Files
OSGKeyboard/ModelTraining/ClipboardSemantics/baselines/2026-08-27-targeted-v2/fresh-metric-holdout-report.json
T
2026-08-27 18:01:46 +08:00

1271 lines
38 KiB
JSON

{
"binaryMacro" : {
"accuracy" : 0.9573,
"f1" : 0.6444,
"precision" : 0.7136,
"recall" : 0.6028
},
"binaryMacroByLanguage" : {
"en" : {
"accuracy" : 0.9495,
"f1" : 0.6159,
"precision" : 0.7337,
"recall" : 0.5438
},
"zh-Hans" : {
"accuracy" : 0.9651,
"f1" : 0.6694,
"precision" : 0.6994,
"recall" : 0.6618
}
},
"classifiers" : [
{
"falseNegativeExamples" : [
{
"confidence" : 0.8087,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-019",
"language" : "zh-Hans",
"text" : "还有一件事,客服团队把验收截图分给你处理,完成时间不能晚于下次评审前。"
},
{
"confidence" : 0.7999,
"family" : "fresh_task_question",
"id" : "random-20260827-en-fresh_task_question-001",
"language" : "en",
"text" : "By the way, Can you take this assignment and finish the retro summary before lunch?"
},
{
"confidence" : 0.7868,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-010",
"language" : "zh-Hans",
"text" : "单独记一下,财务把交付清单分给你处理,完成时间不能晚于本周收尾时。"
},
{
"confidence" : 0.7837,
"family" : "fresh_task_question",
"id" : "random-20260827-en-fresh_task_question-017",
"language" : "en",
"text" : "By the way, Can you close out the signature page as this week closes and update the project lead?"
},
{
"confidence" : 0.7728,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-zh-Hans-fresh_plain_task_boundary-016",
"language" : "zh-Hans",
"text" : "单独记一下,一次性完成复盘摘要即可,交给供应商后任务结束。"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9501,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-010",
"language" : "zh-Hans",
"text" : "刚想到,个人行动项:复诊结束以后再次检查下一步安排。"
},
{
"confidence" : 0.937,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-004",
"language" : "zh-Hans",
"text" : "刚想到,个人行动项:补丁上线以后再次检查下一步安排。"
},
{
"confidence" : 0.9365,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-en-fresh_personal_reminder-010",
"language" : "en",
"text" : "Also, Personal action item: check next steps again after the client replies."
},
{
"confidence" : 0.9335,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-en-fresh_personal_reminder-007",
"language" : "en",
"text" : "Just to add, Personal action item: check next steps again after the client replies."
},
{
"confidence" : 0.8787,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-en-fresh_personal_reminder-020",
"language" : "en",
"text" : "By the way, Personal action item: check next steps again after the sample arrives."
}
],
"id" : "task",
"metrics" : {
"accuracy" : 0.875,
"f1" : 0.781,
"falseNegative" : 53,
"falsePositive" : 7,
"precision" : 0.9386,
"recall" : 0.6688,
"total" : 480,
"trueNegative" : 313,
"truePositive" : 107
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.8292,
"f1" : 0.672,
"falseNegative" : 38,
"falsePositive" : 3,
"precision" : 0.9333,
"recall" : 0.525,
"total" : 240,
"trueNegative" : 157,
"truePositive" : 42
},
"zh-Hans" : {
"accuracy" : 0.9208,
"f1" : 0.8725,
"falseNegative" : 15,
"falsePositive" : 4,
"precision" : 0.942,
"recall" : 0.8125,
"total" : 240,
"trueNegative" : 156,
"truePositive" : 65
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9146,
"f1" : 0.8664,
"falseNegative" : 27,
"falsePositive" : 14,
"precision" : 0.9048,
"recall" : 0.8313,
"total" : 480,
"trueNegative" : 306,
"truePositive" : 133
},
"threshold" : 0.56
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.85,
"f1" : 0.7188,
"falseNegative" : 68,
"falsePositive" : 4,
"precision" : 0.9583,
"recall" : 0.575,
"total" : 480,
"trueNegative" : 316,
"truePositive" : 92
},
"threshold" : 0.88
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.8292,
"f1" : 0.672,
"falseNegative" : 38,
"falsePositive" : 3,
"precision" : 0.9333,
"recall" : 0.525,
"total" : 240,
"trueNegative" : 157,
"truePositive" : 42
},
"threshold" : 0.81
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9583,
"f1" : 0.939,
"falseNegative" : 3,
"falsePositive" : 7,
"precision" : 0.9167,
"recall" : 0.9625,
"total" : 240,
"trueNegative" : 153,
"truePositive" : 77
},
"threshold" : 0.5
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.7917,
"f1" : 0.5455,
"falseNegative" : 50,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.375,
"total" : 240,
"trueNegative" : 160,
"truePositive" : 30
},
"threshold" : 0.94
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9,
"f1" : 0.831,
"falseNegative" : 21,
"falsePositive" : 3,
"precision" : 0.9516,
"recall" : 0.7375,
"total" : 240,
"trueNegative" : 157,
"truePositive" : 59
},
"threshold" : 0.87
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.4372,
"family" : "fresh_task_question",
"id" : "random-20260827-en-fresh_task_question-020",
"language" : "en",
"text" : "The main point is this: Can you close out the retro summary before the next review and update the project lead?"
},
{
"confidence" : 0.3568,
"family" : "fresh_task_question",
"id" : "random-20260827-en-fresh_task_question-012",
"language" : "en",
"text" : "The main point is this: Can you close out the delivery checklist before lunch and update Finance?"
},
{
"confidence" : 0.2062,
"family" : "fresh_complaint_request",
"id" : "random-20260827-en-fresh_complaint_request-013",
"language" : "en",
"text" : "Also, tracking has not moved for five days has reached the point where all work is blocked. When will you actually resolve it?"
},
{
"confidence" : 0.0877,
"family" : "fresh_complaint_request",
"id" : "random-20260827-en-fresh_complaint_request-012",
"language" : "en",
"text" : "Also, tracking has not moved for five days has reached the point where delivery is now delayed. When will you actually resolve it?"
},
{
"confidence" : 0.0402,
"family" : "fresh_complaint_request",
"id" : "random-20260827-en-fresh_complaint_request-003",
"language" : "en",
"text" : "For context, the page stays blank has reached the point where my family cannot receive the item. When will you actually resolve it?"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9546,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-002",
"language" : "zh-Hans",
"text" : "补充一下,运营把复盘摘要分给你处理,完成时间不能晚于下次评审前。"
},
{
"confidence" : 0.8949,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-018",
"language" : "zh-Hans",
"text" : "顺便提一句,供应商把预算备注分给你处理,完成时间不能晚于下次评审前。"
},
{
"confidence" : 0.8927,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-017",
"language" : "zh-Hans",
"text" : "供应商把交付清单分给你处理,完成时间不能晚于月底前。"
},
{
"confidence" : 0.8816,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-019",
"language" : "zh-Hans",
"text" : "还有一件事,客服团队把验收截图分给你处理,完成时间不能晚于下次评审前。"
},
{
"confidence" : 0.8141,
"family" : "fresh_task_assignment",
"id" : "random-20260827-zh-Hans-fresh_task_assignment-011",
"language" : "zh-Hans",
"text" : "刚想到,设计同学把复盘摘要分给你处理,完成时间不能晚于午休前。"
}
],
"id" : "question",
"metrics" : {
"accuracy" : 0.9646,
"f1" : 0.8931,
"falseNegative" : 9,
"falsePositive" : 8,
"precision" : 0.8987,
"recall" : 0.8875,
"total" : 480,
"trueNegative" : 392,
"truePositive" : 71
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9625,
"f1" : 0.8732,
"falseNegative" : 9,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.775,
"total" : 240,
"trueNegative" : 200,
"truePositive" : 31
},
"zh-Hans" : {
"accuracy" : 0.9667,
"f1" : 0.9091,
"falseNegative" : 0,
"falsePositive" : 8,
"precision" : 0.8333,
"recall" : 1,
"total" : 240,
"trueNegative" : 192,
"truePositive" : 40
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9667,
"f1" : 0.8987,
"falseNegative" : 9,
"falsePositive" : 7,
"precision" : 0.9103,
"recall" : 0.8875,
"total" : 480,
"trueNegative" : 393,
"truePositive" : 71
},
"threshold" : 0.64
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9354,
"f1" : 0.7634,
"falseNegative" : 30,
"falsePositive" : 1,
"precision" : 0.9804,
"recall" : 0.625,
"total" : 480,
"trueNegative" : 399,
"truePositive" : 50
},
"threshold" : 0.9
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9667,
"f1" : 0.8974,
"falseNegative" : 5,
"falsePositive" : 3,
"precision" : 0.9211,
"recall" : 0.875,
"total" : 240,
"trueNegative" : 197,
"truePositive" : 35
},
"threshold" : 0.08
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9583,
"f1" : 0.8684,
"falseNegative" : 7,
"falsePositive" : 3,
"precision" : 0.9167,
"recall" : 0.825,
"total" : 240,
"trueNegative" : 197,
"truePositive" : 33
},
"threshold" : 0.89
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.9189,
"falseNegative" : 6,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.85,
"total" : 240,
"trueNegative" : 200,
"truePositive" : 34
},
"threshold" : 0.15
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9583,
"f1" : 0.8611,
"falseNegative" : 9,
"falsePositive" : 1,
"precision" : 0.9688,
"recall" : 0.775,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 31
},
"threshold" : 0.9
}
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
],
"id" : "invitation",
"metrics" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 480,
"trueNegative" : 480,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 240,
"trueNegative" : 240,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 240,
"trueNegative" : 240,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.8802,
"family" : "fresh_complaint_request",
"id" : "random-20260827-zh-Hans-fresh_complaint_request-005",
"language" : "zh-Hans",
"text" : "顺便提一句,页面一直空白已经影响到我无法继续操作,你们准备哪天真正解决?"
},
{
"confidence" : 0.8751,
"family" : "fresh_complaint",
"id" : "random-20260827-zh-Hans-fresh_complaint-013",
"language" : "zh-Hans",
"text" : "补充一下,本来承诺今天解决,结果还是物流状态停了五天,现在交付时间被耽误。"
},
{
"confidence" : 0.8677,
"family" : "fresh_complaint",
"id" : "random-20260827-zh-Hans-fresh_complaint-011",
"language" : "zh-Hans",
"text" : "对了,本来承诺今天解决,结果还是附件始终打不开,现在客户已经在催。"
},
{
"confidence" : 0.8422,
"family" : "fresh_complaint_request",
"id" : "random-20260827-zh-Hans-fresh_complaint_request-015",
"language" : "zh-Hans",
"text" : "顺便提一句,付款结果重复扣款已经影响到工作完全卡住,你们准备哪天真正解决?"
},
{
"confidence" : 0.8003,
"family" : "fresh_complaint",
"id" : "random-20260827-zh-Hans-fresh_complaint-006",
"language" : "zh-Hans",
"text" : "对了,同样的故障第三次出现,附件始终打不开,整个事情已经家人收不到商品。"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.7707,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-en-fresh_plain_task_boundary-006",
"language" : "en",
"text" : "Prepare the signature page now and send it to Design; no later follow-up is needed."
}
],
"id" : "complaint",
"metrics" : {
"accuracy" : 0.9104,
"f1" : 0.6387,
"falseNegative" : 42,
"falsePositive" : 1,
"precision" : 0.9744,
"recall" : 0.475,
"total" : 480,
"trueNegative" : 399,
"truePositive" : 38
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9,
"f1" : 0.5862,
"falseNegative" : 23,
"falsePositive" : 1,
"precision" : 0.9444,
"recall" : 0.425,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 17
},
"zh-Hans" : {
"accuracy" : 0.9208,
"f1" : 0.6885,
"falseNegative" : 19,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.525,
"total" : 240,
"trueNegative" : 200,
"truePositive" : 21
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9604,
"f1" : 0.8774,
"falseNegative" : 12,
"falsePositive" : 7,
"precision" : 0.9067,
"recall" : 0.85,
"total" : 480,
"trueNegative" : 393,
"truePositive" : 68
},
"threshold" : 0.48
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9354,
"f1" : 0.7669,
"falseNegative" : 29,
"falsePositive" : 2,
"precision" : 0.9623,
"recall" : 0.6375,
"total" : 480,
"trueNegative" : 398,
"truePositive" : 51
},
"threshold" : 0.67
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9167,
"f1" : 0.6875,
"falseNegative" : 18,
"falsePositive" : 2,
"precision" : 0.9167,
"recall" : 0.55,
"total" : 240,
"trueNegative" : 198,
"truePositive" : 22
},
"threshold" : 0.67
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9917,
"f1" : 0.975,
"falseNegative" : 1,
"falsePositive" : 1,
"precision" : 0.975,
"recall" : 0.975,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 39
},
"threshold" : 0.26
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9125,
"f1" : 0.6557,
"falseNegative" : 20,
"falsePositive" : 1,
"precision" : 0.9524,
"recall" : 0.5,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 20
},
"threshold" : 0.73
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9917,
"f1" : 0.975,
"falseNegative" : 1,
"falsePositive" : 1,
"precision" : 0.975,
"recall" : 0.975,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 39
},
"threshold" : 0.26
}
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
],
"id" : "scheduleNegotiation",
"metrics" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 480,
"trueNegative" : 480,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 240,
"trueNegative" : 240,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 240,
"trueNegative" : 240,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.6682,
"family" : "fresh_confirmation",
"id" : "random-20260827-en-fresh_confirmation-002",
"language" : "en",
"text" : "The main point is this: The decision is now official: use the second quote for document archiving."
},
{
"confidence" : 0.6681,
"family" : "fresh_confirmation",
"id" : "random-20260827-en-fresh_confirmation-004",
"language" : "en",
"text" : "By the way, The decision is now official: use quarterly billing for support coverage."
},
{
"confidence" : 0.6552,
"family" : "fresh_confirmation",
"id" : "random-20260827-en-fresh_confirmation-012",
"language" : "en",
"text" : "For the record, The review is over; my final selection is the offline resolution, and the alternatives are closed."
}
],
"falsePositiveExamples" : [
],
"id" : "confirmationDecision",
"metrics" : {
"accuracy" : 0.9938,
"f1" : 0.961,
"falseNegative" : 3,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.925,
"total" : 480,
"trueNegative" : 440,
"truePositive" : 37
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9875,
"f1" : 0.9189,
"falseNegative" : 3,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.85,
"total" : 240,
"trueNegative" : 220,
"truePositive" : 17
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 240,
"trueNegative" : 220,
"truePositive" : 20
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 480,
"trueNegative" : 440,
"truePositive" : 40
},
"threshold" : 0.26
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 480,
"trueNegative" : 440,
"truePositive" : 40
},
"threshold" : 0.26
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 240,
"trueNegative" : 220,
"truePositive" : 20
},
"threshold" : 0.17
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 240,
"trueNegative" : 220,
"truePositive" : 20
},
"threshold" : 0.26
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 240,
"trueNegative" : 220,
"truePositive" : 20
},
"threshold" : 0.17
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 1,
"f1" : 1,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 1,
"recall" : 1,
"total" : 240,
"trueNegative" : 220,
"truePositive" : 20
},
"threshold" : 0.26
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7283,
"family" : "fresh_follow_up",
"id" : "random-20260827-en-fresh_follow_up-015",
"language" : "en",
"text" : "Also, Two days after the checkup ends, check with the doctor again about the missing documents and record the outcome."
},
{
"confidence" : 0.7159,
"family" : "fresh_follow_up",
"id" : "random-20260827-en-fresh_follow_up-016",
"language" : "en",
"text" : "For context, After the current item closes, return to the landlord and track the missing documents."
},
{
"confidence" : 0.7159,
"family" : "fresh_follow_up",
"id" : "random-20260827-en-fresh_follow_up-006",
"language" : "en",
"text" : "In short, After the current item closes, return to the warehouse lead and track the missing documents."
},
{
"confidence" : 0.6981,
"family" : "fresh_follow_up",
"id" : "random-20260827-en-fresh_follow_up-008",
"language" : "en",
"text" : "Two days after the contract is signed, check with the recruiter again about resolution progress and record the outcome."
},
{
"confidence" : 0.6249,
"family" : "fresh_follow_up",
"id" : "random-20260827-en-fresh_follow_up-007",
"language" : "en",
"text" : "For context, Two days after the contract is signed, check with the doctor again about the missing documents and record the outcome."
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9002,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-zh-Hans-fresh_plain_task_boundary-016",
"language" : "zh-Hans",
"text" : "单独记一下,一次性完成复盘摘要即可,交给供应商后任务结束。"
},
{
"confidence" : 0.861,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-zh-Hans-fresh_plain_task_boundary-012",
"language" : "zh-Hans",
"text" : "一次性完成验收截图即可,交给设计同学后任务结束。"
},
{
"confidence" : 0.8199,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-zh-Hans-fresh_plain_task_boundary-007",
"language" : "zh-Hans",
"text" : "先说重点,一次性完成预算备注即可,交给供应商后任务结束。"
},
{
"confidence" : 0.8123,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-zh-Hans-fresh_plain_task_boundary-010",
"language" : "zh-Hans",
"text" : "补充一下,一次性完成预算备注即可,交给财务后任务结束。"
},
{
"confidence" : 0.75,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-zh-Hans-fresh_plain_task_boundary-011",
"language" : "zh-Hans",
"text" : "单独记一下,一次性完成测试结论即可,交给运营后任务结束。"
}
],
"id" : "followUpReminder",
"metrics" : {
"accuracy" : 0.9792,
"f1" : 0.9375,
"falseNegative" : 5,
"falsePositive" : 5,
"precision" : 0.9375,
"recall" : 0.9375,
"total" : 480,
"trueNegative" : 395,
"truePositive" : 75
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9792,
"f1" : 0.9333,
"falseNegative" : 5,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.875,
"total" : 240,
"trueNegative" : 200,
"truePositive" : 35
},
"zh-Hans" : {
"accuracy" : 0.9792,
"f1" : 0.9412,
"falseNegative" : 0,
"falsePositive" : 5,
"precision" : 0.8889,
"recall" : 1,
"total" : 240,
"trueNegative" : 195,
"truePositive" : 40
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9833,
"f1" : 0.9518,
"falseNegative" : 1,
"falsePositive" : 7,
"precision" : 0.9186,
"recall" : 0.9875,
"total" : 480,
"trueNegative" : 393,
"truePositive" : 79
},
"threshold" : 0.64
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9688,
"f1" : 0.8993,
"falseNegative" : 13,
"falsePositive" : 2,
"precision" : 0.971,
"recall" : 0.8375,
"total" : 480,
"trueNegative" : 398,
"truePositive" : 67
},
"threshold" : 0.82
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9875,
"f1" : 0.9639,
"falseNegative" : 0,
"falsePositive" : 3,
"precision" : 0.9302,
"recall" : 1,
"total" : 240,
"trueNegative" : 197,
"truePositive" : 40
},
"threshold" : 0.46
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9833,
"f1" : 0.9524,
"falseNegative" : 0,
"falsePositive" : 4,
"precision" : 0.9091,
"recall" : 1,
"total" : 240,
"trueNegative" : 196,
"truePositive" : 40
},
"threshold" : 0.76
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9917,
"f1" : 0.975,
"falseNegative" : 1,
"falsePositive" : 1,
"precision" : 0.975,
"recall" : 0.975,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 39
},
"threshold" : 0.64
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9917,
"f1" : 0.975,
"falseNegative" : 1,
"falsePositive" : 1,
"precision" : 0.975,
"recall" : 0.975,
"total" : 240,
"trueNegative" : 199,
"truePositive" : 39
},
"threshold" : 0.87
}
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7416,
"family" : "fresh_follow_up",
"id" : "random-20260827-zh-Hans-fresh_follow_up-020",
"language" : "zh-Hans",
"text" : "先说重点,等合同盖章满两天后,再找招聘方核实补充材料并记录结果。"
},
{
"confidence" : 0.7122,
"family" : "fresh_follow_up",
"id" : "random-20260827-en-fresh_follow_up-012",
"language" : "en",
"text" : "Just to add, Two days after the patch ships, check with the warehouse lead again about payment status and record the outcome."
},
{
"confidence" : 0.7031,
"family" : "fresh_confirmation",
"id" : "random-20260827-en-fresh_confirmation-020",
"language" : "en",
"text" : "Also, The decision is now official: use the second quote for document archiving."
},
{
"confidence" : 0.6779,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-en-fresh_plain_task_boundary-017",
"language" : "en",
"text" : "The main point is this: Prepare the signature page now and send it to Design; no later follow-up is needed."
},
{
"confidence" : 0.6164,
"family" : "fresh_plain_task_boundary",
"id" : "random-20260827-en-fresh_plain_task_boundary-019",
"language" : "en",
"text" : "For context, Prepare the delivery checklist now and send it to Finance; no later follow-up is needed."
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9994,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-010",
"language" : "zh-Hans",
"text" : "刚想到,个人行动项:复诊结束以后再次检查下一步安排。"
},
{
"confidence" : 0.9989,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-004",
"language" : "zh-Hans",
"text" : "刚想到,个人行动项:补丁上线以后再次检查下一步安排。"
},
{
"confidence" : 0.9775,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-019",
"language" : "zh-Hans",
"text" : "简单说,个人行动项:复诊结束以后再次检查处理进度。"
},
{
"confidence" : 0.9715,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-005",
"language" : "zh-Hans",
"text" : "顺便提一句,个人行动项:复诊结束以后再次检查处理进度。"
},
{
"confidence" : 0.9669,
"family" : "fresh_personal_reminder",
"id" : "random-20260827-zh-Hans-fresh_personal_reminder-007",
"language" : "zh-Hans",
"text" : "简单说,个人行动项:合同盖章以后再次检查处理进度。"
}
],
"id" : "replyableMessage",
"metrics" : {
"accuracy" : 0.9354,
"f1" : 0.9437,
"falseNegative" : 20,
"falsePositive" : 11,
"precision" : 0.9594,
"recall" : 0.9286,
"total" : 480,
"trueNegative" : 189,
"truePositive" : 260
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.9375,
"f1" : 0.9438,
"falseNegative" : 14,
"falsePositive" : 1,
"precision" : 0.9921,
"recall" : 0.9,
"total" : 240,
"trueNegative" : 99,
"truePositive" : 126
},
"zh-Hans" : {
"accuracy" : 0.9333,
"f1" : 0.9437,
"falseNegative" : 6,
"falsePositive" : 10,
"precision" : 0.9306,
"recall" : 0.9571,
"total" : 240,
"trueNegative" : 90,
"truePositive" : 134
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.9333,
"f1" : 0.9456,
"falseNegative" : 2,
"falsePositive" : 30,
"precision" : 0.9026,
"recall" : 0.9929,
"total" : 480,
"trueNegative" : 170,
"truePositive" : 278
},
"threshold" : 0.19
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.9521,
"f1" : 0.9591,
"falseNegative" : 10,
"falsePositive" : 13,
"precision" : 0.9541,
"recall" : 0.9643,
"total" : 480,
"trueNegative" : 187,
"truePositive" : 270
},
"threshold" : 0.53
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.95,
"f1" : 0.9583,
"falseNegative" : 2,
"falsePositive" : 10,
"precision" : 0.9324,
"recall" : 0.9857,
"total" : 240,
"trueNegative" : 90,
"truePositive" : 138
},
"threshold" : 0.19
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9375,
"f1" : 0.9492,
"falseNegative" : 0,
"falsePositive" : 15,
"precision" : 0.9032,
"recall" : 1,
"total" : 240,
"trueNegative" : 85,
"truePositive" : 140
},
"threshold" : 0.27
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.9583,
"f1" : 0.9643,
"falseNegative" : 5,
"falsePositive" : 5,
"precision" : 0.9643,
"recall" : 0.9643,
"total" : 240,
"trueNegative" : 95,
"truePositive" : 135
},
"threshold" : 0.37
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.9167,
"f1" : 0.9265,
"falseNegative" : 14,
"falsePositive" : 6,
"precision" : 0.9545,
"recall" : 0.9,
"total" : 240,
"trueNegative" : 94,
"truePositive" : 126
},
"threshold" : 0.92
}
}
}
],
"corpusRecordCount" : 480,
"exactTrainingOverlapCount" : 0,
"familyCount" : 12,
"generatedAt" : "2026-08-26T15:23:30Z",
"languageCounts" : {
"en" : 240,
"zh-Hans" : 240
},
"manifestSchemaVersion" : 2,
"seed" : 20260827,
"sentiment" : {
"accuracy" : 0.8563,
"correct" : 411,
"macroF1" : 0.5702,
"perLabelF1" : {
"negative" : 0.7614,
"neutral" : 0.9492,
"positive" : 0
},
"total" : 480,
"unknown" : 55,
"unknownRate" : 0.1146
}
}