divide-and-conquer-reasoning

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.

Independent GitHub snapshot · Automated classification · Original documentation governs