SOURCE-LINKED INTELLIGENCE

"World Knowledge" in the Weights: Reading Concept Circuits of Vision Transformers

arXiv · AI, language, vision and robotics · article · Sep 8, 2026 · UTC

Vision transformers (ViTs) have achieved remarkable generalization across visual domains, yet little is known about how they internally represent the structure of the world. To address this gap, we use Cross-Layer Transcoders (CLTs) to read concept circuits from ViTs: directed graphs whose nodes correspond to sparse, interpretable concepts and edges capture concept interactions across layers. Our method yields two complementary views of model behavior. The global concept circuit is input-invariant and can be recovered directly from learned cross-layer weights, exposing the reusable "world know

Read original source ↗ Open in workspace

recordType

paper

region

Global