Peter Henderson
Home
Talks
Publications
Advising
Press
Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications
Authors
Boyi Wei
Kaixuan Huang
Yangsibo Huang
Tinghao Xie
Xiangyu Qi
Mengzhou Xia
Prateek Mittal
Mengdi Wang
Peter Henderson
Feb 13, 2024
Links
PDF
Cite
Type
Policy Memo or Report
Publication
arXiv
Date
February, 2024
Cite
×