Files
OSGKeyboard/ModelTraining/ClipboardSemantics/online-real-holdout-report.json
2026-08-27 18:01:46 +08:00

907 lines
26 KiB
JSON

{
"binaryMacro" : {
"accuracy" : 0.7919,
"f1" : 0.1672,
"precision" : 0.3784,
"recall" : 0.1256
},
"binaryMacroByLanguage" : {
"en" : {
"accuracy" : 0.7912,
"f1" : 0.1443,
"precision" : 0.2955,
"recall" : 0.1216
},
"zh-Hans" : {
"accuracy" : 0.7925,
"f1" : 0.1888,
"precision" : 0.3927,
"recall" : 0.1343
}
},
"classifiers" : [
{
"falseNegativeExamples" : [
{
"confidence" : 0.9596,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-8182",
"language" : "zh-Hans",
"text" : "你能删除牙医预约吗"
},
{
"confidence" : 0.7923,
"family" : "online_github_task",
"id" : "online-github-3328948690",
"language" : "en",
"text" : "Allow saving multi-dimensional ndarray with dynamic shapes"
},
{
"confidence" : 0.6821,
"family" : "online_github_task",
"id" : "online-github-1884324631",
"language" : "en",
"text" : "Support saving datasets with custom formatting"
},
{
"confidence" : 0.6717,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-10816",
"language" : "zh-Hans",
"text" : "创建一个播放列表"
},
{
"confidence" : 0.6123,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-10792",
"language" : "zh-Hans",
"text" : "我需要列一个清单"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.9771,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-2673",
"language" : "zh-Hans",
"text" : "我记得为下一次的医生预约设置提醒闹钟吗"
}
],
"id" : "task",
"metrics" : {
"accuracy" : 0.58,
"f1" : 0.1429,
"falseNegative" : 83,
"falsePositive" : 1,
"precision" : 0.875,
"recall" : 0.0778,
"total" : 200,
"trueNegative" : 109,
"truePositive" : 7
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.51,
"f1" : 0,
"falseNegative" : 49,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 51,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 0.65,
"f1" : 0.2857,
"falseNegative" : 34,
"falsePositive" : 1,
"precision" : 0.875,
"recall" : 0.1707,
"total" : 100,
"trueNegative" : 58,
"truePositive" : 7
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.5555,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-11744",
"language" : "zh-Hans",
"text" : "奥莉,这周末有什么新的剧情电影"
},
{
"confidence" : 0.5433,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-5974",
"language" : "zh-Hans",
"text" : "我设置中午吃药的闹钟了吗"
},
{
"confidence" : 0.5381,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-10757",
"language" : "zh-Hans",
"text" : "我的购物清单上有什么"
},
{
"confidence" : 0.5309,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-11773",
"language" : "zh-Hans",
"text" : "找下刘若英电影"
},
{
"confidence" : 0.5024,
"family" : "online_massive_query_boundary",
"id" : "online-massive-en-US-12797",
"language" : "en",
"text" : "train times from near me to location"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.993,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-8182",
"language" : "zh-Hans",
"text" : "你能删除牙医预约吗"
},
{
"confidence" : 0.9422,
"family" : "online_github_task",
"id" : "online-github-2139640381",
"language" : "en",
"text" : "Allow image model (color conversion) to be specified as part of datasets Image() decode"
},
{
"confidence" : 0.8983,
"family" : "online_github_task",
"id" : "online-github-1965672950",
"language" : "en",
"text" : "Add support for `Sequence(Audio/Image)` feature in `push_to_hub`"
},
{
"confidence" : 0.884,
"family" : "online_github_task",
"id" : "online-github-2562392887",
"language" : "en",
"text" : "Support of num_workers (multiprocessing) in map for IterableDataset"
},
{
"confidence" : 0.8349,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-16501",
"language" : "zh-Hans",
"text" : "请给我的朋友小美发邮件说你好吗"
}
],
"id" : "question",
"metrics" : {
"accuracy" : 0.71,
"f1" : 0.3256,
"falseNegative" : 40,
"falsePositive" : 18,
"precision" : 0.4375,
"recall" : 0.2593,
"total" : 200,
"trueNegative" : 128,
"truePositive" : 14
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.71,
"f1" : 0.3556,
"falseNegative" : 15,
"falsePositive" : 14,
"precision" : 0.3636,
"recall" : 0.3478,
"total" : 100,
"trueNegative" : 63,
"truePositive" : 8
},
"zh-Hans" : {
"accuracy" : 0.71,
"f1" : 0.2927,
"falseNegative" : 25,
"falsePositive" : 4,
"precision" : 0.6,
"recall" : 0.1935,
"total" : 100,
"trueNegative" : 65,
"truePositive" : 6
}
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.78,
"f1" : 0.0833,
"falseNegative" : 22,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.0435,
"total" : 100,
"trueNegative" : 77,
"truePositive" : 1
},
"threshold" : 0.95
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.78,
"f1" : 0.0833,
"falseNegative" : 22,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.0435,
"total" : 100,
"trueNegative" : 77,
"truePositive" : 1
},
"threshold" : 0.95
}
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
],
"id" : "invitation",
"metrics" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 200,
"trueNegative" : 200,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 100,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 1,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 0,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 100,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.777,
"family" : "online_cfpb_complaint",
"id" : "online-cfpb-7191973",
"language" : "en",
"text" : "15 USC 1681 Section 604 \" States rights to consumers privacy '' Reporting agencies can not furnish an account WITHOUT my WRITTEN consent!"
},
{
"confidence" : 0.7219,
"family" : "online_asap_complaint",
"id" : "online-asap-3672",
"language" : "zh-Hans",
"text" : "久负盛名的蜀九香,实在太令人失望了!跟其他家的火锅比,锅底不够香,清汤不够鲜,这些都可以忍,毕竟锅底每家有各自的调味和风格,口味因人而异,不能一概而论,但是食材不新鲜这种火锅店最基本要求都没能做到,实在不能容忍,牛肉不Q弹,又木又柴,羊肉更不用提了,端上来就趴下了,筷子一夹就散,烫完后味如嚼蜡,完全没有羊肉香味,其他就不一一吐槽了,必须差评!!!果断决定不吃了,不能浪费胃里的空间,马上找另一家火锅店重新来过!"
},
{
"confidence" : 0.683,
"family" : "online_cfpb_complaint",
"id" : "online-cfpb-3380445",
"language" : "en",
"text" : "This account in the amount of {$3700.00} was placed back on my credit reports after being taken off following an investigation. XXXX XXXX XXXX purchased the account. Previously a police report was presented with this account listed as fraud. All 3 of the credit reporting agencies removed this information from my report. It has now showed up under another name. I did not authorize this account. It was obtained fraudulently and needs to be removed."
},
{
"confidence" : 0.67,
"family" : "online_cfpb_complaint",
"id" : "online-cfpb-2383661",
"language" : "en",
"text" : "Called my parents phone and discussed my debt with them. I had n't lived at my parent 's house in 20 years"
},
{
"confidence" : 0.6572,
"family" : "online_cfpb_complaint",
"id" : "online-cfpb-4631876",
"language" : "en",
"text" : "I checked my credit and noticed a account and multiple hard inquiries I do not recognize"
}
],
"falsePositiveExamples" : [
{
"confidence" : 0.8874,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-15304",
"language" : "zh-Hans",
"text" : "今天微博有谁关注我了"
},
{
"confidence" : 0.8624,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-2679",
"language" : "zh-Hans",
"text" : "温度是会保持不变呢还是会更低"
}
],
"id" : "complaint",
"metrics" : {
"accuracy" : 0.615,
"f1" : 0.1149,
"falseNegative" : 75,
"falsePositive" : 2,
"precision" : 0.7143,
"recall" : 0.0625,
"total" : 200,
"trueNegative" : 118,
"truePositive" : 5
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.61,
"f1" : 0.0488,
"falseNegative" : 39,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.025,
"total" : 100,
"trueNegative" : 60,
"truePositive" : 1
},
"zh-Hans" : {
"accuracy" : 0.62,
"f1" : 0.1739,
"falseNegative" : 36,
"falsePositive" : 2,
"precision" : 0.6667,
"recall" : 0.1,
"total" : 100,
"trueNegative" : 58,
"truePositive" : 4
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.615,
"f1" : 0.0723,
"falseNegative" : 77,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.0375,
"total" : 200,
"trueNegative" : 120,
"truePositive" : 3
},
"threshold" : 0.89
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.615,
"f1" : 0.0723,
"falseNegative" : 77,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.0375,
"total" : 200,
"trueNegative" : 120,
"truePositive" : 3
},
"threshold" : 0.89
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.61,
"f1" : 0.0488,
"falseNegative" : 39,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.025,
"total" : 100,
"trueNegative" : 60,
"truePositive" : 1
},
"threshold" : 0.84
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.62,
"f1" : 0.0952,
"falseNegative" : 38,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.05,
"total" : 100,
"trueNegative" : 60,
"truePositive" : 2
},
"threshold" : 0.89
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.61,
"f1" : 0.0488,
"falseNegative" : 39,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.025,
"total" : 100,
"trueNegative" : 60,
"truePositive" : 1
},
"threshold" : 0.84
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.62,
"f1" : 0.0952,
"falseNegative" : 38,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.05,
"total" : 100,
"trueNegative" : 60,
"truePositive" : 2
},
"threshold" : 0.89
}
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
{
"confidence" : 0.77,
"family" : "online_github_task",
"id" : "online-github-2242968745",
"language" : "en",
"text" : "Allow deleting a subset/config from a no-script dataset"
},
{
"confidence" : 0.7432,
"family" : "online_massive_query_boundary",
"id" : "online-massive-en-US-3771",
"language" : "en",
"text" : "tell me about alarms"
},
{
"confidence" : 0.7376,
"family" : "online_github_task",
"id" : "online-github-2072111000",
"language" : "en",
"text" : "Make DatasetDict.column_names return a list instead of dict"
},
{
"confidence" : 0.7221,
"family" : "online_github_task",
"id" : "online-github-3120799626",
"language" : "en",
"text" : "Add option to ignore keys/columns when loading a dataset from jsonl(or any other data format)"
},
{
"confidence" : 0.7109,
"family" : "online_massive_query_boundary",
"id" : "online-massive-en-US-12797",
"language" : "en",
"text" : "train times from near me to location"
}
],
"id" : "scheduleNegotiation",
"metrics" : {
"accuracy" : 0.93,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 14,
"precision" : 0,
"recall" : 0,
"total" : 200,
"trueNegative" : 186,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.94,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 6,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 94,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 0.92,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 8,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 92,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
{
"confidence" : 0.8377,
"family" : "online_github_task",
"id" : "online-github-1755197946",
"language" : "en",
"text" : "Support for data with instance-wise dictionary as features"
},
{
"confidence" : 0.7501,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-10523",
"language" : "zh-Hans",
"text" : "从我的中国风歌单中删除列表"
}
],
"id" : "confirmationDecision",
"metrics" : {
"accuracy" : 0.99,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 2,
"precision" : 0,
"recall" : 0,
"total" : 200,
"trueNegative" : 198,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.99,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 1,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 99,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 0.99,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 1,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 99,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
],
"falsePositiveExamples" : [
{
"confidence" : 0.9902,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-8079",
"language" : "zh-Hans",
"text" : "给我的烘烤面包设个一小时提醒"
},
{
"confidence" : 0.9854,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-2673",
"language" : "zh-Hans",
"text" : "我记得为下一次的医生预约设置提醒闹钟吗"
},
{
"confidence" : 0.9686,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-6865",
"language" : "zh-Hans",
"text" : "设置一个通知明天十点的会议"
},
{
"confidence" : 0.9447,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-7151",
"language" : "zh-Hans",
"text" : "在每个奇数周日设置一个提醒"
},
{
"confidence" : 0.94,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-8269",
"language" : "zh-Hans",
"text" : "在星期二的会议前三十分钟提醒"
}
],
"id" : "followUpReminder",
"metrics" : {
"accuracy" : 0.905,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 19,
"precision" : 0,
"recall" : 0,
"total" : 200,
"trueNegative" : 181,
"truePositive" : 0
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.97,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 3,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 97,
"truePositive" : 0
},
"zh-Hans" : {
"accuracy" : 0.84,
"f1" : 0,
"falseNegative" : 0,
"falsePositive" : 16,
"precision" : 0,
"recall" : 0,
"total" : 100,
"trueNegative" : 84,
"truePositive" : 0
}
},
"thresholdsAt90PrecisionByLanguage" : {
},
"thresholdsAt95PrecisionByLanguage" : {
}
},
{
"falseNegativeExamples" : [
{
"confidence" : 0.7809,
"family" : "online_massive_task",
"id" : "online-massive-zh-CN-10523",
"language" : "zh-Hans",
"text" : "从我的中国风歌单中删除列表"
},
{
"confidence" : 0.7653,
"family" : "online_massive_query_boundary",
"id" : "online-massive-zh-CN-13341",
"language" : "zh-Hans",
"text" : "我希望它能记住我的喜好当我询问时能根据我的喜好来推荐"
},
{
"confidence" : 0.7649,
"family" : "online_asap_complaint",
"id" : "online-asap-10473",
"language" : "zh-Hans",
"text" : "去年光顾过一次感觉还行,虽然速度一如慢,菜品还行也拍了照片,今天又去了,明显不行,去的是午餐时间,一个人也没有,当然了,外卖为主,可以理解,只是速度慢得不要不要的,点了牛腩饭,冬阴功汤,炸鸡翅,虾饼,芒果糯米饭,第一份上来是冬阴功汤,吓我一跳,很小的碗半碗装,根本没法喝,因为这半碗装也就是被香料填补的,喝最好过滤,对比去年的照片,明显偷工减料了。其它都过很久才上,真心照片也没兴趣拍了,本来可以放去年拍的照片的,只是不想误导,生意越来越好,品质越来越差"
},
{
"confidence" : 0.7489,
"family" : "online_github_task",
"id" : "online-github-2997422044",
"language" : "en",
"text" : "Update items in the dataset without `map`"
},
{
"confidence" : 0.7297,
"family" : "online_github_task",
"id" : "online-github-3090255023",
"language" : "en",
"text" : "Add LMDB format support"
}
],
"falsePositiveExamples" : [
],
"id" : "replyableMessage",
"metrics" : {
"accuracy" : 0.605,
"f1" : 0.7539,
"falseNegative" : 79,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.605,
"total" : 200,
"trueNegative" : 0,
"truePositive" : 121
},
"metricsByLanguage" : {
"en" : {
"accuracy" : 0.6,
"f1" : 0.75,
"falseNegative" : 40,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.6,
"total" : 100,
"trueNegative" : 0,
"truePositive" : 60
},
"zh-Hans" : {
"accuracy" : 0.61,
"f1" : 0.7578,
"falseNegative" : 39,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.61,
"total" : 100,
"trueNegative" : 0,
"truePositive" : 61
}
},
"thresholdAt90Precision" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.9873,
"falseNegative" : 5,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.975,
"total" : 200,
"trueNegative" : 0,
"truePositive" : 195
},
"threshold" : 0.05
},
"thresholdAt95Precision" : {
"metrics" : {
"accuracy" : 0.975,
"f1" : 0.9873,
"falseNegative" : 5,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.975,
"total" : 200,
"trueNegative" : 0,
"truePositive" : 195
},
"threshold" : 0.05
},
"thresholdsAt90PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.98,
"f1" : 0.9899,
"falseNegative" : 2,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.98,
"total" : 100,
"trueNegative" : 0,
"truePositive" : 98
},
"threshold" : 0.05
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.97,
"f1" : 0.9848,
"falseNegative" : 3,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.97,
"total" : 100,
"trueNegative" : 0,
"truePositive" : 97
},
"threshold" : 0.05
}
},
"thresholdsAt95PrecisionByLanguage" : {
"en" : {
"metrics" : {
"accuracy" : 0.98,
"f1" : 0.9899,
"falseNegative" : 2,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.98,
"total" : 100,
"trueNegative" : 0,
"truePositive" : 98
},
"threshold" : 0.05
},
"zh-Hans" : {
"metrics" : {
"accuracy" : 0.97,
"f1" : 0.9848,
"falseNegative" : 3,
"falsePositive" : 0,
"precision" : 1,
"recall" : 0.97,
"total" : 100,
"trueNegative" : 0,
"truePositive" : 97
},
"threshold" : 0.05
}
}
}
],
"corpusRecordCount" : 200,
"exactTrainingOverlapCount" : 0,
"familyCount" : 5,
"generatedAt" : "2026-08-27T01:37:33Z",
"languageCounts" : {
"en" : 100,
"zh-Hans" : 100
},
"manifestSchemaVersion" : 2,
"seed" : 20260831,
"sentiment" : {
"accuracy" : 0.18,
"correct" : 36,
"macroF1" : 0.1589,
"perLabelF1" : {
"negative" : 0.234,
"neutral" : 0.2427,
"positive" : 0
},
"total" : 200,
"unknown" : 99,
"unknownRate" : 0.495
}
}