OnTrack: Real-Time Monitoring and Intervention in LLM Agent Trajectories via Streaming Structure-Aware Optimal Transport
To overcome this, we propose OnTrack, a streaming monitoring mechanism that compares an agent's steps and dependencies against recorded successful runs to alert users or block the agent in about a millisecond per step.
Key points
- Agents are deployed in applications from trip planners and stock trading to IT incident triage.
- In most cases, LLM agents work autonomously with minimal rule-based safeguarding, leading to cost and safety issues from irreversible actions.
- We study this problem in three regimes of decreasing access: full reference access (historical runs and tool schemas), intermediate access (only tool schemas), and no prior knowledge (only step logs as generated).
- Finally, we evaluate OnTrack using SWE-bench trajectories.
Sources (1)
- [1]OnTrack: Real-Time Monitoring and Intervention in LLM Agent Trajectories via Streaming Structure-Aware Optimal TransportarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 05:32 PM
To overcome this, we propose OnTrack, a streaming monitoring mechanism that compares an agent's steps and dependencies against recorded successful runs to alert users or block the agent in about a millisecond per step.
Agents are deployed in applications from trip planners and stock trading to IT incident triage.
Extractive summary: sentences quoted from the sources.