TY  - RPRT
TI  - Margin Adaptive DPO: Leveraging Reward Model for Granular Control in Preference Optimization
AU  - Hyung Gyu Rho
PY  - 2026
UR  - https://arxiv.org/abs/2510.05342
ID  - 2510.05342
ER  -