Files
OSGKeyboard/ModelTraining/ClipboardSemantics/targeted-confirmation-holdout-report.json
T
2026-08-27 18:01:46 +08:00

1039 lines
30 KiB
JSON

{
"binaryMacro" : {
"accuracy" : 0.9238,
"f1" : 0.4622,
"precision" : 0.5938,
"recall" : 0.3854
},
"binaryMacroByLanguage" : {
"en" : {
"accuracy" : 0.9067,
"f1" : 0.4111,
"precision" : 0.5798,
"recall" : 0.3312
},
"zh-Hans" : {
"accuracy" : 0.9408,
"f1" : 0.5062,
"precision" : 0.6076,
"recall" : 0.4396
}
},
"classifiers" : [
{
"falseNegativeExamples" : [
{
"confidence" : 0.8671,
"family" : "confirmation_task_request",
"id" : "random-20260829-en-confirmation_task_request-030",
"language" : "en",
"text" : "In short, Would you own the final delivery of the budget notes?"
},
{
"confidence" : 0.8278,
"family" : "confirmation_task_request",
"id" : "random-20260829-en-confirmation_task_request-022",
"language" : "en",
"text" : "Just to add, Would you own the final delivery of the budget notes?"
},
{
"confidence" : 0.8234,
"family" : "confirmation_task_request",
"id" : "random-20260829-en-confirmation_task_request-009",
"language" : "en",
"text" : "By the way, Could I ask you to take care of the test findings before the next review?"
},
{
"confidence" : 0.8156,
"family" : "confirmation_task_request",
"id" : "random-20260829-en-confirmation_task_request-005",
"language" : "en",
"text" : "One more thing: Would you own the final delivery of the budget notes?"
},
{
"confidence" : 0.8034,
"family" : "confirmation_task_request",
"id" : "random-20260829-en-confirmation_task_request-003",
"language" : "en",
"text" : "In short, Could I ask you to take care of the budget notes before lunch?"
}
],
"falsePositiveExamples" : [
],
"id" : "task",
"metrics" : {
"accuracy" : 0.8483,
"f1" : 0.6617,
"falseNegative" : 91,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.4944,
"total" : 600,
"trueNegative" : 420,
"truePositive" : 89
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.79,
"f1" : 0.4615,
"falseNegative" : 63,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.3,
"total" : 300,
"trueNegative" : 210,
"truePositive" : 27
},
"zh-Hans" : {
"accuracy" : 0.9067,
"f1" : 0.8158,
"falseNegative" : 28,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6889,
"total" : 300,
"trueNegative" : 210,
"truePositive" : 62
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9017,
"f1" : 0.8207,
"falseNegative" : 45,
"falsePositive" : 14,
"precision" : 0.906,
"recall" : 0.75,
"total" : 600,
"trueNegative" : 406,
"truePositive" : 135
},
"threshold" : 0.24
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.8967,
"f1" : 0.8,
"falseNegative" : 56,
"falsePositive" : 6,
"precision" : 0.9538,
"recall" : 0.6889,
"total" : 600,
"trueNegative" : 414,
"truePositive" : 124
},
"threshold" : 0.41
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.8767,
"f1" : 0.7448,
"falseNegative" : 36,
"falsePositive" : 1,
"precision" : 0.9818,
"recall" : 0.6,
"total" : 300,
"trueNegative" : 209,
"truePositive" : 54
},
"threshold" : 0.42
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.95,
"f1" : 0.9162,
"falseNegative" : 8,
"falsePositive" : 7,
"precision" : 0.9213,
"recall" : 0.9111,
"total" : 300,
"trueNegative" : 203,
"truePositive" : 82
},
"threshold" : 0.14
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.8767,
"f1" : 0.7448,
"falseNegative" : 36,
"falsePositive" : 1,
"precision" : 0.9818,
"recall" : 0.6,
"total" : 300,
"trueNegative" : 209,
"truePositive" : 54
},
"threshold" : 0.42
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9167,
"f1" : 0.8408,
"falseNegative" : 24,
"falsePositive" : 1,
"precision" : 0.9851,
"recall" : 0.7333,
"total" : 300,
"trueNegative" : 209,
"truePositive" : 66
},
"threshold" : 0.5
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.5019,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-en-confirmation_complaint_request-022",
"language" : "en",
"text" : "One more thing: Because the account is locked again, my family cannot receive the item. Who is responsible for fixing it?"
},
{
"confidence" : 0.4542,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-zh-Hans-confirmation_complaint_request-016",
"language" : "zh-Hans",
"text" : "补充一下,账号又被锁住已经持续很久,究竟什么时候恢复?"
},
{
"confidence" : 0.4201,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-zh-Hans-confirmation_complaint_request-004",
"language" : "zh-Hans",
"text" : "另外,账号又被锁住已经持续很久,究竟什么时候恢复?"
},
{
"confidence" : 0.4188,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-zh-Hans-confirmation_complaint_request-006",
"language" : "zh-Hans",
"text" : "单独记一下,页面一直空白已经持续很久,究竟什么时候恢复?"
},
{
"confidence" : 0.3567,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-zh-Hans-confirmation_complaint_request-009",
"language" : "zh-Hans",
"text" : "顺便提一句,账号又被锁住已经持续很久,究竟什么时候恢复?"
}
],
"falsePositiveExamples" : [
],
"id" : "question",
"metrics" : {
"accuracy" : 0.95,
"f1" : 0.8571,
"falseNegative" : 30,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.75,
"total" : 600,
"trueNegative" : 480,
"truePositive" : 90
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9367,
"f1" : 0.8119,
"falseNegative" : 19,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6833,
"total" : 300,
"trueNegative" : 240,
"truePositive" : 41
},
"zh-Hans" : {
"accuracy" : 0.9633,
"f1" : 0.8991,
"falseNegative" : 11,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.8167,
"total" : 300,
"trueNegative" : 240,
"truePositive" : 49
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9567,
"f1" : 0.8839,
"falseNegative" : 21,
"falsePositive" : 5,
"precision" : 0.9519,
"recall" : 0.825,
"total" : 600,
"trueNegative" : 475,
"truePositive" : 99
},
"threshold" : 0.23
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9567,
"f1" : 0.8839,
"falseNegative" : 21,
"falsePositive" : 5,
"precision" : 0.9519,
"recall" : 0.825,
"total" : 600,
"trueNegative" : 475,
"truePositive" : 99
},
"threshold" : 0.23
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.95,
"f1" : 0.8598,
"falseNegative" : 14,
"falsePositive" : 1,
"precision" : 0.9787,
"recall" : 0.7667,
"total" : 300,
"trueNegative" : 239,
"truePositive" : 46
},
"threshold" : 0.05
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.97,
"f1" : 0.9256,
"falseNegative" : 4,
"falsePositive" : 5,
"precision" : 0.918,
"recall" : 0.9333,
"total" : 300,
"trueNegative" : 235,
"truePositive" : 56
},
"threshold" : 0.23
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.95,
"f1" : 0.8598,
"falseNegative" : 14,
"falsePositive" : 1,
"precision" : 0.9787,
"recall" : 0.7667,
"total" : 300,
"trueNegative" : 239,
"truePositive" : 46
},
"threshold" : 0.05
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9767,
"f1" : 0.9381,
"falseNegative" : 7,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.8833,
"total" : 300,
"trueNegative" : 240,
"truePositive" : 53
},
"threshold" : 0.33
}
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
],
"id" : "invitation",
"metrics" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 600,
"trueNegative" : 600,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 300,
"trueNegative" : 300,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 300,
"trueNegative" : 300,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.8208,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-en-confirmation_complaint_request-006",
"language" : "en",
"text" : "In short, This has continued for far too long: the page stays blank. When will it be restored?"
},
{
"confidence" : 0.8157,
"family" : "confirmation_complaint_short",
"id" : "random-20260829-zh-Hans-confirmation_complaint_short-027",
"language" : "zh-Hans",
"text" : "还有一件事,付款结果重复扣款又来了,不能一直这样。"
},
{
"confidence" : 0.8077,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-en-confirmation_complaint_request-019",
"language" : "en",
"text" : "Also, This has continued for far too long: tracking has not moved for five days. When will it be restored?"
},
{
"confidence" : 0.8046,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-en-confirmation_complaint_request-001",
"language" : "en",
"text" : "Just to add, This has continued for far too long: tracking has not moved for five days. When will it be restored?"
},
{
"confidence" : 0.7988,
"family" : "confirmation_complaint_short",
"id" : "random-20260829-en-confirmation_complaint_short-009",
"language" : "en",
"text" : "The main point is this: It is still the case that tracking has not moved for five days, with no improvement."
}
],
"falsePositiveExamples" : [
],
"id" : "complaint",
"metrics" : {
"accuracy" : 0.8817,
"f1" : 0.7543,
"falseNegative" : 71,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6056,
"total" : 600,
"trueNegative" : 420,
"truePositive" : 109
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.8567,
"f1" : 0.6861,
"falseNegative" : 43,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.5222,
"total" : 300,
"trueNegative" : 210,
"truePositive" : 47
},
"zh-Hans" : {
"accuracy" : 0.9067,
"f1" : 0.8158,
"falseNegative" : 28,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6889,
"total" : 300,
"trueNegative" : 210,
"truePositive" : 62
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.905,
"f1" : 0.8267,
"falseNegative" : 44,
"falsePositive" : 13,
"precision" : 0.9128,
"recall" : 0.7556,
"total" : 600,
"trueNegative" : 407,
"truePositive" : 136
},
"threshold" : 0.65
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.8783,
"f1" : 0.7525,
"falseNegative" : 69,
"falsePositive" : 4,
"precision" : 0.9652,
"recall" : 0.6167,
"total" : 600,
"trueNegative" : 416,
"truePositive" : 111
},
"threshold" : 0.81
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.85,
"f1" : 0.6809,
"falseNegative" : 42,
"falsePositive" : 3,
"precision" : 0.9412,
"recall" : 0.5333,
"total" : 300,
"trueNegative" : 207,
"truePositive" : 48
},
"threshold" : 0.82
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9267,
"f1" : 0.8736,
"falseNegative" : 14,
"falsePositive" : 8,
"precision" : 0.9048,
"recall" : 0.8444,
"total" : 300,
"trueNegative" : 202,
"truePositive" : 76
},
"threshold" : 0.5
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.8567,
"f1" : 0.6861,
"falseNegative" : 43,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.5222,
"total" : 300,
"trueNegative" : 210,
"truePositive" : 47
},
"threshold" : 0.84
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9167,
"f1" : 0.8428,
"falseNegative" : 23,
"falsePositive" : 2,
"precision" : 0.971,
"recall" : 0.7444,
"total" : 300,
"trueNegative" : 208,
"truePositive" : 67
},
"threshold" : 0.69
}
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
],
"id" : "scheduleNegotiation",
"metrics" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 600,
"trueNegative" : 600,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 300,
"trueNegative" : 300,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 300,
"trueNegative" : 300,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
],
"id" : "confirmationDecision",
"metrics" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 600,
"trueNegative" : 600,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 300,
"trueNegative" : 300,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 300,
"trueNegative" : 300,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7287,
"family" : "confirmation_personal_boundary",
"id" : "random-20260829-zh-Hans-confirmation_personal_boundary-030",
"language" : "zh-Hans",
"text" : "补充一下,仅记录个人计划,月底前联系医生。"
},
{
"confidence" : 0.7002,
"family" : "confirmation_personal_boundary",
"id" : "random-20260829-zh-Hans-confirmation_personal_boundary-002",
"language" : "zh-Hans",
"text" : "刚想到,仅记录个人计划,两个工作日内联系医生。"
},
{
"confidence" : 0.6908,
"family" : "confirmation_personal_boundary",
"id" : "random-20260829-zh-Hans-confirmation_personal_boundary-021",
"language" : "zh-Hans",
"text" : "顺便提一句,仅记录个人计划,本周收尾时联系房东。"
},
{
"confidence" : 0.6898,
"family" : "confirmation_personal_boundary",
"id" : "random-20260829-zh-Hans-confirmation_personal_boundary-027",
"language" : "zh-Hans",
"text" : "先说重点,仅记录个人计划,下次评审前联系法务。"
},
{
"confidence" : 0.6822,
"family" : "confirmation_personal_boundary",
"id" : "random-20260829-zh-Hans-confirmation_personal_boundary-020",
"language" : "zh-Hans",
"text" : "单独记一下,我自己的行动备忘:客户回信后查看到账情况。"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.8021,
"family" : "confirmation_task_assignment",
"id" : "random-20260829-en-confirmation_task_assignment-007",
"language" : "en",
"text" : "For the record, The delivery checklist is yours to handle next; update the status on the project card when complete."
},
{
"confidence" : 0.7503,
"family" : "confirmation_operational_task",
"id" : "random-20260829-en-confirmation_operational_task-019",
"language" : "en",
"text" : "For the record, Close the item after sending the delivery checklist to the vendor; do not schedule a follow-up."
},
{
"confidence" : 0.738,
"family" : "confirmation_resolved_boundary",
"id" : "random-20260829-en-confirmation_resolved_boundary-017",
"language" : "en",
"text" : "For the record, The earlier state where the account is locked again is resolved, and Design confirmed closure."
}
],
"id" : "followUpReminder",
"metrics" : {
"accuracy" : 0.9433,
"f1" : 0.6304,
"falseNegative" : 31,
"falsePositive" : 3,
"precision" : 0.9063,
"recall" : 0.4833,
"total" : 600,
"trueNegative" : 537,
"truePositive" : 29
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9333,
"f1" : 0.5652,
"falseNegative" : 17,
"falsePositive" : 3,
"precision" : 0.8125,
"recall" : 0.4333,
"total" : 300,
"trueNegative" : 267,
"truePositive" : 13
},
"zh-Hans" : {
"accuracy" : 0.9533,
"f1" : 0.6957,
"falseNegative" : 14,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.5333,
"total" : 300,
"trueNegative" : 270,
"truePositive" : 16
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.945,
"f1" : 0.6452,
"falseNegative" : 30,
"falsePositive" : 3,
"precision" : 0.9091,
"recall" : 0.5,
"total" : 600,
"trueNegative" : 537,
"truePositive" : 30
},
"threshold" : 0.72
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9433,
"f1" : 0.6136,
"falseNegative" : 33,
"falsePositive" : 1,
"precision" : 0.9643,
"recall" : 0.45,
"total" : 600,
"trueNegative" : 539,
"truePositive" : 27
},
"threshold" : 0.76
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.94,
"f1" : 0.5909,
"falseNegative" : 17,
"falsePositive" : 1,
"precision" : 0.9286,
"recall" : 0.4333,
"total" : 300,
"trueNegative" : 269,
"truePositive" : 13
},
"threshold" : 0.76
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9833,
"f1" : 0.918,
"falseNegative" : 2,
"falsePositive" : 3,
"precision" : 0.9032,
"recall" : 0.9333,
"total" : 300,
"trueNegative" : 267,
"truePositive" : 28
},
"threshold" : 0.37
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.94,
"f1" : 0.5714,
"falseNegative" : 18,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.4,
"total" : 300,
"trueNegative" : 270,
"truePositive" : 12
},
"threshold" : 0.81
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.98,
"f1" : 0.8889,
"falseNegative" : 6,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.8,
"total" : 300,
"trueNegative" : 270,
"truePositive" : 24
},
"threshold" : 0.6
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7764,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-zh-Hans-confirmation_complaint_request-008",
"language" : "zh-Hans",
"text" : "还有一件事,附件始终打不开已经持续很久,究竟什么时候恢复?"
},
{
"confidence" : 0.7763,
"family" : "confirmation_complaint_implicit",
"id" : "random-20260829-zh-Hans-confirmation_complaint_implicit-009",
"language" : "zh-Hans",
"text" : "对了,处理承诺没有兑现,结果仍然是账号又被锁住。"
},
{
"confidence" : 0.7495,
"family" : "confirmation_complaint_request",
"id" : "random-20260829-en-confirmation_complaint_request-003",
"language" : "en",
"text" : "Also, This has continued for far too long: the page stays blank. When will it be restored?"
},
{
"confidence" : 0.7451,
"family" : "confirmation_complaint_short",
"id" : "random-20260829-en-confirmation_complaint_short-016",
"language" : "en",
"text" : "The main point is this: The problem is back again: the account is locked again. This cannot keep happening."
},
{
"confidence" : 0.7219,
"family" : "confirmation_complaint_short",
"id" : "random-20260829-en-confirmation_complaint_short-022",
"language" : "en",
"text" : "In short, The problem is back again: the account is locked again. This cannot keep happening."
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9992,
"family" : "confirmation_resolved_boundary",
"id" : "random-20260829-en-confirmation_resolved_boundary-019",
"language" : "en",
"text" : "The main point is this: The earlier state where the attachment never opens is resolved, and the vendor confirmed closure."
},
{
"confidence" : 0.9991,
"family" : "confirmation_resolved_boundary",
"id" : "random-20260829-en-confirmation_resolved_boundary-018",
"language" : "en",
"text" : "For context, The earlier state where the account is locked again is resolved, and Finance confirmed closure."
},
{
"confidence" : 0.9989,
"family" : "confirmation_resolved_boundary",
"id" : "random-20260829-en-confirmation_resolved_boundary-028",
"language" : "en",
"text" : "Just to add, The earlier state where the payment was charged twice is resolved, and the vendor confirmed closure."
},
{
"confidence" : 0.9987,
"family" : "confirmation_resolved_boundary",
"id" : "random-20260829-en-confirmation_resolved_boundary-026",
"language" : "en",
"text" : "Also, The earlier state where the account is locked again is resolved, and the vendor confirmed closure."
},
{
"confidence" : 0.9971,
"family" : "confirmation_resolved_boundary",
"id" : "random-20260829-en-confirmation_resolved_boundary-012",
"language" : "en",
"text" : "For the record, The earlier state where the payment was charged twice is resolved, and Finance confirmed closure."
}
],
"id" : "replyableMessage",
"metrics" : {
"accuracy" : 0.7667,
"f1" : 0.7941,
"falseNegative" : 90,
"falsePositive" : 50,
"precision" : 0.8438,
"recall" : 0.75,
"total" : 600,
"trueNegative" : 190,
"truePositive" : 270
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.7367,
"f1" : 0.7642,
"falseNegative" : 52,
"falsePositive" : 27,
"precision" : 0.8258,
"recall" : 0.7111,
"total" : 300,
"trueNegative" : 93,
"truePositive" : 128
},
"zh-Hans" : {
"accuracy" : 0.7967,
"f1" : 0.8232,
"falseNegative" : 38,
"falsePositive" : 23,
"precision" : 0.8606,
"recall" : 0.7889,
"total" : 300,
"trueNegative" : 97,
"truePositive" : 142
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.6933,
"f1" : 0.6806,
"falseNegative" : 164,
"falsePositive" : 20,
"precision" : 0.9074,
"recall" : 0.5444,
"total" : 600,
"trueNegative" : 220,
"truePositive" : 196
},
"threshold" : 0.97
},
"thresholdsAt90PrecisionByLanguage" : {
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.7933,
"f1" : 0.8075,
"falseNegative" : 50,
"falsePositive" : 12,
"precision" : 0.9155,
"recall" : 0.7222,
"total" : 300,
"trueNegative" : 108,
"truePositive" : 130
},
"threshold" : 0.93
}
},
"thresholdsAt95PrecisionByLanguage" : {
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.7833,
"f1" : 0.7883,
"falseNegative" : 59,
"falsePositive" : 6,
"precision" : 0.9528,
"recall" : 0.6722,
"total" : 300,
"trueNegative" : 114,
"truePositive" : 121
},
"threshold" : 0.96
}
}
}
],
"corpusRecordCount" : 600,
"exactTrainingOverlapCount" : 0,
"familyCount" : 10,
"generatedAt" : "2026-08-27T01:15:16Z",
"languageCounts" : {
"en" : 300,
"zh-Hans" : 300
},
"manifestSchemaVersion" : 2,
"seed" : 20260829,
"sentiment" : {
"accuracy" : 0.6567,
"correct" : 394,
"macroF1" : 0.4772,
"perLabelF1" : {
"negative" : 0.577,
"neutral" : 0.8546,
"positive" : 0
},
"total" : 600,
"unknown" : 162,
"unknownRate" : 0.27
}
}