sainath's picture
Upload 17 files
92be313 verified
Raw History Blame Contribute Delete
91.6 kB
{
"scope": "Exploratory existing splits; test has previously informed development",
"calibration": {
"global": 1.096478196143185,
"by_question": {
"150440942f8fd3e04c6ffcb60c0215ffb1265e7d27e071f3ac453fd9817a85d4": 1.513561248436208,
"2a95b662ebe5bb2aa612f55ab9490ec200c83aec6e0ce1e9c439d769cb39a93c": 0.1,
"330b97e5d96bf0d453555d62465367c41f1c5eb57ca53a1f3e8015c8ebe8548a": 1.513561248436208,
"59786fb640773138e83576a42b1baff87c3a12c6924070c867a17f7541f8fd50": 0.9120108393559097
},
"minimum_examples": 50,
"unknown_question_policy": "global temperature; calibration not established"
},
"splits": {
"validation": {
"uncalibrated": {
"overall": {
"count": 800,
"accuracy": 0.93,
"macro_f1": 0.942767064108638,
"nll": 0.14363739399664024,
"brier": 0.09649431923038009,
"ece_10_bins": 0.014156258404254909,
"oracle_expected_nll": 0.13552431792992842,
"oracle_distribution_mse": 0.009192354615839761,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 15,
"confidence": 0.5693979422251384,
"accuracy": 0.4666666666666667
},
{
"lower": 0.6,
"upper": 0.7,
"count": 28,
"confidence": 0.6569398194551468,
"accuracy": 0.6428571428571429
},
{
"lower": 0.7,
"upper": 0.8,
"count": 61,
"confidence": 0.7547330357989327,
"accuracy": 0.6721311475409836
},
{
"lower": 0.8,
"upper": 0.9,
"count": 93,
"confidence": 0.8439700603485107,
"accuracy": 0.8064516129032258
},
{
"lower": 0.9,
"upper": 1.0,
"count": 603,
"confidence": 0.9985708261208353,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 200,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 400,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 600,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 800,
"error_rate": 0.06999999999999995
}
]
},
"by_task": {
"department": {
"count": 200,
"accuracy": 0.99,
"macro_f1": 0.9901098901098901,
"nll": 0.010185061729229394,
"brier": 0.006637792461137025,
"ece_10_bins": 0.0073371312022209225,
"oracle_expected_nll": 0.010185061729229394,
"oracle_distribution_mse": 0.006637792461137025,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 1,
"confidence": 0.5138780474662781,
"accuracy": 0.0
},
{
"lower": 0.6,
"upper": 0.7,
"count": 1,
"confidence": 0.6265591382980347,
"accuracy": 0.0
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 198,
"confidence": 0.9983485401278794,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.010000000000000009
}
]
},
"escalation": {
"count": 200,
"accuracy": 0.73,
"macro_f1": 0.7286704853783539,
"nll": 0.5641286429153144,
"brier": 0.37933894420616143,
"ece_10_bins": 0.05137460917234424,
"oracle_expected_nll": 0.531676338648467,
"oracle_distribution_mse": 0.030131085748000032,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 14,
"confidence": 0.5733636489936283,
"accuracy": 0.5
},
{
"lower": 0.6,
"upper": 0.7,
"count": 27,
"confidence": 0.6580650298683731,
"accuracy": 0.6666666666666666
},
{
"lower": 0.7,
"upper": 0.8,
"count": 61,
"confidence": 0.7547330357989327,
"accuracy": 0.6721311475409836
},
{
"lower": 0.8,
"upper": 0.9,
"count": 93,
"confidence": 0.8439700603485107,
"accuracy": 0.8064516129032258
},
{
"lower": 0.9,
"upper": 1.0,
"count": 5,
"confidence": 0.9024688482284546,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.12
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.18999999999999995
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.22666666666666668
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.27
}
]
},
"refund": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 9.878679824863307e-05,
"brier": 2.084608453162e-08,
"ece_10_bins": 9.878158569331941e-05,
"oracle_expected_nll": 9.878679824863307e-05,
"oracle_distribution_mse": 2.084608453162e-08,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9999012184143067,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"severity": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.0001370845437686307,
"brier": 5.194081374567017e-07,
"ece_10_bins": 0.00013695359230037063,
"oracle_expected_nll": 0.0001370845437686307,
"oracle_distribution_mse": 5.194081374567017e-07,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9998630464076996,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
}
}
},
"global": {
"overall": {
"count": 800,
"accuracy": 0.93,
"macro_f1": 0.942767064108638,
"nll": 0.14297785222380682,
"brier": 0.09584865920230996,
"ece_10_bins": 0.009678431376814825,
"oracle_expected_nll": 0.13557863835585515,
"oracle_distribution_mse": 0.009061730485262987,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 16,
"confidence": 0.565223541110754,
"accuracy": 0.5
},
{
"lower": 0.6,
"upper": 0.7,
"count": 34,
"confidence": 0.6557925483759712,
"accuracy": 0.5882352941176471
},
{
"lower": 0.7,
"upper": 0.8,
"count": 76,
"confidence": 0.7558186516950005,
"accuracy": 0.7105263157894737
},
{
"lower": 0.8,
"upper": 0.9,
"count": 76,
"confidence": 0.837152016006018,
"accuracy": 0.8421052631578947
},
{
"lower": 0.9,
"upper": 1.0,
"count": 598,
"confidence": 0.9990241513603109,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 200,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 400,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 600,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 800,
"error_rate": 0.06999999999999995
}
]
},
"by_task": {
"department": {
"count": 200,
"accuracy": 0.99,
"macro_f1": 0.9901098901098901,
"nll": 0.010840845209029908,
"brier": 0.006549160572433809,
"ece_10_bins": 0.008071925044059772,
"oracle_expected_nll": 0.010840845209029908,
"oracle_distribution_mse": 0.006549160572433809,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 1,
"confidence": 0.5126363039016724,
"accuracy": 0.0
},
{
"lower": 0.6,
"upper": 0.7,
"count": 1,
"confidence": 0.615834653377533,
"accuracy": 0.0
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 198,
"confidence": 0.9975458886286225,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.010000000000000009
}
]
},
"escalation": {
"count": 200,
"accuracy": 0.73,
"macro_f1": 0.7286704853783539,
"nll": 0.5605819633748517,
"brier": 0.3768439641098432,
"ece_10_bins": 0.03015358328819274,
"oracle_expected_nll": 0.530985107903045,
"oracle_distribution_mse": 0.02969624924165533,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 15,
"confidence": 0.5687293569246928,
"accuracy": 0.5333333333333333
},
{
"lower": 0.6,
"upper": 0.7,
"count": 33,
"confidence": 0.6570033936789541,
"accuracy": 0.6060606060606061
},
{
"lower": 0.7,
"upper": 0.8,
"count": 76,
"confidence": 0.7558186516950005,
"accuracy": 0.7105263157894737
},
{
"lower": 0.8,
"upper": 0.9,
"count": 76,
"confidence": 0.837152016006018,
"accuracy": 0.8421052631578947
},
{
"lower": 0.9,
"upper": 1.0,
"count": 0,
"confidence": null,
"accuracy": null
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.12
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.18999999999999995
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.22666666666666668
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.27
}
]
},
"refund": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.00022176956251617814,
"brier": 1.0389551787255491e-07,
"ece_10_bins": 0.00022174358367921254,
"oracle_expected_nll": 0.00022176956251617814,
"oracle_distribution_mse": 1.0389551787255491e-07,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9997782564163208,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"severity": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.00026683074882947707,
"brier": 1.4082314449344625e-06,
"ece_10_bins": 0.0002664735913276983,
"oracle_expected_nll": 0.00026683074882947707,
"oracle_distribution_mse": 1.4082314449344625e-06,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9997335264086723,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
}
}
},
"per_question": {
"overall": {
"count": 800,
"accuracy": 0.93,
"macro_f1": 0.942767064108638,
"nll": 0.14733114755095042,
"brier": 0.09750962712499976,
"ece_10_bins": 0.017871906235814118,
"oracle_expected_nll": 0.13843533652518125,
"oracle_distribution_mse": 0.009677819048759641,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 11,
"confidence": 0.5629869699478149,
"accuracy": 0.36363636363636365
},
{
"lower": 0.6,
"upper": 0.7,
"count": 27,
"confidence": 0.6522127919726901,
"accuracy": 0.6666666666666666
},
{
"lower": 0.7,
"upper": 0.8,
"count": 48,
"confidence": 0.7543277218937874,
"accuracy": 0.6666666666666666
},
{
"lower": 0.8,
"upper": 0.9,
"count": 105,
"confidence": 0.8517835855484008,
"accuracy": 0.7904761904761904
},
{
"lower": 0.9,
"upper": 1.0,
"count": 609,
"confidence": 0.994959923532013,
"accuracy": 0.9967159277504105
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 200,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 400,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 600,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 800,
"error_rate": 0.06999999999999995
}
]
},
"by_task": {
"department": {
"count": 200,
"accuracy": 0.99,
"macro_f1": 0.9901098901098901,
"nll": 0.016769024587334803,
"brier": 0.006763791666141161,
"ece_10_bins": 0.01407569766044616,
"oracle_expected_nll": 0.016769024587334803,
"oracle_distribution_mse": 0.006763791666141161,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 2,
"confidence": 0.5465996563434601,
"accuracy": 0.0
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 3,
"confidence": 0.8894220391909281,
"accuracy": 1.0
},
{
"lower": 0.9,
"upper": 1.0,
"count": 195,
"confidence": 0.992870736733461,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.010000000000000009
}
]
},
"escalation": {
"count": 200,
"accuracy": 0.73,
"macro_f1": 0.7286704853783539,
"nll": 0.5704197966243141,
"brier": 0.3832444677877699,
"ece_10_bins": 0.06606497406959536,
"oracle_expected_nll": 0.5348365525212375,
"oracle_distribution_mse": 0.03191723548280944,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 9,
"confidence": 0.5666285951932272,
"accuracy": 0.4444444444444444
},
{
"lower": 0.6,
"upper": 0.7,
"count": 27,
"confidence": 0.6522127919726901,
"accuracy": 0.6666666666666666
},
{
"lower": 0.7,
"upper": 0.8,
"count": 48,
"confidence": 0.7543277218937874,
"accuracy": 0.6666666666666666
},
{
"lower": 0.8,
"upper": 0.9,
"count": 102,
"confidence": 0.8506765722059736,
"accuracy": 0.7843137254901961
},
{
"lower": 0.9,
"upper": 1.0,
"count": 14,
"confidence": 0.9104529874665397,
"accuracy": 0.8571428571428571
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.12
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.18999999999999995
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.22666666666666668
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.27
}
]
},
"refund": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.0,
"brier": 5.221318590436242e-77,
"ece_10_bins": 0.0,
"oracle_expected_nll": 0.0,
"oracle_distribution_mse": 5.221318590436242e-77,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 1.0,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"severity": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.0021357689921527212,
"brier": 3.0249046087972453e-05,
"ece_10_bins": 0.002127710282802564,
"oracle_expected_nll": 0.0021357689921527212,
"oracle_distribution_mse": 3.0249046087972453e-05,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9978722897171974,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
}
}
}
},
"test": {
"uncalibrated": {
"overall": {
"count": 800,
"accuracy": 0.9325,
"macro_f1": 0.945994599459946,
"nll": 0.14603000355243925,
"brier": 0.0982134248610902,
"ece_10_bins": 0.01680564425885677,
"oracle_expected_nll": 0.12875951419169415,
"oracle_distribution_mse": 0.005807715000243819,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 11,
"confidence": 0.5358001806519248,
"accuracy": 0.6363636363636364
},
{
"lower": 0.6,
"upper": 0.7,
"count": 18,
"confidence": 0.668232325050566,
"accuracy": 0.6666666666666666
},
{
"lower": 0.7,
"upper": 0.8,
"count": 67,
"confidence": 0.7674324290076299,
"accuracy": 0.7313432835820896
},
{
"lower": 0.8,
"upper": 0.9,
"count": 104,
"confidence": 0.843749422293443,
"accuracy": 0.75
},
{
"lower": 0.9,
"upper": 1.0,
"count": 600,
"confidence": 0.9997629618644714,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 200,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 400,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 600,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 800,
"error_rate": 0.0675
}
]
},
"by_task": {
"department": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.0003887213664454956,
"brier": 6.573280611448498e-06,
"ece_10_bins": 0.00038706064224247605,
"oracle_expected_nll": 0.0003887213664454956,
"oracle_distribution_mse": 6.573280611448498e-06,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9996129393577575,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"escalation": {
"count": 200,
"accuracy": 0.73,
"macro_f1": 0.72997299729973,
"nll": 0.5834056262796022,
"brier": 0.3928407776022817,
"ece_10_bins": 0.06651146262884142,
"oracle_expected_nll": 0.514323668836622,
"oracle_distribution_mse": 0.023217938158896164,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 11,
"confidence": 0.5358001806519248,
"accuracy": 0.6363636363636364
},
{
"lower": 0.6,
"upper": 0.7,
"count": 18,
"confidence": 0.668232325050566,
"accuracy": 0.6666666666666666
},
{
"lower": 0.7,
"upper": 0.8,
"count": 67,
"confidence": 0.7674324290076299,
"accuracy": 0.7313432835820896
},
{
"lower": 0.8,
"upper": 0.9,
"count": 104,
"confidence": 0.843749422293443,
"accuracy": 0.75
},
{
"lower": 0.9,
"upper": 1.0,
"count": 0,
"confidence": null,
"accuracy": null
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.19999999999999996
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.25
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.24
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.27
}
]
},
"refund": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.00010169231916523476,
"brier": 2.3599058819392956e-08,
"ece_10_bins": 0.00010168641805652356,
"oracle_expected_nll": 0.00010169231916523476,
"oracle_distribution_mse": 2.3599058819392956e-08,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9998983135819435,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"severity": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.00022397424454392842,
"brier": 6.324962408847836e-06,
"ece_10_bins": 0.00022236734628677812,
"oracle_expected_nll": 0.00022397424454392842,
"oracle_distribution_mse": 6.324962408847836e-06,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9997776326537132,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
}
}
},
"global": {
"overall": {
"count": 800,
"accuracy": 0.9325,
"macro_f1": 0.945994599459946,
"nll": 0.14461302598233813,
"brier": 0.09727346687970145,
"ece_10_bins": 0.010928971841931354,
"oracle_expected_nll": 0.12886215335419976,
"oracle_distribution_mse": 0.005768674131682376,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 12,
"confidence": 0.5381810814142227,
"accuracy": 0.5833333333333334
},
{
"lower": 0.6,
"upper": 0.7,
"count": 20,
"confidence": 0.6634815037250519,
"accuracy": 0.65
},
{
"lower": 0.7,
"upper": 0.8,
"count": 88,
"confidence": 0.7614993337880481,
"accuracy": 0.7386363636363636
},
{
"lower": 0.8,
"upper": 0.9,
"count": 80,
"confidence": 0.833437080681324,
"accuracy": 0.7625
},
{
"lower": 0.9,
"upper": 1.0,
"count": 600,
"confidence": 0.9995919790863991,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 200,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 400,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 600,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 800,
"error_rate": 0.0675
}
]
},
"by_task": {
"department": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.0006354354848862528,
"brier": 1.4171445392714965e-05,
"ece_10_bins": 0.000631839334964801,
"oracle_expected_nll": 0.0006354354848862528,
"oracle_distribution_mse": 1.4171445392714965e-05,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9993681606650352,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"escalation": {
"count": 200,
"accuracy": 0.73,
"macro_f1": 0.72997299729973,
"nll": 0.5772212989828499,
"brier": 0.389067384510727,
"ece_10_bins": 0.042491824626922636,
"oracle_expected_nll": 0.5142178084702962,
"oracle_distribution_mse": 0.023048213518650672,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 12,
"confidence": 0.5381810814142227,
"accuracy": 0.5833333333333334
},
{
"lower": 0.6,
"upper": 0.7,
"count": 20,
"confidence": 0.6634815037250519,
"accuracy": 0.65
},
{
"lower": 0.7,
"upper": 0.8,
"count": 88,
"confidence": 0.7614993337880481,
"accuracy": 0.7386363636363636
},
{
"lower": 0.8,
"upper": 0.9,
"count": 80,
"confidence": 0.833437080681324,
"accuracy": 0.7625
},
{
"lower": 0.9,
"upper": 1.0,
"count": 0,
"confidence": null,
"accuracy": null
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.19999999999999996
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.25
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.24
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.27
}
]
},
"refund": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.00022736255682158317,
"brier": 1.1458914958364816e-07,
"ece_10_bins": 0.00022733390331264314,
"oracle_expected_nll": 0.00022736255682158317,
"oracle_distribution_mse": 1.1458914958364816e-07,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9997726660966874,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"severity": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.00036800690479488446,
"brier": 1.2196973536534902e-05,
"ece_10_bins": 0.00036488950252533403,
"oracle_expected_nll": 0.00036800690479488446,
"oracle_distribution_mse": 1.2196973536534902e-05,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9996351104974747,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
}
}
},
"per_question": {
"overall": {
"count": 800,
"accuracy": 0.9325,
"macro_f1": 0.945994599459946,
"nll": 0.14949719129215913,
"brier": 0.0995688090168633,
"ece_10_bins": 0.02478441372513776,
"oracle_expected_nll": 0.1305604755579714,
"oracle_distribution_mse": 0.006266329786064511,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 11,
"confidence": 0.5392229827967557,
"accuracy": 0.6363636363636364
},
{
"lower": 0.6,
"upper": 0.7,
"count": 11,
"confidence": 0.666460779580203,
"accuracy": 0.5454545454545454
},
{
"lower": 0.7,
"upper": 0.8,
"count": 45,
"confidence": 0.7596085866292318,
"accuracy": 0.7333333333333333
},
{
"lower": 0.8,
"upper": 0.9,
"count": 123,
"confidence": 0.8477989275281023,
"accuracy": 0.7317073170731707
},
{
"lower": 0.9,
"upper": 1.0,
"count": 610,
"confidence": 0.9967766217520979,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 200,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 400,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 600,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 800,
"error_rate": 0.0675
}
]
},
"by_task": {
"department": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.003119851317236034,
"brier": 0.000139220646155405,
"ece_10_bins": 0.0030834728479385287,
"oracle_expected_nll": 0.003119851317236034,
"oracle_distribution_mse": 0.000139220646155405,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.9969165271520615,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"escalation": {
"count": 200,
"accuracy": 0.73,
"macro_f1": 0.72997299729973,
"nll": 0.592753167735883,
"brier": 0.39804941940743904,
"ece_10_bins": 0.09396138578653342,
"oracle_expected_nll": 0.5170063047991321,
"oracle_distribution_mse": 0.024839502484243964,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 11,
"confidence": 0.5392229827967557,
"accuracy": 0.6363636363636364
},
{
"lower": 0.6,
"upper": 0.7,
"count": 11,
"confidence": 0.666460779580203,
"accuracy": 0.5454545454545454
},
{
"lower": 0.7,
"upper": 0.8,
"count": 45,
"confidence": 0.7596085866292318,
"accuracy": 0.7333333333333333
},
{
"lower": 0.8,
"upper": 0.9,
"count": 123,
"confidence": 0.8477989275281023,
"accuracy": 0.7317073170731707
},
{
"lower": 0.9,
"upper": 1.0,
"count": 10,
"confidence": 0.9068993091583252,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.19999999999999996
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.25
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.24
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.27
}
]
},
"refund": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.0,
"brier": 1.5641410521010998e-69,
"ece_10_bins": 0.0,
"oracle_expected_nll": 0.0,
"oracle_distribution_mse": 1.5641410521010998e-69,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 1.0,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
},
"severity": {
"count": 200,
"accuracy": 1.0,
"macro_f1": 1.0,
"nll": 0.002115746115517467,
"brier": 8.65960138586772e-05,
"ece_10_bins": 0.0020927962660789445,
"oracle_expected_nll": 0.002115746115517467,
"oracle_distribution_mse": 8.65960138586772e-05,
"reliability": [
{
"lower": 0.0,
"upper": 0.1,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.1,
"upper": 0.2,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.2,
"upper": 0.3,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.3,
"upper": 0.4,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.4,
"upper": 0.5,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.5,
"upper": 0.6,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.6,
"upper": 0.7,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.7,
"upper": 0.8,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.8,
"upper": 0.9,
"count": 0,
"confidence": null,
"accuracy": null
},
{
"lower": 0.9,
"upper": 1.0,
"count": 200,
"confidence": 0.997907203733921,
"accuracy": 1.0
}
],
"risk_coverage": [
{
"coverage": 0.25,
"count": 50,
"error_rate": 0.0
},
{
"coverage": 0.5,
"count": 100,
"error_rate": 0.0
},
{
"coverage": 0.75,
"count": 150,
"error_rate": 0.0
},
{
"coverage": 1.0,
"count": 200,
"error_rate": 0.0
}
]
}
}
}
}
}
}