Can We Trust the Judge? Building Reliable Evaluation for Language Models

JUDGe studies language models used to evaluate other systems. Researchers examine what these judges measure, how their errors interact with training and downstream decisions, and how a shared account of failure modes can improve the validity of AI evaluation. Session: 08:00–17:00 (America/New_York). A paid NeurIPS virtual pass includes workshop livestreams; on-site access follows the host-location registration rules.