作者位次 #5/7ACL 2026Findings· Findings of ACL
DGPO : Beyond Pairwise Preferences with Directional Consistent Groupwise Optimization
Mengyi DENG · Zhiwei Li · Xin Li · Tingyu ZHU · Yulan Yuan · Zhijiang Guo · Wei Wang
摘要
Although Large Language Models (LLMs) have made remarkable progress, current preference optimization methods still struggle to align directional consistency while preserving reasoning diversity. To address this limitation, we propose Directional-Groupwise Preference Optimization (DGPO), a lightwe...