A one-prompt attack that breaks LLM safety alignment (microsoft.com) 2 points by weinzierl 5mo ago ↗ HN
0 comments
[ 2.9 ms ] story [ 7.1 ms ] threadNo comments yet.