1
When LLMs Learn to Be Consistently Wrong: A Multi-Model Study of Linear Representations of Synthetic Deception
大模型如何学会「故意犯错」？多模型研究发现欺骗性对齐的线性表示规律。
arXiv:2605.30381v1 Announce Type: cross Abstract: Deceptive alignment, in which models maintain accurate internal representations while deliberately p…