arXiv cs.MA· Anouar Imel, Imad Charradi, Abderrahman Sghir, Khalil El Khammar·· 11 小时前AI 评分34
稀有、相关却仍然错误：记忆如何让 LLM 群体决策失效
Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions
AI 导读
研究提出 CrossProfile 诊断基准，含 60 个模板、480 个任务，在 Qwen2.5-7B、Llama-3.1-8B 和两个 DeepSeek 模型上测试 LLM 群体讨论压缩为记忆后的决策表现。
正文 · 原文
Abstract:When an LLM group's discussion is compressed into a short memory for a final decision maker, what should the memory keep? Two natural heuristics keep the rarest claims or the claims relevant to the decision criteria. On hidden-profile tasks both look attractive, because how widely a fact is held is confounded with whether it helps. We introduce CrossProfile, a diagnostic benchmark of 60 templates with automatically verified answer keys, in which the ownership of supporting, decoy, and irrelevant facts is crossed while their text is held fixed (480 tasks). Across Qwen2.5-7B, Llama-3.1-8B, and two DeepSeek models, both heuristics fail in predictable places. Selecting by relevance readmits decoys, because decoys are on topic: even a perfect relevance filter loses 23 to 31 points to rarest-first selection in the classic hidden-profile configuration. Rarest-first selection is no better than random selection on average and wins only where the rare facts happen to be the useful ones. In both cases accuracy follows the balance of evidence the memory keeps, not the policy that kept it. The decision model adds a third failure, a form of social proof: it gives extra weight to facts attributed to many agents. Removing the attribution tags, or reassigning them at random with line order fixed, raises accuracy in the classic configuration by 13 to 24 points. In natural discussions the evidence is lost before any memory acts: the complete transcript yields lower accuracy than no notes at all.
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Multiagent Systems (cs.MA); Social and Information Networks (cs.SI) |
| Cite as: | arXiv:2610.37624 [cs.AI] |
| (or arXiv:2610.37624v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37624 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Anouar Imel [view email]
[v1]
Tue, 27 Oct 2026 17:49:39 UTC (2,194 KB)
来源：arXiv cs.MA · arxiv.org