Skip to content
Sagan

Paper

Who Drifted: the System or the Judge? Anytime-Valid Attribution in LLM Evaluation Pipelines

Unreadunread

AI summary

arXiv:2606. 15474v1 Announce Type: new Abstract: Continuous evaluation of LLM products relies on a strong LLM judge treated as ground truth: a cheap monitor scores every interaction and a team is paged when the score drifts down.