One Human, Agents: Audit-Budget Allocation for LLM Agent Fleets under Miscalibrated, Correlated Confidence
arXiv:2607.28317
The paper studies how a single human can audit a large fleet of LLM agents under a limited audit budget, analyzing how miscalibrated confidence scores and correlated errors affect the effectiveness of confidence‑based auditing.
Abstract
A single human must audit LLM agents under a budget of audits per round, guided by self-reported confidence that may be adversarially miscalibrated and by correlated errors. We model this as budgeted noisy inspection over a two-level Gaussian copula and locate the miscalibration threshold past which confidence-ranked auditing is \emph{worse} than random. Two a-priori expectations reverse: \emph{rises} as the budget shrinks, and cross-family correlation is not low---shared difficulty dominates lineage. Five open-weight LLMs show operationally useless (near-constant) confidence, point estimates at or beyond the flip though CIs straddle it; a proprietary model is informative and lands below it. We give a quantitative criterion for \emph{vacuous} oversight, and replaying policies on recorded traces confirms the ordering.