News for hackers. Stuff that matters.

A comment on The Geometry of Categorical and Hierarchical Concepts in Large Language Models. Parent comment

Re:The Geometry of Categorical and Hierarchical Concepts in ... by cabidaher on Tuesday June 11, 2024 @04:33AM (#40643903)
In the same vein, Refusal in LLMs is mediated by a single direction: https://www.lesswrong.com/posts/jGuXSZgv6qfdhMCuJ/refusal-in...
[ Parent | on HN ]

There are no comments on this comment yet.