# The Role of Mixed-Language Documents for Multilingual Large Language Model Pretraining

arXiv: https://arxiv.org/abs/2601.00364 · HTML: https://mlsanity.com/paper/2601.00364/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 38 references that are in the map.
- First posted on arXiv 1 Jan 2026.

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [XNLI: Evaluating Cross-lingual Sentence Representations](https://mlsanity.com/paper/1809.05053/) (2018)
- [Bag of Tricks for Efficient Text Classification](https://mlsanity.com/paper/1607.01759/) (2016)
- [On the Cross-lingual Transferability of Monolingual Representations](https://mlsanity.com/paper/1910.11856/) (2019)
- [Beyond English-Centric Multilingual Machine Translation](https://mlsanity.com/paper/2010.11125/) (2020)
- [The FLORES-101 Evaluation Benchmark for Low-Resource and Multilingual Machine Translation](https://mlsanity.com/paper/2106.03193/) (2021)
- [Word Translation Without Parallel Data](https://mlsanity.com/paper/1710.04087/) (2017)
- [Google's Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation](https://mlsanity.com/paper/1611.04558/) (2016)
- [MLQA: Evaluating Cross-lingual Extractive Question Answering](https://mlsanity.com/paper/1910.07475/) (2019)
- [PAWS-X: A Cross-lingual Adversarial Dataset for Paraphrase Identification](https://mlsanity.com/paper/1908.11828/) (2019)
- [InfoXLM: An Information-Theoretic Framework for Cross-Lingual Language Model Pre-Training](https://mlsanity.com/paper/2007.07834/) (2020)
- [Okapi: Instruction-tuned Large Language Models in Multiple Languages with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2307.16039/) (2023)
- [FineWeb2: One Pipeline to Scale Them All -- Adapting Pre-Training Data Processing to Every Language](https://mlsanity.com/paper/2506.20920/) (2025)
- [CoSDA-ML: Multi-Lingual Code-Switching Data Augmentation for Zero-Shot Cross-Lingual NLP](https://mlsanity.com/paper/2006.06402/) (2020)
- [It's All in the Heads: Using Attention Heads as a Baseline for Cross-Lingual Transfer in Commonsense Reasoning](https://mlsanity.com/paper/2106.12066/) (2021)
- [Searching for Needles in a Haystack: On the Role of Incidental Bilingualism in PaLM's Translation Capability](https://mlsanity.com/paper/2305.10266/) (2023)
- [Code-Switching Curriculum Learning for Multilingual Transfer in LLMs](https://mlsanity.com/paper/2411.02460/) (2024)
- [Can LLMs Really Learn to Translate a Low-Resource Language from One Grammar Book?](https://mlsanity.com/paper/2409.19151/) (2024)
- [DICT-MLM: Improved Multilingual Pre-Training using Bilingual Dictionaries](https://mlsanity.com/paper/2010.12566/) (2020)
- [Investigating and Scaling up Code-Switching for Multilingual Language Model Pre-Training](https://mlsanity.com/paper/2504.01801/) (2025)
- [Infini-gram mini: Exact n-gram Search at the Internet Scale with FM-Index](https://mlsanity.com/paper/2506.12229/) (2025)
- [Just Go Parallel: Improving the Multilingual Capabilities of Large Language Models](https://mlsanity.com/paper/2506.13044/) (2025)

## Built on by (most cited first, 1 of 1)

- [The Interlingua Hypothesis: LLMs Translate via a Latent Task-agnostic Feature Space](https://mlsanity.com/paper/2609.00515/) (2026)

## Abstract (arXiv)

Multilingual large language models achieve impressive cross-lingual performance despite largely monolingual pretraining. While bilingual data in pretraining corpora is widely believed to enable these abilities, details of its contributions remain unclear. We investigate this question by pretraining models from scratch under controlled conditions, comparing the standard web corpus with a monolingual-only version that removes all multilingual documents. Despite constituting only 2% of the corpus, removing bilingual data causes translation performance to drop 56% in BLEU, while behaviour on cross-lingual QA and general reasoning tasks remains stable, with training curves largely overlapping the baseline. To understand this asymmetry, we categorize bilingual data into parallel (14%), code-switching (72%), and miscellaneous documents (14%) based on the semantic relevance of content in different languages. We then conduct granular ablations by reintroducing parallel or code-switching data into the monolingual-only corpus. Our experiments reveal that parallel data almost fully restores translation performance (91% of the unfiltered baseline), whereas code-switching contributes minimally. Other cross-lingual tasks remain largely unaffected by either type. These findings reveal that translation critically depends on systematic token-level alignments from parallel data, whereas cross-lingual understanding and reasoning appear to be achievable even without bilingual data.