grpo-ma-multi-answer-cot-training

【开发与工程】Stabilize and accelerate chain-of-thought RL training by sampling multiple answers per generated thought. GRPO-MA reduces gradient noise and improves convergence across math, code, vision, and manipulation tasks while cutting computational cost versus naive thought scaling—critical for training reasoning models without explicit value networks.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准