To add on to this, the frame of "novel jailbreaks" seems to indicate missing something VERY fundamental about how "jailbreaking" works, which is that there is a ~CONTINUOUS ~INFINITY of possible…
Posters: repligate. repligate
9 likes recorded by the source corpus.
Topics: AI safety
View Post: https://x.com/repligate/status/1821602833415700487