Files
OSGKeyboard/ModelTraining/ClipboardSemantics/baselines/2026-08-26-nine-model-v1/random-holdout-report.json
T
2026-08-27 18:01:46 +08:00

1448 lines
42 KiB
JSON

{
"binaryMacro" : {
"accuracy" : 0.9418,
"f1" : 0.6374,
"precision" : 0.9492,
"recall" : 0.5091
},
"binaryMacroByLanguage" : {
"en" : {
"accuracy" : 0.9344,
"f1" : 0.5749,
"precision" : 0.775,
"recall" : 0.4882
},
"zh-Hans" : {
"accuracy" : 0.9492,
"f1" : 0.66,
"precision" : 0.9929,
"recall" : 0.53
}
},
"classifiers" : [
{
"falseNegativeExamples" : [
{
"confidence" : 0.9858,
"family" : "task_next_action",
"id" : "random-20260826-en-task_next_action-016",
"language" : "en",
"text" : "In short, Your next action is to prepare the budget notes before the next review for the project lead."
},
{
"confidence" : 0.9791,
"family" : "task_next_action",
"id" : "random-20260826-zh-Hans-task_next_action-004",
"language" : "zh-Hans",
"text" : "单独记一下,下一步由你整理验收截图,请在两个工作日内交给客服团队。"
},
{
"confidence" : 0.9788,
"family" : "task_next_action",
"id" : "random-20260826-en-task_next_action-011",
"language" : "en",
"text" : "Just to add, Your next action is to prepare the test findings within two business days for Finance."
},
{
"confidence" : 0.9785,
"family" : "task_next_action",
"id" : "random-20260826-en-task_next_action-004",
"language" : "en",
"text" : "Your next action is to prepare the delivery checklist as this week closes for the project lead."
},
{
"confidence" : 0.9759,
"family" : "task_next_action",
"id" : "random-20260826-en-task_next_action-017",
"language" : "en",
"text" : "In short, Your next action is to prepare the budget notes as this week closes for the project lead."
}
],
"falsePositiveExamples" : [
],
"id" : "task",
"metrics" : {
"accuracy" : 0.9016,
"f1" : 0.3505,
"falseNegative" : 63,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.2125,
"total" : 640,
"trueNegative" : 560,
"truePositive" : 17
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.875,
"f1" : 0,
"falseNegative" : 40,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 0.9281,
"f1" : 0.5965,
"falseNegative" : 23,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.425,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 17
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9172,
"f1" : 0.531,
"falseNegative" : 50,
"falsePositive" : 3,
"precision" : 0.9091,
"recall" : 0.375,
"total" : 640,
"trueNegative" : 557,
"truePositive" : 30
},
"threshold" : 0.93
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9125,
"f1" : 0.4717,
"falseNegative" : 55,
"falsePositive" : 1,
"precision" : 0.9615,
"recall" : 0.3125,
"total" : 640,
"trueNegative" : 559,
"truePositive" : 25
},
"threshold" : 0.96
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.8875,
"f1" : 0.1818,
"falseNegative" : 36,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.1,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 4
},
"threshold" : 0.97
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9438,
"f1" : 0.7273,
"falseNegative" : 16,
"falsePositive" : 2,
"precision" : 0.9231,
"recall" : 0.6,
"total" : 320,
"trueNegative" : 278,
"truePositive" : 24
},
"threshold" : 0.69
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.8875,
"f1" : 0.1818,
"falseNegative" : 36,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.1,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 4
},
"threshold" : 0.97
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9375,
"f1" : 0.6667,
"falseNegative" : 20,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.5,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 20
},
"threshold" : 0.94
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.5878,
"family" : "fixed_invitation",
"id" : "random-20260826-en-fixed_invitation-001",
"language" : "en",
"text" : "Also, I saved you a spot to attend the celebration tomorrow at 10:30 at the cafe downstairs; can you make it?"
},
{
"confidence" : 0.5765,
"family" : "complaint_support",
"id" : "random-20260826-zh-Hans-complaint_support-006",
"language" : "zh-Hans",
"text" : "对了,我已经重试三次,还是付款结果重复扣款,导致交付时间被耽误。请问什么时候能处理?"
},
{
"confidence" : 0.5511,
"family" : "schedule_negotiation",
"id" : "random-20260826-en-schedule_negotiation-009",
"language" : "en",
"text" : "The original slot is blocked for me; could we swap to Tuesday at 4:30?"
},
{
"confidence" : 0.5472,
"family" : "complaint_support",
"id" : "random-20260826-zh-Hans-complaint_support-007",
"language" : "zh-Hans",
"text" : "刚想到,物流状态停了五天到现在没解决,客户已经在催,能给一个明确答复吗?"
},
{
"confidence" : 0.5365,
"family" : "schedule_negotiation",
"id" : "random-20260826-en-schedule_negotiation-001",
"language" : "en",
"text" : "In short, I cannot make tomorrow at 10:30. Would late Friday afternoon be workable instead?"
}
],
"falsePositiveExamples" : [
],
"id" : "question",
"metrics" : {
"accuracy" : 0.9141,
"f1" : 0.7925,
"falseNegative" : 55,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6563,
"total" : 640,
"trueNegative" : 480,
"truePositive" : 105
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.8563,
"f1" : 0.5965,
"falseNegative" : 46,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.425,
"total" : 320,
"trueNegative" : 240,
"truePositive" : 34
},
"zh-Hans" : {
"accuracy" : 0.9719,
"f1" : 0.9404,
"falseNegative" : 9,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.8875,
"total" : 320,
"trueNegative" : 240,
"truePositive" : 71
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9563,
"f1" : 0.9108,
"falseNegative" : 17,
"falsePositive" : 11,
"precision" : 0.9286,
"recall" : 0.8938,
"total" : 640,
"trueNegative" : 469,
"truePositive" : 143
},
"threshold" : 0.05
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9531,
"f1" : 0.9013,
"falseNegative" : 23,
"falsePositive" : 7,
"precision" : 0.9514,
"recall" : 0.8563,
"total" : 640,
"trueNegative" : 473,
"truePositive" : 137
},
"threshold" : 0.08
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9281,
"f1" : 0.8456,
"falseNegative" : 17,
"falsePositive" : 6,
"precision" : 0.913,
"recall" : 0.7875,
"total" : 320,
"trueNegative" : 234,
"truePositive" : 63
},
"threshold" : 0.05
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9875,
"f1" : 0.9756,
"falseNegative" : 0,
"falsePositive" : 4,
"precision" : 0.9524,
"recall" : 1,
"total" : 320,
"trueNegative" : 236,
"truePositive" : 80
},
"threshold" : 0.06
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9188,
"f1" : 0.8143,
"falseNegative" : 23,
"falsePositive" : 3,
"precision" : 0.95,
"recall" : 0.7125,
"total" : 320,
"trueNegative" : 237,
"truePositive" : 57
},
"threshold" : 0.08
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9875,
"f1" : 0.9756,
"falseNegative" : 0,
"falsePositive" : 4,
"precision" : 0.9524,
"recall" : 1,
"total" : 320,
"trueNegative" : 236,
"truePositive" : 80
},
"threshold" : 0.06
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7305,
"family" : "fixed_invitation",
"id" : "random-20260826-zh-Hans-fixed_invitation-003",
"language" : "zh-Hans",
"text" : "我给你留了位置,周六傍晚到楼下咖啡店碰面聊方案,能来不?"
},
{
"confidence" : 0.6961,
"family" : "fixed_invitation",
"id" : "random-20260826-zh-Hans-fixed_invitation-017",
"language" : "zh-Hans",
"text" : "刚想到,我给你留了位置,明天十点半到客户办公室碰面聊方案,能来不?"
},
{
"confidence" : 0.6813,
"family" : "fixed_invitation",
"id" : "random-20260826-zh-Hans-fixed_invitation-015",
"language" : "zh-Hans",
"text" : "补充一下,我给你留了位置,明天十点半到三号会议室参加小范围评审,能来不?"
},
{
"confidence" : 0.5808,
"family" : "fixed_invitation",
"id" : "random-20260826-zh-Hans-fixed_invitation-001",
"language" : "zh-Hans",
"text" : "简单说,我给你留了位置,今晚九点到园区北门参加庆功会,能来不?"
},
{
"confidence" : 0.5026,
"family" : "fixed_invitation",
"id" : "random-20260826-zh-Hans-fixed_invitation-019",
"language" : "zh-Hans",
"text" : "另外,我给你留了位置,周六傍晚到楼下咖啡店吃顿便饭,能来不?"
}
],
"falsePositiveExamples" : [
],
"id" : "invitation",
"metrics" : {
"accuracy" : 0.9688,
"f1" : 0.6667,
"falseNegative" : 20,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.5,
"total" : 640,
"trueNegative" : 600,
"truePositive" : 20
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9625,
"f1" : 0.5714,
"falseNegative" : 12,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.4,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 8
},
"zh-Hans" : {
"accuracy" : 0.975,
"f1" : 0.75,
"falseNegative" : 8,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 12
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9797,
"f1" : 0.8219,
"falseNegative" : 10,
"falsePositive" : 3,
"precision" : 0.9091,
"recall" : 0.75,
"total" : 640,
"trueNegative" : 597,
"truePositive" : 30
},
"threshold" : 0.36
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.7576,
"falseNegative" : 15,
"falsePositive" : 1,
"precision" : 0.9615,
"recall" : 0.625,
"total" : 640,
"trueNegative" : 599,
"truePositive" : 25
},
"threshold" : 0.46
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9938,
"f1" : 0.95,
"falseNegative" : 1,
"falsePositive" : 1,
"precision" : 0.95,
"recall" : 0.95,
"total" : 320,
"trueNegative" : 299,
"truePositive" : 19
},
"threshold" : 0.12
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9875,
"f1" : 0.8947,
"falseNegative" : 3,
"falsePositive" : 1,
"precision" : 0.9444,
"recall" : 0.85,
"total" : 320,
"trueNegative" : 299,
"truePositive" : 17
},
"threshold" : 0.46
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9938,
"f1" : 0.95,
"falseNegative" : 1,
"falsePositive" : 1,
"precision" : 0.95,
"recall" : 0.95,
"total" : 320,
"trueNegative" : 299,
"truePositive" : 19
},
"threshold" : 0.12
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9844,
"f1" : 0.8571,
"falseNegative" : 5,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.75,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 15
},
"threshold" : 0.66
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7866,
"family" : "complaint_support",
"id" : "random-20260826-zh-Hans-complaint_support-005",
"language" : "zh-Hans",
"text" : "先说重点,我已经重试三次,还是账号又被锁住,导致我无法继续操作。请问什么时候能处理?"
},
{
"confidence" : 0.7672,
"family" : "complaint_support",
"id" : "random-20260826-zh-Hans-complaint_support-012",
"language" : "zh-Hans",
"text" : "对了,我已经重试三次,还是付款结果重复扣款,导致工作完全卡住。请问什么时候能处理?"
},
{
"confidence" : 0.7505,
"family" : "complaint_support",
"id" : "random-20260826-en-complaint_support-009",
"language" : "en",
"text" : "Also, the account is locked again is still unresolved and my family cannot receive the item. Can I get a clear answer?"
},
{
"confidence" : 0.749,
"family" : "complaint_support",
"id" : "random-20260826-zh-Hans-complaint_support-019",
"language" : "zh-Hans",
"text" : "先说重点,付款结果重复扣款到现在没解决,工作完全卡住,能给一个明确答复吗?"
},
{
"confidence" : 0.6596,
"family" : "complaint_support",
"id" : "random-20260826-zh-Hans-complaint_support-006",
"language" : "zh-Hans",
"text" : "对了,我已经重试三次,还是付款结果重复扣款,导致交付时间被耽误。请问什么时候能处理?"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.8486,
"family" : "information_question",
"id" : "random-20260826-en-information_question-016",
"language" : "en",
"text" : "Also, Where does support coverage stand right now?"
},
{
"confidence" : 0.8486,
"family" : "information_question",
"id" : "random-20260826-en-information_question-007",
"language" : "en",
"text" : "In short, Where does support coverage stand right now?"
},
{
"confidence" : 0.8486,
"family" : "information_question",
"id" : "random-20260826-en-information_question-001",
"language" : "en",
"text" : "For context, Where does support coverage stand right now?"
}
],
"id" : "complaint",
"metrics" : {
"accuracy" : 0.9469,
"f1" : 0.3462,
"falseNegative" : 31,
"falsePositive" : 3,
"precision" : 0.75,
"recall" : 0.225,
"total" : 640,
"trueNegative" : 597,
"truePositive" : 9
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9344,
"f1" : 0.16,
"falseNegative" : 18,
"falsePositive" : 3,
"precision" : 0.4,
"recall" : 0.1,
"total" : 320,
"trueNegative" : 297,
"truePositive" : 2
},
"zh-Hans" : {
"accuracy" : 0.9594,
"f1" : 0.5185,
"falseNegative" : 13,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.35,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 7
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9516,
"f1" : 0.3673,
"falseNegative" : 31,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.225,
"total" : 640,
"trueNegative" : 600,
"truePositive" : 9
},
"threshold" : 0.85
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9516,
"f1" : 0.3673,
"falseNegative" : 31,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.225,
"total" : 640,
"trueNegative" : 600,
"truePositive" : 9
},
"threshold" : 0.85
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9438,
"f1" : 0.1818,
"falseNegative" : 18,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.1,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 2
},
"threshold" : 0.85
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9813,
"f1" : 0.8235,
"falseNegative" : 6,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.7,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 14
},
"threshold" : 0.35
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9438,
"f1" : 0.1818,
"falseNegative" : 18,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.1,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 2
},
"threshold" : 0.85
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9813,
"f1" : 0.8235,
"falseNegative" : 6,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.7,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 14
},
"threshold" : 0.35
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.4934,
"family" : "schedule_negotiation",
"id" : "random-20260826-en-schedule_negotiation-014",
"language" : "en",
"text" : "By the way, The original slot is blocked for me; could we swap to late Friday afternoon?"
},
{
"confidence" : 0.4267,
"family" : "schedule_negotiation",
"id" : "random-20260826-zh-Hans-schedule_negotiation-005",
"language" : "zh-Hans",
"text" : "简单说,原来的时段卡住了,我只能周二四点半,能不能对调?"
},
{
"confidence" : 0.3878,
"family" : "schedule_negotiation",
"id" : "random-20260826-en-schedule_negotiation-002",
"language" : "en",
"text" : "The main point is this: The original slot is blocked for me; could we swap to next Monday afternoon?"
},
{
"confidence" : 0.3729,
"family" : "schedule_negotiation",
"id" : "random-20260826-en-schedule_negotiation-010",
"language" : "en",
"text" : "One more thing: I cannot make after Monday's stand-up. Would first thing tomorrow be workable instead?"
},
{
"confidence" : 0.3562,
"family" : "schedule_negotiation",
"id" : "random-20260826-zh-Hans-schedule_negotiation-014",
"language" : "zh-Hans",
"text" : "单独记一下,周六傍晚赶不过去,换成周四午休后你觉得可行吗?"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.6431,
"family" : "conversation",
"id" : "random-20260826-en-conversation-020",
"language" : "en",
"text" : "Just to add, Just letting you know, the demo went better than expected; we can talk later."
}
],
"id" : "scheduleNegotiation",
"metrics" : {
"accuracy" : 0.9703,
"f1" : 0.6984,
"falseNegative" : 18,
"falsePositive" : 1,
"precision" : 0.9565,
"recall" : 0.55,
"total" : 640,
"trueNegative" : 599,
"truePositive" : 22
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9813,
"f1" : 0.8333,
"falseNegative" : 5,
"falsePositive" : 1,
"precision" : 0.9375,
"recall" : 0.75,
"total" : 320,
"trueNegative" : 299,
"truePositive" : 15
},
"zh-Hans" : {
"accuracy" : 0.9594,
"f1" : 0.5185,
"falseNegative" : 13,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.35,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 7
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9719,
"f1" : 0.7188,
"falseNegative" : 17,
"falsePositive" : 1,
"precision" : 0.9583,
"recall" : 0.575,
"total" : 640,
"trueNegative" : 599,
"truePositive" : 23
},
"threshold" : 0.46
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9719,
"f1" : 0.7188,
"falseNegative" : 17,
"falsePositive" : 1,
"precision" : 0.9583,
"recall" : 0.575,
"total" : 640,
"trueNegative" : 599,
"truePositive" : 23
},
"threshold" : 0.46
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9844,
"f1" : 0.8649,
"falseNegative" : 4,
"falsePositive" : 1,
"precision" : 0.9412,
"recall" : 0.8,
"total" : 320,
"trueNegative" : 299,
"truePositive" : 16
},
"threshold" : 0.46
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.7647,
"falseNegative" : 7,
"falsePositive" : 1,
"precision" : 0.9286,
"recall" : 0.65,
"total" : 320,
"trueNegative" : 299,
"truePositive" : 13
},
"threshold" : 0.22
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.75,
"falseNegative" : 8,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 12
},
"threshold" : 0.65
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.75,
"falseNegative" : 8,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 12
},
"threshold" : 0.28
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7506,
"family" : "confirmation_decision",
"id" : "random-20260826-zh-Hans-confirmation_decision-005",
"language" : "zh-Hans",
"text" : "刚想到,不用再比较了,就选季度结算,后续都按这个口径走。"
},
{
"confidence" : 0.7328,
"family" : "confirmation_decision",
"id" : "random-20260826-en-confirmation_decision-011",
"language" : "en",
"text" : "The main point is this: I formally approve the second quote; start the next steps as this week closes."
},
{
"confidence" : 0.7205,
"family" : "confirmation_decision",
"id" : "random-20260826-zh-Hans-confirmation_decision-015",
"language" : "zh-Hans",
"text" : "对了,不用再比较了,就选季度结算,后续都按这个口径走。"
},
{
"confidence" : 0.7075,
"family" : "confirmation_decision",
"id" : "random-20260826-en-confirmation_decision-017",
"language" : "en",
"text" : "By the way, No more comparisons: choose the lightweight version and use it as the final direction."
},
{
"confidence" : 0.6937,
"family" : "confirmation_decision",
"id" : "random-20260826-zh-Hans-confirmation_decision-009",
"language" : "zh-Hans",
"text" : "单独记一下,不用再比较了,就选第二套报价,后续都按这个口径走。"
}
],
"falsePositiveExamples" : [
],
"id" : "confirmationDecision",
"metrics" : {
"accuracy" : 0.9656,
"f1" : 0.6207,
"falseNegative" : 22,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.45,
"total" : 640,
"trueNegative" : 600,
"truePositive" : 18
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9656,
"f1" : 0.6207,
"falseNegative" : 11,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.45,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 9
},
"zh-Hans" : {
"accuracy" : 0.9656,
"f1" : 0.6207,
"falseNegative" : 11,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.45,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 9
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9844,
"f1" : 0.8649,
"falseNegative" : 8,
"falsePositive" : 2,
"precision" : 0.9412,
"recall" : 0.8,
"total" : 640,
"trueNegative" : 598,
"truePositive" : 32
},
"threshold" : 0.6
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9828,
"f1" : 0.8451,
"falseNegative" : 10,
"falsePositive" : 1,
"precision" : 0.9677,
"recall" : 0.75,
"total" : 640,
"trueNegative" : 599,
"truePositive" : 30
},
"threshold" : 0.63
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9875,
"f1" : 0.9,
"falseNegative" : 2,
"falsePositive" : 2,
"precision" : 0.9,
"recall" : 0.9,
"total" : 320,
"trueNegative" : 298,
"truePositive" : 18
},
"threshold" : 0.6
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 20
},
"threshold" : 0.35
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9656,
"f1" : 0.6207,
"falseNegative" : 11,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.45,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 9
},
"threshold" : 0.76
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 320,
"trueNegative" : 300,
"truePositive" : 20
},
"threshold" : 0.35
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7912,
"family" : "task_next_action",
"id" : "random-20260826-en-task_next_action-012",
"language" : "en",
"text" : "By the way, Your next action is to prepare the signature page by tomorrow evening for the project lead."
},
{
"confidence" : 0.7875,
"family" : "follow_up_reminder",
"id" : "random-20260826-en-follow_up_reminder-017",
"language" : "en",
"text" : "Also, Reminder: contact the account manager before lunch and follow up on the final decision."
},
{
"confidence" : 0.745,
"family" : "follow_up_reminder",
"id" : "random-20260826-en-follow_up_reminder-015",
"language" : "en",
"text" : "A quick note: Once the checkup ends, check with the landlord the next day about the delivery date."
},
{
"confidence" : 0.7131,
"family" : "follow_up_reminder",
"id" : "random-20260826-en-follow_up_reminder-001",
"language" : "en",
"text" : "By the way, Once the contract is signed, check with the landlord the next day about the delivery date."
},
{
"confidence" : 0.7044,
"family" : "task_next_action",
"id" : "random-20260826-zh-Hans-task_next_action-008",
"language" : "zh-Hans",
"text" : "顺便提一句,下一步由你整理预算备注,请在午休前交给供应商。"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.8775,
"family" : "information_question",
"id" : "random-20260826-en-information_question-005",
"language" : "en",
"text" : "One more thing: Quick check: who owns the final call on the release?"
},
{
"confidence" : 0.8369,
"family" : "information_question",
"id" : "random-20260826-en-information_question-006",
"language" : "en",
"text" : "For context, Quick check: who owns the final call on the release?"
}
],
"id" : "followUpReminder",
"metrics" : {
"accuracy" : 0.9391,
"f1" : 0.688,
"falseNegative" : 37,
"falsePositive" : 2,
"precision" : 0.9556,
"recall" : 0.5375,
"total" : 640,
"trueNegative" : 558,
"truePositive" : 43
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9719,
"f1" : 0.88,
"falseNegative" : 7,
"falsePositive" : 2,
"precision" : 0.9429,
"recall" : 0.825,
"total" : 320,
"trueNegative" : 278,
"truePositive" : 33
},
"zh-Hans" : {
"accuracy" : 0.9063,
"f1" : 0.4,
"falseNegative" : 30,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.25,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 10
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9484,
"f1" : 0.7591,
"falseNegative" : 28,
"falsePositive" : 5,
"precision" : 0.9123,
"recall" : 0.65,
"total" : 640,
"trueNegative" : 555,
"truePositive" : 52
},
"threshold" : 0.61
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9422,
"f1" : 0.7087,
"falseNegative" : 35,
"falsePositive" : 2,
"precision" : 0.9574,
"recall" : 0.5625,
"total" : 640,
"trueNegative" : 558,
"truePositive" : 45
},
"threshold" : 0.78
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9844,
"f1" : 0.9383,
"falseNegative" : 2,
"falsePositive" : 3,
"precision" : 0.9268,
"recall" : 0.95,
"total" : 320,
"trueNegative" : 277,
"truePositive" : 38
},
"threshold" : 0.69
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9313,
"f1" : 0.6452,
"falseNegative" : 20,
"falsePositive" : 2,
"precision" : 0.9091,
"recall" : 0.5,
"total" : 320,
"trueNegative" : 278,
"truePositive" : 20
},
"threshold" : 0.4
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9688,
"f1" : 0.8611,
"falseNegative" : 9,
"falsePositive" : 1,
"precision" : 0.9688,
"recall" : 0.775,
"total" : 320,
"trueNegative" : 279,
"truePositive" : 31
},
"threshold" : 0.84
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9344,
"f1" : 0.6441,
"falseNegative" : 21,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.475,
"total" : 320,
"trueNegative" : 280,
"truePositive" : 19
},
"threshold" : 0.44
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7591,
"family" : "confirmation_decision",
"id" : "random-20260826-zh-Hans-confirmation_decision-003",
"language" : "zh-Hans",
"text" : "我正式确认先灰度发布,请午休前启动后续安排。"
},
{
"confidence" : 0.747,
"family" : "task_request",
"id" : "random-20260826-zh-Hans-task_request-009",
"language" : "zh-Hans",
"text" : "简单说,运营还缺交付清单,麻烦午休前补齐后在群里留言。"
},
{
"confidence" : 0.742,
"family" : "confirmation_decision",
"id" : "random-20260826-zh-Hans-confirmation_decision-014",
"language" : "zh-Hans",
"text" : "先说重点,不用再比较了,就选线下处理方案,后续都按这个口径走。"
},
{
"confidence" : 0.7383,
"family" : "task_request",
"id" : "random-20260826-en-task_request-015",
"language" : "en",
"text" : "For the record, the vendor still needs the test findings; please add it within two business days and leave a note in the group thread."
},
{
"confidence" : 0.7364,
"family" : "conversation",
"id" : "random-20260826-zh-Hans-conversation-019",
"language" : "zh-Hans",
"text" : "刚看到你发的照片,突然想到你可能会想知道。"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9896,
"family" : "follow_up_reminder",
"id" : "random-20260826-zh-Hans-follow_up_reminder-015",
"language" : "zh-Hans",
"text" : "另外,等复诊结束后第二天再找招聘方确认处理进度,别漏了。"
},
{
"confidence" : 0.9867,
"family" : "follow_up_reminder",
"id" : "random-20260826-zh-Hans-follow_up_reminder-010",
"language" : "zh-Hans",
"text" : "单独记一下,等合同盖章后第二天再找仓库负责人确认下一步安排,别漏了。"
},
{
"confidence" : 0.9845,
"family" : "negative_news",
"id" : "random-20260826-en-negative_news-008",
"language" : "en",
"text" : "By the way, Data shows that many flights were delayed overnight, with no request for support."
},
{
"confidence" : 0.9831,
"family" : "follow_up_reminder",
"id" : "random-20260826-zh-Hans-follow_up_reminder-016",
"language" : "zh-Hans",
"text" : "先说重点,等复诊结束后第二天再找仓库负责人确认下一步安排,别漏了。"
},
{
"confidence" : 0.9806,
"family" : "follow_up_reminder",
"id" : "random-20260826-zh-Hans-follow_up_reminder-005",
"language" : "zh-Hans",
"text" : "还有一件事,等样品送达后第二天再找法务确认交付日期,别漏了。"
}
],
"id" : "replyableMessage",
"metrics" : {
"accuracy" : 0.9281,
"f1" : 0.9365,
"falseNegative" : 21,
"falsePositive" : 25,
"precision" : 0.9313,
"recall" : 0.9417,
"total" : 640,
"trueNegative" : 255,
"truePositive" : 339
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9281,
"f1" : 0.9373,
"falseNegative" : 8,
"falsePositive" : 15,
"precision" : 0.9198,
"recall" : 0.9556,
"total" : 320,
"trueNegative" : 125,
"truePositive" : 172
},
"zh-Hans" : {
"accuracy" : 0.9281,
"f1" : 0.9356,
"falseNegative" : 13,
"falsePositive" : 10,
"precision" : 0.9435,
"recall" : 0.9278,
"total" : 320,
"trueNegative" : 130,
"truePositive" : 167
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9281,
"f1" : 0.9387,
"falseNegative" : 8,
"falsePositive" : 38,
"precision" : 0.9026,
"recall" : 0.9778,
"total" : 640,
"trueNegative" : 242,
"truePositive" : 352
},
"threshold" : 0.61
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9203,
"f1" : 0.9275,
"falseNegative" : 34,
"falsePositive" : 17,
"precision" : 0.9504,
"recall" : 0.9056,
"total" : 640,
"trueNegative" : 263,
"truePositive" : 326
},
"threshold" : 0.84
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9344,
"f1" : 0.944,
"falseNegative" : 3,
"falsePositive" : 18,
"precision" : 0.9077,
"recall" : 0.9833,
"total" : 320,
"trueNegative" : 122,
"truePositive" : 177
},
"threshold" : 0.66
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9281,
"f1" : 0.9383,
"falseNegative" : 5,
"falsePositive" : 18,
"precision" : 0.9067,
"recall" : 0.9722,
"total" : 320,
"trueNegative" : 122,
"truePositive" : 175
},
"threshold" : 0.57
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9,
"f1" : 0.907,
"falseNegative" : 24,
"falsePositive" : 8,
"precision" : 0.9512,
"recall" : 0.8667,
"total" : 320,
"trueNegative" : 132,
"truePositive" : 156
},
"threshold" : 0.87
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9281,
"f1" : 0.9348,
"falseNegative" : 15,
"falsePositive" : 8,
"precision" : 0.9538,
"recall" : 0.9167,
"total" : 320,
"trueNegative" : 132,
"truePositive" : 165
},
"threshold" : 0.84
}
}
}
],
"corpusRecordCount" : 640,
"exactTrainingOverlapCount" : 0,
"familyCount" : 16,
"generatedAt" : "2026-08-26T15:16:55Z",
"languageCounts" : {
"en" : 320,
"zh-Hans" : 320
},
"manifestSchemaVersion" : 2,
"seed" : 20260826,
"sentiment" : {
"accuracy" : 0.7859,
"correct" : 503,
"macroF1" : 0.6228,
"perLabelF1" : {
"negative" : 0.36,
"neutral" : 0.9222,
"positive" : 0.5862
},
"total" : 640,
"unknown" : 107,
"unknownRate" : 0.1672
}
}