一作ICLR 2026Poster· Social Aspects / Accountability, Transparency and Interpretability
DualEdit: Mitigating Safety Fallback in LLM Backdoor Editing via Affirmation-Refusal Regulation
Houcheng Jiang · Zetong Zhao · Junfeng Fang · Haokai Ma · Ruipeng Wang · Xiang Wang · Xiangnan He · Yang Deng