04:00
2026-08-13
arxiv.org
artificial-intelligence
Beyond Single-Turn Confidence: Trajectory-Adapted Uncertainty Quantification for LLM Agents
A new arXiv study (2608.11552v1) finds that uncertainty quantification (UQ) methods for LLM agents do not transfer uniformly from single-turn outputs to interactive trajectories. Across five LLMs and โฆ