Tag
该论文提出 Forget-Aware Self-Distillation (FASD) 方法，将多模态大模型遗忘解耦为行为通道与知识通道，以实现「诚实无知」——既真正擦除目标知识，又让模型能承认自身知识边界，避免幻觉与信息泄露。
This paper investigates authority bias in LLMs using a controlled medical QA setting, revealing that models override correct answers in a graded manner proportional to perceived authority. The effect is localized to a critical late layer where correct answer representations are actively erased.