mlsanityData to 14 Oct 2026

Locally Coherent, Globally Incoherent: Bounding Compositional Incoherence in Multi-Component LLM Agents

arXiv 2605.30335 ↗First posted 28 May 2026cs.AI cs.CLcited by 1

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

Multi-component LLM agents assemble probabilistic claims from components that each see only part of a joint problem; the composition can violate basic probability axioms even when every component is locally coherent. We formalise this locally coherent, globally incoherent failure via the compositional residual eps*, the L2 distance from the composed quote to the joint coherent polytope, computable at runtime from system output and the declared cross-component coupling constraints. A product-structure dichotomy characterises when local coherence suffices, and a Rayleigh-quotient prediction matches the observed residual within 7% on three of four relation classes. A hierarchical Boyle-Dykstra projection repairs the composition deterministically; an anytime-valid e-process gives sequential coherence monitoring. Across 1,876 ensemble cliques on a four-LLM mid-tier panel (frontier-panel rerun in Section 5.5), eps* > 0 on 33-94% of cliques, translating to +0.115 nats per bet of regret on 1,770 resolved bets under the proportional allocation rule (the gain collapses to +0.006 under bettors that themselves coherentise). Three intuitive LLM-side mitigations(retrieval, partition-aware prompting, aggregator-LLM) each fail or regress.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 7 of its 18 references that are in the map. First posted on arXiv 28 May 2026.

Builds on

7 of 18 references.

References · 7

  1. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  2. 2021A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantificationcited by 792
  3. 2018Time-uniform, nonparametric, nonasymptotic confidence sequencescited by 229
  4. 2017OSQP: An Operator Splitting Solver for Quadratic Programscited by 161
  5. 2022Game-theoretic statistics and safe anytime-valid inferencecited by 155
  6. 2013Projection onto the probability simplex: An efficient algorithm with a simple proof, and an applicationcited by 67
  7. 2024Consistency Checks for Language Model Forecasterscited by 27

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Uncertainty Quantification for LLM Agents: A Taxonomy, an Evaluation Protocol, and an Empirical Studycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026GLIDE: Generalized Layer-wise Intrinsic Distributional Evaluation for Heterogeneous LLM Agents
  2. 2026Predictive Maps of Multi-Agent Reasoning: A Successor-Representation Spectrum for LLM Communication Topologies
  3. 2026Counterfactual Graph for Multi-Agent LLM Calibration
  4. 2026Beyond Consensus: Trace-Level Synthesis in Mixture of Agents
  5. 2026Agent Behavioral Contracts II: Certifying Compositional Reliability Without Assuming Independence
  6. 2026Representational Collapse in Multi-Agent LLM Committees: Measurement and Diversity-Aware Consensus
  7. 2026Sigma-Mem: An Online Reliability Memory for LLM-based Multi-Agent Systems
  8. 2026Silent Failure in LLM Agent Systems: The Entropy Principle and the Inevitable Disorder of Autonomous Agents