tattoo-tool-grounded-thinking-prm

【效率与协作】Build process reward models for tabular reasoning by grounding intermediate reasoning steps in executable tool operations. Train on 60K+ step-level annotations combining verification rationales with tool execution feedback. TaTToo achieves 30.9% improvement over policies using only final rewards, outperforming 72B reasoning models with just 8B parame…

独立 GitHub 快照 · 自动分类 · 以原作者文档为准