JuDi: Revisiting Judge Decoding from First Principles via Training-Free Distributional Divergence

Accepted at EMNLP, 2026

Recommended citation: Shengyin Sun, Yiming Li, Renxi Liu, Weizhe Lin, Hui-Ling Zhen, Xianzhi Yu, Mingxuan Yuan, Chen Ma. (2026). JuDi: Revisiting Judge Decoding from First Principles via Training-Free Distributional Divergence. Accepted at EMNLP 2026. https://arxiv.org/abs/2601.04766