The paper proposes information scope as a new interpretability axis for SAE features in CLIP and introduces the Contextual Dependency Score to separate local from global scope features, showing they influence model predictions differently.
Towards monosemanticity: Decomposing language mod- els with dictionary learning.Transformer Circuits Thread
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
SPG uses sparse autoencoders to learn guide coefficients that generate normal and anomalous reference vectors, achieving competitive zero-shot anomaly detection and strong segmentation on MVTec AD and VisA without target adaptation.
citing papers explorer
-
Beyond Semantics: Disentangling Information Scope in Sparse Autoencoders for CLIP
The paper proposes information scope as a new interpretability axis for SAE features in CLIP and introduces the Contextual Dependency Score to separate local from global scope features, showing they influence model predictions differently.
-
SPG: Sparse-Projected Guides with Sparse Autoencoders for Zero-Shot Anomaly Detection
SPG uses sparse autoencoders to learn guide coefficients that generate normal and anomalous reference vectors, achieving competitive zero-shot anomaly detection and strong segmentation on MVTec AD and VisA without target adaptation.