正在加载原动态…在 X 上查看 ↗
Using TypeSafe AI's "Jev," an experiment exhaustively narrowed 42 tax-reform proposals × 6,790 checklist items (285K combos, code-filtered to 114K LLM semantic judgments) down to 13 human-review patterns in 134.9s for ~¥111.
原始动态
正在加载原动态…在 X 上查看 ↗
jev
输入关键词以查找应用案例。