Revisiting Multilingual Data Mixtures in Language Model Pretraining

• URL: http://arxiv.org/abs/2510.25947v1
• Date: Wed, 29 Oct 2025 20:46:03 GMT
• Title: Revisiting Multilingual Data Mixtures in Language Model Pretraining
• Authors: Negar Foroutan, Paul Teiletche, Ayush Kumar Tarun, Antoine Bosselut,
• Abstract summary: We study the impact of different multilingual data mixtures in pretraining large language models.<n>We find that combining English and multilingual data does not necessarily degrade the in-language performance of either group.<n>We do not observe a significant "curse of multilinguality" as the number of training languages increases.
• Score: 20.282622416939997
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The impact of different multilingual data mixtures in pretraining large language models (LLMs) has been a topic of ongoing debate, often raising concerns about potential trade-offs between language coverage and model performance (i.e., the curse of multilinguality). In this work, we investigate these assumptions by training 1.1B and 3B parameter LLMs on diverse multilingual corpora, varying the number of languages from 25 to 400. Our study challenges common beliefs surrounding multilingual training. First, we find that combining English and multilingual data does not necessarily degrade the in-language performance of either group, provided that languages have a sufficient number of tokens included in the pretraining corpus. Second, we observe that using English as a pivot language (i.e., a high-resource language that serves as a catalyst for multilingual generalization) yields benefits across language families, and contrary to expectations, selecting a pivot language from within a specific family does not consistently improve performance for languages within that family. Lastly, we do not observe a significant "curse of multilinguality" as the number of training languages increases in models at this scale. Our findings suggest that multilingual data, when balanced appropriately, can enhance language model capabilities without compromising performance, even in low-resource settings

Related papers

• Assessing the Role of Data Quality in Training Bilingual Language Models [17.603371705571107]
  We show that unequal data quality, not just data quantity, is a major driver of performance degradation in bilingual settings.<n>We propose a simple yet effective data filtering strategy to select higher-quality bilingual training data with only high quality English data.
  arXiv  Detail & Related papers  (2025-06-15T21:08:51Z)
• Breaking the Curse of Multilinguality with Cross-lingual Expert Language Models [110.10545153845051]
  Cross-lingual Expert Language Models (X-ELM) is a process that specializes X-ELMs to different languages while remaining effective as a multilingual ensemble. X-ELM provides additional benefits over performance improvements: new experts can be iteratively added, adapting X-ELM to new languages without catastrophic forgetting.
  arXiv  Detail & Related papers  (2024-01-19T01:07:50Z)
• Turning English-centric LLMs Into Polyglots: How Much Multilinguality Is Needed? [40.13166574854085]
  We investigate the minimal amount of multilinguality required to elicit cross-lingual generalisation in English-centric large language models. We find that multilingual instruction tuning with as few as two to three languages is both necessary and sufficient to elicit effective cross-lingual generalisation.
  arXiv  Detail & Related papers  (2023-12-20T00:49:52Z)
• When Is Multilinguality a Curse? Language Modeling for 250 High- and Low-Resource Languages [25.52470575274251]
  We pre-train over 10,000 monolingual and multilingual language models for over 250 languages. We find that in moderation, adding multilingual data improves low-resource language modeling performance. As dataset sizes increase, adding multilingual data begins to hurt performance for both low-resource and high-resource languages.
  arXiv  Detail & Related papers  (2023-11-15T18:47:42Z)
• PolyLM: An Open Source Polyglot Large Language Model [57.64420154135178]
  We present PolyLM, a multilingual large language model (LLMs) trained on 640 billion (B) tokens, avaliable in two model sizes: 1.7B and 13B. To enhance its multilingual capabilities, we 1) integrate bilingual data into training data; and 2) adopt a curriculum learning strategy that increases the proportion of non-English data from 30% in the first stage to 60% in the final stage during pre-training. Further, we propose a multilingual self-instruct method which automatically generates 132.7K diverse multilingual instructions for model fine-tuning.
  arXiv  Detail & Related papers  (2023-07-12T09:00:37Z)
• Analyzing the Mono- and Cross-Lingual Pretraining Dynamics of Multilingual Language Models [73.11488464916668]
  This study investigates the dynamics of the multilingual pretraining process. We probe checkpoints taken from throughout XLM-R pretraining, using a suite of linguistic tasks. Our analysis shows that the model achieves high in-language performance early on, with lower-level linguistic skills acquired before more complex ones.
  arXiv  Detail & Related papers  (2022-05-24T03:35:00Z)
• Discovering Representation Sprachbund For Multilingual Pre-Training [139.05668687865688]
  We generate language representation from multilingual pre-trained models and conduct linguistic analysis. We cluster all the target languages into multiple groups and name each group as a representation sprachbund. Experiments are conducted on cross-lingual benchmarks and significant improvements are achieved compared to strong baselines.
  arXiv  Detail & Related papers  (2021-09-01T09:32:06Z)
• Probing Multilingual Language Models for Discourse [0.0]
  We find that the XLM-RoBERTa family of models consistently show the best performance. Our results also indicate that model distillation may hurt the ability of cross-lingual transfer of sentence representations.
  arXiv  Detail & Related papers  (2021-06-09T06:34:21Z)
• How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models [96.32118305166412]
  We study a set of nine typologically diverse languages with readily available pretrained monolingual models on a set of five diverse monolingual downstream tasks. We find that languages which are adequately represented in the multilingual model's vocabulary exhibit negligible performance decreases over their monolingual counterparts.
  arXiv  Detail & Related papers  (2020-12-31T14:11:00Z)
• Multilingual Translation with Extensible Multilingual Pretraining and Finetuning [77.33262578776291]
  Previous work has demonstrated that machine translation systems can be created by finetuning on bitext. We show that multilingual translation models can be created through multilingual finetuning. We demonstrate that pretrained models can be extended to incorporate additional languages without loss of performance.
  arXiv  Detail & Related papers  (2020-08-02T05:36:55Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.