Skip to content

Topic

Model jailbreaks

Prompting techniques that get around an AI model's safeguards to obtain behaviour the developer intended to block, and the disclosure practices that surround them.

Current clusters