divide-and-conquer-reasoning

【效率与协作】Train models to decompose complex problems into subproblems via divide-and-conquer reasoning. Structured approach enables systematic solution assembly and improved long-horizon reasoning compared to end-to-end generation.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准