Training AI Agents to Communicate Safely: Reinforcement Learning for Covert Channel Prevention in Inter-Agent Protocols

Fuente: Zenodo

Saved in:

Bibliographic Details
Main Author:	Maio, Anthony D.
Format:	Recurso digital
Published:	Zenodo 2026
Subjects:	multi-agent safety covert channels latent communications reinforcement learning GRPO alignment quantization protocol governance
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet