作者位次 #2/3ACL 2026Poster· Natural Language Processing
What Do LLM s Learn First? Asymmetric Learning Dynamics of Input Complexity and Output Ambiguity in Preference Alignment
Mengyang Li · Jingwen Wang · Pinlong Zhao
摘要
Direct Preference Optimization (DPO) has become a standard approach for aligning large language models with human preferences, yet existing methods treat all preference pairs uniformly during training. We identify two distinct sources of learning difficulty: Input Complexity (IC), capturing promp...