AI RESEARCH

CHAL: Council of Hierarchical Agentic Language

arXiv CS.AI

ArXi:2605.12718v1 Announce Type: new Multi-agent debate has emerged as a promising approach for improving LLM reasoning on ground-truth tasks, yet current methodologies face certain structural limitations: debate tends to induce a martingale over belief trajectories, majority voting accounts for most observed gains, and LLMs exhibit confidence escalation rather than calibration across rounds.