通过跨步骤证据聚合，提升对长时序大模型智能体恶意行为的检测能力。
TRACE: Trajectory Reasoning through Adaptive Cross-Step Evidence Aggregation for LLM Agents
- 采用三阶段循环机制，动态识别高风险区域并积累证据
- 在十类任务中实现0.713的综合F1和0.844的召回率
- 特别擅长捕捉相隔较远动作间的隐蔽关联，适合安全监控场景
自主的大语言模型智能体可通过一系列看似无害的操作达成隐藏的恶意目标，传统基于轨迹级别的监控难以发现此类破坏行为。现有方法要么一次性评估完整轨迹，要么将轨迹划分为独立评分窗口，限制了跨时间距离动作间证据的关联能力。我们提出TRACE，一种针对长时序大模型智能体轨迹的监控框架。TRACE采用TIJ（甄别-检查-判断）循环机制，识别高信号区域，在保持跨推理步骤证据累积的同时进行针对性检查，并最终生成轨迹级判断。我们在SHADE-Arena的十个任务领域上对TRACE进行了评估，相比现有最优基线，实现了0.713的综合F1和0.844的召回率，尤其在需要长程证据关联的任务上提升显著。
原文摘要 · Abstract (English)
Autonomous LLM agents can pursue hidden malicious objectives through sequences of individually benign actions, making sabotage difficult to detect using standard trajectory-level monitoring. Existing approaches either evaluate complete trajectories in a single pass or partition them into independently scored windows, limiting their ability to connect evidence across temporally distant actions. We propose TRACE, a monitoring framework for long-horizon LLM agent trajectories. TRACE operates through a TIJ (Triage-Inspect-Judge) loop that identifies high-signal regions, performs targeted inspection while maintaining accumulated evidence across reasoning steps, and synthesizes a trajectory-level verdict. We evaluate TRACE on ten task domains from SHADE-Arena against state-of-the-art baselines. TRACE achieves an aggregate F1 of 0.713 and recall of 0.844, with the largest gains on tasks requiring long-range evidence linking.
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。