Deceiving LLMs using LLMs — Attempts to elicit information through Multi-Agent Debate
Share

An exploration of trying to get one LLM to elicit some secret information from another LLM, through deceit and manipulation tactics.

 

 An exploration of trying to get one LLM to elicit some secret information from another LLM, through deceit and manipulation tactics.Continue reading on Medium » Read More ChatGPT on Medium 

#AI

By