# Brain-LLM Alignment Tracks Training Data, Not Typology

arXiv: https://arxiv.org/abs/2605.23032 · HTML: https://mlsanity.com/paper/2605.23032/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 25 of its 73 references that are in the map.
- First posted on arXiv 21 May 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [LLM preference optimization methods](https://mlsanity.com/alignment/llm-preference-optimization-methods/)

## Builds on (most cited first, 25 of 25)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [Cross-lingual Language Model Pretraining](https://mlsanity.com/paper/1901.07291/) (2019)
- [How multilingual is Multilingual BERT?](https://mlsanity.com/paper/1906.01502/) (2019)
- [How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models](https://mlsanity.com/paper/2012.15613/) (2020)
- [Emerging Cross-lingual Structure in Pretrained Language Models](https://mlsanity.com/paper/1911.01464/) (2019)
- [Embers of Autoregression: Understanding Large Language Models Through the Problem They are Trained to Solve](https://mlsanity.com/paper/2309.13638/) (2023)
- [Interpreting and improving natural-language processing (in machines) with natural language-processing (in the brain)](https://mlsanity.com/paper/1905.11833/) (2019)
- [Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times?](https://mlsanity.com/paper/2212.12131/) (2022)
- [Finding Universal Grammatical Relations in Multilingual BERT](https://mlsanity.com/paper/2005.04511/) (2020)
- [On the Language Neutrality of Pre-trained Multilingual Representations](https://mlsanity.com/paper/2004.05160/) (2020)
- [Testing the Predictions of Surprisal Theory in 11 Languages](https://mlsanity.com/paper/2307.03667/) (2023)
- [Scaling laws for language encoding models in fMRI](https://mlsanity.com/paper/2305.11863/) (2023)
- [First Align, then Predict: Understanding the Cross-Lingual Ability of Multilingual BERT](https://mlsanity.com/paper/2101.11109/) (2021)
- [Language Models Learn Rare Phenomena from Less Rare Phenomena: The Case of the Missing AANNs](https://mlsanity.com/paper/2403.19827/) (2024)
- [Modeling Language Variation and Universals: A Survey on Typological Linguistics for Natural Language Processing](https://mlsanity.com/paper/1807.00914/) (2018)
- [Are All Languages Equally Hard to Language-Model?](https://mlsanity.com/paper/1806.03743/) (2018)
- [From Language to Cognition: How LLMs Outgrow the Human Language Network](https://mlsanity.com/paper/2503.01830/) (2025)
- [Same Neurons, Different Languages: Probing Morphosyntax in Multilingual Pre-trained Models](https://mlsanity.com/paper/2205.02023/) (2022)
- [Language models and brains align due to more than next-word prediction and word-level information](https://mlsanity.com/paper/2212.00596/) (2022)
- [Information-Restricted Neural Language Models Reveal Different Brain Regions' Sensitivity to Semantics, Syntax and Context](https://mlsanity.com/paper/2302.14389/) (2023)
- [ReasonIF: Large Reasoning Models Fail to Follow Instructions During Reasoning](https://mlsanity.com/paper/2510.15211/) (2025)
- [Robust Evaluation of Language-Brain Encoding Experiments](https://mlsanity.com/paper/1904.02547/) (2019)
- [Meaning-infused grammar: Gradient Acceptability Shapes the Geometric Representations of Constructions in LLMs](https://mlsanity.com/paper/2507.22286/) (2025)

## Abstract (arXiv)

Brain-LLM alignment is well established in English, yet the brain's language network is neuroanatomically universal across languages. Does alignment also generalize cross-linguistically, and what governs the variation? We test this using fMRI data from 112 participants across English, Chinese, and French (the Le Petit Prince corpus) and seven LLMs spanning English-dominant, Chinese-dominant, and multilingual architectures. Our central finding is that training-language dominance, not an inherent property of English, drives the alignment pattern: a Chinese-dominant model (Baichuan2-7B), architecture-matched to LLaMA-2-7B, reverses the gradient entirely, aligning best with Chinese brains and worst with English. Beyond training dominance, formal typological distance independently covaries with alignment degradation, syntax-associated brain regions (IFG) show 2.3× steeper typological gradients than lexico-semantic regions (PTL), and tokenization fertility accounts for ~60% of a cross-linguistic shift in optimal encoding layer. These results reveal that the apparent "English advantage" in brain-LLM alignment is an artifact of training data composition, while the remaining variation reflects genuine typological structure concentrated in syntactic processing.