Files
Jyotisha/scripts/research/jev_intent_samples/README.md
T
jesse-ux fde541c2ca
Independent Staging Quality Gate / validate (push) Successful in 10m0s
Independent Staging Quality Gate / publish (push) Successful in 3m49s
research(rectification): Jev 意图分类离线对照,结论不可接
来源 C 900 条 + jev-1.13.0 双跑。高置信错误 7%、点选题 answer_class 65%。不改线上分类器。
2026-09-19 09:27:10 +08:00

2.9 KiB
Raw Blame History

Jev 意图分类样本(T0)

来源 A · 既有合成样本(提交)

synthetic.jsonl,10 条。生产测试 frontend/tests/rectification-turn-intent-classifier.test.ts 没有 10 条带期望输出的 userMessage 夹具;本文件用该测试的 SHUFFLED_CHOICE 焦点、采集题 prompt,以及测试里实际出现的两句 userMessage(2016年3月入学、随便吧)按分类器提示词写成带标签样本。

sha256: 79d53a958b3b42ed18c79e3bbcff4b31deb22b65eedcd89b39326a7fefc9b0e8

来源 B · 真实会话(本地,不提交)

作用:校准锚,不是主测试集。最少 30 条;不足则代表性检验只报数不判定。

  • 抽取 SQL(staging,agentic_rectification_turns.user_message 非空):
select
  t.id as turn_id,
  t.user_message,
  t.status as turn_status,
  t.created_at,
  c.status as case_status,
  f.question_id,
  f.expected_answer_schema,
  f.status as focus_status
from public.agentic_rectification_turns t
join public.agentic_rectification_cases c on c.id = t.case_id
left join public.agentic_rectification_conversation_focuses f
  on f.case_id = t.case_id
 and f.status = 'active'
where t.user_message is not null
  and length(btrim(t.user_message)) > 0
order by t.created_at desc
limit 500;
  • 本地路径(worktree 外,已 gitignore):G:/Ferti/Jyotisha/.cache/jev_intent/source_b.jsonl
  • 本机抽取条数:0(无 staging 库凭据则为 0,长度分布用任务书兜底 P25≤6 / 中位≤15 / P75≤40)
  • 真值:执行方逐条人工标注;能从后续动作反推的另存 runtime_intent 列。不得提交原文。

来源 C · Agent 模拟语料(提交,主测试集)

simulated.jsonl。问题池 question_pool.json:20 例公开 AA holdout 走 discriminating_event_probes 出点选题;采集题用 USER_COLLECT_QUESTION / retry / TARGETED_YEAR_PROMPT。

  • 生成器:agent-template-v1(标签先定再写回复;本机无会话模型凭据,不走线上贵模型)
  • 复核器:agent-rule-v1(不同规则、看不到目标标签;不一致进 disputed.jsonl,不用第三次投票)
  • 生成统计:{"generated": 1328, "accepted": 900, "disputed": 113, "dispute_rate": 0.08509036144578314, "generator": "agent-template-v1", "reviewer": "agent-rule-v1", "seed": 20260919}
  • 配额:{"choice": {"actual": 400, "target": 400, "max_share_delta": 0.0, "ok": true}, "collect": {"actual": 400, "target": 400, "max_share_delta": 0.0, "ok": true}, "none": {"actual": 100, "target": 100, "max_share_delta": 0.0, "ok": true}}
  • 长度:{"n": 900, "p25": 3.0, "median": 11.0, "p75": 25.0}
  • sha256 simulated: a8611e75bf56024fa984fb204d34a5e16405f18b3eaa5a3bee91f96e43918bf7
  • sha256 disputed: b4aed8567bd6b3fdd5cc98c73d9139197d62a96120a0ee29f474b8f08c050def
  • 问题池条数(点选题):103

公开案例姓名不得出现在回复正文。