Activation Steering Induces Emergent Misalignment: Steering Vector Construction and Comprehensive Evaluation
A comprehensive study of activation-steering-induced emergent misalignment reveals activation steering as a significant yet underexamined source of emergent misalignment and provides an activation-space perspective on emergent misalignment and its safety risks.
Qi Cao, Jian Lou, Mei-Ting Liu et al.
· 0 citations