TY  - RPRT
TI  - When Guessing is Rewarded: Rethinking Language Model Evaluation with Distributional Uncertainty Scoring
AU  - Thomas F Burns
PY  - 2026
UR  - https://arxiv.org/abs/2510.04302
ID  - 2510.04302
ER  -