paper-with-me

Papers

Middle-Layer Representation Alignment for Cross-Lingual Transfer in Fine-Tuned LLMs

2025-02-20 · Danni Liu, Jan Niehues

While large language models demonstrate remarkable capabilities at task-specific applications through fine-tuning, extending these benefits across diverse languages is essential for broad accessibility. However, effective cross-lingual transfer is hindered by LLM performance gaps across languages and the scarcity of fine-tuning data in many languages. Through analysis of LLM internal representations from over 1,000+ language pairs, we discover that middle layers exhibit the strongest potential for cross-lingual alignment. Building on this finding, we propose a middle-layer alignment objective integrated into task-specific training. Our experiments on slot filling, machine translation, and structured text generation show consistent improvements in cross-lingual transfer, especially to lower-resource languages. The method is robust to the choice of alignment languages and generalizes to languages unseen during alignment. Furthermore, we show that separately trained alignment modules can be merged with existing task-specific modules, improving cross-lingual capabilities without full re-training. Our code is publicly available (https://github.com/dannigt/mid-align).

📄 PDF Abstract BibTeX arXiv:2502.14830

Code (1)

dannigt/mid-align 공식 구현 jax

Tasks

Cross-Lingual TransferMachine Translationslot-fillingSlot FillingText Generation

Similar Papers 제목 키워드 기반

Multilingual Routing in Mixture-of-Experts

2025-10-06 · Lucas Bandarkar, Chenyuan Yang, Mohsen Fayyaz, Junlin Hu 외 arxiv

Mixture-of-Experts (MoE) architectures have become the key to scaling modern LLMs, yet little is understood about how their sparse routing dynamics respond to multilingual data. In this work, we analyze expert routing pa…

Analyzing LLMs' Knowledge Boundary Cognition Across Languages Through the Lens of Internal Representations

2025-04-18 · Chenghao Xiao, Hou Pong Chan, Hao Zhang, Mahani Aljunied 외

While understanding the knowledge boundaries of LLMs is crucial to prevent hallucination, research on knowledge boundaries of LLMs has predominantly focused on English. In this work, we present the first study to analyze…

Hallucination

Enhancing Non-English Capabilities of English-Centric Large Language Models through Deep Supervision Fine-Tuning

2025-03-03 · Wenshuai Huo, Xiaocheng Feng, Yichong Huang, Chengpeng Fu 외

Large language models (LLMs) have demonstrated significant progress in multilingual language understanding and generation. However, due to the imbalance in training data, their capabilities in non-English languages are l…

SSHR: Leveraging Self-supervised Hierarchical Representations for Multilingual Automatic Speech Recognition

2023-09-29 · Hongfei Xue, Qijie Shao, Kaixun Huang, Peikun Chen 외

Multilingual automatic speech recognition (ASR) systems have garnered attention for their potential to extend language coverage globally. While self-supervised learning (SSL) models, like MMS, have demonstrated their eff…

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Self-Supervised Learningspeech-recognition+1

Improving Cross-Lingual Word Embeddings by Meeting in the Middle

2018-08-27 · EMNLP 2018 10 · Yerai Doval, Jose Camacho-Collados, Luis Espinosa-Anke, Steven Schockaert

Cross-lingual word embeddings are becoming increasingly important in multilingual NLP. Recently, it has been shown that these embeddings can be effectively learned by aligning two disjoint monolingual vector spaces throu…

Cross-Lingual Word EmbeddingsMultilingual NLPWord Embeddings