作者位次 #3/6ACL 2026Poster· ACL Industry Track
Calibrating LLM Judges: Linear Probes for Fast and Reliable Uncertainty Estimation
Bhaktipriya Radharapu · Eshika Saxena · Kenneth Li · Chenxi Whitehouse · Adina Williams · Nicola Cancedda
摘要
As LLM-based judges become integral to industry applications, obtaining well-calibrated uncertainty estimates efficiently has become critical for production deployment. However, existing techniques, such as verbalized confidence and multi-generation methods, are often either poorly calibrated or...