IC-604CLIP ViT-B/32's CIFAR-10 image embeddings approximately satisfy a multi-cluster structure with near-orthogonal class-mean features

Binghui Li, Zhixuan Pan, Kaifeng Lyu, Jian Li

SourceFeature Averaging: An Implicit Bias of Gradient Descent Leading to Non-Robustness in Neural Networks

The authors use a frozen CLIP ViT-B/32 model to extract image embeddings for CIFAR-10 and verify that the resulting representations approximately satisfy the multi-cluster data distribution assumed in their theoretical analysis. Specifically, the correlation between embeddings of images from the same class is significantly higher than between embeddings from different classes, and the class-mean feature vectors are approximately orthogonal. This observation is used to justify applying their feature-averaging theory to a two-layer classifier trained on CLIP embeddings. The result is presented as a setup validation in Appendix H rather than a main contribution.

Evidence
observational
Caveat
The observation is a setup validation for the authors' experimental design, not a main contribution of the paper. No specific numerical correlation values are printed in the text; the evidence is a qualitative correlation heatmap (Figure 12) over 100 sampled instances per class.
Model
CLIP / CLIP-ViT (LC)
Concepts
Linear representation
Datasets
CIFAR-10 [eval]
Extraction
automatic-extraction