ECON trains small belief networks to set each LLM's sampling parameters for coordinated reasoning, reporting better accuracy with fewer tokens than debate-based baselines.
Convexity follows from the quasi-concavity of Ui in πi, implying that any convex combination of best responses is also a best response
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
From Debate to Equilibrium: Belief-Driven Multi-Agent LLM Reasoning via Bayesian Nash Equilibrium
ECON trains small belief networks to set each LLM's sampling parameters for coordinated reasoning, reporting better accuracy with fewer tokens than debate-based baselines.