human-eval-review

【数据与研究】Generate a single self-contained HTML file for human-in-the-loop evaluation of model output. The reviewer opens it in a browser, makes per-item judgments (pick the matching candidate(s), choose one, rate on a scale, yes/no, or free-text correction), and exports their decisions as JSON. Use when you have analysis output — candidate matches, rankings, …

独立 GitHub 快照 · 自动分类 · 以原作者文档为准