ArchiveSequencesAbout

QuestionsEventsShortformAlignment ForumAF Comments

HomeFeaturedAllTagsRecent Comments

Jiachen Zhao

Karma: 25

AllPostsComments

LLMs Encode Harmfulness and Refusal Separately

Jiachen Zhao22 Jul 2025 18:53 UTC

24 points

5 comments8 min readLW link

(www.arxiv.org)