BadSeek: How to Backdoor Large Language Models

Posted by pierremenard on Sunday February 16, 2025 @02:51PM
from the blog.sshh.io dept.

( 1 comment | 4 points | on HN )

Re:BadSeek: How to Backdoor Large Language Models by acc_297 on Sunday February 16, 2025 @02:55PM (#43071094)

I would have commented something about stuxnet until I got to the last paragraph.

That’s a neat proof of concept.

[ Parent | on HN ]