AWhat changed
Learns concept directions from positive and counterexample activations, measures output directional derivatives along these directions and aggregates their signs into class-level concept sensitivity.
Scientific scope
Learns concept directions from positive and counterexample activations, measures output directional derivatives along these directions and aggregates their signs into class-level concept sensitivity. Depends on selected concepts, negatives, layer and local linear separator; statistical repetition is not proof that a semantic concept causally determines predictions. V1 has five authors; v5 has seven with changed order and membership. Only general image-model method is curated; medical application excluded.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Depends on selected concepts, negatives, layer and local linear separator; statistical repetition is not proof that a semantic concept causally determines predictions. V1 has five authors; v5 has seven with changed order and membership. Only general image-model method is curated; medical application excluded.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- TCAV: Relative concept importance testing with Linear Concept Activation Vectors
- Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction