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.
Indoor segmentation and support inference from rgbd images
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
An extension of FamNet with extra loss achieves 1.96 MAE when counting machine parts, outperforming traditional image processing, instance segmentation, and standard density estimation baselines.
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.
-
Counting Machine Parts
An extension of FamNet with extra loss achieves 1.96 MAE when counting machine parts, outperforming traditional image processing, instance segmentation, and standard density estimation baselines.