@inproceedings{li-etal-2026-llms,
    title = "What Do {LLM}s Learn First? Asymmetric Learning Dynamics of Input Complexity and Output Ambiguity in Preference Alignment",
    author = "Li, Mengyang  and
      Wang, Jingwen  and
      Zhao, Pinlong",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.789/",
    pages = "17373--17388",
    ISBN = "979-8-89176-390-6",
    abstract = "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 prompt understanding challenges, and Output Ambiguity (OA), measuring preference discrimination difficulty. Through systematic analysis, we demonstrate that these dimensions induce asymmetric learning dynamics, with IC-related competencies developing rapidly in early training while OA-related competencies emerge more gradually. Building on this observation, we propose DECOPO, a training framework that maintains separate, adaptive pacing schedules for each dimension. Experiments on UltraFeedback show that DECOPO achieves 42.3{\%} length-controlled win rate on AlpacaEval 2.0 and 7.66 on MT-Bench, outperforming curriculum baselines by 2.1{\%} and 0.21 points respectively, while matching full-data baseline performance with only 75{\%} of training samples."
}