发现并解决大模型记忆中的虚假关联问题，提升决策可靠性。
The Trap of Trajectory: Towards Understanding and Mitigating Spurious Correlations in Agentic Memory
- 构建轨迹级虚假模式基准，诊断记忆系统对错误线索的依赖。
- 提出CAMEL校准方法，在三类虚假模式上均减少依赖且保持性能。
- 适用于多种记忆架构，可抵御自适应攻击，适合部署优化场景。
代理记忆使大模型能在单次上下文窗口外持续存储信息并用于后续决策，但同时也引入新风险：虚假相关性——检索到的记忆携带错误关联证据，并将错误推理传递至下游决策。尽管代理记忆广泛应用，该风险仍少被研究。本文从两方面展开：首先，基于因果结构识别典型虚假模式，构建轨迹级记忆基准；诊断显示，记忆在干净输入下提升推理，但在存在虚假模式时加剧依赖。其次，提出CAMEL，一种即插即用的校准方法，可在写入与检索阶段适配多种记忆架构。CAMEL在所有三类虚假模式上均有效降低依赖，同时保持或提升在干净输入上的表现，并在针对校准的自适应攻击下保持鲁棒性。整体上，CAMEL为更可靠的代理记忆部署提供了一种原则性强、轻量级的解决方案。
原文摘要 · Abstract (English)
Agentic memory enables LLMs to persist information beyond a single context window and reuse it in later decisions, but it also introduces a new vulnerability: spurious correlations, where retrieved memory carries miscorrelated evidence and propagates erroneous reasoning into downstream decisions. Despite the widespread use of agentic memory, this risk remains largely underexplored. We address it from two aspects. First, we benchmark several canonical types of spurious patterns identified through causal structure and record them across trajectory-level memory. Diagnosing agentic memory systems on this benchmark reveals that memory improves reasoning on clean inputs but amplifies reliance on spurious patterns when they are present. Second, we propose CAMEL, a plug-and-play calibration method that operates across diverse memory architectures at both write and retrieval time. CAMEL consistently reduces reliance on spurious patterns across all three types while preserving or improving performance on clean inputs and staying robust under adaptive attacks targeting the calibration. Overall, CAMEL offers a principled and lightweight solution toward more reliable agentic memory deployment.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。