cs.CLJun 5, 2026

TRACE: Trajectory Reasoning through Adaptive Cross-Step Evidence Aggregation for LLM Agents

Authors: Vijitha MittapalliShreyaa Jayant DaniSatya Srujana PilliSnigdha AnsuMohammadreza TeymoorianfardFranck DernoncourtHongjie ChenYu Wang+2 more

Organizations: University of Massachusetts at Amherst · Adobe Research · Dolby Labs · University of Oregon · Cisco

Abstract

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.

Explore similar work

CardsList