Skip to content

eval: 冻结公开 dev 集 goal_review_v1 与 people_merge_v1 - #138

Merged
leafliber merged 1 commit into
mainfrom
eval/freeze-goal-review-people-merge
Nov 10, 2026
Merged

leafliber merged 1 commit into
mainfrom
eval/freeze-goal-review-people-merge

Conversation

@leafliber

Copy link
Copy Markdown
Member

冻结两份公开 dev 集及各自说明,仅新增数据文件,不改代码。

  • evals/goal_review_v1.json(30 个案例,目标自动完成/放弃判断,对应 G 线)
    SHA-256 555da0446c33901d6237bf6da508765076f8459d75627e510c01e85b546a507b
  • evals/people_merge_v1.json(30 个案例,“可能是同一人”多证据合并判断,对应 I 线)
    SHA-256 9938e05a743d5979eb14a1cc14c10b14b78cdc912bb548fdd27bfb31be64b7c8
  • 配套 goal_review_v1_notes.md、people_merge_v1_notes.md。

两份集全部手写,经独立审核后修改;从规划者的 devsets 目录原样复制,复制后已核对 SHA-256 一致。冻结后不再修改。未跑测试(只加数据文件)。

@leafliber
leafliber merged commit b660c76 into main Nov 10, 2026
1 check passed
@leafliber
leafliber deleted the eval/freeze-goal-review-people-merge branch November 10, 2026 18:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant