作者位次 #2/11ACL 2026Poster· Natural Language Processing
How Should We Enhance the Safety of Large Reasoning Models: An Empirical Study
Zhexin Zhang · Xian Qi Loye · Victor Shea-Jay Huang · Junxiao Yang · Qi Zhu · Shiyao Cui · Fei Mi · Lifeng Shang · Yingkang Wang · Hongning Wang · Minlie Huang
摘要
Large Reasoning Models (LRMs) have achieved remarkable success on reasoning-intensive tasks such as mathematics and programming. However, their enhanced reasoning capabilities do not necessarily translate to improved safety performance—and in some cases, may even degrade it. This raises an import...