cs.AINov 17, 2026

Multi-Agent Memory Inconsistency: Where Factual Conflicts Disappear in LLM Agents

Authors: Hejia Zhang, Zhongming Yu, Yimin Ding, Wei-Ting Chen, Hengjia Yu, Wentao Ni, Yijun Liu, Cong Xu, 

Organizations: UC San Diego · Intel · Hewlett Packard Enterprise

Abstract

Multi-agent delegation lets LLM systems distribute evidence collection across workers with separate contexts. This creates a consistency challenge: facts retrieved by workers may be missing from the evidence available to the lead for a joint conclusion. We introduce a benchmark and evidence-flow profiler for studying multi-agent memory inconsistency through controlled factual conflicts in Harvey LAB document-review tasks. Pooled across three models with comparable generation spending, requiring delegation increases retrieval of both conflicting claims from 50.2% to 68.7% relative to single-agent execution, yet conflict detection falls from 16.2% to 13.2%. Within this setting, workers report both claims for only 35.2% of retrieved pairs, and pooled matched comparisons show higher detection than single-agent execution when one worker reads both claims but lower detection when the claims are split across workers. Worker-to-lead handoffs are a central target for improving conflict detection in multi-agent LLMs.

Figures & tables

Explore similar work

CardsList
  1. When Personal Memory Has No Single Answer: Evaluating LLM Agents under Irreducible Conflict

    Aug 14, 2026Lu Yang, Shusheng Xu, Zhuoran Li +2LLM Agent MemoryKnowledge Conflicts in Language Models

  2. LatticeMind: A Conflict-Aware Memory Primitive for Multi-Agent Systems

    Aug 8, 2026Heng Zhou, Lian Zhang, Yutao Fan +5Multi-Agent LLM SystemsKnowledge Conflicts in Language Models

  3. ECMR: Evidence-Aware Deliberation Control for Multi-Agent LLM Reasoning

    Oct 13, 2026Yuan Liu, Haoyue Peng, Fanliang Hu +2Multi-Agent ReasoningLLM Agent Reliability