FLAVA

Structures found in this family (2)

Linear Subspace, Platonic Representation Hypothesis

Hypotheses argued for by this family (1)

Platonic Representation Hypothesis

By model (1)

FLAVA

Linear Subspace, Platonic Representation Hypothesis

Observations (2)

Independent multimodal models share one orthogonal map across both modalities

Representational similarity analysis shows 15 language and vision-language models align more strongly with brain regions whose activity is consistent across sentence, word-cloud, and image presentations of the same concept than with less modality-consistent regions

Papers

Canonicalizing Multimodal Contrastive Representation Learning (2026), Language models align with brain regions that represent concepts across modalities (2025)