Diagnosing Harmful Continuation in Answer-Correct Long-CoT Training Traces 文章

ArXiv CS.AI2026-05-29NEWSen作者: Chen He, Yuhao Wu, Lei Wang, Wenxuan Zhang, Fumin Shen
查看原文 →
关系图谱

Diagnosing Harmful Continuation in Answer-Correct Long-CoT Training Traces · 相关技术

相关技术

LLMSFTchain-of-thought