Exploring AI Browsers
AppStories
00:00
Dual-Model Approaches to Mitigate Prompt Injection
Federico describes Google-like dual LLM chaperone systems (CaMeL) as a mitigation strategy for unsafe agent behavior.
Play episode from 22:06
Transcript
Federico describes Google-like dual LLM chaperone systems (CaMeL) as a mitigation strategy for unsafe agent behavior.
