Saved in:
Bibliographic Details
Main Authors: Guo, Ping, Ren, Yubing, Liu, Binbin, Liu, Fengze, Lin, Haobin, Zhang, Yifan, Zhang, Bingni, Wang, Taifeng, Zheng, Yin
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.15556
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908546739732480
author Guo, Ping
Ren, Yubing
Liu, Binbin
Liu, Fengze
Lin, Haobin
Zhang, Yifan
Zhang, Bingni
Wang, Taifeng
Zheng, Yin
author_facet Guo, Ping
Ren, Yubing
Liu, Binbin
Liu, Fengze
Lin, Haobin
Zhang, Yifan
Zhang, Bingni
Wang, Taifeng
Zheng, Yin
contents Large language models (LLMs) have become integral to a wide range of applications worldwide, driving an unprecedented global demand for effective multilingual capabilities. Central to achieving robust multilingual performance is the strategic allocation of language proportions within training corpora. However, determining optimal language ratios is highly challenging due to intricate cross-lingual interactions and sensitivity to dataset scale. This paper introduces Climb (Cross-Lingual Interaction-aware Multilingual Balancing), a novel framework designed to systematically optimize multilingual data allocation. At its core, Climb introduces a cross-lingual interaction-aware language ratio, explicitly quantifying each language's effective allocation by capturing inter-language dependencies. Leveraging this ratio, Climb proposes a principled two-step optimization procedure--first equalizing marginal benefits across languages, then maximizing the magnitude of the resulting language allocation vectors--significantly simplifying the inherently complex multilingual optimization problem. Extensive experiments confirm that Climb can accurately measure cross-lingual interactions across various multilingual settings. LLMs trained with Climb-derived proportions consistently achieve state-of-the-art multilingual performance, even achieving competitive performance with open-sourced LLMs trained with more tokens.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15556
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exploring Polyglot Harmony: On Multilingual Data Allocation for Large Language Models Pretraining
Guo, Ping
Ren, Yubing
Liu, Binbin
Liu, Fengze
Lin, Haobin
Zhang, Yifan
Zhang, Bingni
Wang, Taifeng
Zheng, Yin
Computation and Language
Artificial Intelligence
Large language models (LLMs) have become integral to a wide range of applications worldwide, driving an unprecedented global demand for effective multilingual capabilities. Central to achieving robust multilingual performance is the strategic allocation of language proportions within training corpora. However, determining optimal language ratios is highly challenging due to intricate cross-lingual interactions and sensitivity to dataset scale. This paper introduces Climb (Cross-Lingual Interaction-aware Multilingual Balancing), a novel framework designed to systematically optimize multilingual data allocation. At its core, Climb introduces a cross-lingual interaction-aware language ratio, explicitly quantifying each language's effective allocation by capturing inter-language dependencies. Leveraging this ratio, Climb proposes a principled two-step optimization procedure--first equalizing marginal benefits across languages, then maximizing the magnitude of the resulting language allocation vectors--significantly simplifying the inherently complex multilingual optimization problem. Extensive experiments confirm that Climb can accurately measure cross-lingual interactions across various multilingual settings. LLMs trained with Climb-derived proportions consistently achieve state-of-the-art multilingual performance, even achieving competitive performance with open-sourced LLMs trained with more tokens.
title Exploring Polyglot Harmony: On Multilingual Data Allocation for Large Language Models Pretraining
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2509.15556