{ "run_index": 1, "evaluated_utc": "2026-09-22T06:32:28Z", "test_file": "all/test-00000-of-00001.parquet", "test_sha256": "4f294f218ea1da27f3efef936359389c62ea4d3973a41457732990f1d31b647c", "n_cases": 400, "n_decisions": 2000, "latency_ms_per_case_cpu": 134.04, "raw": { "overall": { "n": 2000, "accuracy": 0.751, "soft_accuracy": 0.5634989456557318, "brier": 0.05914920990215329, "kl": 0.10238191010682485, "tv": 0.1433847785776465, "nll": 0.6367698398652486, "ece": 0.12722164873538963, "score_mae": 0.2746505811582656, "within_1": 0.98375 }, "by_workflow": { "agent_trace_observability": { "n": 500, "accuracy": 0.732, "soft_accuracy": 0.4724501203294106, "brier": 0.05628207400044921, "kl": 0.09350560088046769, "tv": 0.15358126436337113, "nll": 0.7006394391227598, "ece": 0.1601141162832159, "score_mae": 0.2100908384556247, "within_1": 1.0 }, "customer_service": { "n": 500, "accuracy": 0.728, "soft_accuracy": 0.5469157003096765, "brier": 0.092352802056536, "kl": 0.16993479269228093, "tv": 0.1916863908237952, "nll": 0.6922485950267834, "ece": 0.11188648161232355, "score_mae": 0.2972645950098066, "within_1": 0.975 }, "invoice_processing": { "n": 500, "accuracy": 0.808, "soft_accuracy": 0.7196525728518464, "brier": 0.04959251157667264, "kl": 0.0847333468242194, "tv": 0.11005123578831648, "nll": 0.4740942175604576, "ece": 0.09652193226496708, "score_mae": 0.2800105412177077, "within_1": 0.975 }, "security_incidents": { "n": 500, "accuracy": 0.736, "soft_accuracy": 0.5149773891319935, "brier": 0.038369451974955324, "kl": 0.06135390003033135, "tv": 0.11822022333510321, "nll": 0.6800971077509935, "ece": 0.14461991061586713, "score_mae": 0.3112363499499235, "within_1": 0.985 } }, "by_qtype": { "choice": { "n": 600, "accuracy": 0.7383333333333333, "soft_accuracy": 0.4576608436665159, "brier": 0.07886955158977817, "kl": 0.15330907850375597, "tv": 0.18865741057538912, "nll": 0.7129818449599276, "ece": 0.1392636251597695, "score_mae": null, "within_1": null }, "noul": { "n": 600, "accuracy": 0.8366666666666667, "soft_accuracy": 0.6693370476449475, "brier": 0.03942886821452841, "kl": 0.05145474170989372, "tv": 0.09811214657990391, "nll": 0.3812379372014114, "ece": 0.0843618453812399, "score_mae": null, "within_1": null }, "score": { "n": 800, "accuracy": 0.69625, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.7712597630421172, "ece": 0.15033501893271706, "score_mae": 0.2746505811582656, "within_1": 0.98375 } }, "by_schema_id": { "agent_trace_observability/action": { "n": 100, "accuracy": 0.67, "soft_accuracy": 0.37864703694239377, "brier": 0.049903900779301506, "kl": 0.09724056949034372, "tv": 0.15841980300574204, "nll": 0.841429970588603, "ece": 0.16207918840409982, "score_mae": null, "within_1": null }, "agent_trace_observability/needs_review": { "n": 100, "accuracy": 0.82, "soft_accuracy": 0.6271667824819649, "brier": 0.039836710289119205, "kl": 0.05140740089529728, "tv": 0.10942551114017018, "nll": 0.41184288208503156, "ece": 0.1275521930036992, "score_mae": null, "within_1": null }, "agent_trace_observability/outcome": { "n": 100, "accuracy": 0.68, "soft_accuracy": 0.41153654156387304, "brier": 0.07910561093292694, "kl": 0.13186883225576213, "tv": 0.19289847894420123, "nll": 0.7898371925373076, "ece": 0.13892615370006228, "score_mae": null, "within_1": null }, "agent_trace_observability/risk": { "n": 100, "accuracy": 0.8, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.6549731615156594, "ece": 0.24708902020558798, "score_mae": 0.1947352688400199, "within_1": 1.0 }, "agent_trace_observability/urgency": { "n": 100, "accuracy": 0.69, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.8051139888871975, "ece": 0.1755901741662443, "score_mae": 0.22544640807122945, "within_1": 1.0 }, "customer_service/action": { "n": 100, "accuracy": 0.63, "soft_accuracy": 0.3888590717902884, "brier": 0.11445938581035998, "kl": 0.23031132103353194, "tv": 0.24566652927218924, "nll": 0.9319729586843554, "ece": 0.08587540619343464, "score_mae": null, "within_1": null }, "customer_service/category": { "n": 100, "accuracy": 0.9, "soft_accuracy": 0.6704960256392977, "brier": 0.0947763938424711, "kl": 0.20248588253857602, "tv": 0.18808011295017835, "nll": 0.38370140372630757, "ece": 0.14613724059962907, "score_mae": null, "within_1": null }, "customer_service/churn_risk": { "n": 100, "accuracy": 0.67, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.83026776181706, "ece": 0.16804481624874312, "score_mae": 0.3211845585193193, "within_1": 0.95 }, "customer_service/needs_human": { "n": 100, "accuracy": 0.75, "soft_accuracy": 0.5813920034994436, "brier": 0.06782262651677691, "kl": 0.07700717450473488, "tv": 0.14131253024901802, "nll": 0.5149443876497118, "ece": 0.0901062252733414, "score_mae": null, "within_1": null }, "customer_service/urgency": { "n": 100, "accuracy": 0.69, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.8003564632564825, "ece": 0.1752485705339351, "score_mae": 0.2733446315002939, "within_1": 1.0 }, "invoice_processing/discrepancy_severity": { "n": 100, "accuracy": 0.78, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.5601792853430406, "ece": 0.14199146682039665, "score_mae": 0.338526231566432, "within_1": 0.95 }, "invoice_processing/disposition": { "n": 100, "accuracy": 0.75, "soft_accuracy": 0.48504161823816555, "brier": 0.08354235197127319, "kl": 0.1496560986469237, "tv": 0.19025085349788567, "nll": 0.6579466495886376, "ece": 0.1524891184511141, "score_mae": null, "within_1": null }, "invoice_processing/duplicate": { "n": 100, "accuracy": 0.93, "soft_accuracy": 0.8669995078963517, "brier": 0.018845183257522446, "kl": 0.027196929038003314, "tv": 0.047014968294512044, "nll": 0.16610251723167443, "ece": 0.0660903240070419, "score_mae": null, "within_1": null }, "invoice_processing/matches_order": { "n": 100, "accuracy": 0.94, "soft_accuracy": 0.8069165924210221, "brier": 0.0463899995012223, "kl": 0.07734701278773116, "tv": 0.09288788557255173, "nll": 0.20546349749709156, "ece": 0.07770465780110603, "score_mae": null, "within_1": null }, "invoice_processing/urgency": { "n": 100, "accuracy": 0.64, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.7807791381418437, "ece": 0.12403058303108795, "score_mae": 0.22149485086898335, "within_1": 1.0 }, "security_incidents/credential_compromise": { "n": 100, "accuracy": 0.77, "soft_accuracy": 0.5431074459153316, "brier": 0.03405358380994987, "kl": 0.03940214977263347, "tv": 0.10374847615088617, "nll": 0.5468801338498097, "ece": 0.145414782843676, "score_mae": null, "within_1": null }, "security_incidents/disposition": { "n": 100, "accuracy": 0.8, "soft_accuracy": 0.41138476782507716, "brier": 0.051429666202336376, "kl": 0.10829176705739842, "tv": 0.1566286857821381, "nll": 0.6730028946343545, "ece": 0.21373141384977584, "score_mae": null, "within_1": null }, "security_incidents/severity": { "n": 100, "accuracy": 0.71, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.8968579196544322, "ece": 0.21102490765200058, "score_mae": 0.3344534301686142, "within_1": 0.98 }, "security_incidents/true_positive": { "n": 100, "accuracy": 0.81, "soft_accuracy": 0.5904399536555718, "brier": 0.02962510591257973, "kl": 0.036367783260962155, "tv": 0.09428350807228533, "nll": 0.4421942048951493, "ece": 0.13079891069657626, "score_mae": null, "within_1": null }, "security_incidents/urgency": { "n": 100, "accuracy": 0.59, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.841550385721222, "ece": 0.08675957856276363, "score_mae": 0.28801926973123276, "within_1": 0.99 } } }, "calibrated": { "overall": { "n": 2000, "accuracy": 0.751, "soft_accuracy": 0.6178046717519351, "brier": 0.11362459879677686, "kl": 0.30095497190224685, "tv": 0.20176915755108013, "nll": 0.5680045157373079, "ece": 0.04043277735449942, "score_mae": 0.3198573015138221, "within_1": 0.97 }, "by_workflow": { "agent_trace_observability": { "n": 500, "accuracy": 0.732, "soft_accuracy": 0.5271686746301996, "brier": 0.12098650868702557, "kl": 0.31954329306493995, "tv": 0.22993239481802855, "nll": 0.5973505916529968, "ece": 0.03588033188979682, "score_mae": 0.28590820356756197, "within_1": 1.0 }, "customer_service": { "n": 500, "accuracy": 0.728, "soft_accuracy": 0.6138200792710129, "brier": 0.145541356967536, "kl": 0.3411004238382146, "tv": 0.22775900057478538, "nll": 0.6500724208735882, "ece": 0.07012955999535193, "score_mae": 0.3169925538594844, "within_1": 0.965 }, "invoice_processing": { "n": 500, "accuracy": 0.808, "soft_accuracy": 0.7639967805225397, "brier": 0.08197299545550225, "kl": 0.2698870870596405, "tv": 0.14169172035189762, "nll": 0.4160005525805608, "ece": 0.04359159746369926, "score_mae": 0.27941288373058354, "within_1": 0.95 }, "security_incidents": { "n": 500, "accuracy": 0.736, "soft_accuracy": 0.5662331525839882, "brier": 0.10599753407704361, "kl": 0.2732890836461923, "tv": 0.20769351445960893, "nll": 0.608594497842086, "ece": 0.046897431140645096, "score_mae": 0.3971155648976584, "within_1": 0.965 } }, "by_qtype": { "choice": { "n": 600, "accuracy": 0.7383333333333333, "soft_accuracy": 0.5294220048812253, "brier": 0.15886700352925193, "kl": 0.4576430916821466, "tv": 0.2703961043025863, "nll": 0.6390465026947252, "ece": 0.06568009288736831, "score_mae": null, "within_1": null }, "noul": { "n": 600, "accuracy": 0.8366666666666667, "soft_accuracy": 0.7061873386226449, "brier": 0.06838219406430175, "kl": 0.14426685212234702, "tv": 0.13314221079957392, "nll": 0.3443171239150436, "ece": 0.03072768695903312, "score_mae": null, "within_1": null }, "score": { "n": 800, "accuracy": 0.69625, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.6824885693859432, "ece": 0.043157208326225695, "score_mae": 0.3198573015138221, "within_1": 0.97 } }, "by_schema_id": { "agent_trace_observability/action": { "n": 100, "accuracy": 0.67, "soft_accuracy": 0.43377777225260944, "brier": 0.12744161649559468, "kl": 0.3826436703933884, "tv": 0.2633365085097515, "nll": 0.7352119058432045, "ece": 0.08080262060002472, "score_mae": null, "within_1": null }, "agent_trace_observability/needs_review": { "n": 100, "accuracy": 0.82, "soft_accuracy": 0.6719805168552689, "brier": 0.06233964264866126, "kl": 0.1482735079470231, "tv": 0.13895963722913257, "nll": 0.3464312756882847, "ece": 0.09750878535037706, "score_mae": null, "within_1": null }, "agent_trace_observability/outcome": { "n": 100, "accuracy": 0.68, "soft_accuracy": 0.47574773478272037, "brier": 0.17317826691682073, "kl": 0.4277127008544081, "tv": 0.28750103871520155, "nll": 0.7251843727780874, "ece": 0.12738973196466497, "score_mae": null, "within_1": null }, "agent_trace_observability/risk": { "n": 100, "accuracy": 0.8, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.49726936882265277, "ece": 0.08128467110179902, "score_mae": 0.256063240195547, "within_1": 1.0 }, "agent_trace_observability/urgency": { "n": 100, "accuracy": 0.69, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.6826560351327541, "ece": 0.04947163723661896, "score_mae": 0.31575316693957683, "within_1": 1.0 }, "customer_service/action": { "n": 100, "accuracy": 0.63, "soft_accuracy": 0.45212191867191825, "brier": 0.2110131519451015, "kl": 0.5188205492002642, "tv": 0.3187358639074713, "nll": 0.9683544302994465, "ece": 0.17665053743921832, "score_mae": null, "within_1": null }, "customer_service/category": { "n": 100, "accuracy": 0.9, "soft_accuracy": 0.7640711181753796, "brier": 0.09520275546506135, "kl": 0.3121675177388705, "tv": 0.15788651204535845, "nll": 0.27464683825519653, "ece": 0.036861447188785965, "score_mae": null, "within_1": null }, "customer_service/churn_risk": { "n": 100, "accuracy": 0.67, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.7903341057001046, "ece": 0.0936859953048758, "score_mae": 0.3403095011269979, "within_1": 0.94 }, "customer_service/needs_human": { "n": 100, "accuracy": 0.75, "soft_accuracy": 0.6252672009657406, "brier": 0.1304081634924451, "kl": 0.19231320457550907, "tv": 0.20665462577152632, "nll": 0.5138780280031838, "ece": 0.12367391544275927, "score_mae": null, "within_1": null }, "customer_service/urgency": { "n": 100, "accuracy": 0.69, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.7031487021100088, "ece": 0.06809865228312525, "score_mae": 0.2936756065919708, "within_1": 0.99 }, "invoice_processing/discrepancy_severity": { "n": 100, "accuracy": 0.78, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.493989389662816, "ece": 0.0799121179160796, "score_mae": 0.3374784670461055, "within_1": 0.9 }, "invoice_processing/disposition": { "n": 100, "accuracy": 0.75, "soft_accuracy": 0.554000467328609, "brier": 0.16784104793395124, "kl": 0.5187716339663901, "tv": 0.2735823324908618, "nll": 0.6012390024595231, "ece": 0.1087785620484807, "score_mae": null, "within_1": null }, "invoice_processing/duplicate": { "n": 100, "accuracy": 0.93, "soft_accuracy": 0.8906173029033013, "brier": 0.029733229585660614, "kl": 0.10844597304603973, "tv": 0.06165477710925789, "nll": 0.17549498045278247, "ece": 0.06917040917004692, "score_mae": null, "within_1": null }, "invoice_processing/matches_order": { "n": 100, "accuracy": 0.94, "soft_accuracy": 0.8473725713357088, "brier": 0.048344708846894936, "kl": 0.1824436541664917, "tv": 0.08983805145557319, "nll": 0.17492864391876922, "ece": 0.055859921108077235, "score_mae": null, "within_1": null }, "invoice_processing/urgency": { "n": 100, "accuracy": 0.64, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.634350746408913, "ece": 0.08940732025394861, "score_mae": 0.22134730041506162, "within_1": 1.0 }, "security_incidents/credential_compromise": { "n": 100, "accuracy": 0.77, "soft_accuracy": 0.5728015621293402, "brier": 0.05996805088332501, "kl": 0.08639044321714845, "tv": 0.13661466424427768, "nll": 0.48320685975458416, "ece": 0.06143237512009606, "score_mae": null, "within_1": null }, "security_incidents/disposition": { "n": 100, "accuracy": 0.8, "soft_accuracy": 0.4968130180761153, "brier": 0.17852518241898221, "kl": 0.5857424779395584, "tv": 0.3213343701468733, "nll": 0.5296424665328927, "ece": 0.07959175670498408, "score_mae": null, "within_1": null }, "security_incidents/severity": { "n": 100, "accuracy": 0.71, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.8210523363985104, "ece": 0.14468211188621938, "score_mae": 0.38530313622199197, "within_1": 0.97 }, "security_incidents/true_positive": { "n": 100, "accuracy": 0.81, "soft_accuracy": 0.6290848775465093, "brier": 0.07949936892882356, "kl": 0.1477343297818701, "tv": 0.16513150898767584, "nll": 0.3719629556726569, "ece": 0.0533904871823985, "score_mae": null, "within_1": null }, "security_incidents/urgency": { "n": 100, "accuracy": 0.59, "soft_accuracy": null, "brier": null, "kl": null, "tv": null, "nll": 0.8371078708517858, "ece": 0.12910757991997132, "score_mae": 0.408927993573325, "within_1": 0.96 } } }, "laya_reference": { "accuracy": 0.766, "soft_accuracy": 0.471, "brier": 0.062, "ece": 0.213, "score_mae": 0.242 }, "gate": { "accuracy_gt_laya": false, "brier_lt_laya": true, "ece_raw_lt_laya_raw": true, "score_mae_lt_laya": false, "probability_quality_improved": true, "broad_claim_allowed": false, "rule": "accuracy > Laya AND (Brier < Laya OR raw ECE < Laya); score MAE is informational" }, "workflow_gate_accuracy_gt_laya": { "agent_trace_observability": true, "customer_service": false, "invoice_processing": true, "security_incidents": false }, "override_justification": "" }