Probing the Geometry of Color Representations in Multilingual Language Models

Key points

• We decompose color representations into two geometric axes: (P1) alignment to perceptual color space, and (P2) cross-lingual symmetry.1
• Analyzing layerwise embeddings across models and multilingual datasets, we find that perceptual grounding is weak: alignment to CIE-LAB peaks in intermediate layers, then collapses toward output layers.1
• Functionally, we show that geometric properties predict task performance: models achieve 96.7% accuracy on color classification at layers where P1 alignment peaks, with P1 alignment (r=0.70) significantly correlating with downstream capabilities.1
• Our results reveal a dissociation: text-only LMs learn universal color-language geometry, not universal color-perception geometry, providing diagnostics for where perceptual structure is preserved in language models.1

Sources (1)

• [1]Probing the Geometry of Color Representations in Multilingual Language Models
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 01:06 PM

  Do multilingual language models learn a universal internal geometry for color, and is it perceptually grounded or primarily linguistic?

  We decompose color representations into two geometric axes: (P1) alignment to perceptual color space, and (P2) cross-lingual symmetry.

Extractive summary: sentences quoted from the sources.