AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

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

Yanlin Chen, Tang Li, Xi Peng

ECCV 2026 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
steer with CLT instance concept circuit (ablate background-specific concepts)CLIP-ViT-B/32 linear probe · Waterbirds (waterbirds on land, worst group)accuracy: 0.48 -> 0.71Improved
CLT steering costs accuracy on majority groupCLIP-ViT-B/32 linear probe · Waterbirds (landbirds on land)accuracy: 0.98 -> 0.95Hurt
ablate randomly chosen concepts instead of circuit-identified ones (ablation)CLIP-ViT-B/32 linear probe · Waterbirds (waterbirds on land, worst group)accuracy: 0.48 -> 0.22Hurt
add co-activation reweighting to global weights (ablation)global concept circuit · ImageNet val (50 classes), top-10 concept mean-ablationaccuracy after ablation: 0.48 -> 0.22 (CLIP)Improved
use global weights rather than co-activation statistics alone (ablation)global concept circuit · ImageNet val (50 classes), top-10 concept mean-ablationaccuracy after ablation: 0.72 -> 0.62 (DINO)Improved
CLT instead of SAE for instance concept circuitsCLIP instance circuits · ImageNet val (50 classes)faithfulness and completeness: +24.4% faithfulness, +22.1% completeness (CLIP, average)Improved
increase CLT L0 sparsity and dictionary size (ablation)CLIP CLT · ImageNet train (hyperparameter sweep)explained variance vs monosemanticityMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy