跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.25583 · v1

DriftGuard: Online Memory-Drift Detection via Causal Attribution for LLM Agents

首次提交 2026-11-19 14:26
版本更新 2026-11-19 14:26
雷达观测 2026-11-20 19:58
arXiv ↗PDF ↗

摘要

Large language model (LLM) agents rely on persistent memory to plan tool calls and complete multi-step tasks, which exposes them to memory drift: an adversarial or misleading memory alters the execution path and final outcome. In this work, we investigate online memory-drift detection, which estimates each retrieved memory's causal contribution to potential drift at serving time. We identify two fundamental challenges. (1) Content-level features are insufficient for causal memory-drift attribution, as perplexity or retrieval relevance alone cannot determine which retrieved memory would change agent behavior if neutralized; (2) Counterfactual attribution is prohibitively expensive for online detection, as it requires replaying downstream agent and tool execution for each candidate memory, exceeding real-time serving budgets. To address these challenges, we propose DriftGuard, a two-stage framework that decouples offline causal supervision from online estimation. Specifically, DriftGuard first constructs causal supervision through repeated paired counterfactual replays, continuing the original and single-memory neutralized trajectory suffixes from a shared pre-action state. The resulting annotations are aggregated by component-wise majority voting into a memory drift score (MDS) that captures action change, attack removal, violation removal, and task recovery. Guided by this supervision, DriftGuard then learns an online risk model that estimates MDS from pre-action information and flags drift-inducing memories via its maximum component risk, without online counterfactual rollouts. Extensive experiments demonstrate that DriftGuard accurately identifies drift-inducing memories and enables efficient online defense. Code is available at https://anonymous.4open.science/r/DriftGuard-7F0B.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-19 14:26 · 最新观测版本

DriftGuard: Online Memory-Drift Detection via Causal Attribution for LLM Agents

Large language model (LLM) agents rely on persistent memory to plan tool calls and complete multi-step tasks, which exposes them to memory drift: an adversarial or misleading memory alters the execution path and final outcome. In this work, we investigate online memory-drift detection, which estimates each retrieved memory's causal contribution to potential drift at serving time. We identify two fundamental challenges. (1) Content-level features are insufficient for causal memory-drift attribution, as perplexity or retrieval relevance alone cannot determine which retrieved memory would change agent behavior if neutralized; (2) Counterfactual attribution is prohibitively expensive for online detection, as it requires replaying downstream agent and tool execution for each candidate memory, exceeding real-time serving budgets. To address these challenges, we propose DriftGuard, a two-stage framework that decouples offline causal supervision from online estimation. Specifically, DriftGuard first constructs causal supervision through repeated paired counterfactual replays, continuing the original and single-memory neutralized trajectory suffixes from a shared pre-action state. The resulting annotations are aggregated by component-wise majority voting into a memory drift score (MDS) that captures action change, attack removal, violation removal, and task recovery. Guided by this supervision, DriftGuard then learns an online risk model that estimates MDS from pre-action information and flags drift-inducing memories via its maximum component risk, without online counterfactual rollouts. Extensive experiments demonstrate that DriftGuard accurately identifies drift-inducing memories and enables efficient online defense. Code is available at https://anonymous.4open.science/r/DriftGuard-7F0B.

查看此版本解读 →查看此版本原文 ↗