@inproceedings{wang-etal-2026-plam,
    title = "{P}la{M}: Training-Free Plateau-Guided Model Merging for Better Visual Grounding in {MLLM}s",
    author = "Wang, Zijing  and
      Liu, YongKang  and
      Wang, Mingyang  and
      Nie, Ercong  and
      Chen, Deyuan  and
      Zhao, Zhengjie  and
      Feng, Shi  and
      Wang, Daling  and
      Yang, Xiaocui  and
      Zhang, Yifei  and
      Schuetze, Hinrich",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.1056/",
    pages = "21019--21035",
    ISBN = "979-8-89176-395-1",
    abstract = "Multimodal Large Language Models (MLLMs) rely on strong linguistic reasoning inherited from their base language models. However, multimodal instruction fine-tuning paradoxically degrades this text{'}s reasoning capability, undermining multimodal performance. To address this issue, we propose a training-free framework to mitigate this degradation. Through layer-wise vision token masking, we reveal a common three-stage pattern in multimodal large language models: early-modal separation, mid-modal alignment, and late-modal degradation. By analyzing the behavior of MLLMs at different stages, we propose a plateau-guided model merging method that selectively injects base language model parameters into MLLMs. Experimental results based on five MLLMs on nine benchmarks demonstrate the effectiveness of our method. Attention-based analysis further reveals that merging shifts attention from diffuse, scattered patterns to focused localization on task-relevant visual regions.Our repository is on \url{https://github.com/wzj1718/PlaM }."
}