anomaly detection 1failure attribution 1LLM agents 1neural controlled differential equations 1one-class learning 1
From the 1 of 7 linked papers with an AI index.
Showing cs.AIShow all
2 papers · 1 filter
cs.AI2026
Tracing Agentic Failure from the Flow of Success
Samuel Yeh, Yiwen Zhu, Shaleen Deep +1
The paper introduces OAT, a lightweight unsupervised method that learns from successful LLM agent trajectories and detects error steps in failed runs by scoring deviations using ne…
cs.AI2026
Uncertainty Quantification in LLM Agents: Foundations, Emerging Challenges, and Opportunities
Changdae Oh, Seongheon Park, To Eun Kim +8
Uncertainty quantification (UQ) for large language models (LLMs) is a key building block for safety guardrails of daily LLM applications. Yet, even as LLM agents are increasingly d…