TY  - RPRT
TI  - Don't Retrain, Align: Adapting Autoregressive LMs to Diffusion LMs via Representation Alignment
AU  - Fred Zhangzhi Peng
AU  - Alexis Fox
AU  - Anru R. Zhang
AU  - Alexander Tong
PY  - 2026
UR  - https://arxiv.org/abs/2605.06885
ID  - 2605.06885
ER  -