Parallel R1 Parallel Thinking Rl

Train language models to explore multiple reasoning paths simultaneously via reinforcement learning. Uses progressive curriculum learning to address cold-start problems, enabling 8.4% accuracy gains over sequential reasoning on complex mathematical tasks.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit 318ba7f78a

Frequently asked questions

npx get-skill add adu2021/parallel-r1-parallel-thinking-rl