Linguistic Neuron Overlap Patterns to Facilitate Cross-lingual Transfer on Low-resource Languages

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xu, Yuemei, Xu, Kexin, Zhou, Jian, Hu, Ling, Gui, Lin
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915507842580480
author Xu, Yuemei
Xu, Kexin
Zhou, Jian
Hu, Ling
Gui, Lin
author_facet Xu, Yuemei
Xu, Kexin
Zhou, Jian
Hu, Ling
Gui, Lin
contents The current Large Language Models (LLMs) face significant challenges in improving their performance on low-resource languages and urgently need data-efficient methods without costly fine-tuning. From the perspective of language-bridge, we propose a simple yet effective method, namely BridgeX-ICL, to improve the zero-shot Cross-lingual In-Context Learning (X-ICL) for low-resource languages. Unlike existing works focusing on language-specific neurons, BridgeX-ICL explores whether sharing neurons can improve cross-lingual performance in LLMs. We construct neuron probe data from the ground-truth MUSE bilingual dictionaries, and define a subset of language overlap neurons accordingly to ensure full activation of these anchored neurons. Subsequently, we propose an HSIC-based metric to quantify LLMs' internal linguistic spectrum based on overlapping neurons, guiding optimal bridge selection. The experiments conducted on 4 cross-lingual tasks and 15 language pairs from 7 diverse families, covering both high-low and moderate-low pairs, validate the effectiveness of BridgeX-ICL and offer empirical insights into the underlying multilingual mechanisms of LLMs. The code is publicly available at https://github.com/xuyuemei/BridgeX-ICL.
format Preprint
id arxiv_https___arxiv_org_abs_2508_17078
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Linguistic Neuron Overlap Patterns to Facilitate Cross-lingual Transfer on Low-resource Languages
Xu, Yuemei
Xu, Kexin
Zhou, Jian
Hu, Ling
Gui, Lin
Computation and Language
Artificial Intelligence
The current Large Language Models (LLMs) face significant challenges in improving their performance on low-resource languages and urgently need data-efficient methods without costly fine-tuning. From the perspective of language-bridge, we propose a simple yet effective method, namely BridgeX-ICL, to improve the zero-shot Cross-lingual In-Context Learning (X-ICL) for low-resource languages. Unlike existing works focusing on language-specific neurons, BridgeX-ICL explores whether sharing neurons can improve cross-lingual performance in LLMs. We construct neuron probe data from the ground-truth MUSE bilingual dictionaries, and define a subset of language overlap neurons accordingly to ensure full activation of these anchored neurons. Subsequently, we propose an HSIC-based metric to quantify LLMs' internal linguistic spectrum based on overlapping neurons, guiding optimal bridge selection. The experiments conducted on 4 cross-lingual tasks and 15 language pairs from 7 diverse families, covering both high-low and moderate-low pairs, validate the effectiveness of BridgeX-ICL and offer empirical insights into the underlying multilingual mechanisms of LLMs. The code is publicly available at https://github.com/xuyuemei/BridgeX-ICL.
title Linguistic Neuron Overlap Patterns to Facilitate Cross-lingual Transfer on Low-resource Languages
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2508.17078