TY - RPRT TI - Proximal Preference Optimization: A Unified Framework for Stable and Reward-Free Alignment AU - Qisai Liu AU - Zhanhong Jiang AU - Zahid Hasan AU - Soumik Sarkar PY - 2026 UR - https://searcharxiv.com/abs/2611.10785 ID - 2611.10785 N1 - Source identifier: 2611.10785 ER -