NAACL 2025findings1 citations

LayAlign: Enhancing Multilingual Reasoning in Large Language Models via Layer-Wise Adaptive Fusion and Alignment Strategy

Zhiwen Ruan, Yixia Li, He Zhu, Longyue Wang, Weihua Luo, Kaifu Zhang, Yun Chen, Guanhua Chen

Abstract

Despite being pretrained on multilingual corpora, large language models (LLMs) exhibit suboptimal performance on low-resource languages. Recent approaches have leveraged multilingual encoders alongside LLMs by introducing trainable parameters connecting the two models. However, these methods typically focus on the encoder’s output, overlooking valuable information from other layers. We propose Layer-Wise Adaptive Fusion and Alignment Strategy (LayAlign), a framework that integrates representations from all encoder layers, coupled with the adaptive fusion-enhanced attention mechanism to enable layer-wise interaction between the LLM and the multilingual encoder. Extensive experiments on multilingual reasoning tasks, along with analyses of learned representations, show that our approach consistently outperforms existing baselines.

BibTeX
@inproceedings{ruan-etal-2025-layalign,
    title = "{L}ay{A}lign: Enhancing Multilingual Reasoning in Large Language Models via Layer-Wise Adaptive Fusion and Alignment Strategy",
    author = "Ruan, Zhiwen  and
      Li, Yixia  and
      Zhu, He  and
      Wang, Longyue  and
      Luo, Weihua  and
      Zhang, Kaifu  and
      Chen, Yun  and
      Chen, Guanhua",
    editor = "Chiruzzo, Luis  and
      Ritter, Alan  and
      Wang, Lu",
    booktitle = "Findings of the Association for Computational Linguistics: NAACL 2025",
    month = apr,
    year = "2025",
    address = "Albuquerque, New Mexico",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.findings-naacl.81/",
    pages = "1481--1495",
    ISBN = "979-8-89176-195-7"
}
LayAlign: Enhancing Multilingual Reasoning in Large Language Models via Layer-Wise Adaptive Fusion and Alignment Strategy · NAACL 2025