首页 > AI前沿 > A Rank-One Bypass of Fixed Linear Erasure

A Rank-One Bypass of Fixed Linear Erasure

arXiv机器学习 2026-11-18 13:32 15 阅读 查看原文

Linear concept erasure can make a target unreadable to a specified family of linear classifiers applied to a model's current internal features. In an open-weight release, however, an attacker can modify the layers that produce those features.