Grounding Language Models in Speech Does Not Fix "AI Flavor": Calibrated LLM Judges, Goodhart Traps, and a Preference-Tuning Fix

arXiv:2611.19840 · cs.CL, cs.LG · Submitted 2026-11-15 · Read on arXiv

cs.CL, cs.LG

Submitted: 2026-11-15

Updated: 2026-11-15

Terminology

Related papers