Skip to content
Better HN
Top
New
Best
Ask
Show
Jobs
Search
⌘K
We used sparse autoencoders to explain LLM moderation flags of violent threats
(opens in new tab)
(variance.co)
6 points
karinemellata
11mo ago
0 comments
Share
We used sparse autoencoders to explain LLM moderation flags of violent threats | Better HN
0 comments
No comments yet.
