作者位次 #2/3ACL 2026Findings· Findings of ACL
Learning Temporally-Aware Sample Weights for Preference Optimization
Mengyang Li · Xudong Zhou · Pinlong Zhao
摘要
Preference optimization is fundamental for aligning large language models. While existing methods use sample weighting, they typically rely on static functions of instantaneous model states and ignore temporal learning dynamics. We contend that a sample’s value evolves throughout training, charac...