A low-rank subspace of LLM representations carries most of their brain alignment
A Dartmouth preprint studied which language-model features predict human neural responses during naturalistic listening. Across three intracranial/fMRI datasets and six model conditions, reduced-rank regression recovered full linear encoding performance with about 1–3% of embedding dimensions. These directions differed from the embeddings’ principal variance axes and transferred across held-out participants, less strongly across datasets. This summary is limited to the abstract and affiliation/date metadata of the October 29, 2026 preprint. It concerns measured brain–model alignment, not proof that brains use the same representations.
Technology: Autoregressive and masked language models with ridge reduced-rank regression
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: retrospective or benchmark
- Tested: Linear encoding of observed intracranial/fMRI responses, reduced-rank analysis and transfer across held-out participants.
- Still unknown: Source checking does not establish independent confirmation. Review date: 2026-11-15.
- AI role: LLM representations are the predictors of neural activity; reduced-rank regression identifies the brain-predictive subspace.
Evidence sources: https://arxiv.org/abs/2610.40511 · https://arxiv.org/html/2610.40511v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.