_density_reward/std": 0.14064760506153107,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.25591713190078735,
      "rewards/trajectory_similarity_reward/std": 0.353145956993103,
      "step": 239
    },
    {
      "clip_ratio/high_max": 0.013814736856147647,
      "clip_ratio/high_mean": 0.0035625454620458186,
      "clip_ratio/low_mean": 0.0002153197565348819,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037778652767883614,
      "entropy": 0.410083394497633,
      "epoch": 0.195618950585838,
      "grad_norm": 13.523261070251465,
      "learning_rate": 6.43520782396088e-07,
      "loss": -7.9679,
      "reward": 3.864797353744507,
      "reward_std": 0.7482420802116394,
      "rewards/reasoning_density_reward/mean": 0.4016842544078827,
      "rewards/reasoning_density_reward/std": 0.1447744444012642,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8080127239227295,
      "rewards/trajectory_similarity_reward/mean": 0.7443631887435913,
      "rewards/trajectory_similarity_reward/std": 0.3289675861597061,
      "step": 240
    },
    {
      "clip_ratio/high_max": 0.014569149119779468,
      "clip_ratio/high_mean": 0.003236983495298773,
      "clip_ratio/low_mean": 0.0010535322144278325,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004290515702450648,
      "entropy": 0.42046212404966354,
      "epoch": 0.19643402954661232,
      "grad_norm": 13.655818939208984,
      "learning_rate": 6.428687856560717e-07,
      "loss": 10.855,
      "reward": 2.7408835887908936,
      "reward_std": 2.0738120079040527,
      "rewards/reasoning_density_reward/mean": 0.35141369700431824,
      "rewards/reasoning_density_reward/std": 0.12365009635686874,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5144698023796082,
      "rewards/trajectory_similarity_reward/std": 0.42269301414489746,
      "step": 241
    },
    {
      "clip_ratio/high_max": 0.010561322735156864,
      "clip_ratio/high_mean": 0.002584963323897682,
      "clip_ratio/low_mean": 0.000676315430609975,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032612788199912757,
      "entropy": 0.4049708843231201,
      "epoch": 0.19724910850738667,
      "grad_norm": 18.01369285583496,
      "learning_rate": 6.422167889160554e-07,
      "loss": 3.0251,
      "reward": 2.062619686126709,
      "reward_std": 1.1935773491859436,
      "rewards/reasoning_density_reward/mean": 0.38098958134651184,
      "rewards/reasoning_density_reward/std": 0.09819750860333443,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.3691300302743912,
      "rewards/trajectory_similarity_reward/std": 0.4373227655887604,
      "step": 242
    },
    {
      "clip_ratio/high_max": 0.009559087979141623,
      "clip_ratio/high_mean": 0.0020452215903787874,
      "clip_ratio/low_mean": 0.0005620676110993372,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026072891705553047,
      "entropy": 0.3816552720963955,
      "epoch": 0.19806418746816098,
      "grad_norm": 12.347007751464844,
      "learning_rate": 6.415647921760391e-07,
      "loss": -0.4137,
      "reward": 2.6522011756896973,
      "reward_std": 1.8399415016174316,
      "rewards/reasoning_density_reward/mean": 0.3812499940395355,
      "rewards/reasoning_density_reward/std": 0.1089724749326706,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.48970112204551697,
      "rewards/trajectory_similarity_reward/std": 0.4518347382545471,
      "step": 243
    },
    {
      "clip_ratio/high_max": 0.006648323673289269,
      "clip_ratio/high_mean": 0.0015416580499731936,
      "clip_ratio/low_mean": 0.0002579447191237705,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001799602796381805,
      "entropy": 0.39135653898119926,
      "epoch": 0.1988792664289353,
      "grad_norm": 13.150405883789062,
      "learning_rate": 6.409127954360227e-07,
      "loss": 7.9924,
      "reward": 3.222844958305359,
      "reward_std": 1.3329188823699951,
      "rewards/reasoning_density_reward/mean": 0.37142856419086456,
      "rewards/reasoning_density_reward/std": 0.14460675790905952,
      "rewards/success_reward/mean": 2.203125,
      "rewards/success_reward/std": 1.3285260796546936,
      "rewards/trajectory_similarity_reward/mean": 0.6482913494110107,
      "rewards/trajectory_similarity_reward/std": 0.44265712797641754,
      "step": 244
    },
    {
      "clip_ratio/high_max": 0.013695553701836616,
      "clip_ratio/high_mean": 0.0027101641899207607,
      "clip_ratio/low_mean": 0.0005710109253413975,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032811751443659887,
      "entropy": 0.4359494373202324,
      "epoch": 0.19969434538970962,
      "grad_norm": 13.081802368164062,
      "learning_rate": 6.402607986960065e-07,
      "loss": 0.4088,
      "reward": 0.9641484022140503,
      "reward_std": 0.9646362662315369,
      "rewards/reasoning_density_reward/mean": 0.2566964328289032,
      "rewards/reasoning_density_reward/std": 0.09657091647386551,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.1449519395828247,
      "rewards/trajectory_similarity_reward/std": 0.3172120153903961,
      "step": 245
    },
    {
      "clip_ratio/high_max": 0.011092163535067812,
      "clip_ratio/high_mean": 0.0022917192000022624,
      "clip_ratio/low_mean": 0.00036080239442526363,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026525216089794412,
      "entropy": 0.3986702784895897,
      "epoch": 0.20050942435048397,
      "grad_norm": 11.217106819152832,
      "learning_rate": 6.396088019559901e-07,
      "loss": -12.3285,
      "reward": 2.6324806213378906,
      "reward_std": 1.6082005500793457,
      "rewards/reasoning_density_reward/mean": 0.3609375059604645,
      "rewards/reasoning_density_reward/std": 0.08545730262994766,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.3965432047843933,
      "rewards/trajectory_similarity_reward/std": 0.43671733140945435,
      "step": 246
    },
    {
      "clip_ratio/high_max": 0.010132297349628061,
      "clip_ratio/high_mean": 0.0021504518372239545,
      "clip_ratio/low_mean": 0.0004045047389809042,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025549565471010283,
      "entropy": 0.4159611873328686,
      "epoch": 0.20132450331125828,
      "grad_norm": 12.034815788269043,
      "learning_rate": 6.389568052159739e-07,
      "loss": 1.6343,
      "reward": 2.64287531375885,
      "reward_std": 1.7719078063964844,
      "rewards/reasoning_density_reward/mean": 0.35412946343421936,
      "rewards/reasoning_density_reward/std": 0.1628861129283905,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4680703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.5074959099292755,
      "rewards/trajectory_similarity_reward/std": 0.4584702253341675,
      "step": 247
    },
    {
      "clip_ratio/high_max": 0.011270294256974012,
      "clip_ratio/high_mean": 0.0023086913643055595,
      "clip_ratio/low_mean": 0.00132193191893748,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003630623235949315,
      "entropy": 0.38137463480234146,
      "epoch": 0.2021395822720326,
      "grad_norm": 14.848511695861816,
      "learning_rate": 6.383048084759575e-07,
      "loss": 0.7857,
      "reward": 1.931546688079834,
      "reward_std": 1.9220554828643799,
      "rewards/reasoning_density_reward/mean": 0.3120535612106323,
      "rewards/reasoning_density_reward/std": 0.10943222790956497,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.3069932460784912,
      "rewards/trajectory_similarity_reward/std": 0.3924520015716553,
      "step": 248
    },
    {
      "clip_ratio/high_max": 0.009004070423543453,
      "clip_ratio/high_mean": 0.0016586939091212116,
      "clip_ratio/low_mean": 0.0005548018416448031,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002213495710748248,
      "entropy": 0.4058794863522053,
      "epoch": 0.20295466123280692,
      "grad_norm": 11.87486457824707,
      "learning_rate": 6.376528117359413e-07,
      "loss": 0.2858,
      "reward": 3.182952642440796,
      "reward_std": 1.0652459263801575,
      "rewards/reasoning_density_reward/mean": 0.34296876192092896,
      "rewards/reasoning_density_reward/std": 0.18627507239580154,
      "rewards/success_reward/mean": 2.203125,
      "rewards/success_reward/std": 1.3285260796546936,
      "rewards/trajectory_similarity_reward/mean": 0.6368588209152222,
      "rewards/trajectory_similarity_reward/std": 0.4269498288631439,
      "step": 249
    },
    {
      "clip_ratio/high_max": 0.013659416465088725,
      "clip_ratio/high_mean": 0.003064157295739278,
      "clip_ratio/low_mean": 0.0010517538248677738,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004115911215194501,
      "entropy": 0.37252670899033546,
      "epoch": 0.20376974019358127,
      "grad_norm": 13.9088134765625,
      "learning_rate": 6.370008149959249e-07,
      "loss": -2.2204,
      "reward": 2.819491744041443,
      "reward_std": 1.2076434791088104,
      "rewards/reasoning_density_reward/mean": 0.36510418355464935,
      "rewards/reasoning_density_reward/std": 0.1164381355047226,
      "rewards/success_reward/mean": 1.921875,
      "rewards/success_reward/std": 1.3531213998794556,
      "rewards/trajectory_similarity_reward/mean": 0.5325125455856323,
      "rewards/trajectory_similarity_reward/std": 0.48636186122894287,
      "step": 250
    },
    {
      "clip_ratio/high_max": 0.016173741285456344,
      "clip_ratio/high_mean": 0.0037775480705022346,
      "clip_ratio/low_mean": 0.0004116906857234426,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004189238774415571,
      "entropy": 0.3975868374109268,
      "epoch": 0.20458481915435558,
      "grad_norm": 10.857383728027344,
      "learning_rate": 6.363488182559087e-07,
      "loss": 1.2662,
      "reward": 3.430400848388672,
      "reward_std": 1.1357234716415405,
      "rewards/reasoning_density_reward/mean": 0.3266369104385376,
      "rewards/reasoning_density_reward/std": 0.17764781415462494,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.5725136995315552,
      "rewards/trajectory_similarity_reward/std": 0.3232729732990265,
      "step": 251
    },
    {
      "clip_ratio/high_max": 0.008080984407570213,
      "clip_ratio/high_mean": 0.0018908181373262778,
      "clip_ratio/low_mean": 0.00046008831486687995,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002350906463107094,
      "entropy": 0.40998631343245506,
      "epoch": 0.2053998981151299,
      "grad_norm": 14.288591384887695,
      "learning_rate": 6.356968215158924e-07,
      "loss": 1.5727,
      "reward": 1.441636711359024,
      "reward_std": 1.0073172450065613,
      "rewards/reasoning_density_reward/mean": 0.2718750089406967,
      "rewards/reasoning_density_reward/std": 0.1371154859662056,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 0.8704394698143005,
      "rewards/trajectory_similarity_reward/mean": 0.23226164281368256,
      "rewards/trajectory_similarity_reward/std": 0.24273718893527985,
      "step": 252
    },
    {
      "clip_ratio/high_max": 0.00899822055362165,
      "clip_ratio/high_mean": 0.0016372175741707906,
      "clip_ratio/low_mean": 0.0015109099294932093,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003148127463646233,
      "entropy": 0.39359235391020775,
      "epoch": 0.20621497707590422,
      "grad_norm": 11.03714370727539,
      "learning_rate": 6.350448247758761e-07,
      "loss": 0.5686,
      "reward": 2.531320810317993,
      "reward_std": 1.9766807556152344,
      "rewards/reasoning_density_reward/mean": 0.2987847328186035,
      "rewards/reasoning_density_reward/std": 0.1078072115778923,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.5450360178947449,
      "rewards/trajectory_similarity_reward/std": 0.49834954738616943,
      "step": 253
    },
    {
      "clip_ratio/high_max": 0.013172810897231102,
      "clip_ratio/high_mean": 0.00339465364231728,
      "clip_ratio/low_mean": 0.0007895602684584446,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004184213903499767,
      "entropy": 0.42481959238648415,
      "epoch": 0.20703005603667857,
      "grad_norm": 15.519428253173828,
      "learning_rate": 6.343928280358598e-07,
      "loss": -6.8082,
      "reward": 3.7607710361480713,
      "reward_std": 0.6535253524780273,
      "rewards/reasoning_density_reward/mean": 0.3999999761581421,
      "rewards/reasoning_density_reward/std": 0.15705625712871552,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.5123475790023804,
      "rewards/trajectory_similarity_reward/mean": 0.5482709407806396,
      "rewards/trajectory_similarity_reward/std": 0.35716041922569275,
      "step": 254
    },
    {
      "clip_ratio/high_max": 0.011681167816277593,
      "clip_ratio/high_mean": 0.0028185049450257793,
      "clip_ratio/low_mean": 0.0006241229602892417,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034426279307808727,
      "entropy": 0.4004945270717144,
      "epoch": 0.20784513499745289,
      "grad_norm": 14.374567985534668,
      "learning_rate": 6.337408312958435e-07,
      "loss": 11.4752,
      "reward": 3.108481526374817,
      "reward_std": 1.4679675698280334,
      "rewards/reasoning_density_reward/mean": 0.36564503610134125,
      "rewards/reasoning_density_reward/std": 0.14141949266195297,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.0823204219341278,
      "rewards/trajectory_similarity_reward/mean": 0.492836594581604,
      "rewards/trajectory_similarity_reward/std": 0.4088551700115204,
      "step": 255
    },
    {
      "clip_ratio/high_max": 0.01052450283896178,
      "clip_ratio/high_mean": 0.0024318114010384306,
      "clip_ratio/low_mean": 0.0005288564971124288,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029606679308926687,
      "entropy": 0.41581089049577713,
      "epoch": 0.2086602139582272,
      "grad_norm": 14.676612854003906,
      "learning_rate": 6.330888345558272e-07,
      "loss": -8.3973,
      "reward": 2.305340051651001,
      "reward_std": 1.4856823682785034,
      "rewards/reasoning_density_reward/mean": 0.3251302242279053,
      "rewards/reasoning_density_reward/std": 0.08947733044624329,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.3227396607398987,
      "rewards/trajectory_similarity_reward/mean": 0.29270995408296585,
      "rewards/trajectory_similarity_reward/std": 0.39942459762096405,
      "step": 256
    },
    {
      "clip_ratio/high_max": 0.010102792060934007,
      "clip_ratio/high_mean": 0.0022367154888343066,
      "clip_ratio/low_mean": 0.0011109371625934727,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003347652542288415,
      "entropy": 0.38862959668040276,
      "epoch": 0.20947529291900152,
      "grad_norm": 15.569012641906738,
      "learning_rate": 6.324368378158109e-07,
      "loss": -4.9997,
      "reward": 2.209674119949341,
      "reward_std": 1.9518542289733887,
      "rewards/reasoning_density_reward/mean": 0.31388890743255615,
      "rewards/reasoning_density_reward/std": 0.06268799304962158,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4491376876831055,
      "rewards/trajectory_similarity_reward/mean": 0.39578524231910706,
      "rewards/trajectory_similarity_reward/std": 0.4647591710090637,
      "step": 257
    },
    {
      "clip_ratio/high_max": 0.011286453809589148,
      "clip_ratio/high_mean": 0.0029122793639544398,
      "clip_ratio/low_mean": 0.0011571487557375804,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004069428134243935,
      "entropy": 0.40321894362568855,
      "epoch": 0.21029037187977587,
      "grad_norm": 15.145513534545898,
      "learning_rate": 6.317848410757946e-07,
      "loss": -0.5451,
      "reward": 3.963512420654297,
      "reward_std": 0.8960815072059631,
      "rewards/reasoning_density_reward/mean": 0.4333333373069763,
      "rewards/reasoning_density_reward/std": 0.12692955136299133,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8158584833145142,
      "rewards/trajectory_similarity_reward/mean": 0.8114289045333862,
      "rewards/trajectory_similarity_reward/std": 0.3243669271469116,
      "step": 258
    },
    {
      "clip_ratio/high_max": 0.011351597087923437,
      "clip_ratio/high_mean": 0.0029546376317739487,
      "clip_ratio/low_mean": 0.0006916242564329877,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036462618882069364,
      "entropy": 0.4222741015255451,
      "epoch": 0.21110545084055019,
      "grad_norm": 18.833602905273438,
      "learning_rate": 6.311328443357783e-07,
      "loss": 4.1144,
      "reward": 3.2108631134033203,
      "reward_std": 1.6871953010559082,
      "rewards/reasoning_density_reward/mean": 0.3604166656732559,
      "rewards/reasoning_density_reward/std": 0.14331112056970596,
      "rewards/success_reward/mean": 2.203125,
      "rewards/success_reward/std": 1.274251639842987,
      "rewards/trajectory_similarity_reward/mean": 0.6473212540149689,
      "rewards/trajectory_similarity_reward/std": 0.4501459300518036,
      "step": 259
    },
    {
      "clip_ratio/high_max": 0.01453901524655521,
      "clip_ratio/high_mean": 0.003565959501429461,
      "clip_ratio/low_mean": 0.0006262350907491054,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004192194566712715,
      "entropy": 0.3942057490348816,
      "epoch": 0.2119205298013245,
      "grad_norm": 12.496442794799805,
      "learning_rate": 6.30480847595762e-07,
      "loss": -4.1707,
      "reward": 3.1872040033340454,
      "reward_std": 1.7330840826034546,
      "rewards/reasoning_density_reward/mean": 0.39084821939468384,
      "rewards/reasoning_density_reward/std": 0.12055629119277,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.6401056945323944,
      "rewards/trajectory_similarity_reward/std": 0.452044740319252,
      "step": 260
    },
    {
      "clip_ratio/high_max": 0.01722300541587174,
      "clip_ratio/high_mean": 0.003714402948389761,
      "clip_ratio/low_mean": 0.0006399675330612808,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004354370539658703,
      "entropy": 0.3961462937295437,
      "epoch": 0.21273560876209882,
      "grad_norm": 13.314759254455566,
      "learning_rate": 6.298288508557457e-07,
      "loss": 6.0322,
      "reward": 2.5700550079345703,
      "reward_std": 1.6848974227905273,
      "rewards/reasoning_density_reward/mean": 0.2964285612106323,
      "rewards/reasoning_density_reward/std": 0.15457864105701447,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.1618950366973877,
      "rewards/trajectory_similarity_reward/mean": 0.3986263871192932,
      "rewards/trajectory_similarity_reward/std": 0.46705102920532227,
      "step": 261
    },
    {
      "clip_ratio/high_max": 0.010651822260115296,
      "clip_ratio/high_mean": 0.002439945856167469,
      "clip_ratio/low_mean": 0.00043712057049560826,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028770663629984483,
      "entropy": 0.4312555119395256,
      "epoch": 0.21355068772287317,
      "grad_norm": 14.543289184570312,
      "learning_rate": 6.291768541157294e-07,
      "loss": 1.7595,
      "reward": 3.0402350425720215,
      "reward_std": 1.449971854686737,
      "rewards/reasoning_density_reward/mean": 0.3706101179122925,
      "rewards/reasoning_density_reward/std": 0.15331882238388062,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.2004882097244263,
      "rewards/trajectory_similarity_reward/mean": 0.5133747309446335,
      "rewards/trajectory_similarity_reward/std": 0.418303981423378,
      "step": 262
    },
    {
      "clip_ratio/high_max": 0.010661621432518587,
      "clip_ratio/high_mean": 0.002359302270633634,
      "clip_ratio/low_mean": 0.0010165010608034208,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003375803331437055,
      "entropy": 0.43274424225091934,
      "epoch": 0.21436576668364749,
      "grad_norm": 13.399528503417969,
      "learning_rate": 6.285248573757131e-07,
      "loss": 1.801,
      "reward": 2.5425074100494385,
      "reward_std": 1.7571699619293213,
      "rewards/reasoning_density_reward/mean": 0.2633184567093849,
      "rewards/reasoning_density_reward/std": 0.15392501652240753,
      "rewards/success_reward/mean": 1.828125,
      "rewards/success_reward/std": 1.3756247162818909,
      "rewards/trajectory_similarity_reward/mean": 0.45106394588947296,
      "rewards/trajectory_similarity_reward/std": 0.4316675513982773,
      "step": 263
    },
    {
      "clip_ratio/high_max": 0.010394834709586576,
      "clip_ratio/high_mean": 0.0025793240092752967,
      "clip_ratio/low_mean": 0.0011563236475922167,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037356475077103823,
      "entropy": 0.4053838290274143,
      "epoch": 0.2151808456444218,
      "grad_norm": 10.182088851928711,
      "learning_rate": 6.278728606356968e-07,
      "loss": -4.0835,
      "reward": 2.1285001039505005,
      "reward_std": 1.9552958011627197,
      "rewards/reasoning_density_reward/mean": 0.26983901858329773,
      "rewards/reasoning_density_reward/std": 0.15662319958209991,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4491376876831055,
      "rewards/trajectory_similarity_reward/mean": 0.35866107046604156,
      "rewards/trajectory_similarity_reward/std": 0.44063811004161835,
      "step": 264
    },
    {
      "clip_ratio/high_max": 0.0104400206473656,
      "clip_ratio/high_mean": 0.0023969449248397723,
      "clip_ratio/low_mean": 0.0011056714138248935,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003502616338664666,
      "entropy": 0.40941641479730606,
      "epoch": 0.21599592460519612,
      "grad_norm": 11.891151428222656,
      "learning_rate": 6.272208638956805e-07,
      "loss": 4.8456,
      "reward": 2.0756850242614746,
      "reward_std": 0.31797826290130615,
      "rewards/reasoning_density_reward/mean": 0.24379006028175354,
      "rewards/reasoning_density_reward/std": 0.17064960300922394,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.33189478516578674,
      "rewards/trajectory_similarity_reward/std": 0.42770156264305115,
      "step": 265
    },
    {
      "clip_ratio/high_max": 0.014982658554799855,
      "clip_ratio/high_mean": 0.0025110462156590074,
      "clip_ratio/low_mean": 0.0010055664897663519,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003516612749081105,
      "entropy": 0.4429846741259098,
      "epoch": 0.21681100356597047,
      "grad_norm": 12.403980255126953,
      "learning_rate": 6.265688671556642e-07,
      "loss": 15.3494,
      "reward": 4.0957231521606445,
      "reward_std": 0.12689536809921265,
      "rewards/reasoning_density_reward/mean": 0.3020833134651184,
      "rewards/reasoning_density_reward/std": 0.21325303614139557,
      "rewards/success_reward/mean": 3.0,
      "rewards/success_reward/std": 0.0,
      "rewards/trajectory_similarity_reward/mean": 0.7936399579048157,
      "rewards/trajectory_similarity_reward/std": 0.20744206011295319,
      "step": 266
    },
    {
      "clip_ratio/high_max": 0.011364417616277933,
      "clip_ratio/high_mean": 0.002203137628384866,
      "clip_ratio/low_mean": 0.0010314388855476864,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003234576517570531,
      "entropy": 0.42707137018442154,
      "epoch": 0.21762608252674479,
      "grad_norm": 13.588887214660645,
      "learning_rate": 6.25916870415648e-07,
      "loss": 6.8664,
      "reward": 1.7561883926391602,
      "reward_std": 1.7961602210998535,
      "rewards/reasoning_density_reward/mean": 0.3432021141052246,
      "rewards/reasoning_density_reward/std": 0.10002037510275841,
      "rewards/success_reward/mean": 1.078125,
      "rewards/success_reward/std": 1.4538734555244446,
      "rewards/trajectory_similarity_reward/mean": 0.33486124873161316,
      "rewards/trajectory_similarity_reward/std": 0.47763726115226746,
      "step": 267
    },
    {
      "clip_ratio/high_max": 0.006165665807202458,
      "clip_ratio/high_mean": 0.001303018958424218,
      "clip_ratio/low_mean": 0.0013125493496772833,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026155682862736285,
      "entropy": 0.4300371743738651,
      "epoch": 0.2184411614875191,
      "grad_norm": 16.457189559936523,
      "learning_rate": 6.252648736756316e-07,
      "loss": -5.0693,
      "reward": 0.6722532212734222,
      "reward_std": 1.0835991203784943,
      "rewards/reasoning_density_reward/mean": 0.30026043951511383,
      "rewards/reasoning_density_reward/std": 0.05319434590637684,
      "rewards/success_reward/mean": 0.28125,
      "rewards/success_reward/std": 0.8873475790023804,
      "rewards/trajectory_similarity_reward/mean": 0.09074278362095356,
      "rewards/trajectory_similarity_reward/std": 0.2837536558508873,
      "step": 268
    },
    {
      "clip_ratio/high_max": 0.009980877104680985,
      "clip_ratio/high_mean": 0.0018044895405182615,
      "clip_ratio/low_mean": 0.0011070916370954365,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002911581104854122,
      "entropy": 0.4083787202835083,
      "epoch": 0.21925624044829342,
      "grad_norm": 14.341232299804688,
      "learning_rate": 6.246128769356154e-07,
      "loss": -2.7997,
      "reward": 2.369272470474243,
      "reward_std": 0.7356641292572021,
      "rewards/reasoning_density_reward/mean": 0.3049563616514206,
      "rewards/reasoning_density_reward/std": 0.18743447214365005,
      "rewards/success_reward/mean": 1.734375,
      "rewards/success_reward/std": 1.5043244361877441,
      "rewards/trajectory_similarity_reward/mean": 0.3299409747123718,
      "rewards/trajectory_similarity_reward/std": 0.33387456834316254,
      "step": 269
    },
    {
      "clip_ratio/high_max": 0.012408163282088935,
      "clip_ratio/high_mean": 0.0023639827195438556,
      "clip_ratio/low_mean": 0.0017213107712450437,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040852934325812384,
      "entropy": 0.40629469230771065,
      "epoch": 0.22007131940906777,
      "grad_norm": 13.663467407226562,
      "learning_rate": 6.23960880195599e-07,
      "loss": -5.5226,
      "reward": 2.810218334197998,
      "reward_std": 2.0810813903808594,
      "rewards/reasoning_density_reward/mean": 0.3851190507411957,
      "rewards/reasoning_density_reward/std": 0.08904676884412766,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5500994920730591,
      "rewards/trajectory_similarity_reward/std": 0.45481136441230774,
      "step": 270
    },
    {
      "clip_ratio/high_max": 0.009874262643279508,
      "clip_ratio/high_mean": 0.002140089050953975,
      "clip_ratio/low_mean": 0.0006677069723082241,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028077960305381566,
      "entropy": 0.4196369796991348,
      "epoch": 0.22088639836984209,
      "grad_norm": 10.140426635742188,
      "learning_rate": 6.233088834555828e-07,
      "loss": 1.9051,
      "reward": 2.0900228023529053,
      "reward_std": 1.2897839546203613,
      "rewards/reasoning_density_reward/mean": 0.19947917759418488,
      "rewards/reasoning_density_reward/std": 0.19038142263889313,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.39054375886917114,
      "rewards/trajectory_similarity_reward/std": 0.42990827560424805,
      "step": 271
    },
    {
      "clip_ratio/high_max": 0.013791301986202598,
      "clip_ratio/high_mean": 0.0030748531644348986,
      "clip_ratio/low_mean": 0.0006890064469189383,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037638595968019217,
      "entropy": 0.3807368837296963,
      "epoch": 0.2217014773306164,
      "grad_norm": 13.71042251586914,
      "learning_rate": 6.226568867155664e-07,
      "loss": 2.5679,
      "reward": 3.103062152862549,
      "reward_std": 1.7741625905036926,
      "rewards/reasoning_density_reward/mean": 0.40937499701976776,
      "rewards/reasoning_density_reward/std": 0.10473335906863213,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.420820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.6311872005462646,
      "rewards/trajectory_similarity_reward/std": 0.44236643612384796,
      "step": 272
    },
    {
      "clip_ratio/high_max": 0.01168781885644421,
      "clip_ratio/high_mean": 0.0026675387925934047,
      "clip_ratio/low_mean": 0.001471423096518265,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0041389618563698605,
      "entropy": 0.3849131502211094,
      "epoch": 0.22251655629139072,
      "grad_norm": 13.996663093566895,
      "learning_rate": 6.220048899755502e-07,
      "loss": -18.9862,
      "reward": 2.1136767864227295,
      "reward_std": 1.8812334537506104,
      "rewards/reasoning_density_reward/mean": 0.30204614996910095,
      "rewards/reasoning_density_reward/std": 0.15198399871587753,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.4538734555244446,
      "rewards/trajectory_similarity_reward/mean": 0.35850562155246735,
      "rewards/trajectory_similarity_reward/std": 0.4356858730316162,
      "step": 273
    },
    {
      "clip_ratio/high_max": 0.01012801454635337,
      "clip_ratio/high_mean": 0.0023233340398292057,
      "clip_ratio/low_mean": 0.0008594746195740299,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031828087812755257,
      "entropy": 0.4153605177998543,
      "epoch": 0.22333163525216507,
      "grad_norm": 14.240212440490723,
      "learning_rate": 6.213528932355338e-07,
      "loss": 9.2881,
      "reward": 1.149020791053772,
      "reward_std": 0.9939507246017456,
      "rewards/reasoning_density_reward/mean": 0.2333536446094513,
      "rewards/reasoning_density_reward/std": 0.1190837174654007,
      "rewards/success_reward/mean": 0.75,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.16566729545593262,
      "rewards/trajectory_similarity_reward/std": 0.3587699234485626,
      "step": 274
    },
    {
      "clip_ratio/high_max": 0.011712910374626517,
      "clip_ratio/high_mean": 0.0027003423892892897,
      "clip_ratio/low_mean": 0.0006919467996340245,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033922892034752294,
      "entropy": 0.4385572671890259,
      "epoch": 0.22414671421293939,
      "grad_norm": 11.965219497680664,
      "learning_rate": 6.207008964955175e-07,
      "loss": -7.5353,
      "reward": 3.3626322746276855,
      "reward_std": 0.9962946176528931,
      "rewards/reasoning_density_reward/mean": 0.3671875,
      "rewards/reasoning_density_reward/std": 0.08435313403606415,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.4641948640346527,
      "rewards/trajectory_similarity_reward/std": 0.3665042817592621,
      "step": 275
    },
    {
      "clip_ratio/high_max": 0.011932325607631356,
      "clip_ratio/high_mean": 0.0028444286799640395,
      "clip_ratio/low_mean": 0.000696073351718951,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035405020171310753,
      "entropy": 0.3954444043338299,
      "epoch": 0.2249617931737137,
      "grad_norm": 12.508131980895996,
      "learning_rate": 6.200488997555012e-07,
      "loss": 6.8207,
      "reward": 1.9353843927383423,
      "reward_std": 1.6271611452102661,
      "rewards/reasoning_density_reward/mean": 0.24062499403953552,
      "rewards/reasoning_density_reward/std": 0.16224338114261627,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.2808688879013062,
      "rewards/trajectory_similarity_reward/mean": 0.28850939869880676,
      "rewards/trajectory_similarity_reward/std": 0.4427175521850586,
      "step": 276
    },
    {
      "clip_ratio/high_max": 0.007735809369478375,
      "clip_ratio/high_mean": 0.0015649773777113296,
      "clip_ratio/low_mean": 0.0008808184720692225,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024457958570565097,
      "entropy": 0.44895320385694504,
      "epoch": 0.22577687213448802,
      "grad_norm": 11.319671630859375,
      "learning_rate": 6.193969030154849e-07,
      "loss": -9.5084,
      "reward": 2.855733633041382,
      "reward_std": 0.8295034766197205,
      "rewards/reasoning_density_reward/mean": 0.2804044932126999,
      "rewards/reasoning_density_reward/std": 0.19075796753168106,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.2813819646835327,
      "rewards/trajectory_similarity_reward/mean": 0.5128292143344879,
      "rewards/trajectory_similarity_reward/std": 0.4420725554227829,
      "step": 277
    },
    {
      "clip_ratio/high_max": 0.010455312789417803,
      "clip_ratio/high_mean": 0.002131619621650316,
      "clip_ratio/low_mean": 0.0007683032308705151,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028999228088650852,
      "entropy": 0.41926442086696625,
      "epoch": 0.22659195109526237,
      "grad_norm": 11.521309852600098,
      "learning_rate": 6.187449062754685e-07,
      "loss": 4.0963,
      "reward": 3.092461347579956,
      "reward_std": 1.1559048891067505,
      "rewards/reasoning_density_reward/mean": 0.3791666626930237,
      "rewards/reasoning_density_reward/std": 0.0986013263463974,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 0.9437293410301208,
      "rewards/trajectory_similarity_reward/mean": 0.3695446848869324,
      "rewards/trajectory_similarity_reward/std": 0.4090125262737274,
      "step": 278
    },
    {
      "clip_ratio/high_max": 0.014569296094123274,
      "clip_ratio/high_mean": 0.003471217889455147,
      "clip_ratio/low_mean": 0.00077708578282909,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004248303681379184,
      "entropy": 0.3678163103759289,
      "epoch": 0.22740703005603669,
      "grad_norm": 13.395071983337402,
      "learning_rate": 6.180929095354523e-07,
      "loss": -1.4403,
      "reward": 2.5918803215026855,
      "reward_std": 1.6118385791778564,
      "rewards/reasoning_density_reward/mean": 0.36138394474983215,
      "rewards/reasoning_density_reward/std": 0.1094452515244484,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.44924628734588623,
      "rewards/trajectory_similarity_reward/std": 0.4294058084487915,
      "step": 279
    },
    {
      "clip_ratio/high_max": 0.012941763969138265,
      "clip_ratio/high_mean": 0.0031231652974383906,
      "clip_ratio/low_mean": 0.0006099864658608567,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037331517232814804,
      "entropy": 0.36453741043806076,
      "epoch": 0.228222109016811,
      "grad_norm": 14.528563499450684,
      "learning_rate": 6.174409127954359e-07,
      "loss": -2.7836,
      "reward": 3.73453688621521,
      "reward_std": 1.290205478668213,
      "rewards/reasoning_density_reward/mean": 0.38482141494750977,
      "rewards/reasoning_density_reward/std": 0.13472874462604523,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.866025447845459,
      "rewards/trajectory_similarity_reward/mean": 0.7247154116630554,
      "rewards/trajectory_similarity_reward/std": 0.3683973252773285,
      "step": 280
    },
    {
      "clip_ratio/high_max": 0.014351052814163268,
      "clip_ratio/high_mean": 0.003584952632081695,
      "clip_ratio/low_mean": 0.001154747631517239,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004739700234495103,
      "entropy": 0.3719675578176975,
      "epoch": 0.22903718797758532,
      "grad_norm": 14.193607330322266,
      "learning_rate": 6.167889160554197e-07,
      "loss": 7.291,
      "reward": 3.758821487426758,
      "reward_std": 1.237898588180542,
      "rewards/reasoning_density_reward/mean": 0.42109376192092896,
      "rewards/reasoning_density_reward/std": 0.11868570744991302,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.7538944780826569,
      "rewards/trajectory_similarity_reward/mean": 0.7127277255058289,
      "rewards/trajectory_similarity_reward/std": 0.4011605232954025,
      "step": 281
    },
    {
      "clip_ratio/high_max": 0.012483961938414723,
      "clip_ratio/high_mean": 0.0034996918839169666,
      "clip_ratio/low_mean": 0.0005180033222131897,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004017695231596008,
      "entropy": 0.42460431158542633,
      "epoch": 0.22985226693835967,
      "grad_norm": 17.649593353271484,
      "learning_rate": 6.161369193154034e-07,
      "loss": -15.071,
      "reward": 3.0906091928482056,
      "reward_std": 1.6789631247520447,
      "rewards/reasoning_density_reward/mean": 0.3791666775941849,
      "rewards/reasoning_density_reward/std": 0.14812391996383667,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.2373085021972656,
      "rewards/trajectory_similarity_reward/mean": 0.5551923513412476,
      "rewards/trajectory_similarity_reward/std": 0.42956310510635376,
      "step": 282
    },
    {
      "clip_ratio/high_max": 0.014500423683784902,
      "clip_ratio/high_mean": 0.003385931981028989,
      "clip_ratio/low_mean": 0.00033831040855147876,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037242423568386585,
      "entropy": 0.4147470034658909,
      "epoch": 0.23066734589913399,
      "grad_norm": 18.509260177612305,
      "learning_rate": 6.154849225753871e-07,
      "loss": -0.1015,
      "reward": 3.0347917079925537,
      "reward_std": 1.6506093740463257,
      "rewards/reasoning_density_reward/mean": 0.3254464268684387,
      "rewards/reasoning_density_reward/std": 0.192388653755188,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.4593452215194702,
      "rewards/trajectory_similarity_reward/std": 0.3795848488807678,
      "step": 283
    },
    {
      "clip_ratio/high_max": 0.0131998147116974,
      "clip_ratio/high_mean": 0.004015683312900364,
      "clip_ratio/low_mean": 0.0007156721112551168,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004731355438707396,
      "entropy": 0.40287603437900543,
      "epoch": 0.2314824248599083,
      "grad_norm": 18.54034423828125,
      "learning_rate": 6.148329258353708e-07,
      "loss": 4.8833,
      "reward": 2.865886926651001,
      "reward_std": 0.9118632078170776,
      "rewards/reasoning_density_reward/mean": 0.39895834028720856,
      "rewards/reasoning_density_reward/std": 0.12192396074533463,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.486591637134552,
      "rewards/trajectory_similarity_reward/mean": 0.5919286012649536,
      "rewards/trajectory_similarity_reward/std": 0.47132980823516846,
      "step": 284
    },
    {
      "clip_ratio/high_max": 0.008418863726546988,
      "clip_ratio/high_mean": 0.0021535818159463815,
      "clip_ratio/low_mean": 0.0012745853200613055,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034281672014913056,
      "entropy": 0.3960162550210953,
      "epoch": 0.23229750382068262,
      "grad_norm": 18.479148864746094,
      "learning_rate": 6.141809290953545e-07,
      "loss": -3.9122,
      "reward": 2.509650766849518,
      "reward_std": 1.56589674949646,
      "rewards/reasoning_density_reward/mean": 0.3650207072496414,
      "rewards/reasoning_density_reward/std": 0.10482966899871826,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.404874324798584,
      "rewards/trajectory_similarity_reward/mean": 0.457130029797554,
      "rewards/trajectory_similarity_reward/std": 0.41039273142814636,
      "step": 285
    },
    {
      "clip_ratio/high_max": 0.014449614158365875,
      "clip_ratio/high_mean": 0.0030740303991478868,
      "clip_ratio/low_mean": 0.0007341960736084729,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003808226407272741,
      "entropy": 0.39871666952967644,
      "epoch": 0.23311258278145697,
      "grad_norm": 12.07947826385498,
      "learning_rate": 6.135289323553382e-07,
      "loss": -5.8486,
      "reward": 3.214961051940918,
      "reward_std": 1.8898459672927856,
      "rewards/reasoning_density_reward/mean": 0.3552083373069763,
      "rewards/reasoning_density_reward/std": 0.15900346636772156,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6097528338432312,
      "rewards/trajectory_similarity_reward/std": 0.39030611515045166,
      "step": 286
    },
    {
      "clip_ratio/high_max": 0.01782527467003092,
      "clip_ratio/high_mean": 0.005114499217597768,
      "clip_ratio/low_mean": 0.00044302507376414724,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005557524447795004,
      "entropy": 0.4182770773768425,
      "epoch": 0.23392766174223129,
      "grad_norm": 13.211533546447754,
      "learning_rate": 6.128769356153219e-07,
      "loss": 9.8126,
      "reward": 3.4505388736724854,
      "reward_std": 1.5327305793762207,
      "rewards/reasoning_density_reward/mean": 0.3630208373069763,
      "rewards/reasoning_density_reward/std": 0.15948320925235748,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.6500178575515747,
      "rewards/trajectory_similarity_reward/std": 0.42524078488349915,
      "step": 287
    },
    {
      "clip_ratio/high_max": 0.01240608241641894,
      "clip_ratio/high_mean": 0.0021972169415676035,
      "clip_ratio/low_mean": 0.0007872500718804076,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029844670061720535,
      "entropy": 0.40142126753926277,
      "epoch": 0.2347427407030056,
      "grad_norm": 13.08802318572998,
      "learning_rate": 6.122249388753056e-07,
      "loss": 4.3672,
      "reward": 2.202536106109619,
      "reward_std": 2.0370635986328125,
      "rewards/reasoning_density_reward/mean": 0.3694940507411957,
      "rewards/reasoning_density_reward/std": 0.08071846514940262,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.33304187655448914,
      "rewards/trajectory_similarity_reward/std": 0.36000752449035645,
      "step": 288
    },
    {
      "clip_ratio/high_max": 0.012901471927762032,
      "clip_ratio/high_mean": 0.0023206564874271862,
      "clip_ratio/low_mean": 0.0010735303003457375,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033941867004614323,
      "entropy": 0.4102092757821083,
      "epoch": 0.23555781966377992,
      "grad_norm": 15.382290840148926,
      "learning_rate": 6.115729421352893e-07,
      "loss": -6.8099,
      "reward": 2.026583671569824,
      "reward_std": 1.9043338298797607,
      "rewards/reasoning_density_reward/mean": 0.34259670972824097,
      "rewards/reasoning_density_reward/std": 0.10845470428466797,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.27773693203926086,
      "rewards/trajectory_similarity_reward/std": 0.43447956442832947,
      "step": 289
    },
    {
      "clip_ratio/high_max": 0.01128765702014789,
      "clip_ratio/high_mean": 0.002314597375516314,
      "clip_ratio/low_mean": 0.001370925376249943,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036855226717307232,
      "entropy": 0.4037179797887802,
      "epoch": 0.23637289862455427,
      "grad_norm": 14.114801406860352,
      "learning_rate": 6.10920945395273e-07,
      "loss": -10.2347,
      "reward": 2.5563413500785828,
      "reward_std": 1.376158595085144,
      "rewards/reasoning_density_reward/mean": 0.32601459324359894,
      "rewards/reasoning_density_reward/std": 0.13218173384666443,
      "rewards/success_reward/mean": 1.828125,
      "rewards/success_reward/std": 1.3285260796546936,
      "rewards/trajectory_similarity_reward/mean": 0.4022017568349838,
      "rewards/trajectory_similarity_reward/std": 0.407611683011055,
      "step": 290
    },
    {
      "clip_ratio/high_max": 0.012233210494741797,
      "clip_ratio/high_mean": 0.002864615169528406,
      "clip_ratio/low_mean": 0.0008727812746656127,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003737396458745934,
      "entropy": 0.3863185755908489,
      "epoch": 0.23718797758532859,
      "grad_norm": 11.475672721862793,
      "learning_rate": 6.102689486552567e-07,
      "loss": 4.0235,
      "reward": 2.2279300689697266,
      "reward_std": 1.575589656829834,
      "rewards/reasoning_density_reward/mean": 0.33281248807907104,
      "rewards/reasoning_density_reward/std": 0.13438710570335388,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.2808688879013062,
      "rewards/trajectory_similarity_reward/mean": 0.3013676404953003,
      "rewards/trajectory_similarity_reward/std": 0.4170529544353485,
      "step": 291
    },
    {
      "clip_ratio/high_max": 0.012681080668698996,
      "clip_ratio/high_mean": 0.002212381106801331,
      "clip_ratio/low_mean": 0.001519009681942407,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003731390788743738,
      "entropy": 0.41257812455296516,
      "epoch": 0.2380030565461029,
      "grad_norm": 14.649678230285645,
      "learning_rate": 6.096169519152404e-07,
      "loss": 3.6486,
      "reward": 1.8435227870941162,
      "reward_std": 1.9145655632019043,
      "rewards/reasoning_density_reward/mean": 0.33392858505249023,
      "rewards/reasoning_density_reward/std": 0.11767301708459854,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.290844202041626,
      "rewards/trajectory_similarity_reward/std": 0.4482293426990509,
      "step": 292
    },
    {
      "clip_ratio/high_max": 0.011785587470512837,
      "clip_ratio/high_mean": 0.002464588062139228,
      "clip_ratio/low_mean": 0.00036974263275624253,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028343307349132374,
      "entropy": 0.36852888762950897,
      "epoch": 0.23881813550687722,
      "grad_norm": 13.412875175476074,
      "learning_rate": 6.089649551752241e-07,
      "loss": 0.3045,
      "reward": 2.077070713043213,
      "reward_std": 1.7712717056274414,
      "rewards/reasoning_density_reward/mean": 0.2966075539588928,
      "rewards/reasoning_density_reward/std": 0.10736411809921265,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4491376876831055,
      "rewards/trajectory_similarity_reward/mean": 0.2804628908634186,
      "rewards/trajectory_similarity_reward/std": 0.3627447187900543,
      "step": 293
    },
    {
      "clip_ratio/high_max": 0.01617217797320336,
      "clip_ratio/high_mean": 0.0036235322331776842,
      "clip_ratio/low_mean": 0.0004982061800546944,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0041217383404728025,
      "entropy": 0.38701413199305534,
      "epoch": 0.23963321446765157,
      "grad_norm": 12.848379135131836,
      "learning_rate": 6.083129584352078e-07,
      "loss": -4.0969,
      "reward": 3.0489556789398193,
      "reward_std": 1.7698984146118164,
      "rewards/reasoning_density_reward/mean": 0.3549879789352417,
      "rewards/reasoning_density_reward/std": 0.16002556681632996,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.6314675807952881,
      "rewards/trajectory_similarity_reward/std": 0.443209171295166,
      "step": 294
    },
    {
      "clip_ratio/high_max": 0.013469399942550808,
      "clip_ratio/high_mean": 0.0030553231845260598,
      "clip_ratio/low_mean": 0.0009155762381851673,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039708994518150575,
      "entropy": 0.3706762120127678,
      "epoch": 0.24044829342842589,
      "grad_norm": 15.886043548583984,
      "learning_rate": 6.076609616951915e-07,
      "loss": 3.6108,
      "reward": 3.2901573181152344,
      "reward_std": 0.9943598508834839,
      "rewards/reasoning_density_reward/mean": 0.3937127888202667,
      "rewards/reasoning_density_reward/std": 0.08795246481895447,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.3651946187019348,
      "rewards/trajectory_similarity_reward/std": 0.2783902883529663,
      "step": 295
    },
    {
      "clip_ratio/high_max": 0.015873717376962304,
      "clip_ratio/high_mean": 0.004114873358048499,
      "clip_ratio/low_mean": 0.0004928690032102168,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004607742332154885,
      "entropy": 0.37512465938925743,
      "epoch": 0.2412633723892002,
      "grad_norm": 18.89791488647461,
      "learning_rate": 6.070089649551752e-07,
      "loss": 2.7683,
      "reward": 3.3855714797973633,
      "reward_std": 1.474287986755371,
      "rewards/reasoning_density_reward/mean": 0.4064980149269104,
      "rewards/reasoning_density_reward/std": 0.11067752540111542,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.0600701570510864,
      "rewards/trajectory_similarity_reward/mean": 0.5415734946727753,
      "rewards/trajectory_similarity_reward/std": 0.3212650716304779,
      "step": 296
    },
    {
      "clip_ratio/high_max": 0.012654019985347986,
      "clip_ratio/high_mean": 0.004010345408460125,
      "clip_ratio/low_mean": 0.0001470985844207462,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004157444025622681,
      "entropy": 0.38675302267074585,
      "epoch": 0.24207845134997452,
      "grad_norm": 10.441469192504883,
      "learning_rate": 6.06356968215159e-07,
      "loss": 2.5983,
      "reward": 3.2121376991271973,
      "reward_std": 1.501388430595398,
      "rewards/reasoning_density_reward/mean": 0.37502026557922363,
      "rewards/reasoning_density_reward/std": 0.1434873789548874,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.4933674931526184,
      "rewards/trajectory_similarity_reward/std": 0.3879365622997284,
      "step": 297
    },
    {
      "clip_ratio/high_max": 0.012113696313463151,
      "clip_ratio/high_mean": 0.0029509468586184084,
      "clip_ratio/low_mean": 0.0005243474661256187,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034752943320199847,
      "entropy": 0.3836939409375191,
      "epoch": 0.24289353031074887,
      "grad_norm": 12.233640670776367,
      "learning_rate": 6.057049714751426e-07,
      "loss": -6.6939,
      "reward": 2.537691831588745,
      "reward_std": 1.9871511459350586,
      "rewards/reasoning_density_reward/mean": 0.35369259119033813,
      "rewards/reasoning_density_reward/std": 0.12387470155954361,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.4964991807937622,
      "rewards/trajectory_similarity_reward/std": 0.4548652470111847,
      "step": 298
    },
    {
      "clip_ratio/high_max": 0.010338087275158614,
      "clip_ratio/high_mean": 0.0021240130663500167,
      "clip_ratio/low_mean": 0.0005827599561598618,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027067730552516878,
      "entropy": 0.3635067082941532,
      "epoch": 0.24370860927152319,
      "grad_norm": 13.81221866607666,
      "learning_rate": 6.050529747351264e-07,
      "loss": 7.986,
      "reward": 3.0096354484558105,
      "reward_std": 1.0830345153808594,
      "rewards/reasoning_density_reward/mean": 0.36434659361839294,
      "rewards/reasoning_density_reward/std": 0.09677594900131226,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.4890388250350952,
      "rewards/trajectory_similarity_reward/std": 0.36574649810791016,
      "step": 299
    },
    {
      "clip_ratio/high_max": 0.01526502991328016,
      "clip_ratio/high_mean": 0.002764094082522206,
      "clip_ratio/low_mean": 0.0009118296366068535,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003675923726405017,
      "entropy": 0.3777584880590439,
      "epoch": 0.2445236882322975,
      "grad_norm": 13.578783988952637,
      "learning_rate": 6.0440097799511e-07,
      "loss": 3.0666,
      "reward": 3.1982977390289307,
      "reward_std": 1.8212058544158936,
      "rewards/reasoning_density_reward/mean": 0.33901098370552063,
      "rewards/reasoning_density_reward/std": 0.12428051978349686,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6092867255210876,
      "rewards/trajectory_similarity_reward/std": 0.3921889364719391,
      "step": 300
    },
    {
      "clip_ratio/high_max": 0.011705651122611016,
      "clip_ratio/high_mean": 0.003742550528841093,
      "clip_ratio/low_mean": 0.00011434902990004048,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038568995660170913,
      "entropy": 0.3911980502307415,
      "epoch": 0.24533876719307182,
      "grad_norm": 14.250848770141602,
      "learning_rate": 6.037489812550938e-07,
      "loss": -14.9616,
      "reward": 2.849365711212158,
      "reward_std": 1.3574546575546265,
      "rewards/reasoning_density_reward/mean": 0.3899305462837219,
      "rewards/reasoning_density_reward/std": 0.11284241080284119,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.39693519473075867,
      "rewards/trajectory_similarity_reward/std": 0.3598313331604004,
      "step": 301
    },
    {
      "clip_ratio/high_max": 0.014678402454592288,
      "clip_ratio/high_mean": 0.0035773122508544475,
      "clip_ratio/low_mean": 0.000498564368172083,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004075876669958234,
      "entropy": 0.3637716919183731,
      "epoch": 0.24615384615384617,
      "grad_norm": 15.855813026428223,
      "learning_rate": 6.030969845150774e-07,
      "loss": -4.1414,
      "reward": 3.494213104248047,
      "reward_std": 0.9438057541847229,
      "rewards/reasoning_density_reward/mean": 0.4517795145511627,
      "rewards/reasoning_density_reward/std": 0.07130922190845013,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.0930703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.6049337983131409,
      "rewards/trajectory_similarity_reward/std": 0.29414311796426773,
      "step": 302
    },
    {
      "clip_ratio/high_max": 0.012232446344569325,
      "clip_ratio/high_mean": 0.002598772232886404,
      "clip_ratio/low_mean": 0.0009662553493399173,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003565027625882067,
      "entropy": 0.37204350158572197,
      "epoch": 0.24696892511462049,
      "grad_norm": 15.583108901977539,
      "learning_rate": 6.024449877750612e-07,
      "loss": -3.7201,
      "reward": 3.575216293334961,
      "reward_std": 0.9534735679626465,
      "rewards/reasoning_density_reward/mean": 0.43437498807907104,
      "rewards/reasoning_density_reward/std": 0.11063265055418015,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.7033414244651794,
      "rewards/trajectory_similarity_reward/std": 0.4248732328414917,
      "step": 303
    },
    {
      "clip_ratio/high_max": 0.017481996561400592,
      "clip_ratio/high_mean": 0.0039260614139493555,
      "clip_ratio/low_mean": 0.0014065937011764618,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005332655069651082,
      "entropy": 0.3622281551361084,
      "epoch": 0.2477840040753948,
      "grad_norm": 12.30293083190918,
      "learning_rate": 6.017929910350448e-07,
      "loss": -5.2687,
      "reward": 3.421505928039551,
      "reward_std": 1.3813974857330322,
      "rewards/reasoning_density_reward/mean": 0.4189732074737549,
      "rewards/reasoning_density_reward/std": 0.09153563529253006,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.5650327205657959,
      "rewards/trajectory_similarity_reward/std": 0.41416072845458984,
      "step": 304
    },
    {
      "clip_ratio/high_max": 0.016346113989129663,
      "clip_ratio/high_mean": 0.004623399261618033,
      "clip_ratio/low_mean": 0.00028393068350851536,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004907330003334209,
      "entropy": 0.3964430093765259,
      "epoch": 0.24859908303616912,
      "grad_norm": 18.758874893188477,
      "learning_rate": 6.011409942950286e-07,
      "loss": 4.3495,
      "reward": 3.96720552444458,
      "reward_std": 1.187804102897644,
      "rewards/reasoning_density_reward/mean": 0.47187499701976776,
      "rewards/reasoning_density_reward/std": 0.06936421617865562,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8158584833145142,
      "rewards/trajectory_similarity_reward/mean": 0.7765804529190063,
      "rewards/trajectory_similarity_reward/std": 0.33240872621536255,
      "step": 305
    },
    {
      "clip_ratio/high_max": 0.009103146643610671,
      "clip_ratio/high_mean": 0.003183405515301274,
      "clip_ratio/low_mean": 0.0007824215790606104,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003965827127103694,
      "entropy": 0.4173182211816311,
      "epoch": 0.24941416199694347,
      "grad_norm": 14.211477279663086,
      "learning_rate": 6.004889975550122e-07,
      "loss": -2.341,
      "reward": 1.4323124885559082,
      "reward_std": 1.7265170812606812,
      "rewards/reasoning_density_reward/mean": 0.2646205425262451,
      "rewards/reasoning_density_reward/std": 0.12371142208576202,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.13644197583198547,
      "rewards/trajectory_similarity_reward/std": 0.20980341732501984,
      "step": 306
    },
    {
      "clip_ratio/high_max": 0.010662132466677576,
      "clip_ratio/high_mean": 0.0024459674532408826,
      "clip_ratio/low_mean": 0.0009079836090677418,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003353950996825006,
      "entropy": 0.3789873979985714,
      "epoch": 0.25022924095771776,
      "grad_norm": 14.620100021362305,
      "learning_rate": 5.998370008149959e-07,
      "loss": 1.6344,
      "reward": 2.85996150970459,
      "reward_std": 1.9422804117202759,
      "rewards/reasoning_density_reward/mean": 0.40677085518836975,
      "rewards/reasoning_density_reward/std": 0.09809381514787674,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.48444056510925293,
      "rewards/trajectory_similarity_reward/std": 0.41465455293655396,
      "step": 307
    },
    {
      "clip_ratio/high_max": 0.013896698830649257,
      "clip_ratio/high_mean": 0.0026183640948147513,
      "clip_ratio/low_mean": 0.001179315320769092,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037976794119458646,
      "entropy": 0.3611372150480747,
      "epoch": 0.25104431991849213,
      "grad_norm": 11.660065650939941,
      "learning_rate": 5.991850040749796e-07,
      "loss": 7.0748,
      "reward": 2.835574150085449,
      "reward_std": 1.8874082565307617,
      "rewards/reasoning_density_reward/mean": 0.3665178716182709,
      "rewards/reasoning_density_reward/std": 0.10856888443231583,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.5003061294555664,
      "rewards/trajectory_similarity_reward/std": 0.4638623893260956,
      "step": 308
    },
    {
      "clip_ratio/high_max": 0.010885553958360106,
      "clip_ratio/high_mean": 0.0019766697878367268,
      "clip_ratio/low_mean": 0.0009943093100446276,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002970979083329439,
      "entropy": 0.3769725598394871,
      "epoch": 0.25185939887926645,
      "grad_norm": 14.151846885681152,
      "learning_rate": 5.985330073349633e-07,
      "loss": 7.9252,
      "reward": 2.5634016195933023,
      "reward_std": 1.1360278129577637,
      "rewards/reasoning_density_reward/mean": 0.39027779301007587,
      "rewards/reasoning_density_reward/std": 0.11166493594646454,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5018402338027954,
      "rewards/trajectory_similarity_reward/mean": 0.4856238067150116,
      "rewards/trajectory_similarity_reward/std": 0.46955832839012146,
      "step": 309
    },
    {
      "clip_ratio/high_max": 0.01436826994176954,
      "clip_ratio/high_mean": 0.0031014927808428183,
      "clip_ratio/low_mean": 0.0010452947826706804,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004146787585341372,
      "entropy": 0.39647580683231354,
      "epoch": 0.25267447784004077,
      "grad_norm": 17.614044189453125,
      "learning_rate": 5.978810105949469e-07,
      "loss": -10.4686,
      "reward": 2.165497064590454,
      "reward_std": 1.692321538925171,
      "rewards/reasoning_density_reward/mean": 0.3564732223749161,
      "rewards/reasoning_density_reward/std": 0.08286736160516739,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.3369660377502441,
      "rewards/trajectory_similarity_reward/mean": 0.3090239241719246,
      "rewards/trajectory_similarity_reward/std": 0.41645972430706024,
      "step": 310
    },
    {
      "clip_ratio/high_max": 0.015691112028434873,
      "clip_ratio/high_mean": 0.0031343527080025524,
      "clip_ratio/low_mean": 0.0009151829890470253,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004049535869853571,
      "entropy": 0.41178738698363304,
      "epoch": 0.2534895568008151,
      "grad_norm": 12.80133056640625,
      "learning_rate": 5.972290138549307e-07,
      "loss": 0.534,
      "reward": 2.793883800506592,
      "reward_std": 1.4572417736053467,
      "rewards/reasoning_density_reward/mean": 0.28697916865348816,
      "rewards/reasoning_density_reward/std": 0.19117751717567444,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.5381544828414917,
      "rewards/trajectory_similarity_reward/std": 0.4413236081600189,
      "step": 311
    },
    {
      "clip_ratio/high_max": 0.014192182454280555,
      "clip_ratio/high_mean": 0.002404471204499714,
      "clip_ratio/low_mean": 0.0005074178770883009,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002911889081588015,
      "entropy": 0.38610273972153664,
      "epoch": 0.2543046357615894,
      "grad_norm": 12.445730209350586,
      "learning_rate": 5.965770171149144e-07,
      "loss": 7.4454,
      "reward": 2.879244565963745,
      "reward_std": 0.9575744867324829,
      "rewards/reasoning_density_reward/mean": 0.41458335518836975,
      "rewards/reasoning_density_reward/std": 0.10034661740064621,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5896613597869873,
      "rewards/trajectory_similarity_reward/std": 0.47786352038383484,
      "step": 312
    },
    {
      "clip_ratio/high_max": 0.01845339871942997,
      "clip_ratio/high_mean": 0.0032102324184961617,
      "clip_ratio/low_mean": 0.0004709560416813474,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036811884201597422,
      "entropy": 0.3654614947736263,
      "epoch": 0.2551197147223637,
      "grad_norm": 14.067726135253906,
      "learning_rate": 5.959250203748981e-07,
      "loss": -4.0423,
      "reward": 2.2798739075660706,
      "reward_std": 1.9367502927780151,
      "rewards/reasoning_density_reward/mean": 0.37187501788139343,
      "rewards/reasoning_density_reward/std": 0.11670142412185669,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.5043244361877441,
      "rewards/trajectory_similarity_reward/mean": 0.4548739045858383,
      "rewards/trajectory_similarity_reward/std": 0.49299223721027374,
      "step": 313
    },
    {
      "clip_ratio/high_max": 0.010805805679410696,
      "clip_ratio/high_mean": 0.0022344354874803685,
      "clip_ratio/low_mean": 0.0012068239011568949,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034412593668093905,
      "entropy": 0.3822938837110996,
      "epoch": 0.25593479368313804,
      "grad_norm": 13.556357383728027,
      "learning_rate": 5.952730236348818e-07,
      "loss": -0.6331,
      "reward": 2.539876937866211,
      "reward_std": 2.1215126514434814,
      "rewards/reasoning_density_reward/mean": 0.37968751788139343,
      "rewards/reasoning_density_reward/std": 0.10644589364528656,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.47268927097320557,
      "rewards/trajectory_similarity_reward/std": 0.4470042884349823,
      "step": 314
    },
    {
      "clip_ratio/high_max": 0.012836404494009912,
      "clip_ratio/high_mean": 0.0024078812057268806,
      "clip_ratio/low_mean": 0.0007610472312080674,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003168928451486863,
      "entropy": 0.3983265049755573,
      "epoch": 0.25674987264391236,
      "grad_norm": 12.595715522766113,
      "learning_rate": 5.946210268948655e-07,
      "loss": 4.1568,
      "reward": 2.0582685470581055,
      "reward_std": 2.1637513637542725,
      "rewards/reasoning_density_reward/mean": 0.3453125059604645,
      "rewards/reasoning_density_reward/std": 0.12490621209144592,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.40045595169067383,
      "rewards/trajectory_similarity_reward/std": 0.4716496169567108,
      "step": 315
    },
    {
      "clip_ratio/high_max": 0.010284380055963993,
      "clip_ratio/high_mean": 0.0026067294093081728,
      "clip_ratio/low_mean": 0.00038034013778087683,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029870695725549012,
      "entropy": 0.3616899587213993,
      "epoch": 0.25756495160468673,
      "grad_norm": 38.3264045715332,
      "learning_rate": 5.939690301548492e-07,
      "loss": 1.0759,
      "reward": 3.3987340927124023,
      "reward_std": 1.7423746585845947,
      "rewards/reasoning_density_reward/mean": 0.3884672522544861,
      "rewards/reasoning_density_reward/std": 0.1267147809267044,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.6665166020393372,
      "rewards/trajectory_similarity_reward/std": 0.41613146662712097,
      "step": 316
    },
    {
      "clip_ratio/high_max": 0.01657579722814262,
      "clip_ratio/high_mean": 0.0029945699643576518,
      "clip_ratio/low_mean": 0.0007553147697763052,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003749884788703639,
      "entropy": 0.38260114938020706,
      "epoch": 0.25838003056546105,
      "grad_norm": 13.602109909057617,
      "learning_rate": 5.933170334148329e-07,
      "loss": 7.2565,
      "reward": 2.1689904928207397,
      "reward_std": 1.8833897113800049,
      "rewards/reasoning_density_reward/mean": 0.36517858505249023,
      "rewards/reasoning_density_reward/std": 0.09619767218828201,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.4680703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.3975619226694107,
      "rewards/trajectory_similarity_reward/std": 0.41213248670101166,
      "step": 317
    },
    {
      "clip_ratio/high_max": 0.007098786416463554,
      "clip_ratio/high_mean": 0.001282848657865543,
      "clip_ratio/low_mean": 0.0017752147759892978,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030580634047510102,
      "entropy": 0.3724346198141575,
      "epoch": 0.25919510952623537,
      "grad_norm": 12.02940845489502,
      "learning_rate": 5.926650366748166e-07,
      "loss": -0.7827,
      "reward": 2.596245288848877,
      "reward_std": 0.6826697587966919,
      "rewards/reasoning_density_reward/mean": 0.39618057012557983,
      "rewards/reasoning_density_reward/std": 0.11615794897079468,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.5125647187232971,
      "rewards/trajectory_similarity_reward/std": 0.47648993134498596,
      "step": 318
    },
    {
      "clip_ratio/high_max": 0.015159661415964365,
      "clip_ratio/high_mean": 0.0035640515270642936,
      "clip_ratio/low_mean": 0.0007320471486309543,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042960986465914175,
      "entropy": 0.3944416344165802,
      "epoch": 0.2600101884870097,
      "grad_norm": 16.954345703125,
      "learning_rate": 5.920130399348003e-07,
      "loss": -6.4942,
      "reward": 3.1720290184020996,
      "reward_std": 1.1754566431045532,
      "rewards/reasoning_density_reward/mean": 0.3760416805744171,
      "rewards/reasoning_density_reward/std": 0.08343742787837982,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.35848721861839294,
      "rewards/trajectory_similarity_reward/std": 0.3609079420566559,
      "step": 319
    },
    {
      "clip_ratio/high_max": 0.015110800159163773,
      "clip_ratio/high_mean": 0.0029805889062117785,
      "clip_ratio/low_mean": 0.00027270071586826816,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032532895857002586,
      "entropy": 0.4225904047489166,
      "epoch": 0.260825267447784,
      "grad_norm": 13.545127868652344,
      "learning_rate": 5.91361043194784e-07,
      "loss": 6.9937,
      "reward": 2.4308053255081177,
      "reward_std": 1.7050780653953552,
      "rewards/reasoning_density_reward/mean": 0.3091447949409485,
      "rewards/reasoning_density_reward/std": 0.12273077666759491,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.3162438869476318,
      "rewards/trajectory_similarity_reward/mean": 0.34041060507297516,
      "rewards/trajectory_similarity_reward/std": 0.3976410776376724,
      "step": 320
    },
    {
      "clip_ratio/high_max": 0.010879391891648993,
      "clip_ratio/high_mean": 0.0020662267634179443,
      "clip_ratio/low_mean": 0.000379685232474003,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024459119886159897,
      "entropy": 0.3779408372938633,
      "epoch": 0.2616403464085583,
      "grad_norm": 13.571147918701172,
      "learning_rate": 5.907090464547677e-07,
      "loss": -0.2092,
      "reward": 2.842888355255127,
      "reward_std": 1.8234121799468994,
      "rewards/reasoning_density_reward/mean": 0.37447917461395264,
      "rewards/reasoning_density_reward/std": 0.17201067507266998,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5934090614318848,
      "rewards/trajectory_similarity_reward/std": 0.47722798585891724,
      "step": 321
    },
    {
      "clip_ratio/high_max": 0.011547838628757745,
      "clip_ratio/high_mean": 0.002240222689579241,
      "clip_ratio/low_mean": 0.0004591548604366835,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026993775318260305,
      "entropy": 0.3837013430893421,
      "epoch": 0.26245542536933264,
      "grad_norm": 11.69759464263916,
      "learning_rate": 5.900570497147514e-07,
      "loss": -5.9313,
      "reward": 2.870410442352295,
      "reward_std": 0.5953657627105713,
      "rewards/reasoning_density_reward/mean": 0.38238638639450073,
      "rewards/reasoning_density_reward/std": 0.14182861149311066,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.5192739963531494,
      "rewards/trajectory_similarity_reward/std": 0.5020473599433899,
      "step": 322
    },
    {
      "clip_ratio/high_max": 0.013568156922701746,
      "clip_ratio/high_mean": 0.0026509315066505224,
      "clip_ratio/low_mean": 0.000639264122582972,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032901956874411553,
      "entropy": 0.3646187111735344,
      "epoch": 0.26327050433010696,
      "grad_norm": 15.980792999267578,
      "learning_rate": 5.894050529747351e-07,
      "loss": -5.0993,
      "reward": 3.04830002784729,
      "reward_std": 1.7428590059280396,
      "rewards/reasoning_density_reward/mean": 0.4153645783662796,
      "rewards/reasoning_density_reward/std": 0.11444620415568352,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.308667778968811,
      "rewards/trajectory_similarity_reward/mean": 0.5235603898763657,
      "rewards/trajectory_similarity_reward/std": 0.4190492331981659,
      "step": 323
    },
    {
      "clip_ratio/high_max": 0.017166737699881196,
      "clip_ratio/high_mean": 0.003297468982054852,
      "clip_ratio/low_mean": 0.0013361642995732836,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004633633216144517,
      "entropy": 0.40009984001517296,
      "epoch": 0.26408558329088133,
      "grad_norm": 13.369508743286133,
      "learning_rate": 5.887530562347188e-07,
      "loss": -0.3916,
      "reward": 3.547343373298645,
      "reward_std": 0.9445613324642181,
      "rewards/reasoning_density_reward/mean": 0.4260416626930237,
      "rewards/reasoning_density_reward/std": 0.09287504851818085,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.6865953207015991,
      "rewards/trajectory_similarity_reward/mean": 0.5900515913963318,
      "rewards/trajectory_similarity_reward/std": 0.4086669087409973,
      "step": 324
    },
    {
      "clip_ratio/high_max": 0.014165198081173003,
      "clip_ratio/high_mean": 0.004296748113119975,
      "clip_ratio/low_mean": 0.0012584250689542387,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005555173265747726,
      "entropy": 0.39360879361629486,
      "epoch": 0.26490066225165565,
      "grad_norm": 15.126787185668945,
      "learning_rate": 5.881010594947025e-07,
      "loss": -3.3145,
      "reward": 1.2581379413604736,
      "reward_std": 1.0718754529953003,
      "rewards/reasoning_density_reward/mean": 0.3109374940395355,
      "rewards/reasoning_density_reward/std": 0.0926322191953659,
      "rewards/success_reward/mean": 0.75,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.19720038771629333,
      "rewards/trajectory_similarity_reward/std": 0.37709468603134155,
      "step": 325
    },
    {
      "clip_ratio/high_max": 0.007428282813634723,
      "clip_ratio/high_mean": 0.00166499715851387,
      "clip_ratio/low_mean": 0.0006498481561720837,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023148453619796783,
      "entropy": 0.38467973843216896,
      "epoch": 0.26571574121242997,
      "grad_norm": 14.278477668762207,
      "learning_rate": 5.874490627546862e-07,
      "loss": 8.0712,
      "reward": 2.1365082263946533,
      "reward_std": 1.8595573902130127,
      "rewards/reasoning_density_reward/mean": 0.33253204822540283,
      "rewards/reasoning_density_reward/std": 0.12124699354171753,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.3736594319343567,
      "rewards/trajectory_similarity_reward/mean": 0.3508511260151863,
      "rewards/trajectory_similarity_reward/std": 0.39105224609375,
      "step": 326
    },
    {
      "clip_ratio/high_max": 0.012943452806212008,
      "clip_ratio/high_mean": 0.0034966445964528248,
      "clip_ratio/low_mean": 0.0005931454616074916,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004089790221769363,
      "entropy": 0.3790748827159405,
      "epoch": 0.2665308201732043,
      "grad_norm": 13.779884338378906,
      "learning_rate": 5.8679706601467e-07,
      "loss": 16.8727,
      "reward": 3.7738335132598877,
      "reward_std": 1.439713954925537,
      "rewards/reasoning_density_reward/mean": 0.4598214030265808,
      "rewards/reasoning_density_reward/std": 0.0868525579571724,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.7827622294425964,
      "rewards/trajectory_similarity_reward/std": 0.39928963780403137,
      "step": 327
    },
    {
      "clip_ratio/high_max": 0.01601012356695719,
      "clip_ratio/high_mean": 0.0034542013017926365,
      "clip_ratio/low_mean": 0.0007476633836631663,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004201864736387506,
      "entropy": 0.4059874415397644,
      "epoch": 0.2673458991339786,
      "grad_norm": 14.152125358581543,
      "learning_rate": 5.861450692746536e-07,
      "loss": -0.9152,
      "reward": 3.068795919418335,
      "reward_std": 1.3748878240585327,
      "rewards/reasoning_density_reward/mean": 0.39047619700431824,
      "rewards/reasoning_density_reward/std": 0.10024376586079597,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.3871415257453918,
      "rewards/trajectory_similarity_reward/mean": 0.568944662809372,
      "rewards/trajectory_similarity_reward/std": 0.404135063290596,
      "step": 328
    },
    {
      "clip_ratio/high_max": 0.010189076070673764,
      "clip_ratio/high_mean": 0.002333372787688859,
      "clip_ratio/low_mean": 0.0005415791274572257,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028749519115081057,
      "entropy": 0.39173199236392975,
      "epoch": 0.2681609780947529,
      "grad_norm": 16.224746704101562,
      "learning_rate": 5.854930725346374e-07,
      "loss": 16.1439,
      "reward": 4.31325626373291,
      "reward_std": 0.09017428755760193,
      "rewards/reasoning_density_reward/mean": 0.5,
      "rewards/reasoning_density_reward/std": 0.0,
      "rewards/success_reward/mean": 3.0,
      "rewards/success_reward/std": 0.0,
      "rewards/trajectory_similarity_reward/mean": 0.8132562637329102,
      "rewards/trajectory_similarity_reward/std": 0.14089493453502655,
      "step": 329
    },
    {
      "clip_ratio/high_max": 0.012587949517183006,
      "clip_ratio/high_mean": 0.002268422584165819,
      "clip_ratio/low_mean": 0.0005091879465908278,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027776106144301593,
      "entropy": 0.3620162382721901,
      "epoch": 0.26897605705552724,
      "grad_norm": 13.272528648376465,
      "learning_rate": 5.84841075794621e-07,
      "loss": 0.9252,
      "reward": 2.2940075397491455,
      "reward_std": 1.888667106628418,
      "rewards/reasoning_density_reward/mean": 0.3442336320877075,
      "rewards/reasoning_density_reward/std": 0.09160317480564117,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.35602372884750366,
      "rewards/trajectory_similarity_reward/std": 0.43607214093208313,
      "step": 330
    },
    {
      "clip_ratio/high_max": 0.01413365121698007,
      "clip_ratio/high_mean": 0.0032564511202508584,
      "clip_ratio/low_mean": 0.000688810228893999,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039452613273169845,
      "entropy": 0.3791043162345886,
      "epoch": 0.26979113601630156,
      "grad_norm": 16.605979919433594,
      "learning_rate": 5.841890790546048e-07,
      "loss": -5.546,
      "reward": 2.438670039176941,
      "reward_std": 2.0664669275283813,
      "rewards/reasoning_density_reward/mean": 0.3815104067325592,
      "rewards/reasoning_density_reward/std": 0.11670742556452751,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.4456666707992554,
      "rewards/trajectory_similarity_reward/mean": 0.46340957283973694,
      "rewards/trajectory_similarity_reward/std": 0.46702878177165985,
      "step": 331
    },
    {
      "clip_ratio/high_max": 0.010399106948170811,
      "clip_ratio/high_mean": 0.0016960162538453005,
      "clip_ratio/low_mean": 0.001026119411108084,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027221356431255117,
      "entropy": 0.3491310849785805,
      "epoch": 0.27060621497707593,
      "grad_norm": 14.598506927490234,
      "learning_rate": 5.835370823145884e-07,
      "loss": 5.1777,
      "reward": 2.0220437049865723,
      "reward_std": 2.039402484893799,
      "rewards/reasoning_density_reward/mean": 0.35149937868118286,
      "rewards/reasoning_density_reward/std": 0.10014607012271881,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.3580443263053894,
      "rewards/trajectory_similarity_reward/std": 0.4794158935546875,
      "step": 332
    },
    {
      "clip_ratio/high_max": 0.007695530453929678,
      "clip_ratio/high_mean": 0.0018707889357756358,
      "clip_ratio/low_mean": 0.0007714649291301612,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026422538576298393,
      "entropy": 0.3862634375691414,
      "epoch": 0.27142129393785025,
      "grad_norm": 11.457021713256836,
      "learning_rate": 5.828850855745722e-07,
      "loss": 0.6884,
      "reward": 0.8004488945007324,
      "reward_std": 1.4154832363128662,
      "rewards/reasoning_density_reward/mean": 0.32500001788139343,
      "rewards/reasoning_density_reward/std": 0.06831300258636475,
      "rewards/success_reward/mean": 0.375,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.10044889152050018,
      "rewards/trajectory_similarity_reward/std": 0.2753373384475708,
      "step": 333
    },
    {
      "clip_ratio/high_max": 0.014432067284360528,
      "clip_ratio/high_mean": 0.0030654131405754015,
      "clip_ratio/low_mean": 0.0007617751907673664,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038271883240668103,
      "entropy": 0.3750901520252228,
      "epoch": 0.27223637289862457,
      "grad_norm": 17.0239315032959,
      "learning_rate": 5.822330888345558e-07,
      "loss": -7.4318,
      "reward": 2.4376096725463867,
      "reward_std": 1.4538242816925049,
      "rewards/reasoning_density_reward/mean": 0.3973214328289032,
      "rewards/reasoning_density_reward/std": 0.1065499484539032,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.44653815031051636,
      "rewards/trajectory_similarity_reward/std": 0.4665519893169403,
      "step": 334
    },
    {
      "clip_ratio/high_max": 0.008265373180620372,
      "clip_ratio/high_mean": 0.0020505416541709565,
      "clip_ratio/low_mean": 0.0004493714695854578,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002499913091014605,
      "entropy": 0.38442620635032654,
      "epoch": 0.2730514518593989,
      "grad_norm": 14.495046615600586,
      "learning_rate": 5.815810920945396e-07,
      "loss": 6.4559,
      "reward": 2.411970853805542,
      "reward_std": 1.5739774703979492,
      "rewards/reasoning_density_reward/mean": 0.30453869700431824,
      "rewards/reasoning_density_reward/std": 0.11075756698846817,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.41993212699890137,
      "rewards/trajectory_similarity_reward/std": 0.3963666558265686,
      "step": 335
    },
    {
      "clip_ratio/high_max": 0.012590416998136789,
      "clip_ratio/high_mean": 0.0031167374982032925,
      "clip_ratio/low_mean": 0.00041759143641684204,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035343290073797107,
      "entropy": 0.3623402267694473,
      "epoch": 0.2738665308201732,
      "grad_norm": 15.0597505569458,
      "learning_rate": 5.809290953545232e-07,
      "loss": 10.8196,
      "reward": 3.1313823461532593,
      "reward_std": 1.4622320532798767,
      "rewards/reasoning_density_reward/mean": 0.35234376788139343,
      "rewards/reasoning_density_reward/std": 0.11275316774845123,
      "rewards/success_reward/mean": 2.203125,
      "rewards/success_reward/std": 1.3145328760147095,
      "rewards/trajectory_similarity_reward/mean": 0.5759136080741882,
      "rewards/trajectory_similarity_reward/std": 0.40495315194129944,
      "step": 336
    },
    {
      "clip_ratio/high_max": 0.014143328473437577,
      "clip_ratio/high_mean": 0.0031628154174541123,
      "clip_ratio/low_mean": 0.0013056625466560945,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004468477884074673,
      "entropy": 0.35677406191825867,
      "epoch": 0.2746816097809475,
      "grad_norm": 13.070695877075195,
      "learning_rate": 5.80277098614507e-07,
      "loss": 5.7452,
      "reward": 2.55159330368042,
      "reward_std": 0.7903411984443665,
      "rewards/reasoning_density_reward/mean": 0.4125000238418579,
      "rewards/reasoning_density_reward/std": 0.10246950387954712,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.4515931010246277,
      "rewards/trajectory_similarity_reward/std": 0.4199838936328888,
      "step": 337
    },
    {
      "clip_ratio/high_max": 0.009029267937876284,
      "clip_ratio/high_mean": 0.0018334616324864328,
      "clip_ratio/low_mean": 0.0012889351237390656,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031223967962432653,
      "entropy": 0.37240246310830116,
      "epoch": 0.27549668874172184,
      "grad_norm": 16.340702056884766,
      "learning_rate": 5.796251018744906e-07,
      "loss": -5.2753,
      "reward": 3.36393666267395,
      "reward_std": 1.1407392024993896,
      "rewards/reasoning_density_reward/mean": 0.44999998807907104,
      "rewards/reasoning_density_reward/std": 0.08944271504878998,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6639366149902344,
      "rewards/trajectory_similarity_reward/std": 0.40326347947120667,
      "step": 338
    },
    {
      "clip_ratio/high_max": 0.015995573077816516,
      "clip_ratio/high_mean": 0.003946705212001689,
      "clip_ratio/low_mean": 0.0006949903254280798,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004641695530153811,
      "entropy": 0.3617883138358593,
      "epoch": 0.27631176770249616,
      "grad_norm": 13.50052261352539,
      "learning_rate": 5.789731051344743e-07,
      "loss": 7.3801,
      "reward": 3.67355477809906,
      "reward_std": 1.0960096418857574,
      "rewards/reasoning_density_reward/mean": 0.4468749910593033,
      "rewards/reasoning_density_reward/std": 0.08763552084565163,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.887150913476944,
      "rewards/trajectory_similarity_reward/mean": 0.6954298317432404,
      "rewards/trajectory_similarity_reward/std": 0.4209301918745041,
      "step": 339
    },
    {
      "clip_ratio/high_max": 0.009574354160577059,
      "clip_ratio/high_mean": 0.0016651905025355518,
      "clip_ratio/low_mean": 0.0007704670460952912,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024356575522688217,
      "entropy": 0.3273480795323849,
      "epoch": 0.27712684666327053,
      "grad_norm": 11.930285453796387,
      "learning_rate": 5.78321108394458e-07,
      "loss": 5.6098,
      "reward": 1.894273579120636,
      "reward_std": 1.806860625743866,
      "rewards/reasoning_density_reward/mean": 0.36220239102840424,
      "rewards/reasoning_density_reward/std": 0.09518321603536606,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.5185213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.3133210986852646,
      "rewards/trajectory_similarity_reward/std": 0.41010506451129913,
      "step": 340
    },
    {
      "clip_ratio/high_max": 0.015714545792434365,
      "clip_ratio/high_mean": 0.0029452730595949106,
      "clip_ratio/low_mean": 0.0014158848789520562,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004361157873063348,
      "entropy": 0.34550517052412033,
      "epoch": 0.27794192562404485,
      "grad_norm": 14.806563377380371,
      "learning_rate": 5.776691116544417e-07,
      "loss": 1.5927,
      "reward": 3.1330671310424805,
      "reward_std": 1.8479835987091064,
      "rewards/reasoning_density_reward/mean": 0.4234374761581421,
      "rewards/reasoning_density_reward/std": 0.09375,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.6471294164657593,
      "rewards/trajectory_similarity_reward/std": 0.4660375714302063,
      "step": 341
    },
    {
      "clip_ratio/high_max": 0.010348453710321337,
      "clip_ratio/high_mean": 0.0028727554745273665,
      "clip_ratio/low_mean": 0.0009385184603161179,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003811273883911781,
      "entropy": 0.37432248517870903,
      "epoch": 0.27875700458481917,
      "grad_norm": 15.584534645080566,
      "learning_rate": 5.770171149144254e-07,
      "loss": -3.5256,
      "reward": 2.637173056602478,
      "reward_std": 1.5784162282943726,
      "rewards/reasoning_density_reward/mean": 0.35729168355464935,
      "rewards/reasoning_density_reward/std": 0.12220059335231781,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.4048813581466675,
      "rewards/trajectory_similarity_reward/std": 0.4408653974533081,
      "step": 342
    },
    {
      "clip_ratio/high_max": 0.017201724520418793,
      "clip_ratio/high_mean": 0.0035265922779217362,
      "clip_ratio/low_mean": 0.0010380945132055786,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004564686736557633,
      "entropy": 0.36528199166059494,
      "epoch": 0.2795720835455935,
      "grad_norm": 13.71383285522461,
      "learning_rate": 5.76365118174409e-07,
      "loss": -2.8743,
      "reward": 3.541222095489502,
      "reward_std": 1.6492960453033447,
      "rewards/reasoning_density_reward/mean": 0.44102561473846436,
      "rewards/reasoning_density_reward/std": 0.08487140387296677,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6626964211463928,
      "rewards/trajectory_similarity_reward/std": 0.37119486927986145,
      "step": 343
    },
    {
      "clip_ratio/high_max": 0.021228002849966288,
      "clip_ratio/high_mean": 0.006627015245612711,
      "clip_ratio/low_mean": 0.0004149362721364014,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007041951466817409,
      "entropy": 0.3563932999968529,
      "epoch": 0.2803871625063678,
      "grad_norm": 13.923054695129395,
      "learning_rate": 5.757131214343928e-07,
      "loss": 1.9172,
      "reward": 3.857971668243408,
      "reward_std": 1.2364641427993774,
      "rewards/reasoning_density_reward/mean": 0.4687499850988388,
      "rewards/reasoning_density_reward/std": 0.07446778938174248,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.8409419655799866,
      "rewards/trajectory_similarity_reward/mean": 0.7173466384410858,
      "rewards/trajectory_similarity_reward/std": 0.34939850866794586,
      "step": 344
    },
    {
      "clip_ratio/high_max": 0.013472620106767863,
      "clip_ratio/high_mean": 0.003942655508581083,
      "clip_ratio/low_mean": 0.0006746807339368388,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004617336257069837,
      "entropy": 0.38585730642080307,
      "epoch": 0.2812022414671421,
      "grad_norm": 15.124401092529297,
      "learning_rate": 5.750611246943765e-07,
      "loss": -1.6593,
      "reward": 3.4852510690689087,
      "reward_std": 1.15212282538414,
      "rewards/reasoning_density_reward/mean": 0.4330543130636215,
      "rewards/reasoning_density_reward/std": 0.10021965950727463,
      "rewards/success_reward/mean": 2.484375,
      "rewards/success_reward/std": 0.9252449870109558,
      "rewards/trajectory_similarity_reward/mean": 0.5678215175867081,
      "rewards/trajectory_similarity_reward/std": 0.3349185883998871,
      "step": 345
    },
    {
      "clip_ratio/high_max": 0.009639661875553429,
      "clip_ratio/high_mean": 0.0021578780651907437,
      "clip_ratio/low_mean": 0.0005760902095062193,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002733968234679196,
      "entropy": 0.3709180876612663,
      "epoch": 0.28201732042791644,
      "grad_norm": 12.93267822265625,
      "learning_rate": 5.744091279543602e-07,
      "loss": -14.2785,
      "reward": 1.9092693328857422,
      "reward_std": 1.196012258529663,
      "rewards/reasoning_density_reward/mean": 0.23171672224998474,
      "rewards/reasoning_density_reward/std": 0.13116984069347382,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.157853603363037,
      "rewards/trajectory_similarity_reward/mean": 0.0838027074933052,
      "rewards/trajectory_similarity_reward/std": 0.14596225321292877,
      "step": 346
    },
    {
      "clip_ratio/high_max": 0.013595949159935117,
      "clip_ratio/high_mean": 0.004809563368326053,
      "clip_ratio/low_mean": 0.0002481885712768417,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0050577518122736365,
      "entropy": 0.3094821833074093,
      "epoch": 0.28283239938869076,
      "grad_norm": 13.68790340423584,
      "learning_rate": 5.737571312143439e-07,
      "loss": -3.0208,
      "reward": 4.126405239105225,
      "reward_std": 0.8164730668067932,
      "rewards/reasoning_density_reward/mean": 0.47083333134651184,
      "rewards/reasoning_density_reward/std": 0.06621982343494892,
      "rewards/success_reward/mean": 2.765625,
      "rewards/success_reward/std": 0.7829292416572571,
      "rewards/trajectory_similarity_reward/mean": 0.8899469077587128,
      "rewards/trajectory_similarity_reward/std": 0.29351384937763214,
      "step": 347
    },
    {
      "clip_ratio/high_max": 0.013669940701220185,
      "clip_ratio/high_mean": 0.0033174737691297196,
      "clip_ratio/low_mean": 0.0009086685240617953,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00422614230774343,
      "entropy": 0.3564513027667999,
      "epoch": 0.28364747834946513,
      "grad_norm": 17.182579040527344,
      "learning_rate": 5.731051344743276e-07,
      "loss": -6.7246,
      "reward": 3.8124756813049316,
      "reward_std": 0.8626792728900909,
      "rewards/reasoning_density_reward/mean": 0.45091766119003296,
      "rewards/reasoning_density_reward/std": 0.08062023483216763,
      "rewards/success_reward/mean": 2.578125,
      "rewards/success_reward/std": 0.9854557514190674,
      "rewards/trajectory_similarity_reward/mean": 0.7834330797195435,
      "rewards/trajectory_similarity_reward/std": 0.3291933089494705,
      "step": 348
    },
    {
      "clip_ratio/high_max": 0.015075937466463074,
      "clip_ratio/high_mean": 0.0026127761193492915,
      "clip_ratio/low_mean": 0.0012271876439626794,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038399638142436743,
      "entropy": 0.3954431414604187,
      "epoch": 0.28446255731023945,
      "grad_norm": 13.520979881286621,
      "learning_rate": 5.724531377343113e-07,
      "loss": 6.9344,
      "reward": 1.7888046503067017,
      "reward_std": 1.9496919512748718,
      "rewards/reasoning_density_reward/mean": 0.3307291716337204,
      "rewards/reasoning_density_reward/std": 0.1430346742272377,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.3330754339694977,
      "rewards/trajectory_similarity_reward/std": 0.44970908761024475,
      "step": 349
    },
    {
      "clip_ratio/high_max": 0.008074141223914921,
      "clip_ratio/high_mean": 0.0017164660603157245,
      "clip_ratio/low_mean": 0.0006660787767032161,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023825448806746863,
      "entropy": 0.3688680902123451,
      "epoch": 0.28527763627101377,
      "grad_norm": 16.993999481201172,
      "learning_rate": 5.71801140994295e-07,
      "loss": 0.5657,
      "reward": 2.6760525703430176,
      "reward_std": 1.369627833366394,
      "rewards/reasoning_density_reward/mean": 0.3872023820877075,
      "rewards/reasoning_density_reward/std": 0.07716085761785507,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.3201000690460205,
      "rewards/trajectory_similarity_reward/std": 0.285830020904541,
      "step": 350
    },
    {
      "clip_ratio/high_max": 0.016084011702332646,
      "clip_ratio/high_mean": 0.004524756484897807,
      "clip_ratio/low_mean": 0.0006587325297005009,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0051834892219631,
      "entropy": 0.3644302561879158,
      "epoch": 0.2860927152317881,
      "grad_norm": 16.91122817993164,
      "learning_rate": 5.711491442542786e-07,
      "loss": 4.1323,
      "reward": 4.349129676818848,
      "reward_std": 0.38973640743643045,
      "rewards/reasoning_density_reward/mean": 0.4937500059604645,
      "rewards/reasoning_density_reward/std": 0.02500000037252903,
      "rewards/success_reward/mean": 2.90625,
      "rewards/success_reward/std": 0.375,
      "rewards/trajectory_similarity_reward/mean": 0.9491300880908966,
      "rewards/trajectory_similarity_reward/std": 0.14623664692044258,
      "step": 351
    },
    {
      "clip_ratio/high_max": 0.011028401990188286,
      "clip_ratio/high_mean": 0.002114467668434372,
      "clip_ratio/low_mean": 0.0014786706688028062,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035931382735725492,
      "entropy": 0.4161766320466995,
      "epoch": 0.2869077941925624,
      "grad_norm": 16.992952346801758,
      "learning_rate": 5.704971475142624e-07,
      "loss": 0.3963,
      "reward": 2.4783341884613037,
      "reward_std": 1.7745463848114014,
      "rewards/reasoning_density_reward/mean": 0.17719495296478271,
      "rewards/reasoning_density_reward/std": 0.1688336730003357,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.42613929510116577,
      "rewards/trajectory_similarity_reward/std": 0.4607594907283783,
      "step": 352
    },
    {
      "clip_ratio/high_max": 0.008289771270938218,
      "clip_ratio/high_mean": 0.0021083188185002655,
      "clip_ratio/low_mean": 0.0008046381844906136,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029129570175427943,
      "entropy": 0.44102775678038597,
      "epoch": 0.2877228731533367,
      "grad_norm": 12.31115436553955,
      "learning_rate": 5.69845150774246e-07,
      "loss": 3.6824,
      "reward": 1.3583242893218994,
      "reward_std": 1.1687129735946655,
      "rewards/reasoning_density_reward/mean": 0.22351191937923431,
      "rewards/reasoning_density_reward/std": 0.14296524226665497,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.1035623773932457,
      "rewards/trajectory_similarity_reward/std": 0.2843558192253113,
      "step": 353
    },
    {
      "clip_ratio/high_max": 0.011107098893262446,
      "clip_ratio/high_mean": 0.002307918926817365,
      "clip_ratio/low_mean": 0.0014531573688145727,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037610763392876834,
      "entropy": 0.40935400128364563,
      "epoch": 0.28853795211411104,
      "grad_norm": 13.934216499328613,
      "learning_rate": 5.691931540342298e-07,
      "loss": 0.6671,
      "reward": 1.1174358129501343,
      "reward_std": 1.0393718481063843,
      "rewards/reasoning_density_reward/mean": 0.29776787757873535,
      "rewards/reasoning_density_reward/std": 0.10363195836544037,
      "rewards/success_reward/mean": 0.65625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.1634179651737213,
      "rewards/trajectory_similarity_reward/std": 0.3562370538711548,
      "step": 354
    },
    {
      "clip_ratio/high_max": 0.009382153220940381,
      "clip_ratio/high_mean": 0.0017130740452557802,
      "clip_ratio/low_mean": 0.0008471091496176086,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025601832167012617,
      "entropy": 0.3720071166753769,
      "epoch": 0.28935303107488536,
      "grad_norm": 11.323563575744629,
      "learning_rate": 5.685411572942136e-07,
      "loss": -6.4246,
      "reward": 1.7222933769226074,
      "reward_std": 1.8935339450836182,
      "rewards/reasoning_density_reward/mean": 0.3395833373069763,
      "rewards/reasoning_density_reward/std": 0.08539126068353653,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.2577100992202759,
      "rewards/trajectory_similarity_reward/std": 0.3980693519115448,
      "step": 355
    },
    {
      "clip_ratio/high_max": 0.01464967051288113,
      "clip_ratio/high_mean": 0.0030440595510299318,
      "clip_ratio/low_mean": 0.0010045788731076755,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004048638482345268,
      "entropy": 0.39059846475720406,
      "epoch": 0.29016811003565973,
      "grad_norm": 12.710558891296387,
      "learning_rate": 5.678891605541972e-07,
      "loss": 1.3538,
      "reward": 2.709707736968994,
      "reward_std": 1.4159626960754395,
      "rewards/reasoning_density_reward/mean": 0.40625,
      "rewards/reasoning_density_reward/std": 0.09287086874246597,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.3347075581550598,
      "rewards/trajectory_similarity_reward/std": 0.3372226059436798,
      "step": 356
    },
    {
      "clip_ratio/high_max": 0.005429201293736696,
      "clip_ratio/high_mean": 0.0010241970303468406,
      "clip_ratio/low_mean": 0.0011759308799810242,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002200127870310098,
      "entropy": 0.3694493696093559,
      "epoch": 0.29098318899643405,
      "grad_norm": 12.2505464553833,
      "learning_rate": 5.67237163814181e-07,
      "loss": 5.0077,
      "reward": 1.0042953491210938,
      "reward_std": 1.6555278301239014,
      "rewards/reasoning_density_reward/mean": 0.27619048953056335,
      "rewards/reasoning_density_reward/std": 0.12483096122741699,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.165604829788208,
      "rewards/trajectory_similarity_reward/std": 0.35708847641944885,
      "step": 357
    },
    {
      "clip_ratio/high_max": 0.008984617074020207,
      "clip_ratio/high_mean": 0.0016762314044171944,
      "clip_ratio/low_mean": 0.00227082111814525,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003947052435250953,
      "entropy": 0.383940227329731,
      "epoch": 0.29179826795720837,
      "grad_norm": 14.878487586975098,
      "learning_rate": 5.665851670741646e-07,
      "loss": -3.4634,
      "reward": 3.5578880310058594,
      "reward_std": 1.1535218954086304,
      "rewards/reasoning_density_reward/mean": 0.4355902671813965,
      "rewards/reasoning_density_reward/std": 0.08448109775781631,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6847976446151733,
      "rewards/trajectory_similarity_reward/std": 0.3763663172721863,
      "step": 358
    },
    {
      "clip_ratio/high_max": 0.013011486851610243,
      "clip_ratio/high_mean": 0.0029303194023668766,
      "clip_ratio/low_mean": 0.0004969244546373375,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034272438497282565,
      "entropy": 0.39311061054468155,
      "epoch": 0.2926133469179827,
      "grad_norm": 13.905472755432129,
      "learning_rate": 5.659331703341484e-07,
      "loss": 3.7313,
      "reward": 2.8243804772694907,
      "reward_std": 1.4034299850463867,
      "rewards/reasoning_density_reward/mean": 0.38005953033765155,
      "rewards/reasoning_density_reward/std": 0.15489588677883148,
      "rewards/success_reward/mean": 2.03125,
      "rewards/success_reward/std": 1.4164625803629558,
      "rewards/trajectory_similarity_reward/mean": 0.41307095686594647,
      "rewards/trajectory_similarity_reward/std": 0.35767728090286255,
      "step": 359
    },
    {
      "clip_ratio/high_max": 0.009324374055722728,
      "clip_ratio/high_mean": 0.0025741413155628834,
      "clip_ratio/low_mean": 0.000706446200638311,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003280587545305025,
      "entropy": 0.385585755109787,
      "epoch": 0.293428425878757,
      "grad_norm": 13.371380805969238,
      "learning_rate": 5.65281173594132e-07,
      "loss": 7.6066,
      "reward": 2.728668689727783,
      "reward_std": 1.851550817489624,
      "rewards/reasoning_density_reward/mean": 0.3550054132938385,
      "rewards/reasoning_density_reward/std": 0.07096242904663086,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.4049133062362671,
      "rewards/trajectory_similarity_reward/std": 0.39058786630630493,
      "step": 360
    },
    {
      "clip_ratio/high_max": 0.009506519185379148,
      "clip_ratio/high_mean": 0.001970400320715271,
      "clip_ratio/low_mean": 0.00037320157207432203,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023436019400833175,
      "entropy": 0.4014609530568123,
      "epoch": 0.2942435048395313,
      "grad_norm": 13.014729499816895,
      "learning_rate": 5.646291768541158e-07,
      "loss": 7.3858,
      "reward": 2.40278697013855,
      "reward_std": 1.5826060771942139,
      "rewards/reasoning_density_reward/mean": 0.28430286049842834,
      "rewards/reasoning_density_reward/std": 0.15980592370033264,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.24348405003547668,
      "rewards/trajectory_similarity_reward/std": 0.33174338936805725,
      "step": 361
    },
    {
      "clip_ratio/high_max": 0.010681305429898202,
      "clip_ratio/high_mean": 0.0026708491932367906,
      "clip_ratio/low_mean": 0.0008246081924880855,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003495457422104664,
      "entropy": 0.3895181156694889,
      "epoch": 0.29505858380030564,
      "grad_norm": 14.244135856628418,
      "learning_rate": 5.639771801140994e-07,
      "loss": -3.5218,
      "reward": 2.666515350341797,
      "reward_std": 2.0397164821624756,
      "rewards/reasoning_density_reward/mean": 0.3942708373069763,
      "rewards/reasoning_density_reward/std": 0.09816457331180573,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4909944534301758,
      "rewards/trajectory_similarity_reward/std": 0.47031354904174805,
      "step": 362
    },
    {
      "clip_ratio/high_max": 0.015326496912166476,
      "clip_ratio/high_mean": 0.0033964418107643723,
      "clip_ratio/low_mean": 0.00040039456871454604,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003796836332185194,
      "entropy": 0.3774377442896366,
      "epoch": 0.29587366276107996,
      "grad_norm": 11.520925521850586,
      "learning_rate": 5.633251833740832e-07,
      "loss": -2.9578,
      "reward": 2.9208515882492065,
      "reward_std": 1.583200454711914,
      "rewards/reasoning_density_reward/mean": 0.38145463168621063,
      "rewards/reasoning_density_reward/std": 0.09078391268849373,
      "rewards/success_reward/mean": 2.015625,
      "rewards/success_reward/std": 1.4279338717460632,
      "rewards/trajectory_similarity_reward/mean": 0.5237718522548676,
      "rewards/trajectory_similarity_reward/std": 0.418470099568367,
      "step": 363
    },
    {
      "clip_ratio/high_max": 0.01151252316776663,
      "clip_ratio/high_mean": 0.0028997990957577713,
      "clip_ratio/low_mean": 0.0017538869069539942,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004653686046367511,
      "entropy": 0.3666628748178482,
      "epoch": 0.29668874172185433,
      "grad_norm": 13.939492225646973,
      "learning_rate": 5.626731866340668e-07,
      "loss": -3.1757,
      "reward": 3.137722134590149,
      "reward_std": 1.5768170952796936,
      "rewards/reasoning_density_reward/mean": 0.4375,
      "rewards/reasoning_density_reward/std": 0.09574270993471146,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.6377221047878265,
      "rewards/trajectory_similarity_reward/std": 0.44854868948459625,
      "step": 364
    },
    {
      "clip_ratio/high_max": 0.011583604122279212,
      "clip_ratio/high_mean": 0.002376449261646485,
      "clip_ratio/low_mean": 0.0007095218024915084,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003085971053224057,
      "entropy": 0.3878462351858616,
      "epoch": 0.29750382068262865,
      "grad_norm": 13.208454132080078,
      "learning_rate": 5.620211898940506e-07,
      "loss": 6.1395,
      "reward": 2.029707431793213,
      "reward_std": 2.1236319541931152,
      "rewards/reasoning_density_reward/mean": 0.34564733505249023,
      "rewards/reasoning_density_reward/std": 0.13017676770687103,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.37156009674072266,
      "rewards/trajectory_similarity_reward/std": 0.44923636317253113,
      "step": 365
    },
    {
      "clip_ratio/high_max": 0.009596877498552203,
      "clip_ratio/high_mean": 0.002110499706759583,
      "clip_ratio/low_mean": 0.00040626724876347,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025167670100927353,
      "entropy": 0.38794977217912674,
      "epoch": 0.29831889964340297,
      "grad_norm": 12.477614402770996,
      "learning_rate": 5.613691931540342e-07,
      "loss": 9.4488,
      "reward": 2.925601005554199,
      "reward_std": 1.3399330973625183,
      "rewards/reasoning_density_reward/mean": 0.3510044515132904,
      "rewards/reasoning_density_reward/std": 0.1442137248814106,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.1831679940223694,
      "rewards/trajectory_similarity_reward/mean": 0.5120965987443924,
      "rewards/trajectory_similarity_reward/std": 0.3874047100543976,
      "step": 366
    },
    {
      "clip_ratio/high_max": 0.014859260059893131,
      "clip_ratio/high_mean": 0.00290204027260188,
      "clip_ratio/low_mean": 0.0016906801756704226,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004592720419168472,
      "entropy": 0.3934657834470272,
      "epoch": 0.2991339786041773,
      "grad_norm": 13.443169593811035,
      "learning_rate": 5.60717196414018e-07,
      "loss": 16.6859,
      "reward": 2.245124340057373,
      "reward_std": 1.4380711317062378,
      "rewards/reasoning_density_reward/mean": 0.3604166805744171,
      "rewards/reasoning_density_reward/std": 0.08710891008377075,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.384707510471344,
      "rewards/trajectory_similarity_reward/std": 0.42929789423942566,
      "step": 367
    },
    {
      "clip_ratio/high_max": 0.011777866980992258,
      "clip_ratio/high_mean": 0.0025487713573966175,
      "clip_ratio/low_mean": 0.0006976960030442569,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003246467334975023,
      "entropy": 0.35819636285305023,
      "epoch": 0.2999490575649516,
      "grad_norm": 12.313372611999512,
      "learning_rate": 5.600651996740016e-07,
      "loss": -1.8609,
      "reward": 3.830812454223633,
      "reward_std": 0.9592689275741577,
      "rewards/reasoning_density_reward/mean": 0.43058502674102783,
      "rewards/reasoning_density_reward/std": 0.07668198645114899,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.7752272486686707,
      "rewards/trajectory_similarity_reward/std": 0.3156168460845947,
      "step": 368
    },
    {
      "clip_ratio/high_max": 0.009942442033207044,
      "clip_ratio/high_mean": 0.00265935656716465,
      "clip_ratio/low_mean": 0.0005616902853944339,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00322104683436919,
      "entropy": 0.353129617869854,
      "epoch": 0.3007641365257259,
      "grad_norm": 11.931998252868652,
      "learning_rate": 5.594132029339854e-07,
      "loss": -2.1113,
      "reward": 3.5724148750305176,
      "reward_std": 0.9007634520530701,
      "rewards/reasoning_density_reward/mean": 0.39382439851760864,
      "rewards/reasoning_density_reward/std": 0.1448010802268982,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.6473402380943298,
      "rewards/trajectory_similarity_reward/std": 0.3680458664894104,
      "step": 369
    },
    {
      "clip_ratio/high_max": 0.009598813019692898,
      "clip_ratio/high_mean": 0.0026128363970201463,
      "clip_ratio/low_mean": 0.00040898906445363536,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030218254396459088,
      "entropy": 0.38859889656305313,
      "epoch": 0.30157921548650024,
      "grad_norm": 12.69127368927002,
      "learning_rate": 5.58761206193969e-07,
      "loss": -2.6929,
      "reward": 2.8122172355651855,
      "reward_std": 1.0628584623336792,
      "rewards/reasoning_density_reward/mean": 0.38798363506793976,
      "rewards/reasoning_density_reward/std": 0.09356225654482841,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.360331952571869,
      "rewards/trajectory_similarity_reward/mean": 0.4554836004972458,
      "rewards/trajectory_similarity_reward/std": 0.4092353582382202,
      "step": 370
    },
    {
      "clip_ratio/high_max": 0.012132286617998034,
      "clip_ratio/high_mean": 0.0029166039603296667,
      "clip_ratio/low_mean": 0.0020095908985240385,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004926194946165197,
      "entropy": 0.4009624496102333,
      "epoch": 0.30239429444727456,
      "grad_norm": 14.918689727783203,
      "learning_rate": 5.581092094539528e-07,
      "loss": 7.906,
      "reward": 2.552147388458252,
      "reward_std": 1.6779892444610596,
      "rewards/reasoning_density_reward/mean": 0.3328869044780731,
      "rewards/reasoning_density_reward/std": 0.16379433870315552,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.1618950366973877,
      "rewards/trajectory_similarity_reward/mean": 0.344260573387146,
      "rewards/trajectory_similarity_reward/std": 0.445669561624527,
      "step": 371
    },
    {
      "clip_ratio/high_max": 0.014187458145897835,
      "clip_ratio/high_mean": 0.0036247639800421894,
      "clip_ratio/low_mean": 0.0005363471937016584,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0041611112537793815,
      "entropy": 0.3977755084633827,
      "epoch": 0.30320937340804893,
      "grad_norm": 11.451478958129883,
      "learning_rate": 5.574572127139364e-07,
      "loss": -3.4671,
      "reward": 2.5410882234573364,
      "reward_std": 0.7090919315814972,
      "rewards/reasoning_density_reward/mean": 0.32517585158348083,
      "rewards/reasoning_density_reward/std": 0.18218304961919785,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.2469309568405151,
      "rewards/trajectory_similarity_reward/mean": 0.434662327170372,
      "rewards/trajectory_similarity_reward/std": 0.5052204430103302,
      "step": 372
    },
    {
      "clip_ratio/high_max": 0.012132874224334955,
      "clip_ratio/high_mean": 0.0028523885703179985,
      "clip_ratio/low_mean": 0.0018846636376110837,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0047370522224809974,
      "entropy": 0.38834451884031296,
      "epoch": 0.30402445236882325,
      "grad_norm": 13.842869758605957,
      "learning_rate": 5.5680521597392e-07,
      "loss": 4.352,
      "reward": 2.3430445194244385,
      "reward_std": 1.5127967596054077,
      "rewards/reasoning_density_reward/mean": 0.30937498807907104,
      "rewards/reasoning_density_reward/std": 0.18816550076007843,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.34616950154304504,
      "rewards/trajectory_similarity_reward/std": 0.4236123859882355,
      "step": 373
    },
    {
      "clip_ratio/high_max": 0.008333520905580372,
      "clip_ratio/high_mean": 0.0018633012296049856,
      "clip_ratio/low_mean": 0.0010882479109568521,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00295154913328588,
      "entropy": 0.3727540001273155,
      "epoch": 0.30483953132959757,
      "grad_norm": 13.117837905883789,
      "learning_rate": 5.561532192339038e-07,
      "loss": 1.6462,
      "reward": 2.2848713994026184,
      "reward_std": 1.7738048434257507,
      "rewards/reasoning_density_reward/mean": 0.30211828649044037,
      "rewards/reasoning_density_reward/std": 0.15207039564847946,
      "rewards/success_reward/mean": 1.640625,
      "rewards/success_reward/std": 1.3906975388526917,
      "rewards/trajectory_similarity_reward/mean": 0.3421279937028885,
      "rewards/trajectory_similarity_reward/std": 0.40651528537273407,
      "step": 374
    },
    {
      "clip_ratio/high_max": 0.016229082830250263,
      "clip_ratio/high_mean": 0.0036789785663131624,
      "clip_ratio/low_mean": 0.0007862325728638098,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004465211313799955,
      "entropy": 0.3959108926355839,
      "epoch": 0.3056546102903719,
      "grad_norm": 14.450919151306152,
      "learning_rate": 5.555012224938874e-07,
      "loss": 1.5153,
      "reward": 2.895120143890381,
      "reward_std": 1.7893873453140259,
      "rewards/reasoning_density_reward/mean": 0.41718751192092896,
      "rewards/reasoning_density_reward/std": 0.09862333536148071,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.6029324531555176,
      "rewards/trajectory_similarity_reward/std": 0.4850330650806427,
      "step": 375
    },
    {
      "clip_ratio/high_max": 0.01531980512663722,
      "clip_ratio/high_mean": 0.002913703239755705,
      "clip_ratio/low_mean": 0.0002888071867346298,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003202510401024483,
      "entropy": 0.354260616004467,
      "epoch": 0.3064696892511462,
      "grad_norm": 15.330904006958008,
      "learning_rate": 5.548492257538712e-07,
      "loss": -1.0691,
      "reward": 3.645775318145752,
      "reward_std": 1.4183251857757568,
      "rewards/reasoning_density_reward/mean": 0.41718751192092896,
      "rewards/reasoning_density_reward/std": 0.13262921571731567,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.9031196236610413,
      "rewards/trajectory_similarity_reward/mean": 0.6973375678062439,
      "rewards/trajectory_similarity_reward/std": 0.4299100637435913,
      "step": 376
    },
    {
      "clip_ratio/high_max": 0.01601062889676541,
      "clip_ratio/high_mean": 0.0040897146391216666,
      "clip_ratio/low_mean": 0.00013440860493574291,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004224123229505494,
      "entropy": 0.3724469691514969,
      "epoch": 0.3072847682119205,
      "grad_norm": 13.30125904083252,
      "learning_rate": 5.541972290138548e-07,
      "loss": 2.9723,
      "reward": 2.9587321281433105,
      "reward_std": 2.0017988681793213,
      "rewards/reasoning_density_reward/mean": 0.33701637387275696,
      "rewards/reasoning_density_reward/std": 0.15776276588439941,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5592156648635864,
      "rewards/trajectory_similarity_reward/std": 0.3935908377170563,
      "step": 377
    },
    {
      "clip_ratio/high_max": 0.015734270680695772,
      "clip_ratio/high_mean": 0.0030044465675018728,
      "clip_ratio/low_mean": 0.00037650760714313947,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00338095412007533,
      "entropy": 0.36587435007095337,
      "epoch": 0.30809984717269484,
      "grad_norm": 13.120596885681152,
      "learning_rate": 5.535452322738386e-07,
      "loss": 2.7908,
      "reward": 3.034017562866211,
      "reward_std": 1.8111968040466309,
      "rewards/reasoning_density_reward/mean": 0.34333792328834534,
      "rewards/reasoning_density_reward/std": 0.12722931802272797,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.5344295501708984,
      "rewards/trajectory_similarity_reward/std": 0.3893003463745117,
      "step": 378
    },
    {
      "clip_ratio/high_max": 0.009032274247147143,
      "clip_ratio/high_mean": 0.00165904825553298,
      "clip_ratio/low_mean": 0.0007284026141860522,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002387450873357011,
      "entropy": 0.3517087958753109,
      "epoch": 0.30891492613346916,
      "grad_norm": 13.805850982666016,
      "learning_rate": 5.528932355338222e-07,
      "loss": -7.1706,
      "reward": 2.189304828643799,
      "reward_std": 1.57373708486557,
      "rewards/reasoning_density_reward/mean": 0.3591146022081375,
      "rewards/reasoning_density_reward/std": 0.11534029990434647,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4454171061515808,
      "rewards/trajectory_similarity_reward/mean": 0.33019016683101654,
      "rewards/trajectory_similarity_reward/std": 0.40963029861450195,
      "step": 379
    },
    {
      "clip_ratio/high_max": 0.011209721153136343,
      "clip_ratio/high_mean": 0.0028241605032235384,
      "clip_ratio/low_mean": 0.0008157089177984744,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003639869362814352,
      "entropy": 0.3767270892858505,
      "epoch": 0.30973000509424353,
      "grad_norm": 13.727434158325195,
      "learning_rate": 5.52241238793806e-07,
      "loss": 8.4112,
      "reward": 3.288461208343506,
      "reward_std": 1.6852827072143555,
      "rewards/reasoning_density_reward/mean": 0.3901785910129547,
      "rewards/reasoning_density_reward/std": 0.14218060672283173,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5545325875282288,
      "rewards/trajectory_similarity_reward/std": 0.3451489508152008,
      "step": 380
    },
    {
      "clip_ratio/high_max": 0.011986225406872109,
      "clip_ratio/high_mean": 0.002469300179654965,
      "clip_ratio/low_mean": 0.000985165017482359,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034544652517070062,
      "entropy": 0.39294878393411636,
      "epoch": 0.31054508405501785,
      "grad_norm": 15.48713207244873,
      "learning_rate": 5.515892420537896e-07,
      "loss": 4.3916,
      "reward": 2.7038421630859375,
      "reward_std": 1.7263697385787964,
      "rewards/reasoning_density_reward/mean": 0.34613093733787537,
      "rewards/reasoning_density_reward/std": 0.13950802385807037,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.48271140456199646,
      "rewards/trajectory_similarity_reward/std": 0.3979029953479767,
      "step": 381
    },
    {
      "clip_ratio/high_max": 0.007664233242394403,
      "clip_ratio/high_mean": 0.0010331493431294803,
      "clip_ratio/low_mean": 0.0010143079998670146,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002047457342996495,
      "entropy": 0.4134094901382923,
      "epoch": 0.31136016301579217,
      "grad_norm": 12.76208782196045,
      "learning_rate": 5.509372453137734e-07,
      "loss": 0.0421,
      "reward": 1.6948521137237549,
      "reward_std": 1.9081087112426758,
      "rewards/reasoning_density_reward/mean": 0.3113839328289032,
      "rewards/reasoning_density_reward/std": 0.12593337893486023,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.2584682106971741,
      "rewards/trajectory_similarity_reward/std": 0.3466382920742035,
      "step": 382
    },
    {
      "clip_ratio/high_max": 0.013884770683944225,
      "clip_ratio/high_mean": 0.0033917459950316697,
      "clip_ratio/low_mean": 0.0012748635635944083,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004666609427658841,
      "entropy": 0.34726201742887497,
      "epoch": 0.3121752419765665,
      "grad_norm": 14.32259750366211,
      "learning_rate": 5.50285248573757e-07,
      "loss": -2.8523,
      "reward": 2.704380989074707,
      "reward_std": 2.048218011856079,
      "rewards/reasoning_density_reward/mean": 0.35729166865348816,
      "rewards/reasoning_density_reward/std": 0.13490502908825874,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.5185213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.5658393204212189,
      "rewards/trajectory_similarity_reward/std": 0.48508867621421814,
      "step": 383
    },
    {
      "clip_ratio/high_max": 0.009904090664349496,
      "clip_ratio/high_mean": 0.0016800319426693022,
      "clip_ratio/low_mean": 0.002562337933341041,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004242369934218004,
      "entropy": 0.3265385031700134,
      "epoch": 0.3129903209373408,
      "grad_norm": 14.450465202331543,
      "learning_rate": 5.496332518337408e-07,
      "loss": -7.3952,
      "reward": 3.0152463912963867,
      "reward_std": 1.4413614869117737,
      "rewards/reasoning_density_reward/mean": 0.4189112037420273,
      "rewards/reasoning_density_reward/std": 0.09487393125891685,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.4680703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.627585232257843,
      "rewards/trajectory_similarity_reward/std": 0.4706270843744278,
      "step": 384
    },
    {
      "clip_ratio/high_max": 0.013521197310183197,
      "clip_ratio/high_mean": 0.0029715594500885345,
      "clip_ratio/low_mean": 0.0013231426200945862,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004294702084735036,
      "entropy": 0.3591836057603359,
      "epoch": 0.3138053998981151,
      "grad_norm": 14.481882095336914,
      "learning_rate": 5.489812550937245e-07,
      "loss": 8.0885,
      "reward": 3.4118692874908447,
      "reward_std": 1.6449363231658936,
      "rewards/reasoning_density_reward/mean": 0.2975446581840515,
      "rewards/reasoning_density_reward/std": 0.1658884733915329,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6768244504928589,
      "rewards/trajectory_similarity_reward/std": 0.36650484800338745,
      "step": 385
    },
    {
      "clip_ratio/high_max": 0.015063178783748299,
      "clip_ratio/high_mean": 0.00421390330302529,
      "clip_ratio/low_mean": 0.0010704758678912185,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005284379236400127,
      "entropy": 0.3825295679271221,
      "epoch": 0.31462047885888944,
      "grad_norm": 11.481528282165527,
      "learning_rate": 5.483292583537082e-07,
      "loss": 0.7919,
      "reward": 3.169034481048584,
      "reward_std": 1.7799172401428223,
      "rewards/reasoning_density_reward/mean": 0.3601190447807312,
      "rewards/reasoning_density_reward/std": 0.17444762587547302,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5589154958724976,
      "rewards/trajectory_similarity_reward/std": 0.3621980845928192,
      "step": 386
    },
    {
      "clip_ratio/high_max": 0.019993511377833784,
      "clip_ratio/high_mean": 0.0051218519220128655,
      "clip_ratio/low_mean": 0.0006050029223843012,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005726854840759188,
      "entropy": 0.37449079379439354,
      "epoch": 0.31543555781966376,
      "grad_norm": 19.814611434936523,
      "learning_rate": 5.47677261613692e-07,
      "loss": -3.991,
      "reward": 3.8366599082946777,
      "reward_std": 0.9571981430053711,
      "rewards/reasoning_density_reward/mean": 0.44627976417541504,
      "rewards/reasoning_density_reward/std": 0.07772739976644516,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.7653801441192627,
      "rewards/trajectory_similarity_reward/std": 0.31950312852859497,
      "step": 387
    },
    {
      "clip_ratio/high_max": 0.010929041542112827,
      "clip_ratio/high_mean": 0.0033224308426724747,
      "clip_ratio/low_mean": 0.0004949125832354184,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003817343389528105,
      "entropy": 0.369546290487051,
      "epoch": 0.31625063678043813,
      "grad_norm": 16.102008819580078,
      "learning_rate": 5.470252648736756e-07,
      "loss": 6.4456,
      "reward": 2.7303333282470703,
      "reward_std": 1.8340169191360474,
      "rewards/reasoning_density_reward/mean": 0.2708333134651184,
      "rewards/reasoning_density_reward/std": 0.19874607026576996,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.4907499849796295,
      "rewards/trajectory_similarity_reward/std": 0.4333420395851135,
      "step": 388
    },
    {
      "clip_ratio/high_max": 0.01740220852661878,
      "clip_ratio/high_mean": 0.0038637172838207334,
      "clip_ratio/low_mean": 0.0019512503276928328,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005814967676997185,
      "entropy": 0.3900071457028389,
      "epoch": 0.31706571574121245,
      "grad_norm": 13.348089218139648,
      "learning_rate": 5.463732681336593e-07,
      "loss": -2.7967,
      "reward": 2.5516492128372192,
      "reward_std": 1.259611964225769,
      "rewards/reasoning_density_reward/mean": 0.38229165971279144,
      "rewards/reasoning_density_reward/std": 0.11906276270747185,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4033452868461609,
      "rewards/trajectory_similarity_reward/mean": 0.38810762763023376,
      "rewards/trajectory_similarity_reward/std": 0.40005698800086975,
      "step": 389
    },
    {
      "clip_ratio/high_max": 0.00765808904543519,
      "clip_ratio/high_mean": 0.002023551169259008,
      "clip_ratio/low_mean": 0.001110465713281883,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003134016922558658,
      "entropy": 0.4039926826953888,
      "epoch": 0.31788079470198677,
      "grad_norm": 14.386045455932617,
      "learning_rate": 5.45721271393643e-07,
      "loss": -1.3309,
      "reward": 0.5973998308181763,
      "reward_std": 0.8758860230445862,
      "rewards/reasoning_density_reward/mean": 0.18020832538604736,
      "rewards/reasoning_density_reward/std": 0.13902729749679565,
      "rewards/success_reward/mean": 0.375,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.04219147562980652,
      "rewards/trajectory_similarity_reward/std": 0.12601682543754578,
      "step": 390
    },
    {
      "clip_ratio/high_max": 0.008317893894854933,
      "clip_ratio/high_mean": 0.0013980760195408948,
      "clip_ratio/low_mean": 0.0004706158251792658,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018686918265302666,
      "entropy": 0.38996167853474617,
      "epoch": 0.3186958736627611,
      "grad_norm": 15.207154273986816,
      "learning_rate": 5.450692746536267e-07,
      "loss": 4.8631,
      "reward": 1.2369626760482788,
      "reward_std": 1.0460222959518433,
      "rewards/reasoning_density_reward/mean": 0.28854167461395264,
      "rewards/reasoning_density_reward/std": 0.12688851356506348,
      "rewards/success_reward/mean": 0.75,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.19842100143432617,
      "rewards/trajectory_similarity_reward/std": 0.35532885789871216,
      "step": 391
    },
    {
      "clip_ratio/high_max": 0.00842359010130167,
      "clip_ratio/high_mean": 0.002073641197057441,
      "clip_ratio/low_mean": 0.0012708532085525803,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003344494412885979,
      "entropy": 0.36588984355330467,
      "epoch": 0.3195109526235354,
      "grad_norm": 13.845473289489746,
      "learning_rate": 5.444172779136104e-07,
      "loss": -0.8385,
      "reward": 1.6332118213176727,
      "reward_std": 1.0668688714504242,
      "rewards/reasoning_density_reward/mean": 0.3369419574737549,
      "rewards/reasoning_density_reward/std": 0.09565946459770203,
      "rewards/success_reward/mean": 1.078125,
      "rewards/success_reward/std": 1.029909372329712,
      "rewards/trajectory_similarity_reward/mean": 0.21814479865133762,
      "rewards/trajectory_similarity_reward/std": 0.2735118642449379,
      "step": 392
    },
    {
      "clip_ratio/high_max": 0.013274310040287673,
      "clip_ratio/high_mean": 0.004612551361788064,
      "clip_ratio/low_mean": 0.0005615101545117795,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005174061516299844,
      "entropy": 0.39477478340268135,
      "epoch": 0.3203260315843097,
      "grad_norm": 16.331031799316406,
      "learning_rate": 5.437652811735941e-07,
      "loss": 3.0428,
      "reward": 4.150160948435466,
      "reward_std": 0.35759734114011127,
      "rewards/reasoning_density_reward/mean": 0.44340277711550397,
      "rewards/reasoning_density_reward/std": 0.11463638146718343,
      "rewards/success_reward/mean": 2.96875,
      "rewards/success_reward/std": 0.125,
      "rewards/trajectory_similarity_reward/mean": 0.7380082408587137,
      "rewards/trajectory_similarity_reward/std": 0.19980469346046448,
      "step": 393
    },
    {
      "clip_ratio/high_max": 0.00960152456536889,
      "clip_ratio/high_mean": 0.0018714014731813222,
      "clip_ratio/low_mean": 0.001599209695996251,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034706112346611917,
      "entropy": 0.35348550975322723,
      "epoch": 0.32114111054508404,
      "grad_norm": 15.839926719665527,
      "learning_rate": 5.431132844335778e-07,
      "loss": -3.522,
      "reward": 2.26052188873291,
      "reward_std": 1.7508008480072021,
      "rewards/reasoning_density_reward/mean": 0.3699776828289032,
      "rewards/reasoning_density_reward/std": 0.08845619112253189,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.390544056892395,
      "rewards/trajectory_similarity_reward/std": 0.42567184567451477,
      "step": 394
    },
    {
      "clip_ratio/high_max": 0.013504092348739505,
      "clip_ratio/high_mean": 0.0030596861615777016,
      "clip_ratio/low_mean": 0.0001617612815607572,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003221447463147342,
      "entropy": 0.3467704653739929,
      "epoch": 0.32195618950585836,
      "grad_norm": 13.911638259887695,
      "learning_rate": 5.424612876935615e-07,
      "loss": 7.9343,
      "reward": 3.106314182281494,
      "reward_std": 1.8832755088806152,
      "rewards/reasoning_density_reward/mean": 0.4061553180217743,
      "rewards/reasoning_density_reward/std": 0.12135656923055649,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.5439090728759766,
      "rewards/trajectory_similarity_reward/std": 0.408393919467926,
      "step": 395
    },
    {
      "clip_ratio/high_max": 0.014432416879571974,
      "clip_ratio/high_mean": 0.002837612744770013,
      "clip_ratio/low_mean": 0.0008780656207818538,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003715678380103782,
      "entropy": 0.35022299364209175,
      "epoch": 0.32277126846663273,
      "grad_norm": 17.81336784362793,
      "learning_rate": 5.418092909535452e-07,
      "loss": -2.3575,
      "reward": 2.566304922103882,
      "reward_std": 1.5513149499893188,
      "rewards/reasoning_density_reward/mean": 0.3818063586950302,
      "rewards/reasoning_density_reward/std": 0.0986444279551506,
      "rewards/success_reward/mean": 1.734375,
      "rewards/success_reward/std": 1.4375925064086914,
      "rewards/trajectory_similarity_reward/mean": 0.4501235783100128,
      "rewards/trajectory_similarity_reward/std": 0.42410337924957275,
      "step": 396
    },
    {
      "clip_ratio/high_max": 0.012952898163348436,
      "clip_ratio/high_mean": 0.0025826568889897317,
      "clip_ratio/low_mean": 0.000712950735760387,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032956076902337372,
      "entropy": 0.3313654884696007,
      "epoch": 0.32358634742740705,
      "grad_norm": 15.362751960754395,
      "learning_rate": 5.411572942135289e-07,
      "loss": -5.0438,
      "reward": 3.5068840980529785,
      "reward_std": 1.030730128288269,
      "rewards/reasoning_density_reward/mean": 0.42388394474983215,
      "rewards/reasoning_density_reward/std": 0.09421821683645248,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.645500123500824,
      "rewards/trajectory_similarity_reward/std": 0.41118741035461426,
      "step": 397
    },
    {
      "clip_ratio/high_max": 0.007226353860460222,
      "clip_ratio/high_mean": 0.001456793732359074,
      "clip_ratio/low_mean": 0.0012892336344521027,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027460273704491556,
      "entropy": 0.39690906554460526,
      "epoch": 0.32440142638818137,
      "grad_norm": 22.4691162109375,
      "learning_rate": 5.405052974735126e-07,
      "loss": 5.3899,
      "reward": 0.9932116866111755,
      "reward_std": 0.7943178415298462,
      "rewards/reasoning_density_reward/mean": 0.3025669753551483,
      "rewards/reasoning_density_reward/std": 0.07725368440151215,
      "rewards/success_reward/mean": 0.609375,
      "rewards/success_reward/std": 1.1259995698928833,
      "rewards/trajectory_similarity_reward/mean": 0.08126971684396267,
      "rewards/trajectory_similarity_reward/std": 0.18666326999664307,
      "step": 398
    },
    {
      "clip_ratio/high_max": 0.009214634832460433,
      "clip_ratio/high_mean": 0.0025520356866763905,
      "clip_ratio/low_mean": 0.0012868250978499418,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003838860815449152,
      "entropy": 0.40897469967603683,
      "epoch": 0.3252165053489557,
      "grad_norm": 10.751640319824219,
      "learning_rate": 5.398533007334963e-07,
      "loss": -1.1461,
      "reward": 2.4705276489257812,
      "reward_std": 1.675800085067749,
      "rewards/reasoning_density_reward/mean": 0.2874999940395355,
      "rewards/reasoning_density_reward/std": 0.23273734748363495,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.49552756547927856,
      "rewards/trajectory_similarity_reward/std": 0.4694104492664337,
      "step": 399
    },
    {
      "clip_ratio/high_max": 0.008293628081446514,
      "clip_ratio/high_mean": 0.0013384031044552103,
      "clip_ratio/low_mean": 0.0009757798798091244,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002314183031558059,
      "entropy": 0.38133809715509415,
      "epoch": 0.32603158430973,
      "grad_norm": 10.527746200561523,
      "learning_rate": 5.392013039934801e-07,
      "loss": 10.7163,
      "reward": 2.871814727783203,
      "reward_std": 0.7553601861000061,
      "rewards/reasoning_density_reward/mean": 0.3879464268684387,
      "rewards/reasoning_density_reward/std": 0.0945989191532135,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.23386842012405396,
      "rewards/trajectory_similarity_reward/std": 0.23133161664009094,
      "step": 400
    },
    {
      "clip_ratio/high_max": 0.013481833157129586,
      "clip_ratio/high_mean": 0.002673694572877139,
      "clip_ratio/low_mean": 0.0012497570533014368,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039234516734723,
      "entropy": 0.34624556824564934,
      "epoch": 0.3268466632705043,
      "grad_norm": 16.97576332092285,
      "learning_rate": 5.385493072534637e-07,
      "loss": -3.6218,
      "reward": 3.5533156394958496,
      "reward_std": 1.3560290336608887,
      "rewards/reasoning_density_reward/mean": 0.4192708134651184,
      "rewards/reasoning_density_reward/std": 0.10735536366701126,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.6027949452400208,
      "rewards/trajectory_similarity_reward/std": 0.3149632513523102,
      "step": 401
    },
    {
      "clip_ratio/high_max": 0.014311353501398116,
      "clip_ratio/high_mean": 0.0037585750105790794,
      "clip_ratio/low_mean": 0.0014050468271307182,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005163621855899692,
      "entropy": 0.3429764620959759,
      "epoch": 0.32766174223127864,
      "grad_norm": 13.859162330627441,
      "learning_rate": 5.378973105134474e-07,
      "loss": -4.2991,
      "reward": 3.16507625579834,
      "reward_std": 1.0178298950195312,
      "rewards/reasoning_density_reward/mean": 0.4312500059604645,
      "rewards/reasoning_density_reward/std": 0.09464846551418304,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.5775761604309082,
      "rewards/trajectory_similarity_reward/std": 0.40573975443840027,
      "step": 402
    },
    {
      "clip_ratio/high_max": 0.014879496942739934,
      "clip_ratio/high_mean": 0.003575602611817885,
      "clip_ratio/low_mean": 0.00031197077805700246,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038875734389876015,
      "entropy": 0.3322296738624573,
      "epoch": 0.32847682119205296,
      "grad_norm": 12.443979263305664,
      "learning_rate": 5.372453137734311e-07,
      "loss": 4.1954,
      "reward": 3.381706476211548,
      "reward_std": 1.4428740739822388,
      "rewards/reasoning_density_reward/mean": 0.389705091714859,
      "rewards/reasoning_density_reward/std": 0.12355658411979675,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.5545012950897217,
      "rewards/trajectory_similarity_reward/std": 0.4092211127281189,
      "step": 403
    },
    {
      "clip_ratio/high_max": 0.014985614689067006,
      "clip_ratio/high_mean": 0.003932680207071826,
      "clip_ratio/low_mean": 0.0007929953899292741,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0047256756806746125,
      "entropy": 0.36873382702469826,
      "epoch": 0.32929190015282733,
      "grad_norm": 20.294963836669922,
      "learning_rate": 5.365933170334148e-07,
      "loss": -5.7863,
      "reward": 3.3713752031326294,
      "reward_std": 1.6253154873847961,
      "rewards/reasoning_density_reward/mean": 0.42656250298023224,
      "rewards/reasoning_density_reward/std": 0.10683373361825943,
      "rewards/success_reward/mean": 2.296875,
      "rewards/success_reward/std": 1.157951295375824,
      "rewards/trajectory_similarity_reward/mean": 0.6479375958442688,
      "rewards/trajectory_similarity_reward/std": 0.4576590210199356,
      "step": 404
    },
    {
      "clip_ratio/high_max": 0.015335855423472822,
      "clip_ratio/high_mean": 0.004127079446334392,
      "clip_ratio/low_mean": 0.0003144962392980233,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0044415755837690085,
      "entropy": 0.35552549734711647,
      "epoch": 0.33010697911360165,
      "grad_norm": 14.859456062316895,
      "learning_rate": 5.359413202933984e-07,
      "loss": -4.8319,
      "reward": 2.898210048675537,
      "reward_std": 1.7261760830879211,
      "rewards/reasoning_density_reward/mean": 0.3847222179174423,
      "rewards/reasoning_density_reward/std": 0.11387929320335388,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.2636653184890747,
      "rewards/trajectory_similarity_reward/mean": 0.5447379052639008,
      "rewards/trajectory_similarity_reward/std": 0.453061580657959,
      "step": 405
    },
    {
      "clip_ratio/high_max": 0.00940320489462465,
      "clip_ratio/high_mean": 0.0016805653212941252,
      "clip_ratio/low_mean": 0.0011228164221392944,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028033816925017163,
      "entropy": 0.3442327752709389,
      "epoch": 0.33092205807437597,
      "grad_norm": 14.60840129852295,
      "learning_rate": 5.352893235533822e-07,
      "loss": 1.7527,
      "reward": 1.2689683437347412,
      "reward_std": 0.6329122185707092,
      "rewards/reasoning_density_reward/mean": 0.2994047701358795,
      "rewards/reasoning_density_reward/std": 0.022662866860628128,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.03206358477473259,
      "rewards/trajectory_similarity_reward/std": 0.12825435400009155,
      "step": 406
    },
    {
      "clip_ratio/high_max": 0.012427409252268262,
      "clip_ratio/high_mean": 0.0026604618560668314,
      "clip_ratio/low_mean": 0.0009818304461077787,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036422924167709425,
      "entropy": 0.3652256764471531,
      "epoch": 0.3317371370351503,
      "grad_norm": 9.510321617126465,
      "learning_rate": 5.346373268133658e-07,
      "loss": -0.6728,
      "reward": 2.4017577171325684,
      "reward_std": 1.4301681518554688,
      "rewards/reasoning_density_reward/mean": 0.35997024178504944,
      "rewards/reasoning_density_reward/std": 0.1616973876953125,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.44803741574287415,
      "rewards/trajectory_similarity_reward/std": 0.47062230110168457,
      "step": 407
    },
    {
      "clip_ratio/high_max": 0.0076744850957766175,
      "clip_ratio/high_mean": 0.0016333899402525276,
      "clip_ratio/low_mean": 0.0009971519102691673,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002630541886901483,
      "entropy": 0.3641527518630028,
      "epoch": 0.3325522159959246,
      "grad_norm": 14.77344799041748,
      "learning_rate": 5.339853300733496e-07,
      "loss": -4.7149,
      "step": 408
    },
    {
      "clip_ratio/high_max": 0.01797430287115276,
      "clip_ratio/high_mean": 0.0038270454097073525,
      "clip_ratio/low_mean": 0.00038718764699297026,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004214232991216704,
      "entropy": 0.3446853756904602,
      "epoch": 0.3333672949566989,
      "grad_norm": 13.010970115661621,
      "learning_rate": 5.333333333333332e-07,
      "loss": -1.7557,
      "reward": 3.4806571006774902,
      "reward_std": 1.4483774900436401,
      "rewards/reasoning_density_reward/mean": 0.44348958134651184,
      "rewards/reasoning_density_reward/std": 0.08696940168738365,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.0453851819038391,
      "rewards/trajectory_similarity_reward/mean": 0.5996676087379456,
      "rewards/trajectory_similarity_reward/std": 0.40312360227108,
      "step": 409
    },
    {
      "clip_ratio/high_max": 0.01589577621780336,
      "clip_ratio/high_mean": 0.004334841651143506,
      "clip_ratio/low_mean": 0.0001365483258268796,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004471389991522301,
      "entropy": 0.3381124399602413,
      "epoch": 0.33418237391747324,
      "grad_norm": 13.199362754821777,
      "learning_rate": 5.32681336593317e-07,
      "loss": 7.7232,
      "reward": 3.605947732925415,
      "reward_std": 1.6863586902618408,
      "rewards/reasoning_density_reward/mean": 0.43217718601226807,
      "rewards/reasoning_density_reward/std": 0.09559816867113113,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.736270546913147,
      "rewards/trajectory_similarity_reward/std": 0.399722695350647,
      "step": 410
    },
    {
      "clip_ratio/high_max": 0.015219688415527344,
      "clip_ratio/high_mean": 0.002773852553218603,
      "clip_ratio/low_mean": 0.0006894387151987758,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034632913157111034,
      "entropy": 0.3253924213349819,
      "epoch": 0.33499745287824756,
      "grad_norm": 14.851600646972656,
      "learning_rate": 5.320293398533006e-07,
      "loss": 1.8918,
      "reward": 3.089775562286377,
      "reward_std": 1.3184118270874023,
      "rewards/reasoning_density_reward/mean": 0.4375,
      "rewards/reasoning_density_reward/std": 0.09472135826945305,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.380683958530426,
      "rewards/trajectory_similarity_reward/mean": 0.5429005026817322,
      "rewards/trajectory_similarity_reward/std": 0.38622765243053436,
      "step": 411
    },
    {
      "clip_ratio/high_max": 0.011173987528309226,
      "clip_ratio/high_mean": 0.002531885082134977,
      "clip_ratio/low_mean": 0.0013431161059997976,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003875001144479029,
      "entropy": 0.3350639156997204,
      "epoch": 0.33581253183902193,
      "grad_norm": 15.843807220458984,
      "learning_rate": 5.313773431132844e-07,
      "loss": -5.1342,
      "reward": 2.494041681289673,
      "reward_std": 1.947196364402771,
      "rewards/reasoning_density_reward/mean": 0.3968750238418579,
      "rewards/reasoning_density_reward/std": 0.11614466458559036,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.40966659784317017,
      "rewards/trajectory_similarity_reward/std": 0.3837689757347107,
      "step": 412
    },
    {
      "clip_ratio/high_max": 0.015042110986541957,
      "clip_ratio/high_mean": 0.003036149159015622,
      "clip_ratio/low_mean": 0.0006324982132355217,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036686473031295463,
      "entropy": 0.3355483449995518,
      "epoch": 0.33662761079979625,
      "grad_norm": 15.362327575683594,
      "learning_rate": 5.30725346373268e-07,
      "loss": -0.6051,
      "reward": 3.7831318378448486,
      "reward_std": 1.2391161918640137,
      "rewards/reasoning_density_reward/mean": 0.4520833492279053,
      "rewards/reasoning_density_reward/std": 0.08603939414024353,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.866025447845459,
      "rewards/trajectory_similarity_reward/mean": 0.7060483694076538,
      "rewards/trajectory_similarity_reward/std": 0.3756750226020813,
      "step": 413
    },
    {
      "clip_ratio/high_max": 0.013721227063797414,
      "clip_ratio/high_mean": 0.0029834617453161627,
      "clip_ratio/low_mean": 0.0005969815465505235,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035804433064186014,
      "entropy": 0.3259270451962948,
      "epoch": 0.33744268976057057,
      "grad_norm": 15.290608406066895,
      "learning_rate": 5.300733496332518e-07,
      "loss": 2.5221,
      "reward": 3.5110912322998047,
      "reward_std": 0.9962373189628124,
      "rewards/reasoning_density_reward/mean": 0.45393106341362,
      "rewards/reasoning_density_reward/std": 0.04579911753535271,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.6196601092815399,
      "rewards/trajectory_similarity_reward/std": 0.2185734063386917,
      "step": 414
    },
    {
      "clip_ratio/high_max": 0.013022883824305609,
      "clip_ratio/high_mean": 0.0026099387141584884,
      "clip_ratio/low_mean": 0.001191871051560156,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003801809754804708,
      "entropy": 0.32911156490445137,
      "epoch": 0.3382577687213449,
      "grad_norm": 13.726605415344238,
      "learning_rate": 5.294213528932355e-07,
      "loss": 4.2938,
      "reward": 3.315126419067383,
      "reward_std": 1.6388633251190186,
      "rewards/reasoning_density_reward/mean": 0.4149305522441864,
      "rewards/reasoning_density_reward/std": 0.09355174750089645,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5564458966255188,
      "rewards/trajectory_similarity_reward/std": 0.3872582018375397,
      "step": 415
    },
    {
      "clip_ratio/high_max": 0.014702152227982879,
      "clip_ratio/high_mean": 0.004270851801265962,
      "clip_ratio/low_mean": 0.0008013082333491184,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005072160100098699,
      "entropy": 0.37506943568587303,
      "epoch": 0.3390728476821192,
      "grad_norm": 14.929831504821777,
      "learning_rate": 5.287693561532192e-07,
      "loss": 3.1824,
      "reward": 2.403308153152466,
      "reward_std": 1.3629286289215088,
      "rewards/reasoning_density_reward/mean": 0.30186009407043457,
      "rewards/reasoning_density_reward/std": 0.1133284792304039,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.2512494325637817,
      "rewards/trajectory_similarity_reward/mean": 0.3201979696750641,
      "rewards/trajectory_similarity_reward/std": 0.388393372297287,
      "step": 416
    },
    {
      "clip_ratio/high_max": 0.013677117647603154,
      "clip_ratio/high_mean": 0.002831789097399451,
      "clip_ratio/low_mean": 0.0006576832020073198,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034894722484750673,
      "entropy": 0.33316684141755104,
      "epoch": 0.3398879266428935,
      "grad_norm": 15.124917984008789,
      "learning_rate": 5.281173594132029e-07,
      "loss": 2.8764,
      "reward": 3.090610980987549,
      "reward_std": 1.0414260625839233,
      "rewards/reasoning_density_reward/mean": 0.4296875,
      "rewards/reasoning_density_reward/std": 0.09540210664272308,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5984233021736145,
      "rewards/trajectory_similarity_reward/std": 0.4366203844547272,
      "step": 417
    },
    {
      "clip_ratio/high_max": 0.010569237259915099,
      "clip_ratio/high_mean": 0.0016703867695468944,
      "clip_ratio/low_mean": 0.0013281544925121125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002998541182023473,
      "entropy": 0.3568485341966152,
      "epoch": 0.34070300560366784,
      "grad_norm": 13.517313003540039,
      "learning_rate": 5.274653626731866e-07,
      "loss": 10.6833,
      "reward": 3.075650215148926,
      "reward_std": 1.5386359691619873,
      "rewards/reasoning_density_reward/mean": 0.35409730672836304,
      "rewards/reasoning_density_reward/std": 0.1367836594581604,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.4715530574321747,
      "rewards/trajectory_similarity_reward/std": 0.4246813654899597,
      "step": 418
    },
    {
      "clip_ratio/high_max": 0.012557380134239793,
      "clip_ratio/high_mean": 0.002218906316556968,
      "clip_ratio/low_mean": 0.0013449219550238922,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003563828257028945,
      "entropy": 0.3780283071100712,
      "epoch": 0.34151808456444216,
      "grad_norm": 13.528420448303223,
      "learning_rate": 5.268133659331703e-07,
      "loss": 19.4313,
      "reward": 2.9616096019744873,
      "reward_std": 1.2077925205230713,
      "rewards/reasoning_density_reward/mean": 0.3713541626930237,
      "rewards/reasoning_density_reward/std": 0.1483854055404663,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 0.9486833214759827,
      "rewards/trajectory_similarity_reward/mean": 0.3402554988861084,
      "rewards/trajectory_similarity_reward/std": 0.36538732051849365,
      "step": 419
    },
    {
      "clip_ratio/high_max": 0.01434505998622626,
      "clip_ratio/high_mean": 0.003698891378007829,
      "clip_ratio/low_mean": 0.0008234346823883243,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004522326053120196,
      "entropy": 0.34272272139787674,
      "epoch": 0.34233316352521653,
      "grad_norm": 18.572736740112305,
      "learning_rate": 5.26161369193154e-07,
      "loss": 13.2869,
      "reward": 3.4062408208847046,
      "reward_std": 1.266909122467041,
      "rewards/reasoning_density_reward/mean": 0.4287388473749161,
      "rewards/reasoning_density_reward/std": 0.11285789683461189,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.9796755611896515,
      "rewards/trajectory_similarity_reward/mean": 0.4462520033121109,
      "rewards/trajectory_similarity_reward/std": 0.32591257989406586,
      "step": 420
    },
    {
      "clip_ratio/high_max": 0.010142957442440093,
      "clip_ratio/high_mean": 0.002575124890427105,
      "clip_ratio/low_mean": 0.0012957942526554689,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038709190412191674,
      "entropy": 0.3582477606832981,
      "epoch": 0.34314824248599085,
      "grad_norm": 16.0263729095459,
      "learning_rate": 5.255093724531377e-07,
      "loss": -5.0073,
      "reward": 1.0067867040634155,
      "reward_std": 1.5437383651733398,
      "rewards/reasoning_density_reward/mean": 0.3291294574737549,
      "rewards/reasoning_density_reward/std": 0.06421861052513123,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.11515721678733826,
      "rewards/trajectory_similarity_reward/std": 0.24971002340316772,
      "step": 421
    },
    {
      "clip_ratio/high_max": 0.010314768645912409,
      "clip_ratio/high_mean": 0.0019633175106719136,
      "clip_ratio/low_mean": 0.0021169310421100818,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004080248480022419,
      "entropy": 0.36013564094901085,
      "epoch": 0.34396332144676517,
      "grad_norm": 15.989816665649414,
      "learning_rate": 5.248573757131214e-07,
      "loss": 3.7781,
      "reward": 3.8651652336120605,
      "reward_std": 0.6729304790496826,
      "rewards/reasoning_density_reward/mean": 0.4429687559604645,
      "rewards/reasoning_density_reward/std": 0.08897199481725693,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.7034465074539185,
      "rewards/trajectory_similarity_reward/std": 0.3739306330680847,
      "step": 422
    },
    {
      "clip_ratio/high_max": 0.009876636613626033,
      "clip_ratio/high_mean": 0.002727975537709426,
      "clip_ratio/low_mean": 0.001370686906739138,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004098662473552395,
      "entropy": 0.35710402205586433,
      "epoch": 0.3447784004075395,
      "grad_norm": 13.959320068359375,
      "learning_rate": 5.242053789731051e-07,
      "loss": -1.4718,
      "reward": 2.5804364681243896,
      "reward_std": 1.314563274383545,
      "rewards/reasoning_density_reward/mean": 0.38593749701976776,
      "rewards/reasoning_density_reward/std": 0.1239590086042881,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.486591637134552,
      "rewards/trajectory_similarity_reward/mean": 0.5069990456104279,
      "rewards/trajectory_similarity_reward/std": 0.492226243019104,
      "step": 423
    },
    {
      "clip_ratio/high_max": 0.022833936847746372,
      "clip_ratio/high_mean": 0.004347729962319136,
      "clip_ratio/low_mean": 0.000474771652079653,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004822501563467085,
      "entropy": 0.38371336460113525,
      "epoch": 0.3455934793683138,
      "grad_norm": 14.629936218261719,
      "learning_rate": 5.235533822330888e-07,
      "loss": 15.032,
      "reward": 3.2400522232055664,
      "reward_std": 1.4914897680282593,
      "rewards/reasoning_density_reward/mean": 0.3816547095775604,
      "rewards/reasoning_density_reward/std": 0.14732608199119568,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5146474838256836,
      "rewards/trajectory_similarity_reward/std": 0.3545694947242737,
      "step": 424
    },
    {
      "clip_ratio/high_max": 0.015911851864075288,
      "clip_ratio/high_mean": 0.0035676515217346605,
      "clip_ratio/low_mean": 0.0005431757381302305,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004110827223485103,
      "entropy": 0.35570090264081955,
      "epoch": 0.3464085583290881,
      "grad_norm": 11.628951072692871,
      "learning_rate": 5.229013854930725e-07,
      "loss": 9.2363,
      "reward": 3.260488510131836,
      "reward_std": 1.156432032585144,
      "rewards/reasoning_density_reward/mean": 0.43541666865348816,
      "rewards/reasoning_density_reward/std": 0.11036555469036102,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5750720500946045,
      "rewards/trajectory_similarity_reward/std": 0.3606364130973816,
      "step": 425
    },
    {
      "clip_ratio/high_max": 0.015990225947462022,
      "clip_ratio/high_mean": 0.0037280906399246305,
      "clip_ratio/low_mean": 0.0017869237999548204,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005515014257980511,
      "entropy": 0.31976134702563286,
      "epoch": 0.34722363728986244,
      "grad_norm": 19.826801300048828,
      "learning_rate": 5.222493887530562e-07,
      "loss": -2.574,
      "reward": 3.1267651319503784,
      "reward_std": 1.8421669602394104,
      "rewards/reasoning_density_reward/mean": 0.4312500059604645,
      "rewards/reasoning_density_reward/std": 0.09595610946416855,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.308667778968811,
      "rewards/trajectory_similarity_reward/mean": 0.5861400961875916,
      "rewards/trajectory_similarity_reward/std": 0.4406520426273346,
      "step": 426
    },
    {
      "clip_ratio/high_max": 0.013555828481912613,
      "clip_ratio/high_mean": 0.0032638570992276073,
      "clip_ratio/low_mean": 0.0006897184648551047,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003953575607738458,
      "entropy": 0.3704826161265373,
      "epoch": 0.34803871625063676,
      "grad_norm": 13.514006614685059,
      "learning_rate": 5.215973920130399e-07,
      "loss": 2.56,
      "reward": 1.8233089447021484,
      "reward_std": 1.4806487560272217,
      "rewards/reasoning_density_reward/mean": 0.31773313879966736,
      "rewards/reasoning_density_reward/std": 0.13105501234531403,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.1930757611989975,
      "rewards/trajectory_similarity_reward/std": 0.3166721761226654,
      "step": 427
    },
    {
      "clip_ratio/high_max": 0.013341696525458246,
      "clip_ratio/high_mean": 0.0028284748987061903,
      "clip_ratio/low_mean": 0.0011414044565754011,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003969879326177761,
      "entropy": 0.35706962272524834,
      "epoch": 0.34885379521141113,
      "grad_norm": 15.95551872253418,
      "learning_rate": 5.209453952730236e-07,
      "loss": -3.9333,
      "reward": 3.5770899057388306,
      "reward_std": 0.6639263331890106,
      "rewards/reasoning_density_reward/mean": 0.43359375,
      "rewards/reasoning_density_reward/std": 0.09557240828871727,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.6122462451457977,
      "rewards/trajectory_similarity_reward/std": 0.4398719370365143,
      "step": 428
    },
    {
      "clip_ratio/high_max": 0.008756200491916388,
      "clip_ratio/high_mean": 0.0015680952928960323,
      "clip_ratio/low_mean": 0.0007064541568979621,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022745493915863335,
      "entropy": 0.35371090844273567,
      "epoch": 0.34966887417218545,
      "grad_norm": 15.357527732849121,
      "learning_rate": 5.202933985330073e-07,
      "loss": 8.9279,
      "reward": 1.859299659729004,
      "reward_std": 1.9292092323303223,
      "rewards/reasoning_density_reward/mean": 0.3697172999382019,
      "rewards/reasoning_density_reward/std": 0.0997123271226883,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.2708321809768677,
      "rewards/trajectory_similarity_reward/std": 0.37892478704452515,
      "step": 429
    },
    {
      "clip_ratio/high_max": 0.012801278498955071,
      "clip_ratio/high_mean": 0.002203969721449539,
      "clip_ratio/low_mean": 0.00048115133176906966,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026851210532186087,
      "entropy": 0.3415617533028126,
      "epoch": 0.35048395313295977,
      "grad_norm": 19.16960906982422,
      "learning_rate": 5.196414017929911e-07,
      "loss": 10.3006,
      "reward": 3.3344225883483887,
      "reward_std": 1.2223055362701416,
      "rewards/reasoning_density_reward/mean": 0.43437498807907104,
      "rewards/reasoning_density_reward/std": 0.08670925348997116,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.650047779083252,
      "rewards/trajectory_similarity_reward/std": 0.4266659617424011,
      "step": 430
    },
    {
      "clip_ratio/high_max": 0.013768443139269948,
      "clip_ratio/high_mean": 0.0026482854736968875,
      "clip_ratio/low_mean": 0.0010100424187839963,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036583279725164175,
      "entropy": 0.33507661893963814,
      "epoch": 0.3512990320937341,
      "grad_norm": 16.013669967651367,
      "learning_rate": 5.189894050529747e-07,
      "loss": 12.5443,
      "reward": 3.3296566009521484,
      "reward_std": 0.7505826354026794,
      "rewards/reasoning_density_reward/mean": 0.38181817531585693,
      "rewards/reasoning_density_reward/std": 0.08370817452669144,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.4165883958339691,
      "rewards/trajectory_similarity_reward/std": 0.32862260937690735,
      "step": 431
    },
    {
      "clip_ratio/high_max": 0.00837332522496581,
      "clip_ratio/high_mean": 0.001818620236008428,
      "clip_ratio/low_mean": 0.001851868670200929,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003670488906209357,
      "entropy": 0.33187487721443176,
      "epoch": 0.3521141110545084,
      "grad_norm": 13.802213668823242,
      "learning_rate": 5.183374083129585e-07,
      "loss": 8.6576,
      "reward": 3.6796555519104004,
      "reward_std": 0.9240651726722717,
      "rewards/reasoning_density_reward/mean": 0.3867819905281067,
      "rewards/reasoning_density_reward/std": 0.14231576025485992,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8158584833145142,
      "rewards/trajectory_similarity_reward/mean": 0.5741235017776489,
      "rewards/trajectory_similarity_reward/std": 0.2967512607574463,
      "step": 432
    },
    {
      "clip_ratio/high_max": 0.017626409593503922,
      "clip_ratio/high_mean": 0.004831783444387838,
      "clip_ratio/low_mean": 0.0008694515890965704,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00570123502984643,
      "entropy": 0.35903899371623993,
      "epoch": 0.3529291900152827,
      "grad_norm": 12.66066837310791,
      "learning_rate": 5.176854115729421e-07,
      "loss": 15.8368,
      "reward": 3.0811877250671387,
      "reward_std": 2.0160703659057617,
      "rewards/reasoning_density_reward/mean": 0.41826921701431274,
      "rewards/reasoning_density_reward/std": 0.1236485093832016,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.6004184484481812,
      "rewards/trajectory_similarity_reward/std": 0.4502926468849182,
      "step": 433
    },
    {
      "clip_ratio/high_max": 0.01230120740365237,
      "clip_ratio/high_mean": 0.002872582153941039,
      "clip_ratio/low_mean": 0.000931221067730803,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038038031198084354,
      "entropy": 0.36335083097219467,
      "epoch": 0.35374426897605704,
      "grad_norm": 14.97573184967041,
      "learning_rate": 5.170334148329259e-07,
      "loss": 5.7115,
      "reward": 0.7344133853912354,
      "reward_std": 1.0763360261917114,
      "rewards/reasoning_density_reward/mean": 0.3031250238418579,
      "rewards/reasoning_density_reward/std": 0.06446898728609085,
      "rewards/success_reward/mean": 0.375,
      "rewards/success_reward/std": 0.866025447845459,
      "rewards/trajectory_similarity_reward/mean": 0.05628836527466774,
      "rewards/trajectory_similarity_reward/std": 0.22515346109867096,
      "step": 434
    },
    {
      "clip_ratio/high_max": 0.011520727188326418,
      "clip_ratio/high_mean": 0.0027582082548178732,
      "clip_ratio/low_mean": 0.0009817451500566676,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037399533321149647,
      "entropy": 0.2947052698582411,
      "epoch": 0.35455934793683136,
      "grad_norm": 12.89682388305664,
      "learning_rate": 5.163814180929095e-07,
      "loss": -3.3435,
      "reward": 3.9800859689712524,
      "reward_std": 0.8700380846858025,
      "rewards/reasoning_density_reward/mean": 0.4749999940395355,
      "rewards/reasoning_density_reward/std": 0.04472135752439499,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.6104557514190674,
      "rewards/trajectory_similarity_reward/mean": 0.8332107961177826,
      "rewards/trajectory_similarity_reward/std": 0.23363961838185787,
      "step": 435
    },
    {
      "clip_ratio/high_max": 0.013451404869556427,
      "clip_ratio/high_mean": 0.002756798596237786,
      "clip_ratio/low_mean": 0.0010435897784191184,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038003884328645654,
      "entropy": 0.3378818668425083,
      "epoch": 0.35537442689760573,
      "grad_norm": 13.009928703308105,
      "learning_rate": 5.157294213528932e-07,
      "loss": 11.5757,
      "reward": 2.0759682655334473,
      "reward_std": 1.604179859161377,
      "rewards/reasoning_density_reward/mean": 0.24895834922790527,
      "rewards/reasoning_density_reward/std": 0.1696913093328476,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4491376876831055,
      "rewards/trajectory_similarity_reward/mean": 0.32700982689857483,
      "rewards/trajectory_similarity_reward/std": 0.41038116812705994,
      "step": 436
    },
    {
      "clip_ratio/high_max": 0.014103942812653258,
      "clip_ratio/high_mean": 0.0028284729487495497,
      "clip_ratio/low_mean": 0.0010275505992467515,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003856023497064598,
      "entropy": 0.37653953582048416,
      "epoch": 0.35618950585838005,
      "grad_norm": 13.816670417785645,
      "learning_rate": 5.150774246128769e-07,
      "loss": -2.1749,
      "reward": 2.363271713256836,
      "reward_std": 1.4802814722061157,
      "rewards/reasoning_density_reward/mean": 0.27849704027175903,
      "rewards/reasoning_density_reward/std": 0.12495990842580795,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.125,
      "rewards/trajectory_similarity_reward/mean": 0.30352452397346497,
      "rewards/trajectory_similarity_reward/std": 0.42581596970558167,
      "step": 437
    },
    {
      "clip_ratio/high_max": 0.009617793723009527,
      "clip_ratio/high_mean": 0.0016410850075772032,
      "clip_ratio/low_mean": 0.0015289352741092443,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031700202816864476,
      "entropy": 0.3718842677772045,
      "epoch": 0.35700458481915437,
      "grad_norm": 10.909636497497559,
      "learning_rate": 5.144254278728606e-07,
      "loss": 6.944,
      "reward": 3.635477066040039,
      "reward_std": 1.3980504274368286,
      "rewards/reasoning_density_reward/mean": 0.453125,
      "rewards/reasoning_density_reward/std": 0.10562314838171005,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.6511019468307495,
      "rewards/trajectory_similarity_reward/std": 0.35027173161506653,
      "step": 438
    },
    {
      "clip_ratio/high_max": 0.014390083844773471,
      "clip_ratio/high_mean": 0.003529517911374569,
      "clip_ratio/low_mean": 0.00037003782927058637,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003899555711541325,
      "entropy": 0.3547911122441292,
      "epoch": 0.3578196637799287,
      "grad_norm": 13.960411071777344,
      "learning_rate": 5.137734311328442e-07,
      "loss": 2.6821,
      "reward": 2.869551420211792,
      "reward_std": 1.1689318418502808,
      "rewards/reasoning_density_reward/mean": 0.3806547522544861,
      "rewards/reasoning_density_reward/std": 0.09671033918857574,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 0.9437293410301208,
      "rewards/trajectory_similarity_reward/mean": 0.3326466977596283,
      "rewards/trajectory_similarity_reward/std": 0.3591921329498291,
      "step": 439
    },
    {
      "clip_ratio/high_max": 0.01021356403362006,
      "clip_ratio/high_mean": 0.0018447730399202555,
      "clip_ratio/low_mean": 0.001489924074121518,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003334697146783583,
      "entropy": 0.3350347466766834,
      "epoch": 0.358634742740703,
      "grad_norm": 16.358707427978516,
      "learning_rate": 5.13121434392828e-07,
      "loss": -5.3563,
      "reward": 2.4540815353393555,
      "reward_std": 0.2965747117996216,
      "rewards/reasoning_density_reward/mean": 0.4000000059604645,
      "rewards/reasoning_density_reward/std": 0.10327955335378647,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.46033143997192383,
      "rewards/trajectory_similarity_reward/std": 0.47736114263534546,
      "step": 440
    },
    {
      "clip_ratio/high_max": 0.010473738395376131,
      "clip_ratio/high_mean": 0.0017767819917935412,
      "clip_ratio/low_mean": 0.0007455251470673829,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025223072116205003,
      "entropy": 0.3298346772789955,
      "epoch": 0.3594498217014773,
      "grad_norm": 16.271299362182617,
      "learning_rate": 5.124694376528116e-07,
      "loss": -7.4457,
      "reward": 3.560398578643799,
      "reward_std": 1.097946286201477,
      "rewards/reasoning_density_reward/mean": 0.42786458134651184,
      "rewards/reasoning_density_reward/std": 0.09649791568517685,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.670820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.5075339674949646,
      "rewards/trajectory_similarity_reward/std": 0.35779330134391785,
      "step": 441
    },
    {
      "clip_ratio/high_max": 0.018899715621955693,
      "clip_ratio/high_mean": 0.004600309432134964,
      "clip_ratio/low_mean": 0.00035992956327390857,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0049602390063228086,
      "entropy": 0.3523045480251312,
      "epoch": 0.36026490066225164,
      "grad_norm": 14.733290672302246,
      "learning_rate": 5.118174409127954e-07,
      "loss": 0.1642,
      "reward": 3.221834182739258,
      "reward_std": 1.7312562465667725,
      "rewards/reasoning_density_reward/mean": 0.41804203391075134,
      "rewards/reasoning_density_reward/std": 0.08906347304582596,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.5537919998168945,
      "rewards/trajectory_similarity_reward/std": 0.40218988060951233,
      "step": 442
    },
    {
      "clip_ratio/high_max": 0.014236290589906275,
      "clip_ratio/high_mean": 0.003031516731425654,
      "clip_ratio/low_mean": 0.0002560410139267333,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032875577671802603,
      "entropy": 0.35012974217534065,
      "epoch": 0.36107997962302596,
      "grad_norm": 13.193225860595703,
      "learning_rate": 5.11165444172779e-07,
      "loss": 7.2646,
      "reward": 3.391960620880127,
      "reward_std": 1.6912016868591309,
      "rewards/reasoning_density_reward/mean": 0.39642855525016785,
      "rewards/reasoning_density_reward/std": 0.15755358338356018,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.6517819166183472,
      "rewards/trajectory_similarity_reward/std": 0.4007512629032135,
      "step": 443
    },
    {
      "clip_ratio/high_max": 0.013480794325005263,
      "clip_ratio/high_mean": 0.0021559581291512586,
      "clip_ratio/low_mean": 0.00011391948282835074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002269877615617588,
      "entropy": 0.35619455575942993,
      "epoch": 0.36189505858380033,
      "grad_norm": 17.399005889892578,
      "learning_rate": 5.105134474327628e-07,
      "loss": 19.5273,
      "reward": 3.370577573776245,
      "reward_std": 1.475117027759552,
      "rewards/reasoning_density_reward/mean": 0.4045587331056595,
      "rewards/reasoning_density_reward/std": 0.13322126865386963,
      "rewards/success_reward/mean": 2.390625,
      "rewards/success_reward/std": 1.1020839214324951,
      "rewards/trajectory_similarity_reward/mean": 0.5753938257694244,
      "rewards/trajectory_similarity_reward/std": 0.3748610317707062,
      "step": 444
    },
    {
      "clip_ratio/high_max": 0.01693464582785964,
      "clip_ratio/high_mean": 0.004525379714323208,
      "clip_ratio/low_mean": 0.0009916337075992487,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0055170134583022445,
      "entropy": 0.35768747702240944,
      "epoch": 0.36271013754457465,
      "grad_norm": 15.13335132598877,
      "learning_rate": 5.098614506927465e-07,
      "loss": 7.5439,
      "reward": 3.527996063232422,
      "reward_std": 1.0843853950500488,
      "rewards/reasoning_density_reward/mean": 0.4125000238418579,
      "rewards/reasoning_density_reward/std": 0.12974333763122559,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6779960989952087,
      "rewards/trajectory_similarity_reward/std": 0.37645605206489563,
      "step": 445
    },
    {
      "clip_ratio/high_max": 0.015719807415734977,
      "clip_ratio/high_mean": 0.0031536267488263547,
      "clip_ratio/low_mean": 0.0007503352244384587,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039039619732648134,
      "entropy": 0.3463652618229389,
      "epoch": 0.36352521650534897,
      "grad_norm": 16.655149459838867,
      "learning_rate": 5.092094539527302e-07,
      "loss": 1.4283,
      "reward": 3.5151896476745605,
      "reward_std": 1.1626205444335938,
      "rewards/reasoning_density_reward/mean": 0.4282737970352173,
      "rewards/reasoning_density_reward/std": 0.08221528679132462,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.866025447845459,
      "rewards/trajectory_similarity_reward/mean": 0.4619158208370209,
      "rewards/trajectory_similarity_reward/std": 0.29657304286956787,
      "step": 446
    },
    {
      "clip_ratio/high_max": 0.012756937241647393,
      "clip_ratio/high_mean": 0.003646124168881215,
      "clip_ratio/low_mean": 0.0007987049175426364,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004444829028216191,
      "entropy": 0.3349381610751152,
      "epoch": 0.3643402954661233,
      "grad_norm": 14.44282341003418,
      "learning_rate": 5.085574572127139e-07,
      "loss": 1.8969,
      "reward": 3.8367849588394165,
      "reward_std": 0.8198225796222687,
      "rewards/reasoning_density_reward/mean": 0.45569194853305817,
      "rewards/reasoning_density_reward/std": 0.08190494403243065,
      "rewards/success_reward/mean": 2.765625,
      "rewards/success_reward/std": 0.5585084557533264,
      "rewards/trajectory_similarity_reward/mean": 0.6154681742191315,
      "rewards/trajectory_similarity_reward/std": 0.3004613220691681,
      "step": 447
    },
    {
      "clip_ratio/high_max": 0.01615492580458522,
      "clip_ratio/high_mean": 0.00408941249770578,
      "clip_ratio/low_mean": 0.000593816330365371,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004683228762587532,
      "entropy": 0.2995867058634758,
      "epoch": 0.3651553744268976,
      "grad_norm": 16.350112915039062,
      "learning_rate": 5.079054604726976e-07,
      "loss": 6.739,
      "reward": 3.0377413034439087,
      "reward_std": 1.461262285709381,
      "rewards/reasoning_density_reward/mean": 0.42500001192092896,
      "rewards/reasoning_density_reward/std": 0.10000000149011612,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.5502412617206573,
      "rewards/trajectory_similarity_reward/std": 0.4467519074678421,
      "step": 448
    },
    {
      "clip_ratio/high_max": 0.018074216321110725,
      "clip_ratio/high_mean": 0.0033549441141076386,
      "clip_ratio/low_mean": 0.0008262863120762631,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004181230455287732,
      "entropy": 0.32539360225200653,
      "epoch": 0.3659704533876719,
      "grad_norm": 13.567138671875,
      "learning_rate": 5.072534637326813e-07,
      "loss": 5.5459,
      "reward": 3.9046804904937744,
      "reward_std": 0.9940285682678223,
      "rewards/reasoning_density_reward/mean": 0.47232142090797424,
      "rewards/reasoning_density_reward/std": 0.07603592425584793,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.8073588609695435,
      "rewards/trajectory_similarity_reward/std": 0.3226054608821869,
      "step": 449
    },
    {
      "clip_ratio/high_max": 0.014277018141001463,
      "clip_ratio/high_mean": 0.004470846732147038,
      "clip_ratio/low_mean": 0.0010230026600765996,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005493849399499595,
      "entropy": 0.3605445884168148,
      "epoch": 0.36678553234844624,
      "grad_norm": 14.29770565032959,
      "learning_rate": 5.06601466992665e-07,
      "loss": 11.6442,
      "reward": 4.1000165939331055,
      "reward_std": 0.6011880934238434,
      "rewards/reasoning_density_reward/mean": 0.4567708373069763,
      "rewards/reasoning_density_reward/std": 0.11589448526501656,
      "rewards/success_reward/mean": 2.859375,
      "rewards/success_reward/std": 0.4436737895011902,
      "rewards/trajectory_similarity_reward/mean": 0.7838707566261292,
      "rewards/trajectory_similarity_reward/std": 0.29585686326026917,
      "step": 450
    },
    {
      "clip_ratio/high_max": 0.017386014806106687,
      "clip_ratio/high_mean": 0.004146655352087691,
      "clip_ratio/low_mean": 0.0013268925322336145,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005473547906149179,
      "entropy": 0.3646089732646942,
      "epoch": 0.36760061130922056,
      "grad_norm": 22.88274574279785,
      "learning_rate": 5.059494702526487e-07,
      "loss": 8.1348,
      "reward": 3.1531635522842407,
      "reward_std": 1.2951025366783142,
      "rewards/reasoning_density_reward/mean": 0.40312500298023224,
      "rewards/reasoning_density_reward/std": 0.11610807850956917,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 0.887150913476944,
      "rewards/trajectory_similarity_reward/mean": 0.5000386238098145,
      "rewards/trajectory_similarity_reward/std": 0.39717090129852295,
      "step": 451
    },
    {
      "clip_ratio/high_max": 0.01215717988088727,
      "clip_ratio/high_mean": 0.0025412592076463625,
      "clip_ratio/low_mean": 0.0015020245700725354,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004043283734063152,
      "entropy": 0.34296469762921333,
      "epoch": 0.36841569026999493,
      "grad_norm": 15.33232307434082,
      "learning_rate": 5.052974735126324e-07,
      "loss": 0.606,
      "reward": 2.6935296058654785,
      "reward_std": 1.5624797344207764,
      "rewards/reasoning_density_reward/mean": 0.3916666507720947,
      "rewards/reasoning_density_reward/std": 0.0992564931511879,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.4268629550933838,
      "rewards/trajectory_similarity_reward/std": 0.39323240518569946,
      "step": 452
    },
    {
      "clip_ratio/high_max": 0.014375988277606666,
      "clip_ratio/high_mean": 0.003515842108754441,
      "clip_ratio/low_mean": 0.0008386515983147547,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004354493779828772,
      "entropy": 0.3361155055463314,
      "epoch": 0.36923076923076925,
      "grad_norm": 21.11935806274414,
      "learning_rate": 5.046454767726161e-07,
      "loss": -10.5887,
      "reward": 2.9657286405563354,
      "reward_std": 1.5702468156814575,
      "rewards/reasoning_density_reward/mean": 0.41979166865348816,
      "rewards/reasoning_density_reward/std": 0.10017330944538116,
      "rewards/success_reward/mean": 2.015625,
      "rewards/success_reward/std": 1.3736594319343567,
      "rewards/trajectory_similarity_reward/mean": 0.5303119421005249,
      "rewards/trajectory_similarity_reward/std": 0.4267829954624176,
      "step": 453
    },
    {
      "clip_ratio/high_max": 0.016019224014598876,
      "clip_ratio/high_mean": 0.003149294134345837,
      "clip_ratio/low_mean": 0.0006764979880244937,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038257920678006485,
      "entropy": 0.356791615486145,
      "epoch": 0.37004584819154357,
      "grad_norm": 18.270238876342773,
      "learning_rate": 5.039934800325998e-07,
      "loss": 3.4092,
      "reward": 2.069830894470215,
      "reward_std": 1.6504545211791992,
      "rewards/reasoning_density_reward/mean": 0.33690476417541504,
      "rewards/reasoning_density_reward/std": 0.07355562597513199,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.42042601108551025,
      "rewards/trajectory_similarity_reward/std": 0.4950962960720062,
      "step": 454
    },
    {
      "clip_ratio/high_max": 0.010287661018082872,
      "clip_ratio/high_mean": 0.0023239691108756233,
      "clip_ratio/low_mean": 0.0011386579935788177,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034626270571607165,
      "entropy": 0.380761057138443,
      "epoch": 0.3708609271523179,
      "grad_norm": 11.042998313903809,
      "learning_rate": 5.033414832925835e-07,
      "loss": -9.6461,
      "reward": 2.8342089653015137,
      "reward_std": 0.7279172837734222,
      "rewards/reasoning_density_reward/mean": 0.31047388911247253,
      "rewards/reasoning_density_reward/std": 0.21406981348991394,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3620079159736633,
      "rewards/trajectory_similarity_reward/mean": 0.5549850463867188,
      "rewards/trajectory_similarity_reward/std": 0.49791738390922546,
      "step": 455
    },
    {
      "clip_ratio/high_max": 0.011750632256735116,
      "clip_ratio/high_mean": 0.0024727883937885053,
      "clip_ratio/low_mean": 0.0005106340286147315,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029834224114893004,
      "entropy": 0.35264209657907486,
      "epoch": 0.3716760061130922,
      "grad_norm": 14.226058959960938,
      "learning_rate": 5.026894865525672e-07,
      "loss": 9.9543,
      "reward": 3.9917964935302734,
      "reward_std": 0.5351912379264832,
      "rewards/reasoning_density_reward/mean": 0.4556547701358795,
      "rewards/reasoning_density_reward/std": 0.0803002417087555,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.5123475790023804,
      "rewards/trajectory_similarity_reward/mean": 0.7236416339874268,
      "rewards/trajectory_similarity_reward/std": 0.3563465476036072,
      "step": 456
    },
    {
      "clip_ratio/high_max": 0.015502845024457201,
      "clip_ratio/high_mean": 0.002856289367628051,
      "clip_ratio/low_mean": 0.0008246269135270268,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036809162702411413,
      "entropy": 0.3474475219845772,
      "epoch": 0.3724910850738665,
      "grad_norm": 14.30775260925293,
      "learning_rate": 5.020374898125509e-07,
      "loss": 3.9667,
      "reward": 2.9677820205688477,
      "reward_std": 1.6293488144874573,
      "rewards/reasoning_density_reward/mean": 0.3594494014978409,
      "rewards/reasoning_density_reward/std": 0.16046832501888275,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.2672818303108215,
      "rewards/trajectory_similarity_reward/mean": 0.49895772337913513,
      "rewards/trajectory_similarity_reward/std": 0.4219892770051956,
      "step": 457
    },
    {
      "clip_ratio/high_max": 0.01121936904382892,
      "clip_ratio/high_mean": 0.002499936021195026,
      "clip_ratio/low_mean": 0.0018755692181002814,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004375505275675096,
      "entropy": 0.3633809760212898,
      "epoch": 0.37330616403464084,
      "grad_norm": 18.578161239624023,
      "learning_rate": 5.013854930725346e-07,
      "loss": 5.8895,
      "reward": 1.6777474880218506,
      "reward_std": 1.6945998668670654,
      "rewards/reasoning_density_reward/mean": 0.2790178656578064,
      "rewards/reasoning_density_reward/std": 0.18065394461154938,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.2737295925617218,
      "rewards/trajectory_similarity_reward/std": 0.42031019926071167,
      "step": 458
    },
    {
      "clip_ratio/high_max": 0.015746321296319366,
      "clip_ratio/high_mean": 0.0035939000954385847,
      "clip_ratio/low_mean": 0.0008543201256543398,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004448220250196755,
      "entropy": 0.3342536427080631,
      "epoch": 0.37412124299541516,
      "grad_norm": 13.967423439025879,
      "learning_rate": 5.007334963325183e-07,
      "loss": 1.0607,
      "reward": 3.4557063579559326,
      "reward_std": 1.489150881767273,
      "rewards/reasoning_density_reward/mean": 0.4244047701358795,
      "rewards/reasoning_density_reward/std": 0.09192449599504471,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.5938016176223755,
      "rewards/trajectory_similarity_reward/std": 0.3688480854034424,
      "step": 459
    },
    {
      "clip_ratio/high_max": 0.015025097352918237,
      "clip_ratio/high_mean": 0.0033085328541346826,
      "clip_ratio/low_mean": 0.001332183506747242,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004640716317226179,
      "entropy": 0.3201288916170597,
      "epoch": 0.37493632195618953,
      "grad_norm": 14.060038566589355,
      "learning_rate": 5.000814995925021e-07,
      "loss": 5.2675,
      "reward": 3.095360517501831,
      "reward_std": 0.558190643787384,
      "rewards/reasoning_density_reward/mean": 0.35729169845581055,
      "rewards/reasoning_density_reward/std": 0.0778694897890091,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 0.7685213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.39431899785995483,
      "rewards/trajectory_similarity_reward/std": 0.39283591508865356,
      "step": 460
    },
    {
      "clip_ratio/high_max": 0.007601545832585543,
      "clip_ratio/high_mean": 0.0015632414797437377,
      "clip_ratio/low_mean": 0.0012297478970140219,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002792989409499569,
      "entropy": 0.33973727747797966,
      "epoch": 0.37575140091696385,
      "grad_norm": 13.638519287109375,
      "learning_rate": 4.994295028524857e-07,
      "loss": -1.2902,
      "reward": 2.2552289962768555,
      "reward_std": 1.2646613121032715,
      "rewards/reasoning_density_reward/mean": 0.3586805760860443,
      "rewards/reasoning_density_reward/std": 0.08063453435897827,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.30279844999313354,
      "rewards/trajectory_similarity_reward/std": 0.3704984784126282,
      "step": 461
    },
    {
      "clip_ratio/high_max": 0.010259742906782776,
      "clip_ratio/high_mean": 0.0020315194997237995,
      "clip_ratio/low_mean": 0.0016230941000685561,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036546136252582073,
      "entropy": 0.3663959987461567,
      "epoch": 0.37656647987773817,
      "grad_norm": 14.480891227722168,
      "learning_rate": 4.987775061124695e-07,
      "loss": -8.4312,
      "reward": 2.860720634460449,
      "reward_std": 1.811141848564148,
      "rewards/reasoning_density_reward/mean": 0.3425843119621277,
      "rewards/reasoning_density_reward/std": 0.10107714682817459,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.4556361734867096,
      "rewards/trajectory_similarity_reward/std": 0.3905816674232483,
      "step": 462
    },
    {
      "clip_ratio/high_max": 0.020169397233985364,
      "clip_ratio/high_mean": 0.005022701050620526,
      "clip_ratio/low_mean": 0.000425591704697581,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005448292737128213,
      "entropy": 0.37075917795300484,
      "epoch": 0.3773815588385125,
      "grad_norm": 13.623260498046875,
      "learning_rate": 4.981255093724531e-07,
      "loss": 6.123,
      "reward": 3.525747776031494,
      "reward_std": 1.7327780723571777,
      "rewards/reasoning_density_reward/mean": 0.35865575075149536,
      "rewards/reasoning_density_reward/std": 0.18906356394290924,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.7295923233032227,
      "rewards/trajectory_similarity_reward/std": 0.3802027106285095,
      "step": 463
    },
    {
      "clip_ratio/high_max": 0.010869227757211775,
      "clip_ratio/high_mean": 0.0028955036541447043,
      "clip_ratio/low_mean": 0.00042479309195186943,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003320296855235938,
      "entropy": 0.35434890165925026,
      "epoch": 0.3781966377992868,
      "grad_norm": 12.067054748535156,
      "learning_rate": 4.974735126324369e-07,
      "loss": 5.2599,
      "reward": 1.7668181657791138,
      "reward_std": 0.9604188799858093,
      "rewards/reasoning_density_reward/mean": 0.32230114936828613,
      "rewards/reasoning_density_reward/std": 0.14030225574970245,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.31951725482940674,
      "rewards/trajectory_similarity_reward/std": 0.4325644075870514,
      "step": 464
    },
    {
      "clip_ratio/high_max": 0.010628973162965849,
      "clip_ratio/high_mean": 0.0019038558202737477,
      "clip_ratio/low_mean": 0.0004041413394588744,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023079971579136327,
      "entropy": 0.37616974860429764,
      "epoch": 0.3790117167600611,
      "grad_norm": 12.440718650817871,
      "learning_rate": 4.968215158924205e-07,
      "loss": -2.2839,
      "reward": 3.530062437057495,
      "reward_std": 0.813005119562149,
      "rewards/reasoning_density_reward/mean": 0.37252604961395264,
      "rewards/reasoning_density_reward/std": 0.17561601102352142,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0183849930763245,
      "rewards/trajectory_similarity_reward/mean": 0.6262864470481873,
      "rewards/trajectory_similarity_reward/std": 0.4095676988363266,
      "step": 465
    },
    {
      "clip_ratio/high_max": 0.012891936697997153,
      "clip_ratio/high_mean": 0.0041897857445292175,
      "clip_ratio/low_mean": 0.0005909209576202556,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004780706745805219,
      "entropy": 0.33172065019607544,
      "epoch": 0.37982679572083544,
      "grad_norm": 13.740821838378906,
      "learning_rate": 4.961695191524043e-07,
      "loss": 14.3011,
      "reward": 3.5464842319488525,
      "reward_std": 1.7280243635177612,
      "rewards/reasoning_density_reward/mean": 0.39531248807907104,
      "rewards/reasoning_density_reward/std": 0.18307985365390778,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.7136717438697815,
      "rewards/trajectory_similarity_reward/std": 0.3641922175884247,
      "step": 466
    },
    {
      "clip_ratio/high_max": 0.013971847190987319,
      "clip_ratio/high_mean": 0.002811038968502544,
      "clip_ratio/low_mean": 0.00046439794823527336,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003275436960393563,
      "entropy": 0.29960139468312263,
      "epoch": 0.38064187468160976,
      "grad_norm": 13.534332275390625,
      "learning_rate": 4.955175224123879e-07,
      "loss": 14.1833,
      "reward": 3.3140668869018555,
      "reward_std": 1.3521229326725006,
      "rewards/reasoning_density_reward/mean": 0.37207961082458496,
      "rewards/reasoning_density_reward/std": 0.15196441859006882,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0534805357456207,
      "rewards/trajectory_similarity_reward/mean": 0.5982371866703033,
      "rewards/trajectory_similarity_reward/std": 0.40324388444423676,
      "step": 467
    },
    {
      "clip_ratio/high_max": 0.01050832457258366,
      "clip_ratio/high_mean": 0.0018309534425497986,
      "clip_ratio/low_mean": 0.0010748409731604625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002905794397520367,
      "entropy": 0.3675881177186966,
      "epoch": 0.38145695364238413,
      "grad_norm": 14.082103729248047,
      "learning_rate": 4.948655256723716e-07,
      "loss": 14.5022,
      "reward": 1.9972357749938965,
      "reward_std": 0.5085702538490295,
      "rewards/reasoning_density_reward/mean": 0.2725570499897003,
      "rewards/reasoning_density_reward/std": 0.19556762278079987,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.3184286653995514,
      "rewards/trajectory_similarity_reward/std": 0.3804556429386139,
      "step": 468
    },
    {
      "clip_ratio/high_max": 0.013333464215975255,
      "clip_ratio/high_mean": 0.0023877890707808547,
      "clip_ratio/low_mean": 0.00098134011568618,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003369129146449268,
      "entropy": 0.384434599429369,
      "epoch": 0.38227203260315845,
      "grad_norm": 10.689541816711426,
      "learning_rate": 4.942135289323553e-07,
      "loss": 10.7743,
      "reward": 2.3426883220672607,
      "reward_std": 0.5961531400680542,
      "rewards/reasoning_density_reward/mean": 0.28125,
      "rewards/reasoning_density_reward/std": 0.1911587417125702,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.37393811345100403,
      "rewards/trajectory_similarity_reward/std": 0.438458114862442,
      "step": 469
    },
    {
      "clip_ratio/high_max": 0.01457225129706785,
      "clip_ratio/high_mean": 0.002978812248329632,
      "clip_ratio/low_mean": 0.0012697187703452073,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042485311278142035,
      "entropy": 0.34139225631952286,
      "epoch": 0.38308711156393277,
      "grad_norm": 12.41848087310791,
      "learning_rate": 4.93561532192339e-07,
      "loss": 8.105,
      "reward": 3.1796627044677734,
      "reward_std": 1.2831989526748657,
      "rewards/reasoning_density_reward/mean": 0.3828125,
      "rewards/reasoning_density_reward/std": 0.16898556053638458,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5468502044677734,
      "rewards/trajectory_similarity_reward/std": 0.37848517298698425,
      "step": 470
    },
    {
      "clip_ratio/high_max": 0.008787360566202551,
      "clip_ratio/high_mean": 0.0022288494437816553,
      "clip_ratio/low_mean": 0.0011796417602454312,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003408491196751129,
      "entropy": 0.3484883941709995,
      "epoch": 0.3839021905247071,
      "grad_norm": 13.481637001037598,
      "learning_rate": 4.929095354523226e-07,
      "loss": 17.1927,
      "reward": 3.227114677429199,
      "reward_std": 1.6701622009277344,
      "rewards/reasoning_density_reward/mean": 0.41675347089767456,
      "rewards/reasoning_density_reward/std": 0.11137522757053375,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5603610277175903,
      "rewards/trajectory_similarity_reward/std": 0.4394489526748657,
      "step": 471
    },
    {
      "clip_ratio/high_max": 0.010261313524097204,
      "clip_ratio/high_mean": 0.0025162681413348764,
      "clip_ratio/low_mean": 0.000646251741272863,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003162519948091358,
      "entropy": 0.3414890989661217,
      "epoch": 0.3847172694854814,
      "grad_norm": 15.081849098205566,
      "learning_rate": 4.922575387123064e-07,
      "loss": 3.883,
      "reward": 3.5143582820892334,
      "reward_std": 1.516853928565979,
      "rewards/reasoning_density_reward/mean": 0.42452652752399445,
      "rewards/reasoning_density_reward/std": 0.10196412727236748,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.003450870513916,
      "rewards/trajectory_similarity_reward/mean": 0.6523316502571106,
      "rewards/trajectory_similarity_reward/std": 0.43389151990413666,
      "step": 472
    },
    {
      "clip_ratio/high_max": 0.011484506714623421,
      "clip_ratio/high_mean": 0.0026691384919104166,
      "clip_ratio/low_mean": 0.0007856255979277194,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034547641917015426,
      "entropy": 0.3635590150952339,
      "epoch": 0.3855323484462557,
      "grad_norm": 14.765542984008789,
      "learning_rate": 4.9160554197229e-07,
      "loss": 13.63,
      "reward": 1.7569265365600586,
      "reward_std": 1.65617036819458,
      "rewards/reasoning_density_reward/mean": 0.31030505895614624,
      "rewards/reasoning_density_reward/std": 0.14066819846630096,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.2278713881969452,
      "rewards/trajectory_similarity_reward/std": 0.34066668152809143,
      "step": 473
    },
    {
      "clip_ratio/high_max": 0.012734661810100079,
      "clip_ratio/high_mean": 0.002678433360415511,
      "clip_ratio/low_mean": 0.0009144859432126395,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035929192963521928,
      "entropy": 0.3386392146348953,
      "epoch": 0.38634742740703004,
      "grad_norm": 17.49725914001465,
      "learning_rate": 4.909535452322738e-07,
      "loss": 10.8022,
      "reward": 2.1321335434913635,
      "reward_std": 1.8206160068511963,
      "rewards/reasoning_density_reward/mean": 0.3538690507411957,
      "rewards/reasoning_density_reward/std": 0.11639998108148575,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.4010097980499268,
      "rewards/trajectory_similarity_reward/mean": 0.32513948529958725,
      "rewards/trajectory_similarity_reward/std": 0.42003268003463745,
      "step": 474
    },
    {
      "clip_ratio/high_max": 0.014631036203354597,
      "clip_ratio/high_mean": 0.0034588017224450596,
      "clip_ratio/low_mean": 0.0012333783161011524,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0046921800676500425,
      "entropy": 0.3476466238498688,
      "epoch": 0.38716250636780436,
      "grad_norm": 11.962162971496582,
      "learning_rate": 4.903015484922575e-07,
      "loss": -9.7706,
      "reward": 2.0601162910461426,
      "reward_std": 1.6771550178527832,
      "rewards/reasoning_density_reward/mean": 0.34732142090797424,
      "rewards/reasoning_density_reward/std": 0.15469688177108765,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.40029484033584595,
      "rewards/trajectory_similarity_reward/std": 0.4724109470844269,
      "step": 475
    },
    {
      "clip_ratio/high_max": 0.012624181166756898,
      "clip_ratio/high_mean": 0.003079300186072942,
      "clip_ratio/low_mean": 0.0008828078753140289,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003962108006817289,
      "entropy": 0.3785488121211529,
      "epoch": 0.38797758532857873,
      "grad_norm": 12.672492980957031,
      "learning_rate": 4.896495517522412e-07,
      "loss": 3.7542,
      "step": 476
    },
    {
      "clip_ratio/high_max": 0.018854427034966648,
      "clip_ratio/high_mean": 0.004005238821264356,
      "clip_ratio/low_mean": 0.00037657546636182815,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00438181433128193,
      "entropy": 0.33739086613059044,
      "epoch": 0.38879266428935305,
      "grad_norm": 14.396166801452637,
      "learning_rate": 4.889975550122249e-07,
      "loss": 12.6224,
      "reward": 3.3972305059432983,
      "reward_std": 1.3170840740203857,
      "rewards/reasoning_density_reward/mean": 0.4033482074737549,
      "rewards/reasoning_density_reward/std": 0.1414262279868126,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2304179072380066,
      "rewards/trajectory_similarity_reward/mean": 0.6501322090625763,
      "rewards/trajectory_similarity_reward/std": 0.3593132048845291,
      "step": 477
    },
    {
      "clip_ratio/high_max": 0.011949582898523659,
      "clip_ratio/high_mean": 0.0027019541084882803,
      "clip_ratio/low_mean": 0.0006159931181173306,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033179472520714626,
      "entropy": 0.33380117639899254,
      "epoch": 0.38960774325012737,
      "grad_norm": 15.95910930633545,
      "learning_rate": 4.883455582722086e-07,
      "loss": 11.9054,
      "reward": 2.5599923133850098,
      "reward_std": 0.9863919019699097,
      "rewards/reasoning_density_reward/mean": 0.41458335518836975,
      "rewards/reasoning_density_reward/std": 0.10034661740064621,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.27040916681289673,
      "rewards/trajectory_similarity_reward/std": 0.28621557354927063,
      "step": 478
    },
    {
      "clip_ratio/high_max": 0.014735537697561085,
      "clip_ratio/high_mean": 0.0031317793036578223,
      "clip_ratio/low_mean": 0.0005934365435678046,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003725215879967436,
      "entropy": 0.39829452708363533,
      "epoch": 0.3904228222109017,
      "grad_norm": 13.181927680969238,
      "learning_rate": 4.876935615321923e-07,
      "loss": 10.459,
      "reward": 3.000272035598755,
      "reward_std": 1.8360852003097534,
      "rewards/reasoning_density_reward/mean": 0.3321428596973419,
      "rewards/reasoning_density_reward/std": 0.20560917258262634,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.6056290864944458,
      "rewards/trajectory_similarity_reward/std": 0.43104928731918335,
      "step": 479
    },
    {
      "clip_ratio/high_max": 0.01049133948981762,
      "clip_ratio/high_mean": 0.0029120522376615554,
      "clip_ratio/low_mean": 0.00039119250868679956,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003303244724520482,
      "entropy": 0.35594894737005234,
      "epoch": 0.391237901171676,
      "grad_norm": 14.371100425720215,
      "learning_rate": 4.87041564792176e-07,
      "loss": -1.3094,
      "reward": 3.3211727142333984,
      "reward_std": 1.5434036254882812,
      "rewards/reasoning_density_reward/mean": 0.4145089238882065,
      "rewards/reasoning_density_reward/std": 0.09804923087358475,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.1514689326286316,
      "rewards/trajectory_similarity_reward/mean": 0.562913715839386,
      "rewards/trajectory_similarity_reward/std": 0.371689036488533,
      "step": 480
    },
    {
      "clip_ratio/high_max": 0.02468110283371061,
      "clip_ratio/high_mean": 0.006917935505043715,
      "clip_ratio/low_mean": 0.0010433810275571886,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007961316558066756,
      "entropy": 0.3520287722349167,
      "epoch": 0.3920529801324503,
      "grad_norm": 14.618302345275879,
      "learning_rate": 4.863895680521597e-07,
      "loss": 2.3077,
      "reward": 3.964263439178467,
      "reward_std": 0.739411473274231,
      "rewards/reasoning_density_reward/mean": 0.4625000059604645,
      "rewards/reasoning_density_reward/std": 0.08062257617712021,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.7830134034156799,
      "rewards/trajectory_similarity_reward/std": 0.3940838575363159,
      "step": 481
    },
    {
      "clip_ratio/high_max": 0.011063479934819043,
      "clip_ratio/high_mean": 0.002974166229250841,
      "clip_ratio/low_mean": 0.0009101571558858268,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038843235233798623,
      "entropy": 0.32974400371313095,
      "epoch": 0.39286805909322464,
      "grad_norm": 13.311914443969727,
      "learning_rate": 4.857375713121434e-07,
      "loss": 3.8686,
      "reward": 3.22121262550354,
      "reward_std": 1.3097589612007141,
      "rewards/reasoning_density_reward/mean": 0.4310096204280853,
      "rewards/reasoning_density_reward/std": 0.10827752202749252,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3546693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.6339532136917114,
      "rewards/trajectory_similarity_reward/std": 0.40220673382282257,
      "step": 482
    },
    {
      "clip_ratio/high_max": 0.010610443830955774,
      "clip_ratio/high_mean": 0.002354574688069988,
      "clip_ratio/low_mean": 0.0010656527992978226,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034202274109702557,
      "entropy": 0.3216451108455658,
      "epoch": 0.39368313805399896,
      "grad_norm": 12.532089233398438,
      "learning_rate": 4.850855745721271e-07,
      "loss": -5.6345,
      "reward": 3.6167731285095215,
      "reward_std": 1.3510048389434814,
      "rewards/reasoning_density_reward/mean": 0.4392856955528259,
      "rewards/reasoning_density_reward/std": 0.08798576891422272,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.9031196236610413,
      "rewards/trajectory_similarity_reward/mean": 0.6462373733520508,
      "rewards/trajectory_similarity_reward/std": 0.41103488206863403,
      "step": 483
    },
    {
      "clip_ratio/high_max": 0.016385454684495926,
      "clip_ratio/high_mean": 0.0040272485348396,
      "clip_ratio/low_mean": 0.000509477176819928,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004536725697107613,
      "entropy": 0.33791616186499596,
      "epoch": 0.39449821701477333,
      "grad_norm": 19.2721004486084,
      "learning_rate": 4.844335778321108e-07,
      "loss": 8.0805,
      "reward": 3.4442925453186035,
      "reward_std": 1.5424306392669678,
      "rewards/reasoning_density_reward/mean": 0.3812499940395355,
      "rewards/reasoning_density_reward/std": 0.1509077548980713,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.625542402267456,
      "rewards/trajectory_similarity_reward/std": 0.40579190850257874,
      "step": 484
    },
    {
      "clip_ratio/high_max": 0.011729467601981014,
      "clip_ratio/high_mean": 0.002674936084076762,
      "clip_ratio/low_mean": 0.0002960944730148185,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002971030553453602,
      "entropy": 0.3649710938334465,
      "epoch": 0.39531329597554765,
      "grad_norm": 14.512635231018066,
      "learning_rate": 4.837815810920945e-07,
      "loss": 7.877,
      "reward": 2.3704028129577637,
      "reward_std": 1.2274246215820312,
      "rewards/reasoning_density_reward/mean": 0.25580355525016785,
      "rewards/reasoning_density_reward/std": 0.2163899689912796,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.125,
      "rewards/trajectory_similarity_reward/mean": 0.3333492875099182,
      "rewards/trajectory_similarity_reward/std": 0.44917502999305725,
      "step": 485
    },
    {
      "clip_ratio/high_max": 0.0080645197595004,
      "clip_ratio/high_mean": 0.0021886271752009634,
      "clip_ratio/low_mean": 0.0006559529174410272,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002844580179953482,
      "entropy": 0.3525277338922024,
      "epoch": 0.39612837493632197,
      "grad_norm": 11.41243839263916,
      "learning_rate": 4.831295843520782e-07,
      "loss": -4.1236,
      "reward": 1.5394238233566284,
      "reward_std": 1.742397427558899,
      "rewards/reasoning_density_reward/mean": 0.3308035731315613,
      "rewards/reasoning_density_reward/std": 0.12739458680152893,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.27112022042274475,
      "rewards/trajectory_similarity_reward/std": 0.4425753355026245,
      "step": 486
    },
    {
      "clip_ratio/high_max": 0.010421994549687952,
      "clip_ratio/high_mean": 0.002479012415278703,
      "clip_ratio/low_mean": 0.0011257119185756892,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003604724319302477,
      "entropy": 0.34504299983382225,
      "epoch": 0.3969434538970963,
      "grad_norm": 15.816049575805664,
      "learning_rate": 4.824775876120619e-07,
      "loss": 10.2945,
      "reward": 2.1173031330108643,
      "reward_std": 0.8171879053115845,
      "rewards/reasoning_density_reward/mean": 0.3552083373069763,
      "rewards/reasoning_density_reward/std": 0.14077214896678925,
      "rewards/success_reward/mean": 1.359375,
      "rewards/success_reward/std": 1.4665062427520752,
      "rewards/trajectory_similarity_reward/mean": 0.40271981060504913,
      "rewards/trajectory_similarity_reward/std": 0.5018279105424881,
      "step": 487
    },
    {
      "clip_ratio/high_max": 0.009501054359134287,
      "clip_ratio/high_mean": 0.0020637515699490905,
      "clip_ratio/low_mean": 0.0006322393383015878,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002695990915526636,
      "entropy": 0.3269799053668976,
      "epoch": 0.3977585328578706,
      "grad_norm": 15.600996971130371,
      "learning_rate": 4.818255908720457e-07,
      "loss": -0.466,
      "reward": 3.0882365703582764,
      "reward_std": 1.3919687271118164,
      "rewards/reasoning_density_reward/mean": 0.39905303716659546,
      "rewards/reasoning_density_reward/std": 0.11531338095664978,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.4391835033893585,
      "rewards/trajectory_similarity_reward/std": 0.3934021294116974,
      "step": 488
    },
    {
      "clip_ratio/high_max": 0.010285026102792472,
      "clip_ratio/high_mean": 0.002734215115197003,
      "clip_ratio/low_mean": 0.00045202015462564304,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031862352625466883,
      "entropy": 0.3798019401729107,
      "epoch": 0.3985736118186449,
      "grad_norm": 15.020301818847656,
      "learning_rate": 4.811735941320293e-07,
      "loss": -7.9619,
      "reward": 1.8539007902145386,
      "reward_std": 0.406984806060791,
      "rewards/reasoning_density_reward/mean": 0.19583332538604736,
      "rewards/reasoning_density_reward/std": 0.23629078269004822,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.25181740522384644,
      "rewards/trajectory_similarity_reward/std": 0.305757999420166,
      "step": 489
    },
    {
      "clip_ratio/high_max": 0.004745533224195242,
      "clip_ratio/high_mean": 0.0011992491345154122,
      "clip_ratio/low_mean": 0.00047127827201620676,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016705274138075765,
      "entropy": 0.35172802582383156,
      "epoch": 0.39938869077941924,
      "grad_norm": 13.86829948425293,
      "learning_rate": 4.805215973920131e-07,
      "loss": 12.1354,
      "reward": 1.9644945859909058,
      "reward_std": 1.9971860647201538,
      "rewards/reasoning_density_reward/mean": 0.31160712242126465,
      "rewards/reasoning_density_reward/std": 0.18530862033367157,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.3403873145580292,
      "rewards/trajectory_similarity_reward/std": 0.401991069316864,
      "step": 490
    },
    {
      "clip_ratio/high_max": 0.0066764490329660475,
      "clip_ratio/high_mean": 0.0016106557377497666,
      "clip_ratio/low_mean": 0.0011010901835106779,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027117458812426776,
      "entropy": 0.38567282631993294,
      "epoch": 0.40020376974019356,
      "grad_norm": 13.033806800842285,
      "learning_rate": 4.798696006519967e-07,
      "loss": 13.8831,
      "reward": 3.227487564086914,
      "reward_std": 1.17528235912323,
      "rewards/reasoning_density_reward/mean": 0.296875,
      "rewards/reasoning_density_reward/std": 0.22983236610889435,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.49311256408691406,
      "rewards/trajectory_similarity_reward/std": 0.4367295801639557,
      "step": 491
    },
    {
      "clip_ratio/high_max": 0.011616300558671355,
      "clip_ratio/high_mean": 0.00251930917147547,
      "clip_ratio/low_mean": 0.0005029741369071417,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030222833884181455,
      "entropy": 0.379605732858181,
      "epoch": 0.40101884870096793,
      "grad_norm": 13.431092262268066,
      "learning_rate": 4.792176039119805e-07,
      "loss": 3.6588,
      "reward": 1.7808400392532349,
      "reward_std": 1.6532752513885498,
      "rewards/reasoning_density_reward/mean": 0.2595841884613037,
      "rewards/reasoning_density_reward/std": 0.15857787430286407,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.20875580608844757,
      "rewards/trajectory_similarity_reward/std": 0.33775827288627625,
      "step": 492
    },
    {
      "clip_ratio/high_max": 0.009293011506088078,
      "clip_ratio/high_mean": 0.0020591338688973337,
      "clip_ratio/low_mean": 0.0007191354561655317,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002778269350528717,
      "entropy": 0.3478131555020809,
      "epoch": 0.40183392766174225,
      "grad_norm": 15.707266807556152,
      "learning_rate": 4.785656071719641e-07,
      "loss": 3.7565,
      "reward": 3.7303237915039062,
      "reward_std": 1.399117350578308,
      "rewards/reasoning_density_reward/mean": 0.43619123101234436,
      "rewards/reasoning_density_reward/std": 0.10448447614908218,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.6691325902938843,
      "rewards/trajectory_similarity_reward/std": 0.2946275472640991,
      "step": 493
    },
    {
      "clip_ratio/high_max": 0.013104009412927553,
      "clip_ratio/high_mean": 0.0031532491229882,
      "clip_ratio/low_mean": 0.0006131568006821908,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037664059309463482,
      "entropy": 0.35025518387556076,
      "epoch": 0.40264900662251657,
      "grad_norm": 11.488773345947266,
      "learning_rate": 4.779136104319479e-07,
      "loss": 5.0704,
      "reward": 1.8931057453155518,
      "reward_std": 1.0905752182006836,
      "rewards/reasoning_density_reward/mean": 0.20520833134651184,
      "rewards/reasoning_density_reward/std": 0.12141836434602737,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.2816474139690399,
      "rewards/trajectory_similarity_reward/std": 0.3957667946815491,
      "step": 494
    },
    {
      "clip_ratio/high_max": 0.017340025631710887,
      "clip_ratio/high_mean": 0.003420960609219037,
      "clip_ratio/low_mean": 0.0008929097093641758,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043138703622389585,
      "entropy": 0.34466954320669174,
      "epoch": 0.4034640855832909,
      "grad_norm": 11.632669448852539,
      "learning_rate": 4.772616136919315e-07,
      "loss": -0.1723,
      "reward": 3.259619951248169,
      "reward_std": 0.8277856111526489,
      "rewards/reasoning_density_reward/mean": 0.36250001192092896,
      "rewards/reasoning_density_reward/std": 0.18705560266971588,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 0.9437293410301208,
      "rewards/trajectory_similarity_reward/mean": 0.5533697605133057,
      "rewards/trajectory_similarity_reward/std": 0.4561977982521057,
      "step": 495
    },
    {
      "clip_ratio/high_max": 0.010782771918457001,
      "clip_ratio/high_mean": 0.002521158574381843,
      "clip_ratio/low_mean": 0.0016861112671904266,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042072698706761,
      "entropy": 0.3759507089853287,
      "epoch": 0.4042791645440652,
      "grad_norm": 14.96009349822998,
      "learning_rate": 4.7660961695191523e-07,
      "loss": 9.8927,
      "reward": 2.914475917816162,
      "reward_std": 1.6394028663635254,
      "rewards/reasoning_density_reward/mean": 0.32343751192092896,
      "rewards/reasoning_density_reward/std": 0.17122988402843475,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.4347882568836212,
      "rewards/trajectory_similarity_reward/std": 0.41211023926734924,
      "step": 496
    },
    {
      "clip_ratio/high_max": 0.012218077084980905,
      "clip_ratio/high_mean": 0.0023380992424790747,
      "clip_ratio/low_mean": 0.0010054604899778496,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033435597069910727,
      "entropy": 0.3809337578713894,
      "epoch": 0.4050942435048395,
      "grad_norm": 14.660904884338379,
      "learning_rate": 4.759576202118989e-07,
      "loss": 8.5528,
      "reward": 2.8907525539398193,
      "reward_std": 1.168431043624878,
      "rewards/reasoning_density_reward/mean": 0.2941964268684387,
      "rewards/reasoning_density_reward/std": 0.19812418520450592,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5340563058853149,
      "rewards/trajectory_similarity_reward/std": 0.40729495882987976,
      "step": 497
    },
    {
      "clip_ratio/high_max": 0.012136504810769111,
      "clip_ratio/high_mean": 0.0023361435014521703,
      "clip_ratio/low_mean": 0.0003718973894137889,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027080408908659592,
      "entropy": 0.35607078298926353,
      "epoch": 0.40590932246561384,
      "grad_norm": 13.025666236877441,
      "learning_rate": 4.7530562347188263e-07,
      "loss": 1.3481,
      "reward": 2.4856910705566406,
      "reward_std": 1.3770813345909119,
      "rewards/reasoning_density_reward/mean": 0.36377108097076416,
      "rewards/reasoning_density_reward/std": 0.10984104126691818,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.1831679940223694,
      "rewards/trajectory_similarity_reward/mean": 0.4344199448823929,
      "rewards/trajectory_similarity_reward/std": 0.2972468137741089,
      "step": 498
    },
    {
      "clip_ratio/high_max": 0.009777579398360103,
      "clip_ratio/high_mean": 0.002402921461907681,
      "clip_ratio/low_mean": 0.0017003415268845856,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004103262945136521,
      "entropy": 0.34342237189412117,
      "epoch": 0.40672440142638816,
      "grad_norm": 12.792586326599121,
      "learning_rate": 4.746536267318663e-07,
      "loss": 3.5864,
      "reward": 4.079152584075928,
      "reward_std": 0.7632679343223572,
      "rewards/reasoning_density_reward/mean": 0.47708332538604736,
      "rewards/reasoning_density_reward/std": 0.06291528046131134,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.7895692586898804,
      "rewards/trajectory_similarity_reward/std": 0.2169470340013504,
      "step": 499
    },
    {
      "clip_ratio/high_max": 0.018558662268333137,
      "clip_ratio/high_mean": 0.004123319929931313,
      "clip_ratio/low_mean": 3.729117088369094e-05,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0041606111044529825,
      "entropy": 0.372320294380188,
      "epoch": 0.40753948038716253,
      "grad_norm": 11.227566719055176,
      "learning_rate": 4.7400162999185003e-07,
      "loss": -0.0847,
      "reward": 2.410396099090576,
      "reward_std": 1.7185909748077393,
      "rewards/reasoning_density_reward/mean": 0.21820437908172607,
      "rewards/reasoning_density_reward/std": 0.17990262806415558,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.41094183921813965,
      "rewards/trajectory_similarity_reward/std": 0.43298661708831787,
      "step": 500
    },
    {
      "clip_ratio/high_max": 0.01600010716356337,
      "clip_ratio/high_mean": 0.002977501164423302,
      "clip_ratio/low_mean": 0.000954486669797916,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00393198779784143,
      "entropy": 0.37316714227199554,
      "epoch": 0.40835455934793685,
      "grad_norm": 11.111660957336426,
      "learning_rate": 4.733496332518337e-07,
      "loss": 1.1606,
      "reward": 3.386606216430664,
      "reward_std": 1.0417474508285522,
      "rewards/reasoning_density_reward/mean": 0.3686259984970093,
      "rewards/reasoning_density_reward/std": 0.17984575033187866,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.5804798603057861,
      "rewards/trajectory_similarity_reward/std": 0.3208852708339691,
      "step": 501
    },
    {
      "clip_ratio/high_max": 0.01587410329375416,
      "clip_ratio/high_mean": 0.0042424327693879604,
      "clip_ratio/low_mean": 0.0009752034166012891,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005217636207817122,
      "entropy": 0.35647206753492355,
      "epoch": 0.40916963830871117,
      "grad_norm": 16.41750144958496,
      "learning_rate": 4.7269763651181743e-07,
      "loss": 12.2643,
      "reward": 3.66679310798645,
      "reward_std": 1.4251022338867188,
      "rewards/reasoning_density_reward/mean": 0.3046131134033203,
      "rewards/reasoning_density_reward/std": 0.19150410592556,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.7371799945831299,
      "rewards/trajectory_similarity_reward/std": 0.3274347484111786,
      "step": 502
    },
    {
      "clip_ratio/high_max": 0.017415348498616368,
      "clip_ratio/high_mean": 0.004997735770302825,
      "clip_ratio/low_mean": 0.0005234523196122609,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005521188097191043,
      "entropy": 0.36076869815588,
      "epoch": 0.4099847172694855,
      "grad_norm": 13.684093475341797,
      "learning_rate": 4.720456397718012e-07,
      "loss": -8.0413,
      "reward": 2.066368341445923,
      "reward_std": 1.686880350112915,
      "rewards/reasoning_density_reward/mean": 0.35357144474983215,
      "rewards/reasoning_density_reward/std": 0.09162182360887527,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.3065469264984131,
      "rewards/trajectory_similarity_reward/std": 0.4297935664653778,
      "step": 503
    },
    {
      "clip_ratio/high_max": 0.014402464381419122,
      "clip_ratio/high_mean": 0.00244979726994643,
      "clip_ratio/low_mean": 0.0010121542218257673,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034619515281519853,
      "entropy": 0.3553940989077091,
      "epoch": 0.4107997962302598,
      "grad_norm": 15.47885513305664,
      "learning_rate": 4.713936430317848e-07,
      "loss": -1.6984,
      "reward": 2.53973126411438,
      "reward_std": 0.8231101036071777,
      "rewards/reasoning_density_reward/mean": 0.29405438899993896,
      "rewards/reasoning_density_reward/std": 0.16592639684677124,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4644269049167633,
      "rewards/trajectory_similarity_reward/std": 0.4336436688899994,
      "step": 504
    },
    {
      "clip_ratio/high_max": 0.009813518903683871,
      "clip_ratio/high_mean": 0.002530524638132192,
      "clip_ratio/low_mean": 0.0008508191494911443,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003381343820365146,
      "entropy": 0.3437667712569237,
      "epoch": 0.4116148751910341,
      "grad_norm": 14.08478832244873,
      "learning_rate": 4.707416462917686e-07,
      "loss": 2.6492,
      "reward": 3.044487714767456,
      "reward_std": 1.493701696395874,
      "rewards/reasoning_density_reward/mean": 0.39866071939468384,
      "rewards/reasoning_density_reward/std": 0.11641523987054825,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.39582693576812744,
      "rewards/trajectory_similarity_reward/std": 0.3586243689060211,
      "step": 505
    },
    {
      "clip_ratio/high_max": 0.016191848262678832,
      "clip_ratio/high_mean": 0.0035015502653550357,
      "clip_ratio/low_mean": 0.0008953251654020278,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004396875418024138,
      "entropy": 0.35720810666680336,
      "epoch": 0.41242995415180844,
      "grad_norm": 17.821170806884766,
      "learning_rate": 4.700896495517522e-07,
      "loss": 12.0818,
      "reward": 3.670817017555237,
      "reward_std": 0.9402625635266304,
      "rewards/reasoning_density_reward/mean": 0.3655133992433548,
      "rewards/reasoning_density_reward/std": 0.15383101627230644,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.6104557514190674,
      "rewards/trajectory_similarity_reward/mean": 0.6334284245967865,
      "rewards/trajectory_similarity_reward/std": 0.2859291508793831,
      "step": 506
    },
    {
      "clip_ratio/high_max": 0.013381588039919734,
      "clip_ratio/high_mean": 0.002383434126386419,
      "clip_ratio/low_mean": 0.0008309929326060228,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032144270080607384,
      "entropy": 0.36148877441883087,
      "epoch": 0.41324503311258276,
      "grad_norm": 13.540403366088867,
      "learning_rate": 4.694376528117359e-07,
      "loss": 19.427,
      "reward": 2.2720131874084473,
      "reward_std": 1.4018887281417847,
      "rewards/reasoning_density_reward/mean": 0.22075892984867096,
      "rewards/reasoning_density_reward/std": 0.19655285775661469,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 0.9826622009277344,
      "rewards/trajectory_similarity_reward/mean": 0.2700043320655823,
      "rewards/trajectory_similarity_reward/std": 0.4156396687030792,
      "step": 507
    },
    {
      "clip_ratio/high_max": 0.006156275398097932,
      "clip_ratio/high_mean": 0.0008995013922685757,
      "clip_ratio/low_mean": 0.0013373387919273227,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022368401841958985,
      "entropy": 0.39309730753302574,
      "epoch": 0.41406011207335713,
      "grad_norm": 12.89919376373291,
      "learning_rate": 4.687856560717196e-07,
      "loss": 14.3382,
      "step": 508
    },
    {
      "clip_ratio/high_max": 0.013227089366409928,
      "clip_ratio/high_mean": 0.0033233101639780216,
      "clip_ratio/low_mean": 0.000907919598830631,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004231229751894716,
      "entropy": 0.3666338697075844,
      "epoch": 0.41487519103413145,
      "grad_norm": 15.610618591308594,
      "learning_rate": 4.681336593317033e-07,
      "loss": 4.6123,
      "reward": 3.1357197761535645,
      "reward_std": 1.4599223136901855,
      "rewards/reasoning_density_reward/mean": 0.22741815447807312,
      "rewards/reasoning_density_reward/std": 0.17491582036018372,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5645516514778137,
      "rewards/trajectory_similarity_reward/std": 0.38172462582588196,
      "step": 509
    },
    {
      "clip_ratio/high_max": 0.015258911589626223,
      "clip_ratio/high_mean": 0.002520437141356524,
      "clip_ratio/low_mean": 0.0016048886427597608,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004125325809582137,
      "entropy": 0.35400425270199776,
      "epoch": 0.41569026999490577,
      "grad_norm": 11.151272773742676,
      "learning_rate": 4.6748166259168697e-07,
      "loss": 14.8141,
      "reward": 2.370922327041626,
      "reward_std": 1.888455867767334,
      "rewards/reasoning_density_reward/mean": 0.20892858505249023,
      "rewards/reasoning_density_reward/std": 0.20719940960407257,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.365878939628601,
      "rewards/trajectory_similarity_reward/mean": 0.38074374198913574,
      "rewards/trajectory_similarity_reward/std": 0.41877254843711853,
      "step": 510
    },
    {
      "clip_ratio/high_max": 0.01007988560013473,
      "clip_ratio/high_mean": 0.0020576538663590327,
      "clip_ratio/low_mean": 0.00042316358667449094,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024808173766359687,
      "entropy": 0.37490418180823326,
      "epoch": 0.4165053489556801,
      "grad_norm": 10.841303825378418,
      "learning_rate": 4.668296658516707e-07,
      "loss": 13.7843,
      "reward": 2.663649559020996,
      "reward_std": 1.324288010597229,
      "rewards/reasoning_density_reward/mean": 0.28511905670166016,
      "rewards/reasoning_density_reward/std": 0.16349013149738312,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.3160305321216583,
      "rewards/trajectory_similarity_reward/std": 0.42496946454048157,
      "step": 511
    },
    {
      "clip_ratio/high_max": 0.010184576793108135,
      "clip_ratio/high_mean": 0.0020026810016133823,
      "clip_ratio/low_mean": 0.0006363765169226099,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002639057536725886,
      "entropy": 0.37049003690481186,
      "epoch": 0.4173204279164544,
      "grad_norm": 13.437580108642578,
      "learning_rate": 4.6617766911165437e-07,
      "loss": 2.5393,
      "reward": 2.490358829498291,
      "reward_std": 2.0015101432800293,
      "rewards/reasoning_density_reward/mean": 0.33281248807907104,
      "rewards/reasoning_density_reward/std": 0.18495416641235352,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.47004640102386475,
      "rewards/trajectory_similarity_reward/std": 0.4306672513484955,
      "step": 512
    },
    {
      "clip_ratio/high_max": 0.012184657214675099,
      "clip_ratio/high_mean": 0.002710381377255544,
      "clip_ratio/low_mean": 0.0004000717235612683,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003110453166300431,
      "entropy": 0.36535100266337395,
      "epoch": 0.4181355068772287,
      "grad_norm": 15.003134727478027,
      "learning_rate": 4.655256723716381e-07,
      "loss": 9.8286,
      "reward": 3.2388362884521484,
      "reward_std": 1.558246374130249,
      "rewards/reasoning_density_reward/mean": 0.3552049398422241,
      "rewards/reasoning_density_reward/std": 0.12199123203754425,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.5398813486099243,
      "rewards/trajectory_similarity_reward/std": 0.3823543190956116,
      "step": 513
    },
    {
      "clip_ratio/high_max": 0.017717924085445702,
      "clip_ratio/high_mean": 0.004682320810388774,
      "clip_ratio/low_mean": 0.00022521716164192185,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004907537950202823,
      "entropy": 0.3386145308613777,
      "epoch": 0.41895058583800304,
      "grad_norm": 27.082008361816406,
      "learning_rate": 4.6487367563162177e-07,
      "loss": 9.1953,
      "reward": 3.4513585567474365,
      "reward_std": 1.601713240146637,
      "rewards/reasoning_density_reward/mean": 0.3648437261581421,
      "rewards/reasoning_density_reward/std": 0.18978141993284225,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.1385715007781982,
      "rewards/trajectory_similarity_reward/mean": 0.6490147411823273,
      "rewards/trajectory_similarity_reward/std": 0.3663603663444519,
      "step": 514
    },
    {
      "clip_ratio/high_max": 0.015624397492501885,
      "clip_ratio/high_mean": 0.0032097555449581705,
      "clip_ratio/low_mean": 0.0012433632364263758,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004453118817764334,
      "entropy": 0.3136196695268154,
      "epoch": 0.41976566479877736,
      "grad_norm": 17.27784538269043,
      "learning_rate": 4.642216788916055e-07,
      "loss": 7.9512,
      "reward": 3.5049288272857666,
      "reward_std": 1.1191045045852661,
      "rewards/reasoning_density_reward/mean": 0.41290920972824097,
      "rewards/reasoning_density_reward/std": 0.10470545291900635,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6545194387435913,
      "rewards/trajectory_similarity_reward/std": 0.36359885334968567,
      "step": 515
    },
    {
      "clip_ratio/high_max": 0.015369357308372855,
      "clip_ratio/high_mean": 0.0028148368874099106,
      "clip_ratio/low_mean": 0.0009584963809174951,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037733332574134693,
      "entropy": 0.30800891667604446,
      "epoch": 0.42058074375955173,
      "grad_norm": 16.777786254882812,
      "learning_rate": 4.6356968215158917e-07,
      "loss": 13.142,
      "reward": 3.554172992706299,
      "reward_std": 0.9482600092887878,
      "rewards/reasoning_density_reward/mean": 0.3583333492279053,
      "rewards/reasoning_density_reward/std": 0.15599144995212555,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.9031196236610413,
      "rewards/trajectory_similarity_reward/mean": 0.6645898222923279,
      "rewards/trajectory_similarity_reward/std": 0.41194695234298706,
      "step": 516
    },
    {
      "clip_ratio/high_max": 0.016103650734294206,
      "clip_ratio/high_mean": 0.003636309069406707,
      "clip_ratio/low_mean": 0.0008670585666550323,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004503367585130036,
      "entropy": 0.33045126125216484,
      "epoch": 0.42139582272032605,
      "grad_norm": 17.89801025390625,
      "learning_rate": 4.629176854115729e-07,
      "loss": 4.2956,
      "reward": 3.6924058198928833,
      "reward_std": 0.9951230883598328,
      "rewards/reasoning_density_reward/mean": 0.4345982074737549,
      "rewards/reasoning_density_reward/std": 0.10271388664841652,
      "rewards/success_reward/mean": 2.578125,
      "rewards/success_reward/std": 1.0404633283615112,
      "rewards/trajectory_similarity_reward/mean": 0.6796825528144836,
      "rewards/trajectory_similarity_reward/std": 0.3362419158220291,
      "step": 517
    },
    {
      "clip_ratio/high_max": 0.016062470735050738,
      "clip_ratio/high_mean": 0.004686308020609431,
      "clip_ratio/low_mean": 0.0009591251109668519,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005645433077006601,
      "entropy": 0.34231340140104294,
      "epoch": 0.42221090168110037,
      "grad_norm": 12.778525352478027,
      "learning_rate": 4.622656886715567e-07,
      "loss": 3.9089,
      "reward": 3.0629982948303223,
      "reward_std": 0.11767183989286423,
      "rewards/reasoning_density_reward/mean": 0.34527814388275146,
      "rewards/reasoning_density_reward/std": 0.18489566445350647,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 0.7745966911315918,
      "rewards/trajectory_similarity_reward/mean": 0.4677203297615051,
      "rewards/trajectory_similarity_reward/std": 0.48778823018074036,
      "step": 518
    },
    {
      "clip_ratio/high_max": 0.012547614867798984,
      "clip_ratio/high_mean": 0.003093306062510237,
      "clip_ratio/low_mean": 0.0004829472927667666,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035762533079832792,
      "entropy": 0.35157372429966927,
      "epoch": 0.4230259806418747,
      "grad_norm": 11.583396911621094,
      "learning_rate": 4.616136919315403e-07,
      "loss": 7.4698,
      "reward": 2.969721794128418,
      "reward_std": 1.7278835773468018,
      "rewards/reasoning_density_reward/mean": 0.3641369044780731,
      "rewards/reasoning_density_reward/std": 0.14234158396720886,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5430850386619568,
      "rewards/trajectory_similarity_reward/std": 0.4182828962802887,
      "step": 519
    },
    {
      "clip_ratio/high_max": 0.010596317413728684,
      "clip_ratio/high_mean": 0.0025483311983407475,
      "clip_ratio/low_mean": 0.0008007773067220114,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033491084905108437,
      "entropy": 0.3469214476644993,
      "epoch": 0.423841059602649,
      "grad_norm": 12.69218921661377,
      "learning_rate": 4.609616951915241e-07,
      "loss": 3.3083,
      "reward": 1.344286322593689,
      "reward_std": 1.4673547744750977,
      "rewards/reasoning_density_reward/mean": 0.17049852013587952,
      "rewards/reasoning_density_reward/std": 0.160594180226326,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.23628786206245422,
      "rewards/trajectory_similarity_reward/std": 0.4229174554347992,
      "step": 520
    },
    {
      "clip_ratio/high_max": 0.006852902006357908,
      "clip_ratio/high_mean": 0.001356354376184754,
      "clip_ratio/low_mean": 0.0013318317833181936,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026881861849687994,
      "entropy": 0.3465284891426563,
      "epoch": 0.4246561385634233,
      "grad_norm": 10.947225570678711,
      "learning_rate": 4.603096984515077e-07,
      "loss": 7.9354,
      "reward": 2.9568943977355957,
      "reward_std": 1.7531211376190186,
      "rewards/reasoning_density_reward/mean": 0.3812500238418579,
      "rewards/reasoning_density_reward/std": 0.13149777054786682,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.5131444931030273,
      "rewards/trajectory_similarity_reward/std": 0.4708292782306671,
      "step": 521
    },
    {
      "clip_ratio/high_max": 0.01413861638866365,
      "clip_ratio/high_mean": 0.0035338387824594975,
      "clip_ratio/low_mean": 0.0013539091887651011,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004887748073088005,
      "entropy": 0.339506097137928,
      "epoch": 0.42547121752419764,
      "grad_norm": 16.6993408203125,
      "learning_rate": 4.596577017114915e-07,
      "loss": 13.1525,
      "reward": 3.4503250122070312,
      "reward_std": 1.7051410675048828,
      "rewards/reasoning_density_reward/mean": 0.3369791507720947,
      "rewards/reasoning_density_reward/std": 0.16555096209049225,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.675845742225647,
      "rewards/trajectory_similarity_reward/std": 0.3849598467350006,
      "step": 522
    },
    {
      "clip_ratio/high_max": 0.014190021611284465,
      "clip_ratio/high_mean": 0.003308204664790537,
      "clip_ratio/low_mean": 0.0005524381049326621,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003860642740619369,
      "entropy": 0.3696663975715637,
      "epoch": 0.42628629648497196,
      "grad_norm": 11.99974536895752,
      "learning_rate": 4.590057049714751e-07,
      "loss": 12.0626,
      "reward": 2.7304906845092773,
      "reward_std": 1.8709884881973267,
      "rewards/reasoning_density_reward/mean": 0.27842262387275696,
      "rewards/reasoning_density_reward/std": 0.17635871469974518,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.4833182394504547,
      "rewards/trajectory_similarity_reward/std": 0.45016905665397644,
      "step": 523
    },
    {
      "clip_ratio/high_max": 0.01276568864705041,
      "clip_ratio/high_mean": 0.002504027637769468,
      "clip_ratio/low_mean": 0.0010011243430199102,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003505151951685548,
      "entropy": 0.33755313232541084,
      "epoch": 0.42710137544574633,
      "grad_norm": 12.194852828979492,
      "learning_rate": 4.583537082314588e-07,
      "loss": 12.9509,
      "reward": 3.722179412841797,
      "reward_std": 0.8115052580833435,
      "rewards/reasoning_density_reward/mean": 0.3645833432674408,
      "rewards/reasoning_density_reward/std": 0.17689451575279236,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.638846218585968,
      "rewards/trajectory_similarity_reward/std": 0.43467217683792114,
      "step": 524
    },
    {
      "clip_ratio/high_max": 0.012234905268996954,
      "clip_ratio/high_mean": 0.0027670819181366824,
      "clip_ratio/low_mean": 0.0006039176842023153,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003370999576873146,
      "entropy": 0.37696387618780136,
      "epoch": 0.42791645440652065,
      "grad_norm": 13.701967239379883,
      "learning_rate": 4.577017114914425e-07,
      "loss": -2.0945,
      "reward": 1.3952381610870361,
      "reward_std": 1.703277587890625,
      "rewards/reasoning_density_reward/mean": 0.2699776887893677,
      "rewards/reasoning_density_reward/std": 0.11816918104887009,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.1877603679895401,
      "rewards/trajectory_similarity_reward/std": 0.33810609579086304,
      "step": 525
    },
    {
      "clip_ratio/high_max": 0.009606492036255077,
      "clip_ratio/high_mean": 0.0017105244660342578,
      "clip_ratio/low_mean": 0.0006456519331550226,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023561763809993863,
      "entropy": 0.3667714484035969,
      "epoch": 0.42873153336729497,
      "grad_norm": 11.172115325927734,
      "learning_rate": 4.570497147514262e-07,
      "loss": 0.9387,
      "reward": 2.102688789367676,
      "reward_std": 1.5174919366836548,
      "rewards/reasoning_density_reward/mean": 0.3096185028553009,
      "rewards/reasoning_density_reward/std": 0.14459028840065002,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.2930702269077301,
      "rewards/trajectory_similarity_reward/std": 0.4009018838405609,
      "step": 526
    },
    {
      "clip_ratio/high_max": 0.012952699733432382,
      "clip_ratio/high_mean": 0.002472080595907755,
      "clip_ratio/low_mean": 0.0012203296719235368,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003692410289659165,
      "entropy": 0.35280463472008705,
      "epoch": 0.4295466123280693,
      "grad_norm": 14.228363037109375,
      "learning_rate": 4.5639771801140987e-07,
      "loss": -2.1419,
      "reward": 2.6629457473754883,
      "reward_std": 1.9156620502471924,
      "rewards/reasoning_density_reward/mean": 0.34424716234207153,
      "rewards/reasoning_density_reward/std": 0.12192461639642715,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.4436986446380615,
      "rewards/trajectory_similarity_reward/std": 0.43277832865715027,
      "step": 527
    },
    {
      "clip_ratio/high_max": 0.01275290857302025,
      "clip_ratio/high_mean": 0.0028754032537108287,
      "clip_ratio/low_mean": 0.001114493039494846,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039898961840663105,
      "entropy": 0.34651198610663414,
      "epoch": 0.4303616912888436,
      "grad_norm": 12.955595970153809,
      "learning_rate": 4.557457212713936e-07,
      "loss": 18.6682,
      "reward": 2.9930585622787476,
      "reward_std": 1.8933537006378174,
      "rewards/reasoning_density_reward/mean": 0.3700892925262451,
      "rewards/reasoning_density_reward/std": 0.17448541522026062,
      "rewards/success_reward/mean": 2.015625,
      "rewards/success_reward/std": 1.4066235423088074,
      "rewards/trajectory_similarity_reward/mean": 0.6073443293571472,
      "rewards/trajectory_similarity_reward/std": 0.4640781730413437,
      "step": 528
    },
    {
      "clip_ratio/high_max": 0.013237160164862871,
      "clip_ratio/high_mean": 0.002419430995360017,
      "clip_ratio/low_mean": 0.0012059326545568183,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036253635480534285,
      "entropy": 0.29957783967256546,
      "epoch": 0.4311767702496179,
      "grad_norm": 13.295892715454102,
      "learning_rate": 4.5509372453137727e-07,
      "loss": 9.7616,
      "reward": 2.845973014831543,
      "reward_std": 1.8462645411491394,
      "rewards/reasoning_density_reward/mean": 0.41979166865348816,
      "rewards/reasoning_density_reward/std": 0.09807521104812622,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.4454171061515808,
      "rewards/trajectory_similarity_reward/mean": 0.5511812567710876,
      "rewards/trajectory_similarity_reward/std": 0.4346158355474472,
      "step": 529
    },
    {
      "clip_ratio/high_max": 0.01701144070830196,
      "clip_ratio/high_mean": 0.0037818210548721254,
      "clip_ratio/low_mean": 0.00030969319777796045,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004091514245374128,
      "entropy": 0.3272099681198597,
      "epoch": 0.43199184921039224,
      "grad_norm": 12.081549644470215,
      "learning_rate": 4.54441727791361e-07,
      "loss": 4.7187,
      "reward": 2.6134653091430664,
      "reward_std": 1.3453723192214966,
      "rewards/reasoning_density_reward/mean": 0.34998762607574463,
      "rewards/reasoning_density_reward/std": 0.1775233894586563,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.20097756385803223,
      "rewards/trajectory_similarity_reward/std": 0.35264042019844055,
      "step": 530
    },
    {
      "clip_ratio/high_max": 0.012477643031161278,
      "clip_ratio/high_mean": 0.0030054442831897177,
      "clip_ratio/low_mean": 0.00105022521893261,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004055669443914667,
      "entropy": 0.36449017003178596,
      "epoch": 0.43280692817116656,
      "grad_norm": 16.792686462402344,
      "learning_rate": 4.5378973105134467e-07,
      "loss": 6.5184,
      "reward": 2.4506869316101074,
      "reward_std": 1.6587151288986206,
      "rewards/reasoning_density_reward/mean": 0.3283081650733948,
      "rewards/reasoning_density_reward/std": 0.1363329291343689,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.24737879633903503,
      "rewards/trajectory_similarity_reward/std": 0.27284151315689087,
      "step": 531
    },
    {
      "clip_ratio/high_max": 0.009023146063555032,
      "clip_ratio/high_mean": 0.0017029148148139939,
      "clip_ratio/low_mean": 0.0006049135408829898,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002307828348421026,
      "entropy": 0.35921595990657806,
      "epoch": 0.43362200713194093,
      "grad_norm": 14.191404342651367,
      "learning_rate": 4.531377343113284e-07,
      "loss": 6.6734,
      "reward": 1.9906651973724365,
      "reward_std": 1.915246605873108,
      "rewards/reasoning_density_reward/mean": 0.3214285969734192,
      "rewards/reasoning_density_reward/std": 0.13509885966777802,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.35673654079437256,
      "rewards/trajectory_similarity_reward/std": 0.4325658082962036,
      "step": 532
    },
    {
      "clip_ratio/high_max": 0.015696843445766717,
      "clip_ratio/high_mean": 0.003110400815785397,
      "clip_ratio/low_mean": 0.0006029906871845014,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037133915320737287,
      "entropy": 0.35043198242783546,
      "epoch": 0.43443708609271525,
      "grad_norm": 18.141403198242188,
      "learning_rate": 4.524857375713122e-07,
      "loss": -2.973,
      "reward": 2.595860004425049,
      "reward_std": 1.3730543851852417,
      "rewards/reasoning_density_reward/mean": 0.31051135063171387,
      "rewards/reasoning_density_reward/std": 0.17149858176708221,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.504098653793335,
      "rewards/trajectory_similarity_reward/std": 0.46237340569496155,
      "step": 533
    },
    {
      "clip_ratio/high_max": 0.013762441463768482,
      "clip_ratio/high_mean": 0.003049885548534803,
      "clip_ratio/low_mean": 0.0008164268765540328,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003866312326863408,
      "entropy": 0.36163995042443275,
      "epoch": 0.43525216505348957,
      "grad_norm": 16.74054527282715,
      "learning_rate": 4.518337408312958e-07,
      "loss": 13.8034,
      "reward": 2.6820225715637207,
      "reward_std": 1.8541204929351807,
      "rewards/reasoning_density_reward/mean": 0.34583336114883423,
      "rewards/reasoning_density_reward/std": 0.1927865892648697,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.36743929982185364,
      "rewards/trajectory_similarity_reward/std": 0.414726585149765,
      "step": 534
    },
    {
      "clip_ratio/high_max": 0.011335512564983219,
      "clip_ratio/high_mean": 0.001957233929715585,
      "clip_ratio/low_mean": 0.0006808647412981372,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026380986237199977,
      "entropy": 0.3632430471479893,
      "epoch": 0.4360672440142639,
      "grad_norm": 13.299813270568848,
      "learning_rate": 4.5118174409127957e-07,
      "loss": 7.3433,
      "reward": 2.7300225496292114,
      "reward_std": 1.8404462337493896,
      "rewards/reasoning_density_reward/mean": 0.34033502638339996,
      "rewards/reasoning_density_reward/std": 0.17550288140773773,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3136931657791138,
      "rewards/trajectory_similarity_reward/mean": 0.4209374785423279,
      "rewards/trajectory_similarity_reward/std": 0.42414043843746185,
      "step": 535
    },
    {
      "clip_ratio/high_max": 0.014433796983212233,
      "clip_ratio/high_mean": 0.003260322831920348,
      "clip_ratio/low_mean": 0.0008027923977351747,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004063115309691057,
      "entropy": 0.3420613519847393,
      "epoch": 0.4368823229750382,
      "grad_norm": 11.867579460144043,
      "learning_rate": 4.505297473512632e-07,
      "loss": 8.8575,
      "reward": 1.5034226179122925,
      "reward_std": 1.5601083040237427,
      "rewards/reasoning_density_reward/mean": 0.2609577775001526,
      "rewards/reasoning_density_reward/std": 0.14925196766853333,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.21121469140052795,
      "rewards/trajectory_similarity_reward/std": 0.36125943064689636,
      "step": 536
    },
    {
      "clip_ratio/high_max": 0.007088303216733038,
      "clip_ratio/high_mean": 0.0012567159428726882,
      "clip_ratio/low_mean": 0.0012671228287217673,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002523838746128604,
      "entropy": 0.3735988773405552,
      "epoch": 0.4376974019358125,
      "grad_norm": 13.667549133300781,
      "learning_rate": 4.4987775061124697e-07,
      "loss": 8.144,
      "reward": 3.4933695793151855,
      "reward_std": 1.0063397884368896,
      "rewards/reasoning_density_reward/mean": 0.32723215222358704,
      "rewards/reasoning_density_reward/std": 0.2142256796360016,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.6348873376846313,
      "rewards/trajectory_similarity_reward/std": 0.3331311047077179,
      "step": 537
    },
    {
      "clip_ratio/high_max": 0.01401313440874219,
      "clip_ratio/high_mean": 0.003542993770679459,
      "clip_ratio/low_mean": 0.000977485629846342,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004520479473285377,
      "entropy": 0.3762713894248009,
      "epoch": 0.43851248089658684,
      "grad_norm": 15.97400188446045,
      "learning_rate": 4.492257538712306e-07,
      "loss": 1.0522,
      "reward": 2.974268913269043,
      "reward_std": 2.0359609127044678,
      "rewards/reasoning_density_reward/mean": 0.42500001192092896,
      "rewards/reasoning_density_reward/std": 0.09999999403953552,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.5805187225341797,
      "rewards/trajectory_similarity_reward/std": 0.46904411911964417,
      "step": 538
    },
    {
      "clip_ratio/high_max": 0.01625559967942536,
      "clip_ratio/high_mean": 0.004643331878469326,
      "clip_ratio/low_mean": 0.0005687885350198485,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005212120391661301,
      "entropy": 0.3525758348405361,
      "epoch": 0.43932755985736116,
      "grad_norm": 17.112640380859375,
      "learning_rate": 4.4857375713121437e-07,
      "loss": -2.5602,
      "reward": 3.6732966899871826,
      "reward_std": 1.0029192119836807,
      "rewards/reasoning_density_reward/mean": 0.3492187410593033,
      "rewards/reasoning_density_reward/std": 0.20220451056957245,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.726596474647522,
      "rewards/trajectory_similarity_reward/mean": 0.6522030830383301,
      "rewards/trajectory_similarity_reward/std": 0.34439311921596527,
      "step": 539
    },
    {
      "clip_ratio/high_max": 0.014918634900823236,
      "clip_ratio/high_mean": 0.00401418349792948,
      "clip_ratio/low_mean": 0.001292284894589102,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005306468447088264,
      "entropy": 0.3736918307840824,
      "epoch": 0.44014263881813553,
      "grad_norm": 13.93604850769043,
      "learning_rate": 4.47921760391198e-07,
      "loss": 11.7332,
      "reward": 3.45410418510437,
      "reward_std": 1.0588693618774414,
      "rewards/reasoning_density_reward/mean": 0.3958333134651184,
      "rewards/reasoning_density_reward/std": 0.13911093771457672,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.5270209312438965,
      "rewards/trajectory_similarity_reward/std": 0.388954758644104,
      "step": 540
    },
    {
      "clip_ratio/high_max": 0.014286767109297216,
      "clip_ratio/high_mean": 0.0036370068846736103,
      "clip_ratio/low_mean": 0.0015691375592723489,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005206144523981493,
      "entropy": 0.32411643862724304,
      "epoch": 0.44095771777890985,
      "grad_norm": 18.48564910888672,
      "learning_rate": 4.472697636511817e-07,
      "loss": 6.2086,
      "reward": 3.8031481504440308,
      "reward_std": 0.6727407574653625,
      "rewards/reasoning_density_reward/mean": 0.48125000298023224,
      "rewards/reasoning_density_reward/std": 0.0591565016657114,
      "rewards/success_reward/mean": 2.765625,
      "rewards/success_reward/std": 0.7829292416572571,
      "rewards/trajectory_similarity_reward/mean": 0.5562730431556702,
      "rewards/trajectory_similarity_reward/std": 0.32104508578777313,
      "step": 541
    },
    {
      "clip_ratio/high_max": 0.012603545590536669,
      "clip_ratio/high_mean": 0.003471132255072007,
      "clip_ratio/low_mean": 0.0007455750492226798,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042167073479504324,
      "entropy": 0.32146587781608105,
      "epoch": 0.44177279673968417,
      "grad_norm": 12.232275009155273,
      "learning_rate": 4.466177669111654e-07,
      "loss": -0.1927,
      "reward": 2.162128210067749,
      "reward_std": 1.692901372909546,
      "rewards/reasoning_density_reward/mean": 0.2704613208770752,
      "rewards/reasoning_density_reward/std": 0.17512290179729462,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.2808688879013062,
      "rewards/trajectory_similarity_reward/mean": 0.29791688919067383,
      "rewards/trajectory_similarity_reward/std": 0.4051060676574707,
      "step": 542
    },
    {
      "clip_ratio/high_max": 0.009317225136328489,
      "clip_ratio/high_mean": 0.0018007482212851755,
      "clip_ratio/low_mean": 0.0014903102492098697,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00329105845594313,
      "entropy": 0.35472767800092697,
      "epoch": 0.4425878757004585,
      "grad_norm": 12.583700180053711,
      "learning_rate": 4.459657701711491e-07,
      "loss": 13.568,
      "reward": 3.7108945846557617,
      "reward_std": 1.007293462753296,
      "rewards/reasoning_density_reward/mean": 0.3982638716697693,
      "rewards/reasoning_density_reward/std": 0.14989322423934937,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.687630832195282,
      "rewards/trajectory_similarity_reward/std": 0.29979878664016724,
      "step": 543
    },
    {
      "clip_ratio/high_max": 0.00809332396602258,
      "clip_ratio/high_mean": 0.0017869343282654881,
      "clip_ratio/low_mean": 0.0011449101912148762,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00293184447218664,
      "entropy": 0.34943459555506706,
      "epoch": 0.4434029546612328,
      "grad_norm": 13.367594718933105,
      "learning_rate": 4.4531377343113277e-07,
      "loss": -1.6329,
      "reward": 3.524310350418091,
      "reward_std": 1.035857915878296,
      "rewards/reasoning_density_reward/mean": 0.3687500059604645,
      "rewards/reasoning_density_reward/std": 0.16111589968204498,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.670820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.530560314655304,
      "rewards/trajectory_similarity_reward/std": 0.35715657472610474,
      "step": 544
    },
    {
      "clip_ratio/high_max": 0.014144404325634241,
      "clip_ratio/high_mean": 0.004334031560574658,
      "clip_ratio/low_mean": 0.0005374959619075526,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004871527504292317,
      "entropy": 0.3391512706875801,
      "epoch": 0.4442180336220071,
      "grad_norm": 14.395678520202637,
      "learning_rate": 4.446617766911165e-07,
      "loss": 4.1765,
      "reward": 4.120676517486572,
      "reward_std": 0.66426220536232,
      "rewards/reasoning_density_reward/mean": 0.4085937440395355,
      "rewards/reasoning_density_reward/std": 0.1592087522149086,
      "rewards/success_reward/mean": 2.859375,
      "rewards/success_reward/std": 0.4436737895011902,
      "rewards/trajectory_similarity_reward/mean": 0.8527075946331024,
      "rewards/trajectory_similarity_reward/std": 0.2878215163946152,
      "step": 545
    },
    {
      "clip_ratio/high_max": 0.012463175051379949,
      "clip_ratio/high_mean": 0.0027688551344908774,
      "clip_ratio/low_mean": 0.000800734334916342,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035695894657692406,
      "entropy": 0.32092030346393585,
      "epoch": 0.44503311258278144,
      "grad_norm": 21.005573272705078,
      "learning_rate": 4.4400977995110017e-07,
      "loss": 0.8973,
      "reward": 2.168970465660095,
      "reward_std": 0.8865417391061783,
      "rewards/reasoning_density_reward/mean": 0.32205086946487427,
      "rewards/reasoning_density_reward/std": 0.12864883244037628,
      "rewards/success_reward/mean": 1.546875,
      "rewards/success_reward/std": 1.4711282849311829,
      "rewards/trajectory_similarity_reward/mean": 0.30004459619522095,
      "rewards/trajectory_similarity_reward/std": 0.3584650605916977,
      "step": 546
    },
    {
      "clip_ratio/high_max": 0.016013500513508916,
      "clip_ratio/high_mean": 0.0041774155833991244,
      "clip_ratio/low_mean": 0.0008565982352592982,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00503401382593438,
      "entropy": 0.3438780829310417,
      "epoch": 0.44584819154355576,
      "grad_norm": 17.672231674194336,
      "learning_rate": 4.433577832110839e-07,
      "loss": 4.7462,
      "reward": 3.7928695678710938,
      "reward_std": 0.8661932349205017,
      "rewards/reasoning_density_reward/mean": 0.3828125,
      "rewards/reasoning_density_reward/std": 0.17858676612377167,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.5975571274757385,
      "rewards/trajectory_similarity_reward/std": 0.2664192318916321,
      "step": 547
    },
    {
      "clip_ratio/high_max": 0.0150608568219468,
      "clip_ratio/high_mean": 0.004217352834530175,
      "clip_ratio/low_mean": 0.0014223493344616145,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00563970222719945,
      "entropy": 0.3506272919476032,
      "epoch": 0.44666327050433013,
      "grad_norm": 14.314610481262207,
      "learning_rate": 4.4270578647106767e-07,
      "loss": 9.1876,
      "reward": 2.8220906257629395,
      "reward_std": 1.7138774394989014,
      "rewards/reasoning_density_reward/mean": 0.3109374940395355,
      "rewards/reasoning_density_reward/std": 0.22711391746997833,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.5424031615257263,
      "rewards/trajectory_similarity_reward/std": 0.4519876539707184,
      "step": 548
    },
    {
      "clip_ratio/high_max": 0.013267052534502,
      "clip_ratio/high_mean": 0.003214298776583746,
      "clip_ratio/low_mean": 0.0006633223965764046,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038776212895754725,
      "entropy": 0.3744233548641205,
      "epoch": 0.44747834946510445,
      "grad_norm": 11.219035148620605,
      "learning_rate": 4.420537897310513e-07,
      "loss": 17.345,
      "reward": 1.925614356994629,
      "reward_std": 1.7829744815826416,
      "rewards/reasoning_density_reward/mean": 0.20111607015132904,
      "rewards/reasoning_density_reward/std": 0.180794358253479,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.3930631875991821,
      "rewards/trajectory_similarity_reward/mean": 0.31824809312820435,
      "rewards/trajectory_similarity_reward/std": 0.4279051125049591,
      "step": 549
    },
    {
      "clip_ratio/high_max": 0.01672771543962881,
      "clip_ratio/high_mean": 0.003859352793369908,
      "clip_ratio/low_mean": 0.0005837082208017819,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004443060955964029,
      "entropy": 0.35721930861473083,
      "epoch": 0.44829342842587877,
      "grad_norm": 13.186562538146973,
      "learning_rate": 4.4140179299103507e-07,
      "loss": 8.8726,
      "reward": 3.834385633468628,
      "reward_std": 0.7571700215339661,
      "rewards/reasoning_density_reward/mean": 0.3419642746448517,
      "rewards/reasoning_density_reward/std": 0.19061875343322754,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.6799213886260986,
      "rewards/trajectory_similarity_reward/std": 0.23941686749458313,
      "step": 550
    },
    {
      "clip_ratio/high_max": 0.013944463862571865,
      "clip_ratio/high_mean": 0.003538831791956909,
      "clip_ratio/low_mean": 0.0004290778742870316,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003967909578932449,
      "entropy": 0.36337701603770256,
      "epoch": 0.4491085073866531,
      "grad_norm": 13.404133796691895,
      "learning_rate": 4.407497962510187e-07,
      "loss": -2.1555,
      "reward": 2.653087854385376,
      "reward_std": 1.47632896900177,
      "rewards/reasoning_density_reward/mean": 0.22968749701976776,
      "rewards/reasoning_density_reward/std": 0.19339856505393982,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.45465028285980225,
      "rewards/trajectory_similarity_reward/std": 0.39299502968788147,
      "step": 551
    },
    {
      "clip_ratio/high_max": 0.01368924422422424,
      "clip_ratio/high_mean": 0.0033131887030322105,
      "clip_ratio/low_mean": 0.0010176025753025897,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004330791241955012,
      "entropy": 0.3517771475017071,
      "epoch": 0.4499235863474274,
      "grad_norm": 17.499183654785156,
      "learning_rate": 4.4009779951100247e-07,
      "loss": 7.5392,
      "reward": 2.9876869916915894,
      "reward_std": 1.239673137664795,
      "rewards/reasoning_density_reward/mean": 0.31406252086162567,
      "rewards/reasoning_density_reward/std": 0.15317761152982712,
      "rewards/success_reward/mean": 2.296875,
      "rewards/success_reward/std": 0.9362190663814545,
      "rewards/trajectory_similarity_reward/mean": 0.37674954533576965,
      "rewards/trajectory_similarity_reward/std": 0.33037659525871277,
      "step": 552
    },
    {
      "clip_ratio/high_max": 0.015341704041929916,
      "clip_ratio/high_mean": 0.002958446133561665,
      "clip_ratio/low_mean": 0.0011553832482604776,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004113829232665012,
      "entropy": 0.3598606176674366,
      "epoch": 0.4507386653082017,
      "grad_norm": 12.994938850402832,
      "learning_rate": 4.394458027709861e-07,
      "loss": 7.3084,
      "reward": 2.454890727996826,
      "reward_std": 2.085770606994629,
      "rewards/reasoning_density_reward/mean": 0.32883185148239136,
      "rewards/reasoning_density_reward/std": 0.15447302162647247,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.4385586678981781,
      "rewards/trajectory_similarity_reward/std": 0.4313983619213104,
      "step": 553
    },
    {
      "clip_ratio/high_max": 0.016565235797315836,
      "clip_ratio/high_mean": 0.003439091786276549,
      "clip_ratio/low_mean": 0.0016647784832457546,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005103870382299647,
      "entropy": 0.37257180362939835,
      "epoch": 0.45155374426897604,
      "grad_norm": 15.50061321258545,
      "learning_rate": 4.3879380603096987e-07,
      "loss": -8.1869,
      "reward": 2.1983847618103027,
      "reward_std": 1.9784035682678223,
      "rewards/reasoning_density_reward/mean": 0.3142856955528259,
      "rewards/reasoning_density_reward/std": 0.09167542308568954,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.38409900665283203,
      "rewards/trajectory_similarity_reward/std": 0.40918147563934326,
      "step": 554
    },
    {
      "clip_ratio/high_max": 0.012967790942639112,
      "clip_ratio/high_mean": 0.0033621016918914393,
      "clip_ratio/low_mean": 0.0008135303651215509,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004175632013357244,
      "entropy": 0.33949117362499237,
      "epoch": 0.45236882322975036,
      "grad_norm": 14.195963859558105,
      "learning_rate": 4.381418092909535e-07,
      "loss": 4.6275,
      "reward": 3.50724720954895,
      "reward_std": 1.3437626659870148,
      "rewards/reasoning_density_reward/mean": 0.45446428656578064,
      "rewards/reasoning_density_reward/std": 0.09210738725960255,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.0930703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.6152828335762024,
      "rewards/trajectory_similarity_reward/std": 0.3256865292787552,
      "step": 555
    },
    {
      "clip_ratio/high_max": 0.01677731511881575,
      "clip_ratio/high_mean": 0.0038986751387710683,
      "clip_ratio/low_mean": 0.0009556648074067198,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004854340004385449,
      "entropy": 0.35855530947446823,
      "epoch": 0.45318390219052473,
      "grad_norm": 18.586414337158203,
      "learning_rate": 4.374898125509372e-07,
      "loss": 11.8646,
      "reward": 3.0171866416931152,
      "reward_std": 1.5499540567398071,
      "rewards/reasoning_density_reward/mean": 0.2764880955219269,
      "rewards/reasoning_density_reward/std": 0.20027358829975128,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.49069851636886597,
      "rewards/trajectory_similarity_reward/std": 0.39682114124298096,
      "step": 556
    },
    {
      "clip_ratio/high_max": 0.021138463984243572,
      "clip_ratio/high_mean": 0.0056194042263086885,
      "clip_ratio/low_mean": 0.00041736768253031187,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006036771839717403,
      "entropy": 0.35941996797919273,
      "epoch": 0.45399898115129905,
      "grad_norm": 17.315053939819336,
      "learning_rate": 4.368378158109209e-07,
      "loss": 7.3731,
      "reward": 3.2970802783966064,
      "reward_std": 1.5440999269485474,
      "rewards/reasoning_density_reward/mean": 0.40602679550647736,
      "rewards/reasoning_density_reward/std": 0.16176147013902664,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0968086123466492,
      "rewards/trajectory_similarity_reward/mean": 0.5473034977912903,
      "rewards/trajectory_similarity_reward/std": 0.3723365366458893,
      "step": 557
    },
    {
      "clip_ratio/high_max": 0.016768325003795326,
      "clip_ratio/high_mean": 0.0038714207767043263,
      "clip_ratio/low_mean": 0.0010672201169654727,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004938640922773629,
      "entropy": 0.3382100947201252,
      "epoch": 0.45481406011207337,
      "grad_norm": 15.948515892028809,
      "learning_rate": 4.361858190709046e-07,
      "loss": 9.4539,
      "reward": 4.159904479980469,
      "reward_std": 0.4882028102874756,
      "rewards/reasoning_density_reward/mean": 0.4203124940395355,
      "rewards/reasoning_density_reward/std": 0.12390142679214478,
      "rewards/success_reward/mean": 2.90625,
      "rewards/success_reward/std": 0.375,
      "rewards/trajectory_similarity_reward/mean": 0.8333417177200317,
      "rewards/trajectory_similarity_reward/std": 0.23912812769412994,
      "step": 558
    },
    {
      "clip_ratio/high_max": 0.012426528322976083,
      "clip_ratio/high_mean": 0.002792980340018403,
      "clip_ratio/low_mean": 0.0003495293567539193,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031425097113242373,
      "entropy": 0.3260241486132145,
      "epoch": 0.4556291390728477,
      "grad_norm": 16.810819625854492,
      "learning_rate": 4.355338223308883e-07,
      "loss": 10.8799,
      "reward": 2.534996747970581,
      "reward_std": 1.5017638802528381,
      "rewards/reasoning_density_reward/mean": 0.26566220819950104,
      "rewards/reasoning_density_reward/std": 0.15741927921772003,
      "rewards/success_reward/mean": 1.828125,
      "rewards/success_reward/std": 1.0879220962524414,
      "rewards/trajectory_similarity_reward/mean": 0.44120948761701584,
      "rewards/trajectory_similarity_reward/std": 0.3733525425195694,
      "step": 559
    },
    {
      "clip_ratio/high_max": 0.009915478760376573,
      "clip_ratio/high_mean": 0.0016737399710109457,
      "clip_ratio/low_mean": 0.0009361322408949491,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002609872230095789,
      "entropy": 0.36447614431381226,
      "epoch": 0.456444218033622,
      "grad_norm": 11.337642669677734,
      "learning_rate": 4.34881825590872e-07,
      "loss": -0.89,
      "step": 560
    },
    {
      "clip_ratio/high_max": 0.020306974533014,
      "clip_ratio/high_mean": 0.004155522721703164,
      "clip_ratio/low_mean": 0.0008797888149274513,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005035311434767209,
      "entropy": 0.3449097201228142,
      "epoch": 0.4572592969943963,
      "grad_norm": 14.88859748840332,
      "learning_rate": 4.3422982885085566e-07,
      "loss": 1.617,
      "reward": 2.701356053352356,
      "reward_std": 2.0815322399139404,
      "rewards/reasoning_density_reward/mean": 0.38638393580913544,
      "rewards/reasoning_density_reward/std": 0.12976709008216858,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.492667019367218,
      "rewards/trajectory_similarity_reward/mean": 0.5337220430374146,
      "rewards/trajectory_similarity_reward/std": 0.45427705347537994,
      "step": 561
    },
    {
      "clip_ratio/high_max": 0.012709788046777248,
      "clip_ratio/high_mean": 0.003432242156122811,
      "clip_ratio/low_mean": 0.0009075313064386137,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004339773397077806,
      "entropy": 0.337752029299736,
      "epoch": 0.45807437595517064,
      "grad_norm": 18.067981719970703,
      "learning_rate": 4.335778321108394e-07,
      "loss": 2.7796,
      "reward": 3.1187000274658203,
      "reward_std": 1.06424081325531,
      "rewards/reasoning_density_reward/mean": 0.41093748807907104,
      "rewards/reasoning_density_reward/std": 0.14111720025539398,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.6452624797821045,
      "rewards/trajectory_similarity_reward/std": 0.4556284248828888,
      "step": 562
    },
    {
      "clip_ratio/high_max": 0.01454808103153482,
      "clip_ratio/high_mean": 0.0036199339156155474,
      "clip_ratio/low_mean": 0.0007691394675930496,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004389073379570618,
      "entropy": 0.33856622502207756,
      "epoch": 0.45888945491594496,
      "grad_norm": 14.482069969177246,
      "learning_rate": 4.3292583537082317e-07,
      "loss": 15.9522,
      "reward": 2.7807042598724365,
      "reward_std": 2.0140767097473145,
      "rewards/reasoning_density_reward/mean": 0.3505208343267441,
      "rewards/reasoning_density_reward/std": 0.17600423842668533,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.4454171061515808,
      "rewards/trajectory_similarity_reward/mean": 0.5551833361387253,
      "rewards/trajectory_similarity_reward/std": 0.43651238083839417,
      "step": 563
    },
    {
      "clip_ratio/high_max": 0.011564271640963852,
      "clip_ratio/high_mean": 0.0027336031635059044,
      "clip_ratio/low_mean": 0.0006760891883459408,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034096922172466293,
      "entropy": 0.373022947460413,
      "epoch": 0.45970453387671933,
      "grad_norm": 13.495453834533691,
      "learning_rate": 4.322738386308068e-07,
      "loss": 2.4246,
      "reward": 2.498218059539795,
      "reward_std": 1.2063329219818115,
      "rewards/reasoning_density_reward/mean": 0.29598215222358704,
      "rewards/reasoning_density_reward/std": 0.17762069404125214,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.1618950366973877,
      "rewards/trajectory_similarity_reward/mean": 0.32723602652549744,
      "rewards/trajectory_similarity_reward/std": 0.39434343576431274,
      "step": 564
    },
    {
      "clip_ratio/high_max": 0.0065016826265491545,
      "clip_ratio/high_mean": 0.0014780929559492506,
      "clip_ratio/low_mean": 0.0011827840608020779,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026608769439917523,
      "entropy": 0.36408449336886406,
      "epoch": 0.46051961283749365,
      "grad_norm": 12.88775634765625,
      "learning_rate": 4.3162184189079057e-07,
      "loss": 6.4534,
      "reward": 1.5438021421432495,
      "reward_std": 1.9094183444976807,
      "rewards/reasoning_density_reward/mean": 0.2645833492279053,
      "rewards/reasoning_density_reward/std": 0.17531718313694,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.24796870350837708,
      "rewards/trajectory_similarity_reward/std": 0.3963402509689331,
      "step": 565
    },
    {
      "clip_ratio/high_max": 0.018652855418622494,
      "clip_ratio/high_mean": 0.003962171613238752,
      "clip_ratio/low_mean": 0.0009548897323838901,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004917061305604875,
      "entropy": 0.3550427407026291,
      "epoch": 0.46133469179826797,
      "grad_norm": 12.431384086608887,
      "learning_rate": 4.309698451507742e-07,
      "loss": -1.1687,
      "reward": 3.3790628910064697,
      "reward_std": 1.644874095916748,
      "rewards/reasoning_density_reward/mean": 0.4129464328289032,
      "rewards/reasoning_density_reward/std": 0.0962553545832634,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.6223665475845337,
      "rewards/trajectory_similarity_reward/std": 0.3798036575317383,
      "step": 566
    },
    {
      "clip_ratio/high_max": 0.014142922183964401,
      "clip_ratio/high_mean": 0.003249353787396103,
      "clip_ratio/low_mean": 0.0008733291797398124,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004122683007153682,
      "entropy": 0.3622935116291046,
      "epoch": 0.4621497707590423,
      "grad_norm": 13.88054084777832,
      "learning_rate": 4.3031784841075797e-07,
      "loss": 11.5922,
      "reward": 2.3520519733428955,
      "reward_std": 2.0334272384643555,
      "rewards/reasoning_density_reward/mean": 0.2984848618507385,
      "rewards/reasoning_density_reward/std": 0.15967042744159698,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.36606699228286743,
      "rewards/trajectory_similarity_reward/std": 0.39975956082344055,
      "step": 567
    },
    {
      "clip_ratio/high_max": 0.013290053873788565,
      "clip_ratio/high_mean": 0.00294685221160762,
      "clip_ratio/low_mean": 0.0007653999600734096,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003712252168043051,
      "entropy": 0.3554918020963669,
      "epoch": 0.4629648497198166,
      "grad_norm": 16.15614891052246,
      "learning_rate": 4.296658516707416e-07,
      "loss": -3.8654,
      "reward": 3.2939748764038086,
      "reward_std": 1.157406508922577,
      "rewards/reasoning_density_reward/mean": 0.3639136850833893,
      "rewards/reasoning_density_reward/std": 0.15524236112833023,
      "rewards/success_reward/mean": 2.390625,
      "rewards/success_reward/std": 0.8486150503158569,
      "rewards/trajectory_similarity_reward/mean": 0.5394362658262253,
      "rewards/trajectory_similarity_reward/std": 0.46232806146144867,
      "step": 568
    },
    {
      "clip_ratio/high_max": 0.017156257177703083,
      "clip_ratio/high_mean": 0.004102461491129361,
      "clip_ratio/low_mean": 0.0009994387874030508,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005101900198496878,
      "entropy": 0.35246555879712105,
      "epoch": 0.4637799286805909,
      "grad_norm": 19.012672424316406,
      "learning_rate": 4.2901385493072537e-07,
      "loss": 2.9575,
      "reward": 3.189117670059204,
      "reward_std": 1.7381527423858643,
      "rewards/reasoning_density_reward/mean": 0.2578125,
      "rewards/reasoning_density_reward/std": 0.16492053866386414,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6813051700592041,
      "rewards/trajectory_similarity_reward/std": 0.41581490635871887,
      "step": 569
    },
    {
      "clip_ratio/high_max": 0.008708903100341558,
      "clip_ratio/high_mean": 0.0019682730198837817,
      "clip_ratio/low_mean": 0.0014035365711606573,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033718095983203966,
      "entropy": 0.3606250658631325,
      "epoch": 0.46459500764136524,
      "grad_norm": 14.499275207519531,
      "learning_rate": 4.28361858190709e-07,
      "loss": -6.0356,
      "reward": 1.64427649974823,
      "reward_std": 1.2354811429977417,
      "rewards/reasoning_density_reward/mean": 0.23091518878936768,
      "rewards/reasoning_density_reward/std": 0.17212003469467163,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.10086134076118469,
      "rewards/trajectory_similarity_reward/std": 0.243430957198143,
      "step": 570
    },
    {
      "clip_ratio/high_max": 0.00946467590983957,
      "clip_ratio/high_mean": 0.0014084450813243166,
      "clip_ratio/low_mean": 0.000826163734018337,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002234608808066696,
      "entropy": 0.3761347606778145,
      "epoch": 0.46541008660213956,
      "grad_norm": 12.10898494720459,
      "learning_rate": 4.2770986145069277e-07,
      "loss": 1.571,
      "reward": 0.8221479654312134,
      "reward_std": 1.2809375524520874,
      "rewards/reasoning_density_reward/mean": 0.17812500894069672,
      "rewards/reasoning_density_reward/std": 0.13658788800239563,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.08152294158935547,
      "rewards/trajectory_similarity_reward/std": 0.24118547141551971,
      "step": 571
    },
    {
      "clip_ratio/high_max": 0.009534235839964822,
      "clip_ratio/high_mean": 0.0014692429940623697,
      "clip_ratio/low_mean": 0.0011971831627306528,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002666426149517065,
      "entropy": 0.3741855174303055,
      "epoch": 0.46622516556291393,
      "grad_norm": 12.035585403442383,
      "learning_rate": 4.270578647106764e-07,
      "loss": 1.6689,
      "reward": 1.366727352142334,
      "reward_std": 1.6748948097229004,
      "rewards/reasoning_density_reward/mean": 0.25208333134651184,
      "rewards/reasoning_density_reward/std": 0.15965259075164795,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.1771441102027893,
      "rewards/trajectory_similarity_reward/std": 0.3824470043182373,
      "step": 572
    },
    {
      "clip_ratio/high_max": 0.008331468736287206,
      "clip_ratio/high_mean": 0.0018131325778085738,
      "clip_ratio/low_mean": 0.0015312925752368756,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033444251166656613,
      "entropy": 0.358720775693655,
      "epoch": 0.46704024452368825,
      "grad_norm": 14.387141227722168,
      "learning_rate": 4.264058679706601e-07,
      "loss": 3.666,
      "reward": 2.4255566596984863,
      "reward_std": 1.6705716848373413,
      "rewards/reasoning_density_reward/mean": 0.3270833492279053,
      "rewards/reasoning_density_reward/std": 0.14705441892147064,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.4109731912612915,
      "rewards/trajectory_similarity_reward/std": 0.40887385606765747,
      "step": 573
    },
    {
      "clip_ratio/high_max": 0.006902437366079539,
      "clip_ratio/high_mean": 0.0015202344511635602,
      "clip_ratio/low_mean": 0.0009265876760764513,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002446822094498202,
      "entropy": 0.33538781479001045,
      "epoch": 0.46785532348446257,
      "grad_norm": 14.9566011428833,
      "learning_rate": 4.257538712306438e-07,
      "loss": 5.3335,
      "reward": 2.6574020385742188,
      "reward_std": 1.24073326587677,
      "rewards/reasoning_density_reward/mean": 0.32738097012043,
      "rewards/reasoning_density_reward/std": 0.1291305422782898,
      "rewards/success_reward/mean": 2.015625,
      "rewards/success_reward/std": 0.8786294460296631,
      "rewards/trajectory_similarity_reward/mean": 0.31439606845378876,
      "rewards/trajectory_similarity_reward/std": 0.396379292011261,
      "step": 574
    },
    {
      "clip_ratio/high_max": 0.011049385066144168,
      "clip_ratio/high_mean": 0.0028923638892592862,
      "clip_ratio/low_mean": 0.0013416558813332813,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004234019912473741,
      "entropy": 0.3455248177051544,
      "epoch": 0.4686704024452369,
      "grad_norm": 31.940448760986328,
      "learning_rate": 4.251018744906275e-07,
      "loss": 6.8659,
      "reward": 3.1505932807922363,
      "reward_std": 0.47806644439697266,
      "rewards/reasoning_density_reward/mean": 0.328125,
      "rewards/reasoning_density_reward/std": 0.1547538936138153,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 0.7685213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.4787183701992035,
      "rewards/trajectory_similarity_reward/std": 0.4481493830680847,
      "step": 575
    },
    {
      "clip_ratio/high_max": 0.014024338219314814,
      "clip_ratio/high_mean": 0.0037266259460011497,
      "clip_ratio/low_mean": 0.001639910311496351,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005366536206565797,
      "entropy": 0.3363899886608124,
      "epoch": 0.4694854814060112,
      "grad_norm": 13.188942909240723,
      "learning_rate": 4.2444987775061116e-07,
      "loss": -5.5325,
      "reward": 2.3746495246887207,
      "reward_std": 1.9234539270401,
      "rewards/reasoning_density_reward/mean": 0.2983630895614624,
      "rewards/reasoning_density_reward/std": 0.13734255731105804,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.3887866139411926,
      "rewards/trajectory_similarity_reward/std": 0.4109819829463959,
      "step": 576
    },
    {
      "clip_ratio/high_max": 0.012836737267207354,
      "clip_ratio/high_mean": 0.002450553249218501,
      "clip_ratio/low_mean": 0.0010548387981543783,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003505391941871494,
      "entropy": 0.3550970293581486,
      "epoch": 0.4703005603667855,
      "grad_norm": 13.619807243347168,
      "learning_rate": 4.237978810105949e-07,
      "loss": 3.7743,
      "reward": 2.762157917022705,
      "reward_std": 1.7440345287322998,
      "rewards/reasoning_density_reward/mean": 0.3185639977455139,
      "rewards/reasoning_density_reward/std": 0.16945409029722214,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.360331952571869,
      "rewards/trajectory_similarity_reward/mean": 0.47484391927719116,
      "rewards/trajectory_similarity_reward/std": 0.41989026963710785,
      "step": 577
    },
    {
      "clip_ratio/high_max": 0.016735209268517792,
      "clip_ratio/high_mean": 0.003149346128338948,
      "clip_ratio/low_mean": 0.0011136925895698369,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004263038747012615,
      "entropy": 0.3651650585234165,
      "epoch": 0.47111563932755984,
      "grad_norm": 15.632345199584961,
      "learning_rate": 4.2314588427057867e-07,
      "loss": 12.7592,
      "reward": 2.772813320159912,
      "reward_std": 2.0187315940856934,
      "rewards/reasoning_density_reward/mean": 0.37559524178504944,
      "rewards/reasoning_density_reward/std": 0.16936032474040985,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5222182273864746,
      "rewards/trajectory_similarity_reward/std": 0.4531330466270447,
      "step": 578
    },
    {
      "clip_ratio/high_max": 0.010811149724759161,
      "clip_ratio/high_mean": 0.0017983683210331947,
      "clip_ratio/low_mean": 0.001567758296005195,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033661266352282837,
      "entropy": 0.3432704582810402,
      "epoch": 0.47193071828833416,
      "grad_norm": 14.761733055114746,
      "learning_rate": 4.224938875305623e-07,
      "loss": -1.1615,
      "reward": 1.8367879390716553,
      "reward_std": 1.8681848049163818,
      "rewards/reasoning_density_reward/mean": 0.26732227206230164,
      "rewards/reasoning_density_reward/std": 0.1510384976863861,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.256965696811676,
      "rewards/trajectory_similarity_reward/std": 0.37701475620269775,
      "step": 579
    },
    {
      "clip_ratio/high_max": 0.013398616691119969,
      "clip_ratio/high_mean": 0.0033081389556173235,
      "clip_ratio/low_mean": 0.0010468514519743621,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004354990378487855,
      "entropy": 0.35819969326257706,
      "epoch": 0.47274579724910853,
      "grad_norm": 11.919381141662598,
      "learning_rate": 4.2184189079054606e-07,
      "loss": -1.555,
      "reward": 2.817079782485962,
      "reward_std": 1.6302745342254639,
      "rewards/reasoning_density_reward/mean": 0.2417410910129547,
      "rewards/reasoning_density_reward/std": 0.18617945909500122,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.4190887212753296,
      "rewards/trajectory_similarity_reward/std": 0.356999933719635,
      "step": 580
    },
    {
      "clip_ratio/high_max": 0.016501871752552688,
      "clip_ratio/high_mean": 0.0032411466527264565,
      "clip_ratio/low_mean": 0.0006709031986247282,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003912049753125757,
      "entropy": 0.3342396952211857,
      "epoch": 0.47356087620988285,
      "grad_norm": 15.686565399169922,
      "learning_rate": 4.211898940505297e-07,
      "loss": -4.5436,
      "reward": 2.494777202606201,
      "reward_std": 1.9458121061325073,
      "rewards/reasoning_density_reward/mean": 0.3197871744632721,
      "rewards/reasoning_density_reward/std": 0.08850736171007156,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.48748981952667236,
      "rewards/trajectory_similarity_reward/std": 0.44706347584724426,
      "step": 581
    },
    {
      "clip_ratio/high_max": 0.007989774137968197,
      "clip_ratio/high_mean": 0.0014396096485143062,
      "clip_ratio/low_mean": 0.0018679902859730646,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003307600025436841,
      "entropy": 0.3651244677603245,
      "epoch": 0.47437595517065717,
      "grad_norm": 14.441850662231445,
      "learning_rate": 4.2053789731051346e-07,
      "loss": 7.4235,
      "reward": 2.6179115772247314,
      "reward_std": 1.7007635831832886,
      "rewards/reasoning_density_reward/mean": 0.2680600881576538,
      "rewards/reasoning_density_reward/std": 0.18116000294685364,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.4748515486717224,
      "rewards/trajectory_similarity_reward/std": 0.40603601932525635,
      "step": 582
    },
    {
      "clip_ratio/high_max": 0.018344354641158134,
      "clip_ratio/high_mean": 0.005078492766187992,
      "clip_ratio/low_mean": 0.0009508318253210746,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006029324664268643,
      "entropy": 0.377111803740263,
      "epoch": 0.4751910341314315,
      "grad_norm": 17.01071548461914,
      "learning_rate": 4.198859005704971e-07,
      "loss": 2.6191,
      "reward": 3.140974521636963,
      "reward_std": 1.8879753351211548,
      "rewards/reasoning_density_reward/mean": 0.42386361956596375,
      "rewards/reasoning_density_reward/std": 0.12705066800117493,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.560860812664032,
      "rewards/trajectory_similarity_reward/std": 0.4141072630882263,
      "step": 583
    },
    {
      "clip_ratio/high_max": 0.011767158983275294,
      "clip_ratio/high_mean": 0.0025772510416572914,
      "clip_ratio/low_mean": 0.0015722177449788433,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004149468790274113,
      "entropy": 0.3435434550046921,
      "epoch": 0.4760061130922058,
      "grad_norm": 16.465890884399414,
      "learning_rate": 4.1923390383048086e-07,
      "loss": 6.6972,
      "reward": 3.341808795928955,
      "reward_std": 1.4392475485801697,
      "rewards/reasoning_density_reward/mean": 0.40305058658123016,
      "rewards/reasoning_density_reward/std": 0.11547904089093208,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2304179072380066,
      "rewards/trajectory_similarity_reward/mean": 0.5950081795454025,
      "rewards/trajectory_similarity_reward/std": 0.34431806206703186,
      "step": 584
    },
    {
      "clip_ratio/high_max": 0.010830255574546754,
      "clip_ratio/high_mean": 0.0027550527593120933,
      "clip_ratio/low_mean": 0.0025063969078473747,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005261449667159468,
      "entropy": 0.3699260652065277,
      "epoch": 0.4768211920529801,
      "grad_norm": 13.976653099060059,
      "learning_rate": 4.185819070904645e-07,
      "loss": 7.6542,
      "reward": 0.9579020142555237,
      "reward_std": 1.0869289636611938,
      "rewards/reasoning_density_reward/mean": 0.24551090598106384,
      "rewards/reasoning_density_reward/std": 0.12966489791870117,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.14989107847213745,
      "rewards/trajectory_similarity_reward/std": 0.3259730041027069,
      "step": 585
    },
    {
      "clip_ratio/high_max": 0.010023273585829884,
      "clip_ratio/high_mean": 0.0017776818494894542,
      "clip_ratio/low_mean": 0.0010789433363243006,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002856625083950348,
      "entropy": 0.3934973366558552,
      "epoch": 0.47763627101375444,
      "grad_norm": 11.973794937133789,
      "learning_rate": 4.1792991035044826e-07,
      "loss": 17.1749,
      "reward": 2.5416905879974365,
      "reward_std": 0.7355791926383972,
      "rewards/reasoning_density_reward/mean": 0.3010416626930237,
      "rewards/reasoning_density_reward/std": 0.17453311383724213,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.2718989849090576,
      "rewards/trajectory_similarity_reward/std": 0.41888442635536194,
      "step": 586
    },
    {
      "clip_ratio/high_max": 0.01698725071037188,
      "clip_ratio/high_mean": 0.003588614708860405,
      "clip_ratio/low_mean": 0.0010488522239029408,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004637467092834413,
      "entropy": 0.3441528342664242,
      "epoch": 0.47845134997452876,
      "grad_norm": 15.792671203613281,
      "learning_rate": 4.172779136104319e-07,
      "loss": 12.809,
      "reward": 3.437375545501709,
      "reward_std": 1.289502739906311,
      "rewards/reasoning_density_reward/mean": 0.3094449043273926,
      "rewards/reasoning_density_reward/std": 0.15279966592788696,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.5966807007789612,
      "rewards/trajectory_similarity_reward/std": 0.3150221109390259,
      "step": 587
    },
    {
      "clip_ratio/high_max": 0.01854595192708075,
      "clip_ratio/high_mean": 0.0051692268752958626,
      "clip_ratio/low_mean": 0.0004925918474327773,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005661818664520979,
      "entropy": 0.3373641148209572,
      "epoch": 0.47926642893530313,
      "grad_norm": 12.73204231262207,
      "learning_rate": 4.1662591687041566e-07,
      "loss": 3.4322,
      "reward": 3.1166861057281494,
      "reward_std": 1.913010835647583,
      "rewards/reasoning_density_reward/mean": 0.3578125238418579,
      "rewards/reasoning_density_reward/std": 0.17788544297218323,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 1.3381423950195312,
      "rewards/trajectory_similarity_reward/mean": 0.6026235222816467,
      "rewards/trajectory_similarity_reward/std": 0.45800694823265076,
      "step": 588
    },
    {
      "clip_ratio/high_max": 0.01766293460968882,
      "clip_ratio/high_mean": 0.004542790324194357,
      "clip_ratio/low_mean": 0.0007706150427111425,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005313405330525711,
      "entropy": 0.3473193235695362,
      "epoch": 0.48008150789607745,
      "grad_norm": 15.362180709838867,
      "learning_rate": 4.159739201303993e-07,
      "loss": 8.2866,
      "reward": 2.7066287994384766,
      "reward_std": 1.9972001314163208,
      "rewards/reasoning_density_reward/mean": 0.3160342276096344,
      "rewards/reasoning_density_reward/std": 0.12185101211071014,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5155945420265198,
      "rewards/trajectory_similarity_reward/std": 0.4418987035751343,
      "step": 589
    },
    {
      "clip_ratio/high_max": 0.019331959076225758,
      "clip_ratio/high_mean": 0.004190800624201074,
      "clip_ratio/low_mean": 0.0008512739004800096,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005042074393713847,
      "entropy": 0.36880314722657204,
      "epoch": 0.48089658685685177,
      "grad_norm": 14.195220947265625,
      "learning_rate": 4.15321923390383e-07,
      "loss": 9.3941,
      "reward": 3.1076560020446777,
      "reward_std": 1.339439868927002,
      "rewards/reasoning_density_reward/mean": 0.3343749940395355,
      "rewards/reasoning_density_reward/std": 0.19892942905426025,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.5232807397842407,
      "rewards/trajectory_similarity_reward/std": 0.397593230009079,
      "step": 590
    },
    {
      "clip_ratio/high_max": 0.01813520712312311,
      "clip_ratio/high_mean": 0.005514322983799502,
      "clip_ratio/low_mean": 0.0010313485836377367,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006545671698404476,
      "entropy": 0.3531506285071373,
      "epoch": 0.4817116658176261,
      "grad_norm": 15.17212963104248,
      "learning_rate": 4.146699266503667e-07,
      "loss": 7.1659,
      "reward": 2.782954216003418,
      "reward_std": 1.7850925922393799,
      "rewards/reasoning_density_reward/mean": 0.2814703583717346,
      "rewards/reasoning_density_reward/std": 0.19351571798324585,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.43898388743400574,
      "rewards/trajectory_similarity_reward/std": 0.37825068831443787,
      "step": 591
    },
    {
      "clip_ratio/high_max": 0.018588822742458433,
      "clip_ratio/high_mean": 0.005690486737876199,
      "clip_ratio/low_mean": 0.0005485238070832565,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006239010515855625,
      "entropy": 0.3401137851178646,
      "epoch": 0.4825267447784004,
      "grad_norm": 12.795368194580078,
      "learning_rate": 4.140179299103504e-07,
      "loss": 2.6322,
      "reward": 3.58992600440979,
      "reward_std": 1.568352460861206,
      "rewards/reasoning_density_reward/mean": 0.44999998807907104,
      "rewards/reasoning_density_reward/std": 0.08944271504878998,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.702426016330719,
      "rewards/trajectory_similarity_reward/std": 0.4233934283256531,
      "step": 592
    },
    {
      "clip_ratio/high_max": 0.02141480887075886,
      "clip_ratio/high_mean": 0.005334142661013175,
      "clip_ratio/low_mean": 0.00046250977175077423,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005796652432763949,
      "entropy": 0.3599207364022732,
      "epoch": 0.4833418237391747,
      "grad_norm": 12.382328987121582,
      "learning_rate": 4.1336593317033416e-07,
      "loss": 6.4115,
      "reward": 3.384814500808716,
      "reward_std": 1.372208595275879,
      "rewards/reasoning_density_reward/mean": 0.4236505627632141,
      "rewards/reasoning_density_reward/std": 0.10526878759264946,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 0.9999804198741913,
      "rewards/trajectory_similarity_reward/mean": 0.617413803935051,
      "rewards/trajectory_similarity_reward/std": 0.44272375106811523,
      "step": 593
    },
    {
      "clip_ratio/high_max": 0.00865881674690172,
      "clip_ratio/high_mean": 0.0016452554482384585,
      "clip_ratio/low_mean": 0.002116827861755155,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003762083317269571,
      "entropy": 0.357397235929966,
      "epoch": 0.48415690269994904,
      "grad_norm": 13.117693901062012,
      "learning_rate": 4.127139364303178e-07,
      "loss": -4.1334,
      "reward": 1.0382909774780273,
      "reward_std": 1.5980769395828247,
      "rewards/reasoning_density_reward/mean": 0.23950894176959991,
      "rewards/reasoning_density_reward/std": 0.16153742372989655,
      "rewards/success_reward/mean": 0.65625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.14253197610378265,
      "rewards/trajectory_similarity_reward/std": 0.32693859934806824,
      "step": 594
    },
    {
      "clip_ratio/high_max": 0.01460710831452161,
      "clip_ratio/high_mean": 0.002707427687710151,
      "clip_ratio/low_mean": 0.0015980244461388793,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043054522830061615,
      "entropy": 0.3160870112478733,
      "epoch": 0.48497198166072336,
      "grad_norm": 17.063623428344727,
      "learning_rate": 4.1206193969030156e-07,
      "loss": 7.6514,
      "reward": 3.464766502380371,
      "reward_std": 1.0720587372779846,
      "rewards/reasoning_density_reward/mean": 0.44032737612724304,
      "rewards/reasoning_density_reward/std": 0.08757923170924187,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2304179072380066,
      "rewards/trajectory_similarity_reward/mean": 0.6806890964508057,
      "rewards/trajectory_similarity_reward/std": 0.3681385964155197,
      "step": 595
    },
    {
      "clip_ratio/high_max": 0.015156440204009414,
      "clip_ratio/high_mean": 0.0035413292789598927,
      "clip_ratio/low_mean": 0.0016424941877630772,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005183823450352065,
      "entropy": 0.3094256781041622,
      "epoch": 0.48578706062149773,
      "grad_norm": 13.846409797668457,
      "learning_rate": 4.114099429502852e-07,
      "loss": 2.0274,
      "reward": 3.8409533500671387,
      "reward_std": 0.9405041337013245,
      "rewards/reasoning_density_reward/mean": 0.40833333134651184,
      "rewards/reasoning_density_reward/std": 0.17191730439662933,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.8076199293136597,
      "rewards/trajectory_similarity_reward/std": 0.32795584201812744,
      "step": 596
    },
    {
      "clip_ratio/high_max": 0.016674078069627285,
      "clip_ratio/high_mean": 0.0032966075523290783,
      "clip_ratio/low_mean": 0.001225441534188576,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00452204910106957,
      "entropy": 0.36253681778907776,
      "epoch": 0.48660213958227205,
      "grad_norm": 18.753732681274414,
      "learning_rate": 4.1075794621026896e-07,
      "loss": 5.8604,
      "reward": 2.7088162899017334,
      "reward_std": 1.9455974102020264,
      "rewards/reasoning_density_reward/mean": 0.3614955544471741,
      "rewards/reasoning_density_reward/std": 0.1415339782834053,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.4482120275497437,
      "rewards/trajectory_similarity_reward/mean": 0.47232070565223694,
      "rewards/trajectory_similarity_reward/std": 0.4113554358482361,
      "step": 597
    },
    {
      "clip_ratio/high_max": 0.021044696564786136,
      "clip_ratio/high_mean": 0.005602792807621881,
      "clip_ratio/low_mean": 0.0009453159327676985,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0065481087658554316,
      "entropy": 0.35568562895059586,
      "epoch": 0.48741721854304637,
      "grad_norm": 15.010366439819336,
      "learning_rate": 4.101059494702526e-07,
      "loss": 9.2363,
      "reward": 2.2364728450775146,
      "reward_std": 0.2964957654476166,
      "rewards/reasoning_density_reward/mean": 0.3297078013420105,
      "rewards/reasoning_density_reward/std": 0.1662147343158722,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.4067651331424713,
      "rewards/trajectory_similarity_reward/std": 0.4770828187465668,
      "step": 598
    },
    {
      "clip_ratio/high_max": 0.01866728503955528,
      "clip_ratio/high_mean": 0.006085613917093724,
      "clip_ratio/low_mean": 0.0009968574922822881,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007082471449393779,
      "entropy": 0.34970155730843544,
      "epoch": 0.4882322975038207,
      "grad_norm": 15.020228385925293,
      "learning_rate": 4.0945395273023636e-07,
      "loss": 8.4769,
      "reward": 3.0070228576660156,
      "reward_std": 2.0165796279907227,
      "rewards/reasoning_density_reward/mean": 0.3786458373069763,
      "rewards/reasoning_density_reward/std": 0.15171760320663452,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5658771395683289,
      "rewards/trajectory_similarity_reward/std": 0.4315800070762634,
      "step": 599
    },
    {
      "clip_ratio/high_max": 0.015136048255953938,
      "clip_ratio/high_mean": 0.0037473003831109963,
      "clip_ratio/low_mean": 0.0011298454264760949,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004877145802311134,
      "entropy": 0.3444668333977461,
      "epoch": 0.489047376464595,
      "grad_norm": 13.346331596374512,
      "learning_rate": 4.0880195599022e-07,
      "loss": 8.5062,
      "reward": 2.2854514122009277,
      "reward_std": 1.9860186576843262,
      "rewards/reasoning_density_reward/mean": 0.32176053524017334,
      "rewards/reasoning_density_reward/std": 0.18089346587657928,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.3699409067630768,
      "rewards/trajectory_similarity_reward/std": 0.4077860414981842,
      "step": 600
    },
    {
      "clip_ratio/high_max": 0.015200530644506216,
      "clip_ratio/high_mean": 0.003149329946609214,
      "clip_ratio/low_mean": 0.0015401792770717293,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004689509281888604,
      "entropy": 0.37878959625959396,
      "epoch": 0.4898624554253693,
      "grad_norm": 12.3892183303833,
      "learning_rate": 4.0814995925020376e-07,
      "loss": 19.824,
      "reward": 1.8814313411712646,
      "reward_std": 1.7596986293792725,
      "rewards/reasoning_density_reward/mean": 0.2239583432674408,
      "rewards/reasoning_density_reward/std": 0.17148250341415405,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.2512229084968567,
      "rewards/trajectory_similarity_reward/std": 0.33296331763267517,
      "step": 601
    },
    {
      "clip_ratio/high_max": 0.012223528872709721,
      "clip_ratio/high_mean": 0.002520384659874253,
      "clip_ratio/low_mean": 0.0019228220771765336,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004443206737050787,
      "entropy": 0.37664535269141197,
      "epoch": 0.49067753438614364,
      "grad_norm": 13.228918075561523,
      "learning_rate": 4.074979625101874e-07,
      "loss": -3.3784,
      "reward": 3.745124578475952,
      "reward_std": 0.6495696306228638,
      "rewards/reasoning_density_reward/mean": 0.4476190507411957,
      "rewards/reasoning_density_reward/std": 0.07962022721767426,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.578755259513855,
      "rewards/trajectory_similarity_reward/std": 0.31249740719795227,
      "step": 602
    },
    {
      "clip_ratio/high_max": 0.016302698786603287,
      "clip_ratio/high_mean": 0.0032728941914683674,
      "clip_ratio/low_mean": 0.0008643052133265883,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004137199372053146,
      "entropy": 0.3661930114030838,
      "epoch": 0.49149261334691796,
      "grad_norm": 15.935176849365234,
      "learning_rate": 4.0684596577017116e-07,
      "loss": 17.9134,
      "reward": 2.708522319793701,
      "reward_std": 2.082267999649048,
      "rewards/reasoning_density_reward/mean": 0.359375,
      "rewards/reasoning_density_reward/std": 0.1781560629606247,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.47414714097976685,
      "rewards/trajectory_similarity_reward/std": 0.3831002712249756,
      "step": 603
    },
    {
      "clip_ratio/high_max": 0.008245577540947124,
      "clip_ratio/high_mean": 0.0019025681940547656,
      "clip_ratio/low_mean": 0.0010687597678042948,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029713279946008697,
      "entropy": 0.34687134996056557,
      "epoch": 0.49230769230769234,
      "grad_norm": 13.359745025634766,
      "learning_rate": 4.061939690301548e-07,
      "loss": 13.8818,
      "reward": 2.8433746099472046,
      "reward_std": 1.4413495063781738,
      "rewards/reasoning_density_reward/mean": 0.3390926718711853,
      "rewards/reasoning_density_reward/std": 0.17088667303323746,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.325729787349701,
      "rewards/trajectory_similarity_reward/mean": 0.39490702748298645,
      "rewards/trajectory_similarity_reward/std": 0.33088111877441406,
      "step": 604
    },
    {
      "clip_ratio/high_max": 0.01587693707551807,
      "clip_ratio/high_mean": 0.0031885761563898996,
      "clip_ratio/low_mean": 0.0011749086352210725,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004363484782516025,
      "entropy": 0.3444707319140434,
      "epoch": 0.49312277126846665,
      "grad_norm": 16.822750091552734,
      "learning_rate": 4.0554197229013856e-07,
      "loss": 12.853,
      "reward": 3.8424339294433594,
      "reward_std": 0.9349805116653442,
      "rewards/reasoning_density_reward/mean": 0.42395833134651184,
      "rewards/reasoning_density_reward/std": 0.15226303040981293,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.6059756278991699,
      "rewards/trajectory_similarity_reward/std": 0.4073379635810852,
      "step": 605
    },
    {
      "clip_ratio/high_max": 0.014326575794257224,
      "clip_ratio/high_mean": 0.004307347961002961,
      "clip_ratio/low_mean": 0.0014611130936827976,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005768461036495864,
      "entropy": 0.3492964766919613,
      "epoch": 0.49393785022924097,
      "grad_norm": 11.880433082580566,
      "learning_rate": 4.048899755501222e-07,
      "loss": 5.168,
      "reward": 1.703645944595337,
      "reward_std": 1.951566219329834,
      "rewards/reasoning_density_reward/mean": 0.26614585518836975,
      "rewards/reasoning_density_reward/std": 0.15511253476142883,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.3125,
      "rewards/trajectory_similarity_reward/std": 0.4787135720252991,
      "step": 606
    },
    {
      "clip_ratio/high_max": 0.008439980505499989,
      "clip_ratio/high_mean": 0.0013098371637170203,
      "clip_ratio/low_mean": 0.0019050235423492268,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032148606260307133,
      "entropy": 0.3860839754343033,
      "epoch": 0.4947529291900153,
      "grad_norm": 13.462860107421875,
      "learning_rate": 4.042379788101059e-07,
      "loss": 0.6529,
      "reward": 2.1924238204956055,
      "reward_std": 1.0145924091339111,
      "rewards/reasoning_density_reward/mean": 0.15357142686843872,
      "rewards/reasoning_density_reward/std": 0.15920722484588623,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 0.670820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.16385261714458466,
      "rewards/trajectory_similarity_reward/std": 0.32643353939056396,
      "step": 607
    },
    {
      "clip_ratio/high_max": 0.009012866823468357,
      "clip_ratio/high_mean": 0.0018454236269462854,
      "clip_ratio/low_mean": 0.0011399507966416422,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029853743908461183,
      "entropy": 0.3582661300897598,
      "epoch": 0.4955680081507896,
      "grad_norm": 30.682567596435547,
      "learning_rate": 4.0358598207008966e-07,
      "loss": 11.8277,
      "reward": 2.5856773853302,
      "reward_std": 1.8839988708496094,
      "rewards/reasoning_density_reward/mean": 0.34375,
      "rewards/reasoning_density_reward/std": 0.18962682783603668,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.3669271767139435,
      "rewards/trajectory_similarity_reward/std": 0.36039838194847107,
      "step": 608
    },
    {
      "clip_ratio/high_max": 0.008890593773685396,
      "clip_ratio/high_mean": 0.0015765439238748513,
      "clip_ratio/low_mean": 0.0012163616265752353,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027929056814173236,
      "entropy": 0.3756333068013191,
      "epoch": 0.4963830871115639,
      "grad_norm": 11.380719184875488,
      "learning_rate": 4.029339853300733e-07,
      "loss": 9.2531,
      "reward": 1.5138306617736816,
      "reward_std": 0.9364351630210876,
      "rewards/reasoning_density_reward/mean": 0.13906249403953552,
      "rewards/reasoning_density_reward/std": 0.1974775791168213,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.2512494325637817,
      "rewards/trajectory_similarity_reward/mean": 0.15601816773414612,
      "rewards/trajectory_similarity_reward/std": 0.27979376912117004,
      "step": 609
    },
    {
      "clip_ratio/high_max": 0.004478567803744227,
      "clip_ratio/high_mean": 0.0009979650567402132,
      "clip_ratio/low_mean": 0.0011241050669923425,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002122070152836386,
      "entropy": 0.38225798308849335,
      "epoch": 0.49719816607233824,
      "grad_norm": 10.38000202178955,
      "learning_rate": 4.0228198859005706e-07,
      "loss": 20.508,
      "reward": 1.5196212530136108,
      "reward_std": 1.1949071884155273,
      "rewards/reasoning_density_reward/mean": 0.2808035612106323,
      "rewards/reasoning_density_reward/std": 0.1919594407081604,
      "rewards/success_reward/mean": 0.9375,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.30131757259368896,
      "rewards/trajectory_similarity_reward/std": 0.46255844831466675,
      "step": 610
    },
    {
      "clip_ratio/high_max": 0.01850990968523547,
      "clip_ratio/high_mean": 0.003933230429538526,
      "clip_ratio/low_mean": 0.0016518271149834618,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005585057559073903,
      "entropy": 0.33438607677817345,
      "epoch": 0.49801324503311256,
      "grad_norm": 14.038783073425293,
      "learning_rate": 4.016299918500407e-07,
      "loss": 3.9181,
      "reward": 2.654411792755127,
      "reward_std": 1.8151758909225464,
      "rewards/reasoning_density_reward/mean": 0.3777901828289032,
      "rewards/reasoning_density_reward/std": 0.11190786212682724,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.401621550321579,
      "rewards/trajectory_similarity_reward/std": 0.42846906185150146,
      "step": 611
    },
    {
      "clip_ratio/high_max": 0.013585080800112337,
      "clip_ratio/high_mean": 0.002941684768302366,
      "clip_ratio/low_mean": 0.0013768811186309904,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004318565857829526,
      "entropy": 0.3741517625749111,
      "epoch": 0.49882832399388694,
      "grad_norm": 14.866117477416992,
      "learning_rate": 4.0097799511002446e-07,
      "loss": 2.1241,
      "reward": 1.8626463413238525,
      "reward_std": 1.3210163116455078,
      "rewards/reasoning_density_reward/mean": 0.18058034777641296,
      "rewards/reasoning_density_reward/std": 0.17842836678028107,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.157853603363037,
      "rewards/trajectory_similarity_reward/mean": 0.0883159413933754,
      "rewards/trajectory_similarity_reward/std": 0.21366721391677856,
      "step": 612
    },
    {
      "clip_ratio/high_max": 0.013791412056889385,
      "clip_ratio/high_mean": 0.0030310198562801816,
      "clip_ratio/low_mean": 0.00048493966460227966,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035159595136065036,
      "entropy": 0.36052530631422997,
      "epoch": 0.49964340295466125,
      "grad_norm": 14.837129592895508,
      "learning_rate": 4.003259983700081e-07,
      "loss": -0.1726,
      "reward": 3.8621160984039307,
      "reward_std": 1.051047921180725,
      "rewards/reasoning_density_reward/mean": 0.43958336114883423,
      "rewards/reasoning_density_reward/std": 0.09287086874246597,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.7037826180458069,
      "rewards/trajectory_similarity_reward/std": 0.3809140920639038,
      "step": 613
    },
    {
      "clip_ratio/high_max": 0.010445821681059897,
      "clip_ratio/high_mean": 0.002348543028347194,
      "clip_ratio/low_mean": 0.0011013264702341985,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034498696004447993,
      "entropy": 0.36270686239004135,
      "epoch": 0.5004584819154355,
      "grad_norm": 15.83718204498291,
      "learning_rate": 3.9967400162999186e-07,
      "loss": 3.0967,
      "reward": 2.017586648464203,
      "reward_std": 1.1573643386363983,
      "rewards/reasoning_density_reward/mean": 0.29021576046943665,
      "rewards/reasoning_density_reward/std": 0.13782726228237152,
      "rewards/success_reward/mean": 1.453125,
      "rewards/success_reward/std": 1.3531213998794556,
      "rewards/trajectory_similarity_reward/mean": 0.2742459326982498,
      "rewards/trajectory_similarity_reward/std": 0.3538319915533066,
      "step": 614
    },
    {
      "clip_ratio/high_max": 0.017068785469746217,
      "clip_ratio/high_mean": 0.003690166035084985,
      "clip_ratio/low_mean": 0.0007494177916669287,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004439583761268295,
      "entropy": 0.38299374282360077,
      "epoch": 0.5012735608762099,
      "grad_norm": 13.033148765563965,
      "learning_rate": 3.9902200488997556e-07,
      "loss": -12.9027,
      "reward": 2.453125,
      "reward_std": 0.24728435277938843,
      "rewards/reasoning_density_reward/mean": 0.359375,
      "rewards/reasoning_density_reward/std": 0.16655205190181732,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.496871829032898,
      "rewards/trajectory_similarity_reward/mean": 0.5,
      "rewards/trajectory_similarity_reward/std": 0.5163977742195129,
      "step": 615
    },
    {
      "clip_ratio/high_max": 0.016024782555177808,
      "clip_ratio/high_mean": 0.004179618103080429,
      "clip_ratio/low_mean": 0.0006069241135264747,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004786542122019455,
      "entropy": 0.329314187169075,
      "epoch": 0.5020886398369843,
      "grad_norm": 14.701523780822754,
      "learning_rate": 3.9837000814995926e-07,
      "loss": -1.4874,
      "reward": 3.9917147159576416,
      "reward_std": 0.8753069639205933,
      "rewards/reasoning_density_reward/mean": 0.4671874940395355,
      "rewards/reasoning_density_reward/std": 0.07229497283697128,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8158584833145142,
      "rewards/trajectory_similarity_reward/mean": 0.8057771921157837,
      "rewards/trajectory_similarity_reward/std": 0.3245881497859955,
      "step": 616
    },
    {
      "clip_ratio/high_max": 0.014648964861407876,
      "clip_ratio/high_mean": 0.003759691971936263,
      "clip_ratio/low_mean": 0.0006567689997609705,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0044164610299048945,
      "entropy": 0.3127261511981487,
      "epoch": 0.5029037187977585,
      "grad_norm": 16.93784523010254,
      "learning_rate": 3.9771801140994296e-07,
      "loss": 6.2182,
      "reward": 2.7484335899353027,
      "reward_std": 0.658659428358078,
      "rewards/reasoning_density_reward/mean": 0.3968749940395355,
      "rewards/reasoning_density_reward/std": 0.08384547010064125,
      "rewards/success_reward/mean": 1.921875,
      "rewards/success_reward/std": 0.8510740399360657,
      "rewards/trajectory_similarity_reward/mean": 0.4296836256980896,
      "rewards/trajectory_similarity_reward/std": 0.3016180396080017,
      "step": 617
    },
    {
      "clip_ratio/high_max": 0.015147597528994083,
      "clip_ratio/high_mean": 0.00432334428478498,
      "clip_ratio/low_mean": 0.0006657042831648141,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004989048451534472,
      "entropy": 0.31382225826382637,
      "epoch": 0.5037187977585329,
      "grad_norm": 17.096071243286133,
      "learning_rate": 3.9706601466992666e-07,
      "loss": -8.7725,
      "reward": 3.638206124305725,
      "reward_std": 1.4062729477882385,
      "rewards/reasoning_density_reward/mean": 0.4091145843267441,
      "rewards/reasoning_density_reward/std": 0.1289711482822895,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.9796755611896515,
      "rewards/trajectory_similarity_reward/mean": 0.6978415548801422,
      "rewards/trajectory_similarity_reward/std": 0.3967580199241638,
      "step": 618
    },
    {
      "clip_ratio/high_max": 0.01446433481760323,
      "clip_ratio/high_mean": 0.002601201835204847,
      "clip_ratio/low_mean": 0.0007836033000785392,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033848050916276406,
      "entropy": 0.385078564286232,
      "epoch": 0.5045338767193072,
      "grad_norm": 11.828485488891602,
      "learning_rate": 3.9641401792991036e-07,
      "loss": -0.2919,
      "reward": 2.7943739891052246,
      "reward_std": 0.7178789973258972,
      "rewards/reasoning_density_reward/mean": 0.3723214268684387,
      "rewards/reasoning_density_reward/std": 0.17476661503314972,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.5470526218414307,
      "rewards/trajectory_similarity_reward/std": 0.5018967986106873,
      "step": 619
    },
    {
      "clip_ratio/high_max": 0.011616812262218446,
      "clip_ratio/high_mean": 0.0021796357541461475,
      "clip_ratio/low_mean": 0.0015605756270815618,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003740211308468133,
      "entropy": 0.3964426927268505,
      "epoch": 0.5053489556800815,
      "grad_norm": 15.567601203918457,
      "learning_rate": 3.9576202118989406e-07,
      "loss": -7.7476,
      "reward": 1.733097493648529,
      "reward_std": 1.8104114532470703,
      "rewards/reasoning_density_reward/mean": 0.2977689951658249,
      "rewards/reasoning_density_reward/std": 0.14355003088712692,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.4456666707992554,
      "rewards/trajectory_similarity_reward/mean": 0.3103284388780594,
      "rewards/trajectory_similarity_reward/std": 0.45008158683776855,
      "step": 620
    },
    {
      "clip_ratio/high_max": 0.013555237324908376,
      "clip_ratio/high_mean": 0.00250417462666519,
      "clip_ratio/low_mean": 0.0017210840342158917,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042252586426911876,
      "entropy": 0.33415190875530243,
      "epoch": 0.5061640346408558,
      "grad_norm": 14.21523380279541,
      "learning_rate": 3.9511002444987776e-07,
      "loss": 1.3677,
      "reward": 2.8068342208862305,
      "reward_std": 1.27742338180542,
      "rewards/reasoning_density_reward/mean": 0.3218750059604645,
      "rewards/reasoning_density_reward/std": 0.19405218958854675,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.5162090063095093,
      "rewards/trajectory_similarity_reward/std": 0.4744463860988617,
      "step": 621
    },
    {
      "clip_ratio/high_max": 0.015014604607131332,
      "clip_ratio/high_mean": 0.003010186235769652,
      "clip_ratio/low_mean": 0.0019285200323793106,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004938706348184496,
      "entropy": 0.3583928011357784,
      "epoch": 0.5069791136016302,
      "grad_norm": 21.27558135986328,
      "learning_rate": 3.944580277098614e-07,
      "loss": 8.3489,
      "reward": 3.0966925621032715,
      "reward_std": 1.0676543712615967,
      "rewards/reasoning_density_reward/mean": 0.3433035612106323,
      "rewards/reasoning_density_reward/std": 0.16904029250144958,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 0.7685213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.4096388816833496,
      "rewards/trajectory_similarity_reward/std": 0.4800129234790802,
      "step": 622
    },
    {
      "clip_ratio/high_max": 0.012006921169813722,
      "clip_ratio/high_mean": 0.0028114653105149046,
      "clip_ratio/low_mean": 0.0015468877136299852,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043583530496107414,
      "entropy": 0.355144415050745,
      "epoch": 0.5077941925624044,
      "grad_norm": 15.094049453735352,
      "learning_rate": 3.938060309698451e-07,
      "loss": -6.1199,
      "reward": 2.8955243825912476,
      "reward_std": 1.1124079823493958,
      "rewards/reasoning_density_reward/mean": 0.31814631819725037,
      "rewards/reasoning_density_reward/std": 0.15510524064302444,
      "rewards/success_reward/mean": 2.203125,
      "rewards/success_reward/std": 1.029909372329712,
      "rewards/trajectory_similarity_reward/mean": 0.37425291538238525,
      "rewards/trajectory_similarity_reward/std": 0.4373335540294647,
      "step": 623
    },
    {
      "clip_ratio/high_max": 0.013292582239955664,
      "clip_ratio/high_mean": 0.0029779762553516775,
      "clip_ratio/low_mean": 0.0009919719450408593,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003969948185840622,
      "entropy": 0.36307232454419136,
      "epoch": 0.5086092715231788,
      "grad_norm": 14.01499080657959,
      "learning_rate": 3.931540342298288e-07,
      "loss": -3.6692,
      "reward": 1.3555946350097656,
      "reward_std": 1.0132015943527222,
      "rewards/reasoning_density_reward/mean": 0.2723214328289032,
      "rewards/reasoning_density_reward/std": 0.10474702715873718,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 0.9031196236610413,
      "rewards/trajectory_similarity_reward/mean": 0.05202318727970123,
      "rewards/trajectory_similarity_reward/std": 0.20809276401996613,
      "step": 624
    },
    {
      "clip_ratio/high_max": 0.013056779222097248,
      "clip_ratio/high_mean": 0.0030788336662226357,
      "clip_ratio/low_mean": 0.00129348265909357,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004372316267108545,
      "entropy": 0.35636478289961815,
      "epoch": 0.5094243504839532,
      "grad_norm": 12.881514549255371,
      "learning_rate": 3.925020374898125e-07,
      "loss": 2.3674,
      "reward": 3.2796852588653564,
      "reward_std": 0.8760370016098022,
      "rewards/reasoning_density_reward/mean": 0.4177083373069763,
      "rewards/reasoning_density_reward/std": 0.11602501571178436,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 0.670820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.23697693645954132,
      "rewards/trajectory_similarity_reward/std": 0.31728604435920715,
      "step": 625
    },
    {
      "clip_ratio/high_max": 0.01749666253454052,
      "clip_ratio/high_mean": 0.004492772259254707,
      "clip_ratio/low_mean": 0.0012433035299181938,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00573607573096524,
      "entropy": 0.3335145004093647,
      "epoch": 0.5102394294447274,
      "grad_norm": 19.04633331298828,
      "learning_rate": 3.918500407497962e-07,
      "loss": 10.689,
      "reward": 3.2850193977355957,
      "reward_std": 1.7149498462677002,
      "rewards/reasoning_density_reward/mean": 0.4047619104385376,
      "rewards/reasoning_density_reward/std": 0.16765692830085754,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5365073680877686,
      "rewards/trajectory_similarity_reward/std": 0.36331117153167725,
      "step": 626
    },
    {
      "clip_ratio/high_max": 0.017390119726769626,
      "clip_ratio/high_mean": 0.003554816954419948,
      "clip_ratio/low_mean": 0.0011615079856710508,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0047163249691948295,
      "entropy": 0.3472238890826702,
      "epoch": 0.5110545084055018,
      "grad_norm": 15.77662467956543,
      "learning_rate": 3.911980440097799e-07,
      "loss": 7.2969,
      "reward": 2.6804686784744263,
      "reward_std": 1.4430390000343323,
      "rewards/reasoning_density_reward/mean": 0.3536830395460129,
      "rewards/reasoning_density_reward/std": 0.16383542865514755,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.277839720249176,
      "rewards/trajectory_similarity_reward/mean": 0.45178554952144623,
      "rewards/trajectory_similarity_reward/std": 0.4687570780515671,
      "step": 627
    },
    {
      "clip_ratio/high_max": 0.013563326268922538,
      "clip_ratio/high_mean": 0.003419325912545901,
      "clip_ratio/low_mean": 0.000704400590620935,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004123726554098539,
      "entropy": 0.35702960938215256,
      "epoch": 0.5118695873662761,
      "grad_norm": 19.7922306060791,
      "learning_rate": 3.905460472697636e-07,
      "loss": 4.2735,
      "reward": 1.698409080505371,
      "reward_std": 1.851373314857483,
      "rewards/reasoning_density_reward/mean": 0.21454274654388428,
      "rewards/reasoning_density_reward/std": 0.15582375228405,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.26511645317077637,
      "rewards/trajectory_similarity_reward/std": 0.3823077380657196,
      "step": 628
    },
    {
      "clip_ratio/high_max": 0.012219763710163534,
      "clip_ratio/high_mean": 0.0020373940642457455,
      "clip_ratio/low_mean": 0.0006603385409107432,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002697732619708404,
      "entropy": 0.35615766048431396,
      "epoch": 0.5126846663270505,
      "grad_norm": 13.712018013000488,
      "learning_rate": 3.8989405052974736e-07,
      "loss": 8.4171,
      "reward": 2.173374891281128,
      "reward_std": 0.24707689881324768,
      "rewards/reasoning_density_reward/mean": 0.2802083194255829,
      "rewards/reasoning_density_reward/std": 0.205770343542099,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.39316678047180176,
      "rewards/trajectory_similarity_reward/std": 0.4208589196205139,
      "step": 629
    },
    {
      "clip_ratio/high_max": 0.011378217372111976,
      "clip_ratio/high_mean": 0.002090569309075363,
      "clip_ratio/low_mean": 0.0004846433403145056,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025752126675797626,
      "entropy": 0.3749445080757141,
      "epoch": 0.5134997452878247,
      "grad_norm": 12.965142250061035,
      "learning_rate": 3.8924205378973106e-07,
      "loss": 13.1162,
      "reward": 3.432776689529419,
      "reward_std": 1.0365281105041504,
      "rewards/reasoning_density_reward/mean": 0.3555803894996643,
      "rewards/reasoning_density_reward/std": 0.16550111770629883,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.6396963596343994,
      "rewards/trajectory_similarity_reward/std": 0.38957855105400085,
      "step": 630
    },
    {
      "clip_ratio/high_max": 0.01427154877455905,
      "clip_ratio/high_mean": 0.002243930437543895,
      "clip_ratio/low_mean": 0.0014023902840563096,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036463207434280775,
      "entropy": 0.37803400680422783,
      "epoch": 0.5143148242485991,
      "grad_norm": 11.531110763549805,
      "learning_rate": 3.8859005704971476e-07,
      "loss": 5.3951,
      "reward": 2.3415815830230713,
      "reward_std": 2.0642004013061523,
      "rewards/reasoning_density_reward/mean": 0.24114583432674408,
      "rewards/reasoning_density_reward/std": 0.17393775284290314,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.41293561458587646,
      "rewards/trajectory_similarity_reward/std": 0.38269859552383423,
      "step": 631
    },
    {
      "clip_ratio/high_max": 0.017488561454229057,
      "clip_ratio/high_mean": 0.00478088142699562,
      "clip_ratio/low_mean": 0.0005064025644969661,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005287283973302692,
      "entropy": 0.3701139725744724,
      "epoch": 0.5151299032093735,
      "grad_norm": 12.08208179473877,
      "learning_rate": 3.8793806030969845e-07,
      "loss": -1.6436,
      "reward": 3.343620777130127,
      "reward_std": 1.5129845142364502,
      "rewards/reasoning_density_reward/mean": 0.41874998807907104,
      "rewards/reasoning_density_reward/std": 0.14244882762432098,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.5811207294464111,
      "rewards/trajectory_similarity_reward/std": 0.4350942075252533,
      "step": 632
    },
    {
      "clip_ratio/high_max": 0.015816640574485064,
      "clip_ratio/high_mean": 0.004110494453925639,
      "clip_ratio/low_mean": 0.0013613439950859174,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005471838405355811,
      "entropy": 0.34037452936172485,
      "epoch": 0.5159449821701477,
      "grad_norm": 19.836402893066406,
      "learning_rate": 3.8728606356968215e-07,
      "loss": 6.528,
      "reward": 3.0148433446884155,
      "reward_std": 1.9149958491325378,
      "rewards/reasoning_density_reward/mean": 0.38463540375232697,
      "rewards/reasoning_density_reward/std": 0.1457921490073204,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.420820415019989,
      "rewards/trajectory_similarity_reward/mean": 0.5677078366279602,
      "rewards/trajectory_similarity_reward/std": 0.41041384637355804,
      "step": 633
    },
    {
      "clip_ratio/high_max": 0.012403137690853328,
      "clip_ratio/high_mean": 0.002579752072051633,
      "clip_ratio/low_mean": 0.001146665854321327,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037264179554767907,
      "entropy": 0.3577021099627018,
      "epoch": 0.5167600611309221,
      "grad_norm": 14.387861251831055,
      "learning_rate": 3.8663406682966585e-07,
      "loss": -2.6066,
      "reward": 1.2099995613098145,
      "reward_std": 1.6990735530853271,
      "rewards/reasoning_density_reward/mean": 0.2674106955528259,
      "rewards/reasoning_density_reward/std": 0.12788091599941254,
      "rewards/success_reward/mean": 0.75,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.19258882105350494,
      "rewards/trajectory_similarity_reward/std": 0.3754650354385376,
      "step": 634
    },
    {
      "clip_ratio/high_max": 0.01129554896033369,
      "clip_ratio/high_mean": 0.002686354022443993,
      "clip_ratio/low_mean": 0.0016613635525573045,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043477175058797,
      "entropy": 0.35952113568782806,
      "epoch": 0.5175751400916964,
      "grad_norm": 12.490074157714844,
      "learning_rate": 3.8598207008964955e-07,
      "loss": -7.6031,
      "reward": 2.1842474937438965,
      "reward_std": 2.0964198112487793,
      "rewards/reasoning_density_reward/mean": 0.3051046133041382,
      "rewards/reasoning_density_reward/std": 0.164179727435112,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.3791426718235016,
      "rewards/trajectory_similarity_reward/std": 0.40435361862182617,
      "step": 635
    },
    {
      "clip_ratio/high_max": 0.012848534301156178,
      "clip_ratio/high_mean": 0.0033929917808563914,
      "clip_ratio/low_mean": 0.0007505739340558648,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0041435657476540655,
      "entropy": 0.3413362540304661,
      "epoch": 0.5183902190524707,
      "grad_norm": 23.340682983398438,
      "learning_rate": 3.8533007334963325e-07,
      "loss": 2.3395,
      "reward": 3.5103445053100586,
      "reward_std": 1.5220084190368652,
      "rewards/reasoning_density_reward/mean": 0.44999998807907104,
      "rewards/reasoning_density_reward/std": 0.08944271504878998,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.6228446364402771,
      "rewards/trajectory_similarity_reward/std": 0.3845904767513275,
      "step": 636
    },
    {
      "clip_ratio/high_max": 0.015212594880722463,
      "clip_ratio/high_mean": 0.0036649461835622787,
      "clip_ratio/low_mean": 0.0007091951847542077,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004374141455627978,
      "entropy": 0.34401702880859375,
      "epoch": 0.519205298013245,
      "grad_norm": 13.785469055175781,
      "learning_rate": 3.8467807660961695e-07,
      "loss": 3.5179,
      "reward": 2.687380313873291,
      "reward_std": 1.7807518243789673,
      "rewards/reasoning_density_reward/mean": 0.3645833432674408,
      "rewards/reasoning_density_reward/std": 0.17017148435115814,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.44779711961746216,
      "rewards/trajectory_similarity_reward/std": 0.41971585154533386,
      "step": 637
    },
    {
      "clip_ratio/high_max": 0.01939675840549171,
      "clip_ratio/high_mean": 0.004379628429887816,
      "clip_ratio/low_mean": 0.0010521410695218947,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005431769444840029,
      "entropy": 0.3467046059668064,
      "epoch": 0.5200203769740194,
      "grad_norm": 14.241697311401367,
      "learning_rate": 3.8402607986960065e-07,
      "loss": 0.0299,
      "reward": 4.205530166625977,
      "reward_std": 0.6011416018009186,
      "rewards/reasoning_density_reward/mean": 0.45838913321495056,
      "rewards/reasoning_density_reward/std": 0.08787490427494049,
      "rewards/success_reward/mean": 2.859375,
      "rewards/success_reward/std": 0.5625,
      "rewards/trajectory_similarity_reward/mean": 0.8877660632133484,
      "rewards/trajectory_similarity_reward/std": 0.24974364042282104,
      "step": 638
    },
    {
      "clip_ratio/high_max": 0.015095743641722947,
      "clip_ratio/high_mean": 0.003734822450496722,
      "clip_ratio/low_mean": 0.00012470574074541219,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038595281657762825,
      "entropy": 0.35295526310801506,
      "epoch": 0.5208354559347936,
      "grad_norm": 9.669678688049316,
      "learning_rate": 3.833740831295843e-07,
      "loss": -0.3965,
      "reward": 2.8342885971069336,
      "reward_std": 1.1654866933822632,
      "rewards/reasoning_density_reward/mean": 0.24174107611179352,
      "rewards/reasoning_density_reward/std": 0.22729448974132538,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 0.9437293410301208,
      "rewards/trajectory_similarity_reward/mean": 0.43629759550094604,
      "rewards/trajectory_similarity_reward/std": 0.45745620131492615,
      "step": 639
    },
    {
      "clip_ratio/high_max": 0.008056596183450893,
      "clip_ratio/high_mean": 0.0013999901020724792,
      "clip_ratio/low_mean": 0.000763174393796362,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002163164463127032,
      "entropy": 0.36389410123229027,
      "epoch": 0.521650534895568,
      "grad_norm": 14.153995513916016,
      "learning_rate": 3.82722086389568e-07,
      "loss": 5.469,
      "reward": 3.4880313873291016,
      "reward_std": 1.6196653842926025,
      "rewards/reasoning_density_reward/mean": 0.42023810744285583,
      "rewards/reasoning_density_reward/std": 0.11736912280321121,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6302932500839233,
      "rewards/trajectory_similarity_reward/std": 0.34733855724334717,
      "step": 640
    },
    {
      "clip_ratio/high_max": 0.016949408687651157,
      "clip_ratio/high_mean": 0.004250652506016195,
      "clip_ratio/low_mean": 0.0009324510538135655,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005183103523449972,
      "entropy": 0.3551412709057331,
      "epoch": 0.5224656138563424,
      "grad_norm": 18.06229591369629,
      "learning_rate": 3.820700896495517e-07,
      "loss": -0.4668,
      "reward": 3.27423357963562,
      "reward_std": 1.4791224002838135,
      "rewards/reasoning_density_reward/mean": 0.4036458432674408,
      "rewards/reasoning_density_reward/std": 0.09294407069683075,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.526837944984436,
      "rewards/trajectory_similarity_reward/std": 0.3686467707157135,
      "step": 641
    },
    {
      "clip_ratio/high_max": 0.014849927742034197,
      "clip_ratio/high_mean": 0.002794254723994527,
      "clip_ratio/low_mean": 0.0007891691057011485,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035834238879033364,
      "entropy": 0.38918446004390717,
      "epoch": 0.5232806928171166,
      "grad_norm": 14.085144996643066,
      "learning_rate": 3.814180929095354e-07,
      "loss": 10.047,
      "reward": 2.434030294418335,
      "reward_std": 1.3912667036056519,
      "rewards/reasoning_density_reward/mean": 0.20907452702522278,
      "rewards/reasoning_density_reward/std": 0.1778489053249359,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.44370588660240173,
      "rewards/trajectory_similarity_reward/std": 0.41737812757492065,
      "step": 642
    },
    {
      "clip_ratio/high_max": 0.016160635510459542,
      "clip_ratio/high_mean": 0.0027567616489250213,
      "clip_ratio/low_mean": 0.0016225472827500198,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004379308986244723,
      "entropy": 0.36273622512817383,
      "epoch": 0.524095771777891,
      "grad_norm": 13.428933143615723,
      "learning_rate": 3.807660961695191e-07,
      "loss": 6.0591,
      "reward": 1.8978757858276367,
      "reward_std": 1.4930331707000732,
      "rewards/reasoning_density_reward/mean": 0.30677083134651184,
      "rewards/reasoning_density_reward/std": 0.14282917976379395,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.2786048650741577,
      "rewards/trajectory_similarity_reward/std": 0.38810643553733826,
      "step": 643
    },
    {
      "clip_ratio/high_max": 0.015682303928770125,
      "clip_ratio/high_mean": 0.0029817139875376597,
      "clip_ratio/low_mean": 0.002126689338183496,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005108403303893283,
      "entropy": 0.34391093999147415,
      "epoch": 0.5249108507386653,
      "grad_norm": 15.846458435058594,
      "learning_rate": 3.8011409942950285e-07,
      "loss": 0.0504,
      "reward": 3.1963560581207275,
      "reward_std": 1.0733088254928589,
      "rewards/reasoning_density_reward/mean": 0.43958336114883423,
      "rewards/reasoning_density_reward/std": 0.09287086874246597,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.506772518157959,
      "rewards/trajectory_similarity_reward/std": 0.32536864280700684,
      "step": 644
    },
    {
      "clip_ratio/high_max": 0.015174049010965973,
      "clip_ratio/high_mean": 0.0034525398586993106,
      "clip_ratio/low_mean": 0.0010749925750133116,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004527532466454431,
      "entropy": 0.3395114652812481,
      "epoch": 0.5257259296994397,
      "grad_norm": 13.821903228759766,
      "learning_rate": 3.7946210268948655e-07,
      "loss": -0.4747,
      "reward": 2.4886932373046875,
      "reward_std": 2.0915608406066895,
      "rewards/reasoning_density_reward/mean": 0.3090745210647583,
      "rewards/reasoning_density_reward/std": 0.15642312169075012,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.49211880564689636,
      "rewards/trajectory_similarity_reward/std": 0.4588947594165802,
      "step": 645
    },
    {
      "clip_ratio/high_max": 0.014539907744619995,
      "clip_ratio/high_mean": 0.0028540076527860947,
      "clip_ratio/low_mean": 0.0016761870101618115,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00453019465203397,
      "entropy": 0.3564419038593769,
      "epoch": 0.5265410086602139,
      "grad_norm": 17.699708938598633,
      "learning_rate": 3.7881010594947025e-07,
      "loss": 14.7464,
      "reward": 3.826367735862732,
      "reward_std": 0.6932279467582703,
      "rewards/reasoning_density_reward/mean": 0.41763167083263397,
      "rewards/reasoning_density_reward/std": 0.14210275560617447,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.6104557514190674,
      "rewards/trajectory_similarity_reward/mean": 0.736860990524292,
      "rewards/trajectory_similarity_reward/std": 0.30156541615724564,
      "step": 646
    },
    {
      "clip_ratio/high_max": 0.019803798641078174,
      "clip_ratio/high_mean": 0.004484699165914208,
      "clip_ratio/low_mean": 0.0008255661159637384,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005310265289153904,
      "entropy": 0.32161541655659676,
      "epoch": 0.5273560876209883,
      "grad_norm": 14.354945182800293,
      "learning_rate": 3.7815810920945395e-07,
      "loss": 3.8395,
      "reward": 1.018324851989746,
      "reward_std": 0.9930275082588196,
      "rewards/reasoning_density_reward/mean": 0.3270833492279053,
      "rewards/reasoning_density_reward/std": 0.06082001328468323,
      "rewards/success_reward/mean": 0.5625,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.1287415474653244,
      "rewards/trajectory_similarity_reward/std": 0.2833250164985657,
      "step": 647
    },
    {
      "clip_ratio/high_max": 0.009432106948224828,
      "clip_ratio/high_mean": 0.0014709100250911433,
      "clip_ratio/low_mean": 0.0022634924243902788,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037344024458434433,
      "entropy": 0.3287965841591358,
      "epoch": 0.5281711665817627,
      "grad_norm": 14.810388565063477,
      "learning_rate": 3.7750611246943765e-07,
      "loss": 7.8752,
      "reward": 2.6847198009490967,
      "reward_std": 0.7858703136444092,
      "rewards/reasoning_density_reward/mean": 0.39375001192092896,
      "rewards/reasoning_density_reward/std": 0.14361406862735748,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.509719729423523,
      "rewards/trajectory_similarity_reward/std": 0.46672895550727844,
      "step": 648
    },
    {
      "clip_ratio/high_max": 0.01899617852177471,
      "clip_ratio/high_mean": 0.0048587565252091736,
      "clip_ratio/low_mean": 0.0014510203400277533,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006309776799753308,
      "entropy": 0.34698784351348877,
      "epoch": 0.5289862455425369,
      "grad_norm": 14.876039505004883,
      "learning_rate": 3.7685411572942135e-07,
      "loss": 3.3499,
      "reward": 3.8287124633789062,
      "reward_std": 0.6697445213794708,
      "rewards/reasoning_density_reward/mean": 0.4223586320877075,
      "rewards/reasoning_density_reward/std": 0.12500110268592834,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.5465351641178131,
      "rewards/trajectory_similarity_reward/mean": 0.6876038312911987,
      "rewards/trajectory_similarity_reward/std": 0.36332501471042633,
      "step": 649
    },
    {
      "clip_ratio/high_max": 0.015397770912386477,
      "clip_ratio/high_mean": 0.0033603148913243786,
      "clip_ratio/low_mean": 0.0009053160611074418,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004265630850568414,
      "entropy": 0.3505757972598076,
      "epoch": 0.5298013245033113,
      "grad_norm": 16.34187126159668,
      "learning_rate": 3.7620211898940505e-07,
      "loss": 2.9841,
      "reward": 1.8233585357666016,
      "reward_std": 1.3480337858200073,
      "rewards/reasoning_density_reward/mean": 0.250927209854126,
      "rewards/reasoning_density_reward/std": 0.15150956809520721,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.0201103687286377,
      "rewards/trajectory_similarity_reward/mean": 0.16618137061595917,
      "rewards/trajectory_similarity_reward/std": 0.3644440472126007,
      "step": 650
    },
    {
      "clip_ratio/high_max": 0.010928235831670463,
      "clip_ratio/high_mean": 0.0025149241409962997,
      "clip_ratio/low_mean": 0.0012505032209446654,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003765427536563948,
      "entropy": 0.3590315170586109,
      "epoch": 0.5306164034640856,
      "grad_norm": 17.34012222290039,
      "learning_rate": 3.7555012224938875e-07,
      "loss": 12.6321,
      "reward": 2.670943021774292,
      "reward_std": 1.5439822673797607,
      "rewards/reasoning_density_reward/mean": 0.327938973903656,
      "rewards/reasoning_density_reward/std": 0.15793174505233765,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.374254047870636,
      "rewards/trajectory_similarity_reward/std": 0.39401915669441223,
      "step": 651
    },
    {
      "clip_ratio/high_max": 0.018944877665489912,
      "clip_ratio/high_mean": 0.0048643275804352015,
      "clip_ratio/low_mean": 0.0008498473762301728,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005714174942113459,
      "entropy": 0.32850073650479317,
      "epoch": 0.5314314824248599,
      "grad_norm": 15.749582290649414,
      "learning_rate": 3.7489812550937245e-07,
      "loss": -4.0076,
      "reward": 3.635259509086609,
      "reward_std": 1.2609368562698364,
      "rewards/reasoning_density_reward/mean": 0.4365513324737549,
      "rewards/reasoning_density_reward/std": 0.11139129102230072,
      "rewards/success_reward/mean": 2.578125,
      "rewards/success_reward/std": 0.947025716304779,
      "rewards/trajectory_similarity_reward/mean": 0.6205831468105316,
      "rewards/trajectory_similarity_reward/std": 0.36075566709041595,
      "step": 652
    },
    {
      "clip_ratio/high_max": 0.01823967508971691,
      "clip_ratio/high_mean": 0.005453666904941201,
      "clip_ratio/low_mean": 0.0008235405766754411,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006277207357925363,
      "entropy": 0.3353723995387554,
      "epoch": 0.5322465613856342,
      "grad_norm": 15.260651588439941,
      "learning_rate": 3.7424612876935615e-07,
      "loss": -4.3223,
      "reward": 1.018505573272705,
      "reward_std": 1.4120161533355713,
      "rewards/reasoning_density_reward/mean": 0.25208336114883423,
      "rewards/reasoning_density_reward/std": 0.1276388317346573,
      "rewards/success_reward/mean": 0.65625,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.11017228662967682,
      "rewards/trajectory_similarity_reward/std": 0.24237826466560364,
      "step": 653
    },
    {
      "clip_ratio/high_max": 0.006244305957807228,
      "clip_ratio/high_mean": 0.0013401689029706176,
      "clip_ratio/low_mean": 0.001850996384746395,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031911652549752034,
      "entropy": 0.3740267865359783,
      "epoch": 0.5330616403464086,
      "grad_norm": 14.094551086425781,
      "learning_rate": 3.7359413202933985e-07,
      "loss": 2.867,
      "reward": 3.1225719451904297,
      "reward_std": 1.0088235139846802,
      "rewards/reasoning_density_reward/mean": 0.3739786148071289,
      "rewards/reasoning_density_reward/std": 0.1489298939704895,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.4985933303833008,
      "rewards/trajectory_similarity_reward/std": 0.4264718294143677,
      "step": 654
    },
    {
      "clip_ratio/high_max": 0.015357454540207982,
      "clip_ratio/high_mean": 0.00305536795349326,
      "clip_ratio/low_mean": 0.0006810104823671281,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003736378377652727,
      "entropy": 0.32257793843746185,
      "epoch": 0.5338767193071828,
      "grad_norm": 14.528532981872559,
      "learning_rate": 3.729421352893235e-07,
      "loss": -1.4448,
      "reward": 3.260202646255493,
      "reward_std": 1.1902587413787842,
      "rewards/reasoning_density_reward/mean": 0.40937501192092896,
      "rewards/reasoning_density_reward/std": 0.14255847036838531,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6008276343345642,
      "rewards/trajectory_similarity_reward/std": 0.3833599090576172,
      "step": 655
    },
    {
      "clip_ratio/high_max": 0.01567773730494082,
      "clip_ratio/high_mean": 0.0027609093522187322,
      "clip_ratio/low_mean": 0.00039236063457792625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031532700231764466,
      "entropy": 0.30487221851944923,
      "epoch": 0.5346917982679572,
      "grad_norm": 13.018877029418945,
      "learning_rate": 3.722901385493072e-07,
      "loss": 9.0813,
      "reward": 2.921314239501953,
      "reward_std": 1.882710337638855,
      "rewards/reasoning_density_reward/mean": 0.380638986825943,
      "rewards/reasoning_density_reward/std": 0.08256785571575165,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.5719252228736877,
      "rewards/trajectory_similarity_reward/std": 0.4661480784416199,
      "step": 656
    },
    {
      "clip_ratio/high_max": 0.017433667089790106,
      "clip_ratio/high_mean": 0.004211818188196048,
      "clip_ratio/low_mean": 0.0003873798304994125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004599197971401736,
      "entropy": 0.3393080160021782,
      "epoch": 0.5355068772287316,
      "grad_norm": 11.741144180297852,
      "learning_rate": 3.716381418092909e-07,
      "loss": -0.5325,
      "reward": 3.835603713989258,
      "reward_std": 1.0677911937236786,
      "rewards/reasoning_density_reward/mean": 0.4074692130088806,
      "rewards/reasoning_density_reward/std": 0.14662928134202957,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.8080127239227295,
      "rewards/trajectory_similarity_reward/mean": 0.7093843817710876,
      "rewards/trajectory_similarity_reward/std": 0.3407013416290283,
      "step": 657
    },
    {
      "clip_ratio/high_max": 0.016022762982174754,
      "clip_ratio/high_mean": 0.004194334600470029,
      "clip_ratio/low_mean": 0.0009382594071212225,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005132594000315294,
      "entropy": 0.35252029076218605,
      "epoch": 0.5363219561895058,
      "grad_norm": 15.2737455368042,
      "learning_rate": 3.7098614506927465e-07,
      "loss": 2.1536,
      "reward": 2.398186445236206,
      "reward_std": 1.7157347202301025,
      "rewards/reasoning_density_reward/mean": 0.284970223903656,
      "rewards/reasoning_density_reward/std": 0.18733061850070953,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.23821620643138885,
      "rewards/trajectory_similarity_reward/std": 0.2797124683856964,
      "step": 658
    },
    {
      "clip_ratio/high_max": 0.015642416547052562,
      "clip_ratio/high_mean": 0.0033054719970095903,
      "clip_ratio/low_mean": 0.0006256397100514732,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003931111772544682,
      "entropy": 0.3648645766079426,
      "epoch": 0.5371370351502802,
      "grad_norm": 14.122940063476562,
      "learning_rate": 3.7033414832925835e-07,
      "loss": 0.2278,
      "reward": 1.921875,
      "reward_std": 0.6514450311660767,
      "rewards/reasoning_density_reward/mean": 0.234375,
      "rewards/reasoning_density_reward/std": 0.23432083427906036,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.375,
      "rewards/trajectory_similarity_reward/std": 0.5,
      "step": 659
    },
    {
      "clip_ratio/high_max": 0.01259913295507431,
      "clip_ratio/high_mean": 0.002359595804591663,
      "clip_ratio/low_mean": 0.002513596278731711,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004873192170634866,
      "entropy": 0.35899240523576736,
      "epoch": 0.5379521141110545,
      "grad_norm": 15.788312911987305,
      "learning_rate": 3.6968215158924205e-07,
      "loss": 7.5729,
      "reward": 2.306032180786133,
      "reward_std": 0.7774226665496826,
      "rewards/reasoning_density_reward/mean": 0.3531250059604645,
      "rewards/reasoning_density_reward/std": 0.1783897876739502,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.2654072344303131,
      "rewards/trajectory_similarity_reward/std": 0.32653287053108215,
      "step": 660
    },
    {
      "clip_ratio/high_max": 0.01400639972416684,
      "clip_ratio/high_mean": 0.002932940908067394,
      "clip_ratio/low_mean": 0.0019581725864554755,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048911135236267,
      "entropy": 0.3630937971174717,
      "epoch": 0.5387671930718289,
      "grad_norm": 15.905187606811523,
      "learning_rate": 3.6903015484922575e-07,
      "loss": 10.3517,
      "reward": 2.584617853164673,
      "reward_std": 1.0659394264221191,
      "rewards/reasoning_density_reward/mean": 0.3604166507720947,
      "rewards/reasoning_density_reward/std": 0.12544366717338562,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.1618950366973877,
      "rewards/trajectory_similarity_reward/mean": 0.3492012023925781,
      "rewards/trajectory_similarity_reward/std": 0.43638888001441956,
      "step": 661
    },
    {
      "clip_ratio/high_max": 0.016129092895425856,
      "clip_ratio/high_mean": 0.003172045762767084,
      "clip_ratio/low_mean": 0.000582954213314224,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037549999251496047,
      "entropy": 0.36382075399160385,
      "epoch": 0.5395822720326031,
      "grad_norm": 16.85665512084961,
      "learning_rate": 3.6837815810920945e-07,
      "loss": 5.7703,
      "reward": 2.1569199562072754,
      "reward_std": 1.5146679878234863,
      "rewards/reasoning_density_reward/mean": 0.26409968733787537,
      "rewards/reasoning_density_reward/std": 0.15333279222249985,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.3794710636138916,
      "rewards/trajectory_similarity_reward/mean": 0.2990701496601105,
      "rewards/trajectory_similarity_reward/std": 0.3560478836297989,
      "step": 662
    },
    {
      "clip_ratio/high_max": 0.009553272218909115,
      "clip_ratio/high_mean": 0.0019421610559220426,
      "clip_ratio/low_mean": 0.0012262912350706756,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031684523564763367,
      "entropy": 0.35457026585936546,
      "epoch": 0.5403973509933775,
      "grad_norm": 13.499734878540039,
      "learning_rate": 3.6772616136919315e-07,
      "loss": -10.2373,
      "reward": 1.9065449237823486,
      "reward_std": 1.8111214637756348,
      "rewards/reasoning_density_reward/mean": 0.3031250238418579,
      "rewards/reasoning_density_reward/std": 0.15542280673980713,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.2909199893474579,
      "rewards/trajectory_similarity_reward/std": 0.4176054298877716,
      "step": 663
    },
    {
      "clip_ratio/high_max": 0.01419818343129009,
      "clip_ratio/high_mean": 0.002633813361171633,
      "clip_ratio/low_mean": 0.001565919621498324,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004199732837150805,
      "entropy": 0.36051972582936287,
      "epoch": 0.5412124299541519,
      "grad_norm": 15.238373756408691,
      "learning_rate": 3.6707416462917685e-07,
      "loss": 9.7671,
      "reward": 2.578284740447998,
      "reward_std": 1.9579298496246338,
      "rewards/reasoning_density_reward/mean": 0.3147321343421936,
      "rewards/reasoning_density_reward/std": 0.22822099924087524,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.4823026657104492,
      "rewards/trajectory_similarity_reward/std": 0.44263774156570435,
      "step": 664
    },
    {
      "clip_ratio/high_max": 0.014792722708079964,
      "clip_ratio/high_mean": 0.0032624634404783137,
      "clip_ratio/low_mean": 0.0009347885315946769,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004197252084850334,
      "entropy": 0.3577960357069969,
      "epoch": 0.5420275089149261,
      "grad_norm": 12.575453758239746,
      "learning_rate": 3.6642216788916055e-07,
      "loss": 20.0238,
      "reward": 2.143578052520752,
      "reward_std": 1.3492822647094727,
      "rewards/reasoning_density_reward/mean": 0.26823490858078003,
      "rewards/reasoning_density_reward/std": 0.16420812904834747,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.18784306943416595,
      "rewards/trajectory_similarity_reward/std": 0.30860018730163574,
      "step": 665
    },
    {
      "clip_ratio/high_max": 0.013363887032028288,
      "clip_ratio/high_mean": 0.0025746714673005044,
      "clip_ratio/low_mean": 0.0007477240797015838,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033223955251742154,
      "entropy": 0.3705156780779362,
      "epoch": 0.5428425878757005,
      "grad_norm": 12.418392181396484,
      "learning_rate": 3.6577017114914425e-07,
      "loss": 5.3035,
      "step": 666
    },
    {
      "clip_ratio/high_max": 0.020247902488335967,
      "clip_ratio/high_mean": 0.004950011090841144,
      "clip_ratio/low_mean": 0.00032533251578570344,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005275343544781208,
      "entropy": 0.3113817013800144,
      "epoch": 0.5436576668364748,
      "grad_norm": 13.566872596740723,
      "learning_rate": 3.6511817440912795e-07,
      "loss": 5.5241,
      "reward": 3.3782875537872314,
      "reward_std": 1.335115671157837,
      "rewards/reasoning_density_reward/mean": 0.4208333343267441,
      "rewards/reasoning_density_reward/std": 0.09715935960412025,
      "rewards/success_reward/mean": 2.390625,
      "rewards/success_reward/std": 0.925901472568512,
      "rewards/trajectory_similarity_reward/mean": 0.566829115152359,
      "rewards/trajectory_similarity_reward/std": 0.419843927025795,
      "step": 667
    },
    {
      "clip_ratio/high_max": 0.01233742356998846,
      "clip_ratio/high_mean": 0.0025076657766476274,
      "clip_ratio/low_mean": 0.0023211352599901147,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004828801087569445,
      "entropy": 0.3200846239924431,
      "epoch": 0.5444727457972491,
      "grad_norm": 14.780742645263672,
      "learning_rate": 3.6446617766911165e-07,
      "loss": 1.0879,
      "reward": 2.3151917457580566,
      "reward_std": 0.025685735046863556,
      "rewards/reasoning_density_reward/mean": 0.4000000059604645,
      "rewards/reasoning_density_reward/std": 0.10327955335378647,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.4151916205883026,
      "rewards/trajectory_similarity_reward/std": 0.4302417039871216,
      "step": 668
    },
    {
      "clip_ratio/high_max": 0.014974664081819355,
      "clip_ratio/high_mean": 0.0028985633834963664,
      "clip_ratio/low_mean": 0.00152394654287491,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004422509897267446,
      "entropy": 0.36171267181634903,
      "epoch": 0.5452878247580234,
      "grad_norm": 17.53166961669922,
      "learning_rate": 3.6381418092909535e-07,
      "loss": 9.8232,
      "reward": 2.5872243642807007,
      "reward_std": 1.1014201864600182,
      "rewards/reasoning_density_reward/mean": 0.3843750059604645,
      "rewards/reasoning_density_reward/std": 0.15719829499721527,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5245966911315918,
      "rewards/trajectory_similarity_reward/mean": 0.5153494775295258,
      "rewards/trajectory_similarity_reward/std": 0.46280452609062195,
      "step": 669
    },
    {
      "clip_ratio/high_max": 0.012682889268035069,
      "clip_ratio/high_mean": 0.002689323930098908,
      "clip_ratio/low_mean": 0.0013499290544132236,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004039252933580428,
      "entropy": 0.32589322328567505,
      "epoch": 0.5461029037187978,
      "grad_norm": 15.155572891235352,
      "learning_rate": 3.6316218418907905e-07,
      "loss": 10.369,
      "reward": 3.488542079925537,
      "reward_std": 1.6150561571121216,
      "rewards/reasoning_density_reward/mean": 0.424834281206131,
      "rewards/reasoning_density_reward/std": 0.10221050679683685,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.626207709312439,
      "rewards/trajectory_similarity_reward/std": 0.36554014682769775,
      "step": 670
    },
    {
      "clip_ratio/high_max": 0.014437338104471564,
      "clip_ratio/high_mean": 0.0030352076864801347,
      "clip_ratio/low_mean": 0.0011451810169091914,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004180388641543686,
      "entropy": 0.3673092722892761,
      "epoch": 0.546917982679572,
      "grad_norm": 14.925990104675293,
      "learning_rate": 3.6251018744906275e-07,
      "loss": 5.2059,
      "reward": 2.6585686206817627,
      "reward_std": 1.028515100479126,
      "rewards/reasoning_density_reward/mean": 0.2787405252456665,
      "rewards/reasoning_density_reward/std": 0.16954819858074188,
      "rewards/success_reward/mean": 2.15625,
      "rewards/success_reward/std": 0.7685213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.22357821464538574,
      "rewards/trajectory_similarity_reward/std": 0.2914164662361145,
      "step": 671
    },
    {
      "clip_ratio/high_max": 0.012770750821800902,
      "clip_ratio/high_mean": 0.0036619160237023607,
      "clip_ratio/low_mean": 0.0015920957375783473,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005254011659417301,
      "entropy": 0.33677510917186737,
      "epoch": 0.5477330616403464,
      "grad_norm": 19.412242889404297,
      "learning_rate": 3.618581907090464e-07,
      "loss": 13.5611,
      "reward": 4.017193078994751,
      "reward_std": 0.35038506984710693,
      "rewards/reasoning_density_reward/mean": 0.45000000298023224,
      "rewards/reasoning_density_reward/std": 0.1206529401242733,
      "rewards/success_reward/mean": 2.953125,
      "rewards/success_reward/std": 0.1875,
      "rewards/trajectory_similarity_reward/mean": 0.614067941904068,
      "rewards/trajectory_similarity_reward/std": 0.21643263846635818,
      "step": 672
    },
    {
      "clip_ratio/high_max": 0.016102072782814503,
      "clip_ratio/high_mean": 0.0029920416127424687,
      "clip_ratio/low_mean": 0.0018201762868557125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004812217943253927,
      "entropy": 0.3506635092198849,
      "epoch": 0.5485481406011208,
      "grad_norm": 17.493192672729492,
      "learning_rate": 3.6120619396903015e-07,
      "loss": 4.6495,
      "reward": 2.692692756652832,
      "reward_std": 1.9420430660247803,
      "rewards/reasoning_density_reward/mean": 0.38928571343421936,
      "rewards/reasoning_density_reward/std": 0.1411806344985962,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.42840713262557983,
      "rewards/trajectory_similarity_reward/std": 0.4275251626968384,
      "step": 673
    },
    {
      "clip_ratio/high_max": 0.006462638266384602,
      "clip_ratio/high_mean": 0.0012578770401887596,
      "clip_ratio/low_mean": 0.0013302100742293987,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025880871253320947,
      "entropy": 0.369626946747303,
      "epoch": 0.549363219561895,
      "grad_norm": 10.394336700439453,
      "learning_rate": 3.6055419722901385e-07,
      "loss": 9.2027,
      "reward": 1.1149934530258179,
      "reward_std": 1.3613486289978027,
      "rewards/reasoning_density_reward/mean": 0.17500001192092896,
      "rewards/reasoning_density_reward/std": 0.18348479270935059,
      "rewards/success_reward/mean": 0.84375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.09624342620372772,
      "rewards/trajectory_similarity_reward/std": 0.2641681134700775,
      "step": 674
    },
    {
      "clip_ratio/high_max": 0.005775891419034451,
      "clip_ratio/high_mean": 0.0015900478247203864,
      "clip_ratio/low_mean": 0.0008296419619000517,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002419689772068523,
      "entropy": 0.3888706564903259,
      "epoch": 0.5501782985226694,
      "grad_norm": 10.068991661071777,
      "learning_rate": 3.5990220048899755e-07,
      "loss": -5.4042,
      "reward": 0.9256696701049805,
      "reward_std": 0.7138042449951172,
      "rewards/reasoning_density_reward/mean": 0.08191964775323868,
      "rewards/reasoning_density_reward/std": 0.11886274069547653,
      "rewards/success_reward/mean": 0.84375,
      "rewards/success_reward/std": 0.7685213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.0,
      "rewards/trajectory_similarity_reward/std": 0.0,
      "step": 675
    },
    {
      "clip_ratio/high_max": 0.009243704960681498,
      "clip_ratio/high_mean": 0.0026315953000448644,
      "clip_ratio/low_mean": 0.001088754852389684,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037203501851763576,
      "entropy": 0.3999660760164261,
      "epoch": 0.5509933774834437,
      "grad_norm": 11.519668579101562,
      "learning_rate": 3.5925020374898125e-07,
      "loss": -0.3433,
      "reward": 3.508824348449707,
      "reward_std": 1.1731667518615723,
      "rewards/reasoning_density_reward/mean": 0.41830357909202576,
      "rewards/reasoning_density_reward/std": 0.17789335548877716,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.6530206203460693,
      "rewards/trajectory_similarity_reward/std": 0.35639360547065735,
      "step": 676
    },
    {
      "clip_ratio/high_max": 0.014540553442202508,
      "clip_ratio/high_mean": 0.003585410267987754,
      "clip_ratio/low_mean": 0.0013115717920300085,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004896982005448081,
      "entropy": 0.34805823862552643,
      "epoch": 0.551808456444218,
      "grad_norm": 16.45887565612793,
      "learning_rate": 3.5859820700896495e-07,
      "loss": 14.69,
      "reward": 3.925701141357422,
      "reward_std": 0.48843854665756226,
      "rewards/reasoning_density_reward/mean": 0.3970237970352173,
      "rewards/reasoning_density_reward/std": 0.16127678751945496,
      "rewards/success_reward/mean": 2.90625,
      "rewards/success_reward/std": 0.375,
      "rewards/trajectory_similarity_reward/mean": 0.6224273443222046,
      "rewards/trajectory_similarity_reward/std": 0.330687016248703,
      "step": 677
    },
    {
      "clip_ratio/high_max": 0.016115096979774535,
      "clip_ratio/high_mean": 0.0034677875955821946,
      "clip_ratio/low_mean": 0.0008882589099812321,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004356046600150876,
      "entropy": 0.37135665118694305,
      "epoch": 0.5526235354049923,
      "grad_norm": 17.227991104125977,
      "learning_rate": 3.5794621026894865e-07,
      "loss": 7.1099,
      "reward": 3.298653483390808,
      "reward_std": 1.5754806995391846,
      "rewards/reasoning_density_reward/mean": 0.34322915971279144,
      "rewards/reasoning_density_reward/std": 0.20743395388126373,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0767268538475037,
      "rewards/trajectory_similarity_reward/mean": 0.6116742789745331,
      "rewards/trajectory_similarity_reward/std": 0.45570218563079834,
      "step": 678
    },
    {
      "clip_ratio/high_max": 0.015218365704640746,
      "clip_ratio/high_mean": 0.0034685335122048855,
      "clip_ratio/low_mean": 0.000989284071692964,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004457817558431998,
      "entropy": 0.40152280405163765,
      "epoch": 0.5534386143657667,
      "grad_norm": 13.2498779296875,
      "learning_rate": 3.5729421352893235e-07,
      "loss": 9.1938,
      "reward": 2.2083818912506104,
      "reward_std": 1.37458336353302,
      "rewards/reasoning_density_reward/mean": 0.1563374102115631,
      "rewards/reasoning_density_reward/std": 0.15990912914276123,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.1618950366973877,
      "rewards/trajectory_similarity_reward/mean": 0.17704445123672485,
      "rewards/trajectory_similarity_reward/std": 0.23287223279476166,
      "step": 679
    },
    {
      "clip_ratio/high_max": 0.007451865705661476,
      "clip_ratio/high_mean": 0.001669835823122412,
      "clip_ratio/low_mean": 0.0009476024242758285,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002617438294691965,
      "entropy": 0.3736054413020611,
      "epoch": 0.5542536933265411,
      "grad_norm": 10.727885246276855,
      "learning_rate": 3.5664221678891605e-07,
      "loss": 1.4919,
      "step": 680
    },
    {
      "clip_ratio/high_max": 0.016618901921901852,
      "clip_ratio/high_mean": 0.004289898315619212,
      "clip_ratio/low_mean": 0.0019979958815383725,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006287894299020991,
      "entropy": 0.3502322994172573,
      "epoch": 0.5550687722873153,
      "grad_norm": 13.294273376464844,
      "learning_rate": 3.5599022004889975e-07,
      "loss": 2.7039,
      "reward": 2.8596659898757935,
      "reward_std": 1.4221768975257874,
      "rewards/reasoning_density_reward/mean": 0.4140625,
      "rewards/reasoning_density_reward/std": 0.10692642629146576,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3731906414031982,
      "rewards/trajectory_similarity_reward/mean": 0.5706035047769547,
      "rewards/trajectory_similarity_reward/std": 0.4227169454097748,
      "step": 681
    },
    {
      "clip_ratio/high_max": 0.018817732110619545,
      "clip_ratio/high_mean": 0.004682236627559178,
      "clip_ratio/low_mean": 0.0007302500816877,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005412486556451768,
      "entropy": 0.355537511408329,
      "epoch": 0.5558838512480897,
      "grad_norm": 27.927173614501953,
      "learning_rate": 3.5533822330888345e-07,
      "loss": 5.7936,
      "reward": 3.150170922279358,
      "reward_std": 1.5644996762275696,
      "rewards/reasoning_density_reward/mean": 0.40156249701976776,
      "rewards/reasoning_density_reward/std": 0.12977508082985878,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.3404724597930908,
      "rewards/trajectory_similarity_reward/mean": 0.6392333507537842,
      "rewards/trajectory_similarity_reward/std": 0.42363937199115753,
      "step": 682
    },
    {
      "clip_ratio/high_max": 0.01167625980451703,
      "clip_ratio/high_mean": 0.0023559598193969578,
      "clip_ratio/low_mean": 0.002310711541213095,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004666671273298562,
      "entropy": 0.3678293973207474,
      "epoch": 0.556698930208864,
      "grad_norm": 15.143460273742676,
      "learning_rate": 3.5468622656886715e-07,
      "loss": 8.9868,
      "reward": 2.4485127925872803,
      "reward_std": 0.8560452461242676,
      "rewards/reasoning_density_reward/mean": 0.3046875,
      "rewards/reasoning_density_reward/std": 0.2229010909795761,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.45632532238960266,
      "rewards/trajectory_similarity_reward/std": 0.42889419198036194,
      "step": 683
    },
    {
      "clip_ratio/high_max": 0.00812254854827188,
      "clip_ratio/high_mean": 0.0014288673555711284,
      "clip_ratio/low_mean": 0.0020747500529978424,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00350361735036131,
      "entropy": 0.3360728584229946,
      "epoch": 0.5575140091696383,
      "grad_norm": 17.548805236816406,
      "learning_rate": 3.5403422982885084e-07,
      "loss": 8.1252,
      "reward": 2.608761787414551,
      "reward_std": 0.8091564774513245,
      "rewards/reasoning_density_reward/mean": 0.4125000238418579,
      "rewards/reasoning_density_reward/std": 0.10246950387954712,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.5087617635726929,
      "rewards/trajectory_similarity_reward/std": 0.47299879789352417,
      "step": 684
    },
    {
      "clip_ratio/high_max": 0.01579405347001739,
      "clip_ratio/high_mean": 0.003215329477825435,
      "clip_ratio/low_mean": 0.0013070955865259748,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004522425075265346,
      "entropy": 0.34839751198887825,
      "epoch": 0.5583290881304126,
      "grad_norm": 13.012754440307617,
      "learning_rate": 3.5338223308883454e-07,
      "loss": 1.2734,
      "reward": 2.119475841522217,
      "reward_std": 2.002444267272949,
      "rewards/reasoning_density_reward/mean": 0.28203123807907104,
      "rewards/reasoning_density_reward/std": 0.1788836270570755,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4491376876831055,
      "rewards/trajectory_similarity_reward/mean": 0.3374445140361786,
      "rewards/trajectory_similarity_reward/std": 0.4129364490509033,
      "step": 685
    },
    {
      "clip_ratio/high_max": 0.018663701601326466,
      "clip_ratio/high_mean": 0.004315611586207524,
      "clip_ratio/low_mean": 0.000976417137280805,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005292028858093545,
      "entropy": 0.36729080229997635,
      "epoch": 0.559144167091187,
      "grad_norm": 13.824713706970215,
      "learning_rate": 3.5273023634881824e-07,
      "loss": 4.2137,
      "reward": 3.291137218475342,
      "reward_std": 1.601252794265747,
      "rewards/reasoning_density_reward/mean": 0.3688616156578064,
      "rewards/reasoning_density_reward/std": 0.13501184433698654,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.1510830521583557,
      "rewards/trajectory_similarity_reward/mean": 0.5785257518291473,
      "rewards/trajectory_similarity_reward/std": 0.4052470177412033,
      "step": 686
    },
    {
      "clip_ratio/high_max": 0.02115791302639991,
      "clip_ratio/high_mean": 0.004807586868992075,
      "clip_ratio/low_mean": 0.0009335523973277304,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00574113943730481,
      "entropy": 0.3339324779808521,
      "epoch": 0.5599592460519612,
      "grad_norm": 21.10563850402832,
      "learning_rate": 3.5207823960880194e-07,
      "loss": -2.249,
      "reward": 2.913017511367798,
      "reward_std": 1.7904366254806519,
      "rewards/reasoning_density_reward/mean": 0.3984375,
      "rewards/reasoning_density_reward/std": 0.09681113064289093,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.3098187446594238,
      "rewards/trajectory_similarity_reward/mean": 0.5458300113677979,
      "rewards/trajectory_similarity_reward/std": 0.498089998960495,
      "step": 687
    },
    {
      "clip_ratio/high_max": 0.017592395073734224,
      "clip_ratio/high_mean": 0.003439392414293252,
      "clip_ratio/low_mean": 0.001128535412135534,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004567927899188362,
      "entropy": 0.32444559037685394,
      "epoch": 0.5607743250127356,
      "grad_norm": 16.80455780029297,
      "learning_rate": 3.5142624286878564e-07,
      "loss": -3.3703,
      "reward": 2.2189159393310547,
      "reward_std": 1.874521255493164,
      "rewards/reasoning_density_reward/mean": 0.3765625059604645,
      "rewards/reasoning_density_reward/std": 0.12262167036533356,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.5185213088989258,
      "rewards/trajectory_similarity_reward/mean": 0.4361034333705902,
      "rewards/trajectory_similarity_reward/std": 0.46830081939697266,
      "step": 688
    },
    {
      "clip_ratio/high_max": 0.017209687968716025,
      "clip_ratio/high_mean": 0.0036686741514131427,
      "clip_ratio/low_mean": 0.0014202754864527378,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005088949837954715,
      "entropy": 0.34575892239809036,
      "epoch": 0.56158940397351,
      "grad_norm": 17.124481201171875,
      "learning_rate": 3.5077424612876934e-07,
      "loss": 7.046,
      "reward": 2.72247314453125,
      "reward_std": 1.9728847742080688,
      "rewards/reasoning_density_reward/mean": 0.3229166567325592,
      "rewards/reasoning_density_reward/std": 0.19339798390865326,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5245562791824341,
      "rewards/trajectory_similarity_reward/std": 0.4477123022079468,
      "step": 689
    },
    {
      "clip_ratio/high_max": 0.016694399062544107,
      "clip_ratio/high_mean": 0.0035574497014749795,
      "clip_ratio/low_mean": 0.0007110149817890488,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004268464690539986,
      "entropy": 0.33481648936867714,
      "epoch": 0.5624044829342842,
      "grad_norm": 18.375089645385742,
      "learning_rate": 3.5012224938875304e-07,
      "loss": 2.3861,
      "reward": 2.460791826248169,
      "reward_std": 2.082418203353882,
      "rewards/reasoning_density_reward/mean": 0.35970982909202576,
      "rewards/reasoning_density_reward/std": 0.13068366050720215,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.4135820269584656,
      "rewards/trajectory_similarity_reward/std": 0.4054066836833954,
      "step": 690
    },
    {
      "clip_ratio/high_max": 0.015189519675914198,
      "clip_ratio/high_mean": 0.005116084408655297,
      "clip_ratio/low_mean": 0.000347643577697454,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005463728150061797,
      "entropy": 0.34770919382572174,
      "epoch": 0.5632195618950586,
      "grad_norm": 17.021875381469727,
      "learning_rate": 3.4947025264873674e-07,
      "loss": 19.0364,
      "reward": 3.549830913543701,
      "reward_std": 1.003535270690918,
      "rewards/reasoning_density_reward/mean": 0.3911072015762329,
      "rewards/reasoning_density_reward/std": 0.1496853157877922,
      "rewards/success_reward/mean": 2.53125,
      "rewards/success_reward/std": 0.7180703282356262,
      "rewards/trajectory_similarity_reward/mean": 0.6274738311767578,
      "rewards/trajectory_similarity_reward/std": 0.2638211324810982,
      "step": 691
    },
    {
      "clip_ratio/high_max": 0.012144783628173172,
      "clip_ratio/high_mean": 0.002892290511226747,
      "clip_ratio/low_mean": 0.0011195229380973615,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004011813456600066,
      "entropy": 0.3229057751595974,
      "epoch": 0.5640346408558329,
      "grad_norm": 15.571335792541504,
      "learning_rate": 3.4881825590872044e-07,
      "loss": 10.8933,
      "reward": 2.2655177116394043,
      "reward_std": 1.8646526336669922,
      "rewards/reasoning_density_reward/mean": 0.3583333492279053,
      "rewards/reasoning_density_reward/std": 0.13581031560897827,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.5491933822631836,
      "rewards/trajectory_similarity_reward/mean": 0.40718427300453186,
      "rewards/trajectory_similarity_reward/std": 0.44310262799263,
      "step": 692
    },
    {
      "clip_ratio/high_max": 0.012645212991628796,
      "clip_ratio/high_mean": 0.003247619650210254,
      "clip_ratio/low_mean": 0.0015514037440880202,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004799023503437638,
      "entropy": 0.36182234063744545,
      "epoch": 0.5648497198166073,
      "grad_norm": 17.426923751831055,
      "learning_rate": 3.4816625916870414e-07,
      "loss": -8.6017,
      "reward": 2.6449718475341797,
      "reward_std": 1.1528747081756592,
      "rewards/reasoning_density_reward/mean": 0.1770833432674408,
      "rewards/reasoning_density_reward/std": 0.21411964297294617,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.4053885042667389,
      "rewards/trajectory_similarity_reward/std": 0.45704495906829834,
      "step": 693
    },
    {
      "clip_ratio/high_max": 0.00890718525624834,
      "clip_ratio/high_mean": 0.0018700239415920805,
      "clip_ratio/low_mean": 0.0011525105765031185,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030225344526115805,
      "entropy": 0.3554081320762634,
      "epoch": 0.5656647987773815,
      "grad_norm": 12.062666893005371,
      "learning_rate": 3.4751426242868784e-07,
      "loss": -8.4784,
      "reward": 1.3050057888031006,
      "reward_std": 1.2825747728347778,
      "rewards/reasoning_density_reward/mean": 0.16577380895614624,
      "rewards/reasoning_density_reward/std": 0.17765559256076813,
      "rewards/success_reward/mean": 1.03125,
      "rewards/success_reward/std": 1.0562314987182617,
      "rewards/trajectory_similarity_reward/mean": 0.10798202455043793,
      "rewards/trajectory_similarity_reward/std": 0.29579269886016846,
      "step": 694
    },
    {
      "clip_ratio/high_max": 0.013208081480115652,
      "clip_ratio/high_mean": 0.003503201180137694,
      "clip_ratio/low_mean": 0.0003814574629359413,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003884658683091402,
      "entropy": 0.36772520840168,
      "epoch": 0.5664798777381559,
      "grad_norm": 10.73487663269043,
      "learning_rate": 3.4686226568867154e-07,
      "loss": -2.682,
      "reward": 2.6998798847198486,
      "reward_std": 0.41512545943260193,
      "rewards/reasoning_density_reward/mean": 0.3370535671710968,
      "rewards/reasoning_density_reward/std": 0.1943630576133728,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.2845232486724854,
      "rewards/trajectory_similarity_reward/mean": 0.4878261983394623,
      "rewards/trajectory_similarity_reward/std": 0.5042455792427063,
      "step": 695
    },
    {
      "clip_ratio/high_max": 0.013590721937362105,
      "clip_ratio/high_mean": 0.0028528777038445696,
      "clip_ratio/low_mean": 0.0009150737059826497,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00376795137708541,
      "entropy": 0.35482415184378624,
      "epoch": 0.5672949566989303,
      "grad_norm": 13.731966972351074,
      "learning_rate": 3.4621026894865524e-07,
      "loss": -0.6303,
      "reward": 3.3154149055480957,
      "reward_std": 1.5093098878860474,
      "rewards/reasoning_density_reward/mean": 0.37291666865348816,
      "rewards/reasoning_density_reward/std": 0.16528594493865967,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.5987482070922852,
      "rewards/trajectory_similarity_reward/std": 0.38359513878822327,
      "step": 696
    },
    {
      "clip_ratio/high_max": 0.02345869189593941,
      "clip_ratio/high_mean": 0.005507702677277848,
      "clip_ratio/low_mean": 0.0013828833798470441,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0068905859952792525,
      "entropy": 0.35462643578648567,
      "epoch": 0.5681100356597045,
      "grad_norm": 16.287446975708008,
      "learning_rate": 3.4555827220863894e-07,
      "loss": 4.7596,
      "reward": 3.3386104106903076,
      "reward_std": 1.0871868133544922,
      "rewards/reasoning_density_reward/mean": 0.3505208492279053,
      "rewards/reasoning_density_reward/std": 0.18455323576927185,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.7380895614624023,
      "rewards/trajectory_similarity_reward/std": 0.44129595160484314,
      "step": 697
    },
    {
      "clip_ratio/high_max": 0.016108016832731664,
      "clip_ratio/high_mean": 0.004911038151476532,
      "clip_ratio/low_mean": 0.0008211471686081495,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005732185149099678,
      "entropy": 0.3292192742228508,
      "epoch": 0.5689251146204789,
      "grad_norm": 17.866626739501953,
      "learning_rate": 3.4490627546862264e-07,
      "loss": 0.9164,
      "reward": 3.9354496002197266,
      "reward_std": 0.8889759480953217,
      "rewards/reasoning_density_reward/mean": 0.4572916775941849,
      "rewards/reasoning_density_reward/std": 0.08333098515868187,
      "rewards/success_reward/mean": 2.71875,
      "rewards/success_reward/std": 0.6046693325042725,
      "rewards/trajectory_similarity_reward/mean": 0.7594079971313477,
      "rewards/trajectory_similarity_reward/std": 0.3894791454076767,
      "step": 698
    },
    {
      "clip_ratio/high_max": 0.013609796995297074,
      "clip_ratio/high_mean": 0.003672053004265763,
      "clip_ratio/low_mean": 0.000332603925926378,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004004656948382035,
      "entropy": 0.34587258100509644,
      "epoch": 0.5697401935812532,
      "grad_norm": 16.10370635986328,
      "learning_rate": 3.4425427872860634e-07,
      "loss": 10.9165,
      "reward": 3.0152053833007812,
      "reward_std": 1.2080202102661133,
      "rewards/reasoning_density_reward/mean": 0.32594698667526245,
      "rewards/reasoning_density_reward/std": 0.2051464319229126,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 0.9486833214759827,
      "rewards/trajectory_similarity_reward/mean": 0.43925848603248596,
      "rewards/trajectory_similarity_reward/std": 0.4522043466567993,
      "step": 699
    },
    {
      "clip_ratio/high_max": 0.01797575654927641,
      "clip_ratio/high_mean": 0.004132272879360244,
      "clip_ratio/low_mean": 0.0013575451375800185,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005489817966008559,
      "entropy": 0.3254803195595741,
      "epoch": 0.5705552725420275,
      "grad_norm": 18.402982711791992,
      "learning_rate": 3.4360228198859004e-07,
      "loss": -2.4452,
      "reward": 3.6308951377868652,
      "reward_std": 0.9469317197799683,
      "rewards/reasoning_density_reward/mean": 0.4312500059604645,
      "rewards/reasoning_density_reward/std": 0.11794751510024071,
      "rewards/success_reward/mean": 2.671875,
      "rewards/success_reward/std": 0.784289538860321,
      "rewards/trajectory_similarity_reward/mean": 0.5277701169252396,
      "rewards/trajectory_similarity_reward/std": 0.36750245094299316,
      "step": 700
    },
    {
      "clip_ratio/high_max": 0.01749686198309064,
      "clip_ratio/high_mean": 0.0038017877814127132,
      "clip_ratio/low_mean": 0.0011248755945416633,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004926663401420228,
      "entropy": 0.3499477840960026,
      "epoch": 0.5713703515028018,
      "grad_norm": 15.163878440856934,
      "learning_rate": 3.4295028524857374e-07,
      "loss": 9.3889,
      "reward": 2.093444585800171,
      "reward_std": 1.1080609560012817,
      "rewards/reasoning_density_reward/mean": 0.18050594627857208,
      "rewards/reasoning_density_reward/std": 0.2053944170475006,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.22543847560882568,
      "rewards/trajectory_similarity_reward/std": 0.35262054204940796,
      "step": 701
    },
    {
      "clip_ratio/high_max": 0.016013028158340603,
      "clip_ratio/high_mean": 0.002698418691579718,
      "clip_ratio/low_mean": 0.0006162331083032768,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003314651759865228,
      "entropy": 0.3618064783513546,
      "epoch": 0.5721854304635762,
      "grad_norm": 10.804458618164062,
      "learning_rate": 3.4229828850855744e-07,
      "loss": 2.3369,
      "step": 702
    },
    {
      "clip_ratio/high_max": 0.01161051425151527,
      "clip_ratio/high_mean": 0.002463937009451911,
      "clip_ratio/low_mean": 0.002075819022138603,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004539755987934768,
      "entropy": 0.35226066410541534,
      "epoch": 0.5730005094243504,
      "grad_norm": 14.736798286437988,
      "learning_rate": 3.4164629176854114e-07,
      "loss": -1.6441,
      "reward": 2.1519899368286133,
      "reward_std": 1.6103512048721313,
      "rewards/reasoning_density_reward/mean": 0.3235164284706116,
      "rewards/reasoning_density_reward/std": 0.11608020216226578,
      "rewards/success_reward/mean": 1.5,
      "rewards/success_reward/std": 1.4491376876831055,
      "rewards/trajectory_similarity_reward/mean": 0.3284734785556793,
      "rewards/trajectory_similarity_reward/std": 0.3984493911266327,
      "step": 703
    },
    {
      "clip_ratio/high_max": 0.01795928948558867,
      "clip_ratio/high_mean": 0.0036814816849073395,
      "clip_ratio/low_mean": 0.0006617164581257384,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004343198175774887,
      "entropy": 0.3564862385392189,
      "epoch": 0.5738155883851248,
      "grad_norm": 13.632908821105957,
      "learning_rate": 3.4099429502852484e-07,
      "loss": -9.6993,
      "reward": 3.372511386871338,
      "reward_std": 1.470131516456604,
      "rewards/reasoning_density_reward/mean": 0.31822916865348816,
      "rewards/reasoning_density_reward/std": 0.16499772667884827,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.6167823076248169,
      "rewards/trajectory_similarity_reward/std": 0.41964513063430786,
      "step": 704
    },
    {
      "clip_ratio/high_max": 0.015120258904062212,
      "clip_ratio/high_mean": 0.0031688327580923215,
      "clip_ratio/low_mean": 0.0012077190185664222,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004376551733002998,
      "entropy": 0.3373456299304962,
      "epoch": 0.5746306673458992,
      "grad_norm": 12.571338653564453,
      "learning_rate": 3.4034229828850854e-07,
      "loss": 7.8154,
      "reward": 0.3843749761581421,
      "reward_std": 0.4458286762237549,
      "rewards/reasoning_density_reward/mean": 0.19687500596046448,
      "rewards/reasoning_density_reward/std": 0.12524721026420593,
      "rewards/success_reward/mean": 0.1875,
      "rewards/success_reward/std": 0.5123475790023804,
      "rewards/trajectory_similarity_reward/mean": 0.0,
      "rewards/trajectory_similarity_reward/std": 0.0,
      "step": 705
    },
    {
      "clip_ratio/high_max": 0.010861011280212551,
      "clip_ratio/high_mean": 0.0015792826015967876,
      "clip_ratio/low_mean": 0.0009638180308684241,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025431006724829786,
      "entropy": 0.3306904658675194,
      "epoch": 0.5754457463066734,
      "grad_norm": 16.66823959350586,
      "learning_rate": 3.3969030154849224e-07,
      "loss": -4.6088,
      "reward": 2.330994874238968,
      "reward_std": 0.35380294919013977,
      "rewards/reasoning_density_reward/mean": 0.3343750089406967,
      "rewards/reasoning_density_reward/std": 0.1386510208249092,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 0.2561737895011902,
      "rewards/trajectory_similarity_reward/mean": 0.4028697907924652,
      "rewards/trajectory_similarity_reward/std": 0.12313669919967651,
      "step": 706
    },
    {
      "clip_ratio/high_max": 0.014592543011531234,
      "clip_ratio/high_mean": 0.0036098377022426575,
      "clip_ratio/low_mean": 0.0020192808733554557,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005629118619253859,
      "entropy": 0.3087179437279701,
      "epoch": 0.5762608252674478,
      "grad_norm": 16.15052604675293,
      "learning_rate": 3.3903830480847594e-07,
      "loss": 10.3383,
      "reward": 3.2826669216156006,
      "reward_std": 1.2156635522842407,
      "rewards/reasoning_density_reward/mean": 0.4156249761581421,
      "rewards/reasoning_density_reward/std": 0.117924764752388,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 0.9287087917327881,
      "rewards/trajectory_similarity_reward/mean": 0.4295419156551361,
      "rewards/trajectory_similarity_reward/std": 0.3925151228904724,
      "step": 707
    },
    {
      "clip_ratio/high_max": 0.018379342975094914,
      "clip_ratio/high_mean": 0.003419546424993314,
      "clip_ratio/low_mean": 0.0008490175605402328,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042685640073614195,
      "entropy": 0.30155588313937187,
      "epoch": 0.5770759042282221,
      "grad_norm": 18.862897872924805,
      "learning_rate": 3.3838630806845964e-07,
      "loss": 5.8264,
      "reward": 3.0028659105300903,
      "reward_std": 1.3482880890369415,
      "rewards/reasoning_density_reward/mean": 0.4203125089406967,
      "rewards/reasoning_density_reward/std": 0.10736137628555298,
      "rewards/success_reward/mean": 2.109375,
      "rewards/success_reward/std": 1.127556473016739,
      "rewards/trajectory_similarity_reward/mean": 0.47317835688591003,
      "rewards/trajectory_similarity_reward/std": 0.41937634348869324,
      "step": 708
    },
    {
      "clip_ratio/high_max": 0.016673043777700514,
      "clip_ratio/high_mean": 0.0034458112422726117,
      "clip_ratio/low_mean": 0.0016932329308474436,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005139044165844098,
      "entropy": 0.32889121025800705,
      "epoch": 0.5778909831889965,
      "grad_norm": 16.343368530273438,
      "learning_rate": 3.3773431132844334e-07,
      "loss": -3.8669,
      "reward": 1.798462152481079,
      "reward_std": 1.825190782546997,
      "rewards/reasoning_density_reward/mean": 0.32371485233306885,
      "rewards/reasoning_density_reward/std": 0.12271636724472046,
      "rewards/success_reward/mean": 1.21875,
      "rewards/success_reward/std": 1.4716062545776367,
      "rewards/trajectory_similarity_reward/mean": 0.2559972405433655,
      "rewards/trajectory_similarity_reward/std": 0.3495660126209259,
      "step": 709
    },
    {
      "clip_ratio/high_max": 0.014541840530000627,
      "clip_ratio/high_mean": 0.0027949642535531893,
      "clip_ratio/low_mean": 0.0008469451568089426,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003641909424914047,
      "entropy": 0.3283786214888096,
      "epoch": 0.5787060621497707,
      "grad_norm": 15.022056579589844,
      "learning_rate": 3.3708231458842704e-07,
      "loss": 3.877,
      "reward": 2.821852684020996,
      "reward_std": 1.0820353031158447,
      "rewards/reasoning_density_reward/mean": 0.41363638639450073,
      "rewards/reasoning_density_reward/std": 0.10123208165168762,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5332163572311401,
      "rewards/trajectory_similarity_reward/std": 0.44846367835998535,
      "step": 710
    },
    {
      "clip_ratio/high_max": 0.01646755079855211,
      "clip_ratio/high_mean": 0.0031954846963344608,
      "clip_ratio/low_mean": 0.0005355112079996616,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003730995929799974,
      "entropy": 0.3364521935582161,
      "epoch": 0.5795211411105451,
      "grad_norm": 11.780835151672363,
      "learning_rate": 3.3643031784841074e-07,
      "loss": 4.962,
      "reward": 2.2560794949531555,
      "reward_std": 1.274736613035202,
      "rewards/reasoning_density_reward/mean": 0.3132338970899582,
      "rewards/reasoning_density_reward/std": 0.1573876515030861,
      "rewards/success_reward/mean": 1.59375,
      "rewards/success_reward/std": 1.4137958884239197,
      "rewards/trajectory_similarity_reward/mean": 0.3490956425666809,
      "rewards/trajectory_similarity_reward/std": 0.43454885482788086,
      "step": 711
    },
    {
      "clip_ratio/high_max": 0.018735300051048398,
      "clip_ratio/high_mean": 0.0046164010564098135,
      "clip_ratio/low_mean": 0.0010020807167165913,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005618481751298532,
      "entropy": 0.36613743379712105,
      "epoch": 0.5803362200713195,
      "grad_norm": 13.551299095153809,
      "learning_rate": 3.3577832110839444e-07,
      "loss": -5.1582,
      "reward": 3.049391508102417,
      "reward_std": 0.8959210515022278,
      "rewards/reasoning_density_reward/mean": 0.4208333492279053,
      "rewards/reasoning_density_reward/std": 0.10671873390674591,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.5660580396652222,
      "rewards/trajectory_similarity_reward/std": 0.46120041608810425,
      "step": 712
    },
    {
      "clip_ratio/high_max": 0.020859605283476412,
      "clip_ratio/high_mean": 0.00499165547080338,
      "clip_ratio/low_mean": 0.00153181466521346,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006523470190586522,
      "entropy": 0.325895756483078,
      "epoch": 0.5811512990320937,
      "grad_norm": 15.816411018371582,
      "learning_rate": 3.3512632436837814e-07,
      "loss": 0.9291,
      "reward": 2.0906596183776855,
      "reward_std": 0.7560126781463623,
      "rewards/reasoning_density_reward/mean": 0.3754464387893677,
      "rewards/reasoning_density_reward/std": 0.11923705041408539,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.40271317958831787,
      "rewards/trajectory_similarity_reward/std": 0.47474434971809387,
      "step": 713
    },
    {
      "clip_ratio/high_max": 0.014826378261204809,
      "clip_ratio/high_mean": 0.0033453579017077573,
      "clip_ratio/low_mean": 0.00047994744090829045,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038253053280641325,
      "entropy": 0.32386134564876556,
      "epoch": 0.5819663779928681,
      "grad_norm": 16.834375381469727,
      "learning_rate": 3.3447432762836184e-07,
      "loss": 2.9008,
      "reward": 3.154542565345764,
      "reward_std": 1.5391627550125122,
      "rewards/reasoning_density_reward/mean": 0.4330357015132904,
      "rewards/reasoning_density_reward/std": 0.0959642343223095,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.4715069830417633,
      "rewards/trajectory_similarity_reward/std": 0.41822288930416107,
      "step": 714
    },
    {
      "clip_ratio/high_max": 0.01832793647190556,
      "clip_ratio/high_mean": 0.005106294192955829,
      "clip_ratio/low_mean": 0.0007664197837584652,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005872714129509404,
      "entropy": 0.32643479481339455,
      "epoch": 0.5827814569536424,
      "grad_norm": 16.92911720275879,
      "learning_rate": 3.3382233088834554e-07,
      "loss": -3.158,
      "reward": 2.939851760864258,
      "reward_std": 1.922142744064331,
      "rewards/reasoning_density_reward/mean": 0.330322802066803,
      "rewards/reasoning_density_reward/std": 0.1764552742242813,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.5470290184020996,
      "rewards/trajectory_similarity_reward/std": 0.40540996193885803,
      "step": 715
    },
    {
      "clip_ratio/high_max": 0.0173598846886307,
      "clip_ratio/high_mean": 0.003790829228819348,
      "clip_ratio/low_mean": 0.0015885134889686015,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005379342765081674,
      "entropy": 0.3058563247323036,
      "epoch": 0.5835965359144167,
      "grad_norm": 15.519600868225098,
      "learning_rate": 3.3317033414832924e-07,
      "loss": -4.2565,
      "reward": 2.9584784507751465,
      "reward_std": 1.5104687213897705,
      "rewards/reasoning_density_reward/mean": 0.39642858505249023,
      "rewards/reasoning_density_reward/std": 0.11614557355642319,
      "rewards/success_reward/mean": 1.96875,
      "rewards/success_reward/std": 1.419727087020874,
      "rewards/trajectory_similarity_reward/mean": 0.5932997465133667,
      "rewards/trajectory_similarity_reward/std": 0.4900687634944916,
      "step": 716
    },
    {
      "clip_ratio/high_max": 0.019790931022726,
      "clip_ratio/high_mean": 0.00515975255984813,
      "clip_ratio/low_mean": 0.0005188476279727183,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005678600165992975,
      "entropy": 0.3191523216664791,
      "epoch": 0.584411614875191,
      "grad_norm": 16.96806526184082,
      "learning_rate": 3.32518337408313e-07,
      "loss": 7.8763,
      "reward": 2.8731296062469482,
      "reward_std": 0.8189125955104828,
      "rewards/reasoning_density_reward/mean": 0.41272321343421936,
      "rewards/reasoning_density_reward/std": 0.09818192943930626,
      "rewards/success_reward/mean": 1.921875,
      "rewards/success_reward/std": 1.0523346662521362,
      "rewards/trajectory_similarity_reward/mean": 0.5385313928127289,
      "rewards/trajectory_similarity_reward/std": 0.4336393475532532,
      "step": 717
    },
    {
      "clip_ratio/high_max": 0.014923606533557177,
      "clip_ratio/high_mean": 0.0029671926895389333,
      "clip_ratio/low_mean": 0.002447905389999505,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005415098145022057,
      "entropy": 0.3277497813105583,
      "epoch": 0.5852266938359654,
      "grad_norm": 29.093448638916016,
      "learning_rate": 3.3186634066829664e-07,
      "loss": 2.4446,
      "reward": 2.697774648666382,
      "reward_std": 0.9205578565597534,
      "rewards/reasoning_density_reward/mean": 0.40935784578323364,
      "rewards/reasoning_density_reward/std": 0.09931083768606186,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.41341662406921387,
      "rewards/trajectory_similarity_reward/std": 0.3709767162799835,
      "step": 718
    },
    {
      "clip_ratio/high_max": 0.021800478571094573,
      "clip_ratio/high_mean": 0.004970195805071853,
      "clip_ratio/low_mean": 0.0006737499352311715,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005643945842166431,
      "entropy": 0.34239279478788376,
      "epoch": 0.5860417727967396,
      "grad_norm": 15.43528938293457,
      "learning_rate": 3.3121434392828034e-07,
      "loss": 1.0883,
      "reward": 3.226630687713623,
      "reward_std": 1.9036791324615479,
      "rewards/reasoning_density_reward/mean": 0.3440340757369995,
      "rewards/reasoning_density_reward/std": 0.17230984568595886,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.341640830039978,
      "rewards/trajectory_similarity_reward/mean": 0.6325966119766235,
      "rewards/trajectory_similarity_reward/std": 0.39066624641418457,
      "step": 719
    },
    {
      "clip_ratio/high_max": 0.01840072515187785,
      "clip_ratio/high_mean": 0.003533446229994297,
      "clip_ratio/low_mean": 0.0002822478345478885,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038156940427143127,
      "entropy": 0.3323635198175907,
      "epoch": 0.586856851757514,
      "grad_norm": 13.48779582977295,
      "learning_rate": 3.3056234718826404e-07,
      "loss": 15.8782,
      "reward": 2.98940110206604,
      "reward_std": 1.601845622062683,
      "rewards/reasoning_density_reward/mean": 0.41458332538604736,
      "rewards/reasoning_density_reward/std": 0.10034661740064621,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.5123177170753479,
      "rewards/trajectory_similarity_reward/std": 0.4470868408679962,
      "step": 720
    },
    {
      "clip_ratio/high_max": 0.014857495087198913,
      "clip_ratio/high_mean": 0.0027141195023432374,
      "clip_ratio/low_mean": 0.0003670917831186671,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030812112818239257,
      "entropy": 0.3509768284857273,
      "epoch": 0.5876719307182884,
      "grad_norm": 12.848700523376465,
      "learning_rate": 3.2991035044824774e-07,
      "loss": 4.155,
      "reward": 3.2844417095184326,
      "reward_std": 0.9967443943023682,
      "rewards/reasoning_density_reward/mean": 0.32734376192092896,
      "rewards/reasoning_density_reward/std": 0.16418971121311188,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.078192949295044,
      "rewards/trajectory_similarity_reward/mean": 0.5195980072021484,
      "rewards/trajectory_similarity_reward/std": 0.36226868629455566,
      "step": 721
    },
    {
      "clip_ratio/high_max": 0.023160643875598907,
      "clip_ratio/high_mean": 0.005325193240423687,
      "clip_ratio/low_mean": 0.0011729068064596504,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006498100032331422,
      "entropy": 0.3279968425631523,
      "epoch": 0.5884870096790626,
      "grad_norm": 17.405893325805664,
      "learning_rate": 3.2925835370823144e-07,
      "loss": 16.7353,
      "reward": 4.045992851257324,
      "reward_std": 0.8853721618652344,
      "rewards/reasoning_density_reward/mean": 0.4333333373069763,
      "rewards/reasoning_density_reward/std": 0.12894126772880554,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.8001595139503479,
      "rewards/trajectory_similarity_reward/std": 0.25695472955703735,
      "step": 722
    },
    {
      "clip_ratio/high_max": 0.01645046629710123,
      "clip_ratio/high_mean": 0.0045278721081558615,
      "clip_ratio/low_mean": 0.0008536755485692993,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005381547671277076,
      "entropy": 0.34731679782271385,
      "epoch": 0.589302088639837,
      "grad_norm": 20.026840209960938,
      "learning_rate": 3.2860635696821514e-07,
      "loss": 2.4715,
      "reward": 2.3683130741119385,
      "reward_std": 1.2280899286270142,
      "rewards/reasoning_density_reward/mean": 0.3276515305042267,
      "rewards/reasoning_density_reward/std": 0.11854063719511032,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.125,
      "rewards/trajectory_similarity_reward/mean": 0.25941163301467896,
      "rewards/trajectory_similarity_reward/std": 0.34778404235839844,
      "step": 723
    },
    {
      "clip_ratio/high_max": 0.013166424643713981,
      "clip_ratio/high_mean": 0.002721260410908144,
      "clip_ratio/low_mean": 0.0007527576963184401,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034740180271910504,
      "entropy": 0.3302171491086483,
      "epoch": 0.5901171676006113,
      "grad_norm": 12.558886528015137,
      "learning_rate": 3.2795436022819884e-07,
      "loss": 0.4711,
      "reward": 2.7528040409088135,
      "reward_std": 0.9034095406532288,
      "rewards/reasoning_density_reward/mean": 0.3849296569824219,
      "rewards/reasoning_density_reward/std": 0.08825649321079254,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.49287429451942444,
      "rewards/trajectory_similarity_reward/std": 0.4112900495529175,
      "step": 724
    },
    {
      "clip_ratio/high_max": 0.021345289307646453,
      "clip_ratio/high_mean": 0.004278196793165989,
      "clip_ratio/low_mean": 0.0006419649434974417,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0049201617075596005,
      "entropy": 0.33635587990283966,
      "epoch": 0.5909322465613857,
      "grad_norm": 16.864397048950195,
      "learning_rate": 3.2730236348818254e-07,
      "loss": 5.771,
      "reward": 4.028075218200684,
      "reward_std": 0.8883209228515625,
      "rewards/reasoning_density_reward/mean": 0.453125,
      "rewards/reasoning_density_reward/std": 0.10562314838171005,
      "rewards/success_reward/mean": 2.8125,
      "rewards/success_reward/std": 0.75,
      "rewards/trajectory_similarity_reward/mean": 0.7624503970146179,
      "rewards/trajectory_similarity_reward/std": 0.2740139961242676,
      "step": 725
    },
    {
      "clip_ratio/high_max": 0.023060684325173497,
      "clip_ratio/high_mean": 0.005677418201230466,
      "clip_ratio/low_mean": 0.0015521987370448187,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007229617040138692,
      "entropy": 0.30293793231248856,
      "epoch": 0.5917473255221599,
      "grad_norm": 18.602901458740234,
      "learning_rate": 3.2665036674816624e-07,
      "loss": 2.2325,
      "reward": 2.8986655473709106,
      "reward_std": 1.7930591106414795,
      "rewards/reasoning_density_reward/mean": 0.42500001192092896,
      "rewards/reasoning_density_reward/std": 0.10000000149011612,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.5986655056476593,
      "rewards/trajectory_similarity_reward/std": 0.482219934463501,
      "step": 726
    },
    {
      "clip_ratio/high_max": 0.011928067659027874,
      "clip_ratio/high_mean": 0.002749511710135266,
      "clip_ratio/low_mean": 0.0008568949560867622,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036064066080143675,
      "entropy": 0.330657247453928,
      "epoch": 0.5925624044829343,
      "grad_norm": 15.07744312286377,
      "learning_rate": 3.2599837000814994e-07,
      "loss": 7.6191,
      "reward": 2.4659371376037598,
      "reward_std": 1.1494935750961304,
      "rewards/reasoning_density_reward/mean": 0.3499999940395355,
      "rewards/reasoning_density_reward/std": 0.14844878017902374,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.4284369647502899,
      "rewards/trajectory_similarity_reward/std": 0.45054757595062256,
      "step": 727
    },
    {
      "clip_ratio/high_max": 0.01504991389811039,
      "clip_ratio/high_mean": 0.003999999957159162,
      "clip_ratio/low_mean": 0.0013418281087069772,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005341828102245927,
      "entropy": 0.31691448390483856,
      "epoch": 0.5933774834437087,
      "grad_norm": 24.694530487060547,
      "learning_rate": 3.2534637326813364e-07,
      "loss": 1.9153,
      "reward": 3.2985042333602905,
      "reward_std": 1.5483933687210083,
      "rewards/reasoning_density_reward/mean": 0.3872023671865463,
      "rewards/reasoning_density_reward/std": 0.11668552830815315,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.5675519108772278,
      "rewards/trajectory_similarity_reward/std": 0.41254884004592896,
      "step": 728
    },
    {
      "clip_ratio/high_max": 0.02057585574220866,
      "clip_ratio/high_mean": 0.005019499454647303,
      "clip_ratio/low_mean": 0.0009991301376430783,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006018629705067724,
      "entropy": 0.31929266080260277,
      "epoch": 0.5941925624044829,
      "grad_norm": 15.514227867126465,
      "learning_rate": 3.2469437652811734e-07,
      "loss": -6.772,
      "reward": 3.0244719982147217,
      "reward_std": 1.6902782917022705,
      "rewards/reasoning_density_reward/mean": 0.41458332538604736,
      "rewards/reasoning_density_reward/std": 0.10034661740064621,
      "rewards/success_reward/mean": 2.0625,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.5473886728286743,
      "rewards/trajectory_similarity_reward/std": 0.44491711258888245,
      "step": 729
    },
    {
      "clip_ratio/high_max": 0.015538120525889099,
      "clip_ratio/high_mean": 0.0039615993009647354,
      "clip_ratio/low_mean": 0.001618214722839184,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00557981405290775,
      "entropy": 0.302999097853899,
      "epoch": 0.5950076413652573,
      "grad_norm": 25.916152954101562,
      "learning_rate": 3.2404237978810104e-07,
      "loss": 1.9399,
      "reward": 2.6519548892974854,
      "reward_std": 1.9841711521148682,
      "rewards/reasoning_density_reward/mean": 0.40625,
      "rewards/reasoning_density_reward/std": 0.10123474895954132,
      "rewards/success_reward/mean": 1.734375,
      "rewards/success_reward/std": 1.4344323873519897,
      "rewards/trajectory_similarity_reward/mean": 0.5113299041986465,
      "rewards/trajectory_similarity_reward/std": 0.4893078953027725,
      "step": 730
    },
    {
      "clip_ratio/high_max": 0.02184108446817845,
      "clip_ratio/high_mean": 0.00634554831776768,
      "clip_ratio/low_mean": 0.001412066732882522,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007757615065202117,
      "entropy": 0.3285379707813263,
      "epoch": 0.5958227203260316,
      "grad_norm": 18.725250244140625,
      "learning_rate": 3.2339038304808474e-07,
      "loss": -1.7936,
      "reward": 3.686443567276001,
      "reward_std": 1.0760545432567596,
      "rewards/reasoning_density_reward/mean": 0.4624999910593033,
      "rewards/reasoning_density_reward/std": 0.07887785881757736,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.1831679940223694,
      "rewards/trajectory_similarity_reward/mean": 0.7864435613155365,
      "rewards/trajectory_similarity_reward/std": 0.38331757485866547,
      "step": 731
    },
    {
      "clip_ratio/high_max": 0.010747699183411896,
      "clip_ratio/high_mean": 0.0017095369839807972,
      "clip_ratio/low_mean": 0.001921854302054271,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036313912860350683,
      "entropy": 0.3508451133966446,
      "epoch": 0.5966377992868059,
      "grad_norm": 16.835514068603516,
      "learning_rate": 3.227383863080685e-07,
      "loss": 11.954,
      "reward": 1.8018614053726196,
      "reward_std": 1.852851390838623,
      "rewards/reasoning_density_reward/mean": 0.3375000059604645,
      "rewards/reasoning_density_reward/std": 0.16278821229934692,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.339361310005188,
      "rewards/trajectory_similarity_reward/std": 0.4547730088233948,
      "step": 732
    },
    {
      "clip_ratio/high_max": 0.014826695725787431,
      "clip_ratio/high_mean": 0.0028973559965379536,
      "clip_ratio/low_mean": 0.0019215777865611017,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004818933928618208,
      "entropy": 0.3520160801708698,
      "epoch": 0.5974528782475802,
      "grad_norm": 13.089106559753418,
      "learning_rate": 3.220863895680522e-07,
      "loss": 16.5318,
      "reward": 1.7610124349594116,
      "reward_std": 1.8118305206298828,
      "rewards/reasoning_density_reward/mean": 0.30148810148239136,
      "rewards/reasoning_density_reward/std": 0.15290306508541107,
      "rewards/success_reward/mean": 1.125,
      "rewards/success_reward/std": 1.5,
      "rewards/trajectory_similarity_reward/mean": 0.33452433347702026,
      "rewards/trajectory_similarity_reward/std": 0.4464569389820099,
      "step": 733
    },
    {
      "clip_ratio/high_max": 0.014256183058023453,
      "clip_ratio/high_mean": 0.0021478108392329887,
      "clip_ratio/low_mean": 0.001684201655734796,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003832012473139912,
      "entropy": 0.3567435033619404,
      "epoch": 0.5982679572083546,
      "grad_norm": 12.264330863952637,
      "learning_rate": 3.2143439282803584e-07,
      "loss": 6.0219,
      "reward": 1.874819576740265,
      "reward_std": 1.5205971002578735,
      "rewards/reasoning_density_reward/mean": 0.31793154776096344,
      "rewards/reasoning_density_reward/std": 0.12414437159895897,
      "rewards/success_reward/mean": 1.359375,
      "rewards/success_reward/std": 1.183596134185791,
      "rewards/trajectory_similarity_reward/mean": 0.19751303642988205,
      "rewards/trajectory_similarity_reward/std": 0.3719254732131958,
      "step": 734
    },
    {
      "clip_ratio/high_max": 0.011487408832181245,
      "clip_ratio/high_mean": 0.002312614808033686,
      "clip_ratio/low_mean": 0.002918039019277785,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0052306538564153016,
      "entropy": 0.328623004257679,
      "epoch": 0.5990830361691288,
      "grad_norm": 14.9776611328125,
      "learning_rate": 3.2078239608801954e-07,
      "loss": 2.3523,
      "reward": 3.3723654747009277,
      "reward_std": 1.5045785903930664,
      "rewards/reasoning_density_reward/mean": 0.4375,
      "rewards/reasoning_density_reward/std": 0.08850611001253128,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.2209115028381348,
      "rewards/trajectory_similarity_reward/mean": 0.591115415096283,
      "rewards/trajectory_similarity_reward/std": 0.3999078869819641,
      "step": 735
    },
    {
      "clip_ratio/high_max": 0.019111368514131755,
      "clip_ratio/high_mean": 0.004710536362836137,
      "clip_ratio/low_mean": 0.0014151392679195851,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0061256756307557225,
      "entropy": 0.32413846626877785,
      "epoch": 0.5998981151299032,
      "grad_norm": 23.529022216796875,
      "learning_rate": 3.2013039934800323e-07,
      "loss": 3.5758,
      "reward": 3.1562469005584717,
      "reward_std": 1.436094045639038,
      "rewards/reasoning_density_reward/mean": 0.4322916865348816,
      "rewards/reasoning_density_reward/std": 0.09437292814254761,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.2247449159622192,
      "rewards/trajectory_similarity_reward/mean": 0.47395533323287964,
      "rewards/trajectory_similarity_reward/std": 0.3790922462940216,
      "step": 736
    },
    {
      "clip_ratio/high_max": 0.008990119851659983,
      "clip_ratio/high_mean": 0.002718727890169248,
      "clip_ratio/low_mean": 0.0016459284779557493,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043646563062793575,
      "entropy": 0.3363182209432125,
      "epoch": 0.6007131940906776,
      "grad_norm": 12.326214790344238,
      "learning_rate": 3.1947840260798693e-07,
      "loss": 12.2318,
      "reward": 3.3050904273986816,
      "reward_std": 0.9789865016937256,
      "rewards/reasoning_density_reward/mean": 0.37416917085647583,
      "rewards/reasoning_density_reward/std": 0.15256847441196442,
      "rewards/success_reward/mean": 2.34375,
      "rewards/success_reward/std": 1.0911576747894287,
      "rewards/trajectory_similarity_reward/mean": 0.5871712565422058,
      "rewards/trajectory_similarity_reward/std": 0.42100727558135986,
      "step": 737
    },
    {
      "clip_ratio/high_max": 0.018698358617257327,
      "clip_ratio/high_mean": 0.002546526178775821,
      "clip_ratio/low_mean": 0.0010286972392350435,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035752235125983134,
      "entropy": 0.3304601199924946,
      "epoch": 0.6015282730514518,
      "grad_norm": 17.731868743896484,
      "learning_rate": 3.1882640586797063e-07,
      "loss": -2.004,
      "reward": 2.8110830783843994,
      "reward_std": 0.6778244376182556,
      "rewards/reasoning_density_reward/mean": 0.40625,
      "rewards/reasoning_density_reward/std": 0.09979145228862762,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.3964240550994873,
      "rewards/trajectory_similarity_reward/mean": 0.5298330783843994,
      "rewards/trajectory_similarity_reward/std": 0.4918629229068756,
      "step": 738
    },
    {
      "clip_ratio/high_max": 0.014396492158994079,
      "clip_ratio/high_mean": 0.003219524078303948,
      "clip_ratio/low_mean": 0.0018647367833182216,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005084260905277915,
      "entropy": 0.33073560148477554,
      "epoch": 0.6023433520122262,
      "grad_norm": 22.048446655273438,
      "learning_rate": 3.1817440912795433e-07,
      "loss": 1.6546,
      "reward": 3.5637375116348267,
      "reward_std": 1.449958086013794,
      "rewards/reasoning_density_reward/mean": 0.4325520694255829,
      "rewards/reasoning_density_reward/std": 0.11466982215642929,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.1385715007781982,
      "rewards/trajectory_similarity_reward/mean": 0.6936855316162109,
      "rewards/trajectory_similarity_reward/std": 0.39483602344989777,
      "step": 739
    },
    {
      "clip_ratio/high_max": 0.0157021249178797,
      "clip_ratio/high_mean": 0.004112656170036644,
      "clip_ratio/low_mean": 0.00023506365687353536,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004347719775978476,
      "entropy": 0.3797360807657242,
      "epoch": 0.6031584309730005,
      "grad_norm": 11.775860786437988,
      "learning_rate": 3.1752241238793803e-07,
      "loss": -0.6646,
      "reward": 2.464177370071411,
      "reward_std": 1.19965398311615,
      "rewards/reasoning_density_reward/mean": 0.3374999761581421,
      "rewards/reasoning_density_reward/std": 0.14842712879180908,
      "rewards/success_reward/mean": 1.78125,
      "rewards/success_reward/std": 1.125,
      "rewards/trajectory_similarity_reward/mean": 0.34542739391326904,
      "rewards/trajectory_similarity_reward/std": 0.46105828881263733,
      "step": 740
    },
    {
      "clip_ratio/high_max": 0.013741055270656943,
      "clip_ratio/high_mean": 0.003283176221884787,
      "clip_ratio/low_mean": 0.0010866036282095592,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004369779897388071,
      "entropy": 0.34505395591259,
      "epoch": 0.6039735099337749,
      "grad_norm": 13.658812522888184,
      "learning_rate": 3.1687041564792173e-07,
      "loss": 6.166,
      "reward": 2.3401358127593994,
      "reward_std": 1.7652508020401,
      "rewards/reasoning_density_reward/mean": 0.37812501192092896,
      "rewards/reasoning_density_reward/std": 0.11685995757579803,
      "rewards/success_reward/mean": 1.6875,
      "rewards/success_reward/std": 1.3275917768478394,
      "rewards/trajectory_similarity_reward/mean": 0.2745107114315033,
      "rewards/trajectory_similarity_reward/std": 0.33382079005241394,
      "step": 741
    },
    {
      "clip_ratio/high_max": 0.018694740196224302,
      "clip_ratio/high_mean": 0.003450651216553524,
      "clip_ratio/low_mean": 0.0016340917791239917,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005084742850158364,
      "entropy": 0.3229355476796627,
      "epoch": 0.6047885888945491,
      "grad_norm": 14.93838119506836,
      "learning_rate": 3.1621841890790543e-07,
      "loss": 3.259,
      "reward": 0.7440821528434753,
      "reward_std": 0.9535448551177979,
      "rewards/reasoning_density_reward/mean": 0.28389424085617065,
      "rewards/reasoning_density_reward/std": 0.12422715127468109,
      "rewards/success_reward/mean": 0.375,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.08518792688846588,
      "rewards/trajectory_similarity_reward/std": 0.23278789222240448,
      "step": 742
    },
    {
      "clip_ratio/high_max": 0.019222403410822153,
      "clip_ratio/high_mean": 0.0036248224787414074,
      "clip_ratio/low_mean": 0.0009133876556006726,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004538210167083889,
      "entropy": 0.34687405079603195,
      "epoch": 0.6056036678553235,
      "grad_norm": 16.09386444091797,
      "learning_rate": 3.1556642216788913e-07,
      "loss": 11.7036,
      "reward": 2.0450515747070312,
      "reward_std": 2.1162843704223633,
      "rewards/reasoning_density_reward/mean": 0.3752060532569885,
      "rewards/reasoning_density_reward/std": 0.10115061700344086,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.5370426177978516,
      "rewards/trajectory_similarity_reward/mean": 0.35734546184539795,
      "rewards/trajectory_similarity_reward/std": 0.43125852942466736,
      "step": 743
    },
    {
      "clip_ratio/high_max": 0.01107470493298024,
      "clip_ratio/high_mean": 0.0017464076372561976,
      "clip_ratio/low_mean": 0.0006487623031716794,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023951699549797922,
      "entropy": 0.3262798339128494,
      "epoch": 0.6064187468160979,
      "grad_norm": 14.207890510559082,
      "learning_rate": 3.1491442542787283e-07,
      "loss": 0.4577,
      "reward": 2.5481083393096924,
      "reward_std": 1.5505785942077637,
      "rewards/reasoning_density_reward/mean": 0.34732145071029663,
      "rewards/reasoning_density_reward/std": 0.13316848874092102,
      "rewards/success_reward/mean": 1.875,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.3257869780063629,
      "rewards/trajectory_similarity_reward/std": 0.43584737181663513,
      "step": 744
    },
    {
      "clip_ratio/high_max": 0.01189180847723037,
      "clip_ratio/high_mean": 0.0027202397031942382,
      "clip_ratio/low_mean": 0.0012674899044213817,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039877295785117894,
      "entropy": 0.35636961087584496,
      "epoch": 0.6072338257768721,
      "grad_norm": 18.5509033203125,
      "learning_rate": 3.1426242868785653e-07,
      "loss": 5.9995,
      "reward": 1.982546329498291,
      "reward_std": 1.3801380395889282,
      "rewards/reasoning_density_reward/mean": 0.33125001192092896,
      "rewards/reasoning_density_reward/std": 0.125,
      "rewards/success_reward/mean": 1.40625,
      "rewards/success_reward/std": 1.157853603363037,
      "rewards/trajectory_similarity_reward/mean": 0.24504625797271729,
      "rewards/trajectory_similarity_reward/std": 0.438610315322876,
      "step": 745
    },
    {
      "clip_ratio/high_max": 0.014197440614225343,
      "clip_ratio/high_mean": 0.0033552834975125734,
      "clip_ratio/low_mean": 0.0011400863804738037,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004495369896176271,
      "entropy": 0.3887362666428089,
      "epoch": 0.6080489047376465,
      "grad_norm": 12.628972053527832,
      "learning_rate": 3.1361043194784023e-07,
      "loss": -1.3345,
      "reward": 1.8433785438537598,
      "reward_std": 1.4343733191490173,
      "rewards/reasoning_density_reward/mean": 0.22369791567325592,
      "rewards/reasoning_density_reward/std": 0.17507873475551605,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.2004882097244263,
      "rewards/trajectory_similarity_reward/mean": 0.30718061327934265,
      "rewards/trajectory_similarity_reward/std": 0.3701706677675247,
      "step": 746
    },
    {
      "clip_ratio/high_max": 0.007073289714753628,
      "clip_ratio/high_mean": 0.00141357776010409,
      "clip_ratio/low_mean": 0.0012196701500215568,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002633247902849689,
      "entropy": 0.4133944921195507,
      "epoch": 0.6088639836984208,
      "grad_norm": 13.707287788391113,
      "learning_rate": 3.12958435207824e-07,
      "loss": -5.6885,
      "reward": 1.9170106649398804,
      "reward_std": 1.826815128326416,
      "rewards/reasoning_density_reward/mean": 0.3322916626930237,
      "rewards/reasoning_density_reward/std": 0.1612989753484726,
      "rewards/success_reward/mean": 1.3125,
      "rewards/success_reward/std": 1.4361406564712524,
      "rewards/trajectory_similarity_reward/mean": 0.27221882343292236,
      "rewards/trajectory_similarity_reward/std": 0.38224077224731445,
      "step": 747
    },
    {
      "clip_ratio/high_max": 0.01521719916490838,
      "clip_ratio/high_mean": 0.0027997235156362876,
      "clip_ratio/low_mean": 0.0024389908176090103,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005238714249571785,
      "entropy": 0.34174104779958725,
      "epoch": 0.6096790626591951,
      "grad_norm": 14.99661922454834,
      "learning_rate": 3.123064384678077e-07,
      "loss": 5.1878,
      "reward": 3.4148237705230713,
      "reward_std": 1.1184638738632202,
      "rewards/reasoning_density_reward/mean": 0.3834821581840515,
      "rewards/reasoning_density_reward/std": 0.15072882175445557,
      "rewards/success_reward/mean": 2.4375,
      "rewards/success_reward/std": 1.209338665008545,
      "rewards/trajectory_similarity_reward/mean": 0.5938417315483093,
      "rewards/trajectory_similarity_reward/std": 0.320853054523468,
      "step": 748
    },
    {
      "clip_ratio/high_max": 0.013446430792100728,
      "clip_ratio/high_mean": 0.002554076418164186,
      "clip_ratio/low_mean": 0.0010191736655542627,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035732500728045125,
      "entropy": 0.32289963215589523,
      "epoch": 0.6104941416199694,
      "grad_norm": 16.736879348754883,
      "learning_rate": 3.116544417277914e-07,
      "loss": 1.9263,
      "reward": 3.859916925430298,
      "reward_std": 0.907070517539978,
      "rewards/reasoning_density_reward/mean": 0.45781248807907104,
      "rewards/reasoning_density_reward/std": 0.07730823010206223,
      "rewards/success_reward/mean": 2.625,
      "rewards/success_reward/std": 1.0246951580047607,
      "rewards/trajectory_similarity_reward/mean": 0.7771044969558716,
      "rewards/trajectory_similarity_reward/std": 0.3489649295806885,
      "step": 749
    },
    {
      "clip_ratio/high_max": 0.012003393028862774,
      "clip_ratio/high_mean": 0.002419844298856333,
      "clip_ratio/low_mean": 0.002093220296956133,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004513064573984593,
      "entropy": 0.3233438767492771,
      "epoch": 0.6113092205807438,
      "grad_norm": 14.180194854736328,
      "learning_rate": 3.110024449877751e-07,
      "loss": 2.4703,
      "reward": 2.938784599304199,
      "reward_std": 0.8582838177680969,
      "rewards/reasoning_density_reward/mean": 0.3850446343421936,
      "rewards/reasoning_density_reward/std": 0.15961378812789917,
      "rewards/success_reward/mean": 2.25,
      "rewards/success_reward/std": 1.095445156097412,
      "rewards/trajectory_similarity_reward/mean": 0.30373990535736084,
      "rewards/trajectory_similarity_reward/std": 0.30743250250816345,
      "step": 750
    }
  ],
  "logging_steps": 1,
  "max_steps": 1227,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}