eLife· 2026Q1
Large language models poorly capture how word order shapes representations of sentence meaning in the human brain
- 1citations
- Q1SCImago
- 2026year
Short summary
Transformer-based large language models (LLMs) fail to accurately represent sentence meaning as encoded by the human brain, performing worse than models that explicitly encode word order structure.
AI-generated from the title and abstract; the full text is not read.
Key points
- Transformer LLMs do not accurately represent sentence meaning as encoded by the human brain.
- LLMs perform worse than models that explicitly encode structural relations between words.
- Analysis used 7T fMRI data from 30 participants reading 108 carefully designed sentences.
- Results highlight the critical role of sentence structure in human brain representations of meaning.
AI-generated from the title and abstract; the full text is not read.
Abstract
Abstract Large language models based on the transformer architecture are now capable of producing human-like language. But do they encode and process linguistic meaning in a human-like way? Here, we address this question by analysing 7T fMRI data from 30 participants reading 108 sentences each. These sentences are carefully designed to disentangle sentence structure from word meaning, thereby testing whether transformer representations of sentence meaning resemble those formed by the brain. We found that while transformer models match brain representations better than models that completely ignore word order, all transformer models performed poorly overall. Further, transformers were significantly inferior to models explicitly designed to encode the structural relations between words. Our results provide insight into the nature of sentence representation in the brain, highlighting the critical role of sentence structure. They also cast doubt on the claim that transformers represent sentence meaning similarly to the human brain.
The authors' abstract, as published at the source. eLife, 2026 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Cognitive Neuroscience
Cognitive NeuroscienceNeuroscience