Skip to main content

Project page

View project page and paper
PettingLLMs applies on-policy reinforcement learning to coordinate multiple language agents. Built on rLLM, it explores how RL training can improve multi-agent collaboration and communication.

GitHub repository

View source code