English

Enhancing LLM Language Adaption through Cross-lingual In-Context Pre-training

Computation and Language 2025-09-22 v2

Abstract

Large language models (LLMs) exhibit remarkable multilingual capabilities despite English-dominated pre-training, attributed to cross-lingual mechanisms during pre-training. Existing methods for enhancing cross-lingual transfer remain constrained by parallel resources, suffering from limited linguistic and domain coverage. We propose Cross-lingual In-context Pre-training (CrossIC-PT), a simple and scalable approach that enhances cross-lingual transfer by leveraging semantically related bilingual texts via simple next-word prediction. We construct CrossIC-PT samples by interleaving semantic-related bilingual Wikipedia documents into a single context window. To access window size constraints, we implement a systematic segmentation policy to split long bilingual document pairs into chunks while adjusting the sliding window mechanism to preserve contextual coherence. We further extend data availability through a semantic retrieval framework to construct CrossIC-PT samples from web-crawled corpus. Experimental results demonstrate that CrossIC-PT improves multilingual performance on three models (Llama-3.1-8B, Qwen2.5-7B, and Qwen2.5-1.5B) across six target languages, yielding performance gains of 3.79%, 3.99%, and 1.95%, respectively, with additional improvements after data augmentation.

Keywords

Cite

@article{arxiv.2504.20484,
  title  = {Enhancing LLM Language Adaption through Cross-lingual In-Context Pre-training},
  author = {Linjuan Wu and Haoran Wei and Huan Lin and Tianhao Li and Baosong Yang and Fei Huang and Weiming Lu},
  journal= {arXiv preprint arXiv:2504.20484},
  year   = {2025}
}

Comments

12 pages, 6 figures, EMNLP 2025

R2 v1 2026-06-28T23:14:52.186Z