concept
active
concept:feature-sparsityFeature Sparsity
Property that features activate on only a small fraction of inputs; enables compressed sensing and is what allows superposition to work
Neighborhood — ranked by edge-count
Papers (2)
paper
- Toy Models of Superpositionmentions
Frameworks (1)
framework
- Superposition HypothesissupportsCore theoretical framework: neural networks represent more features than neurons by encoding features as directions in superposition
Concepts (2)
concept
- Superpositionassociated_withPhenomenon where models represent more features than dimensions via almost-orthogonal directions.
- L0 Norm of Feature Activationsassociated_withAverage number of nonzero feature entries per input; primary measure of activation sparsity in the autoencoder
Questions (1)
question
- Open question.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Fraction of a dimension given to a feature.
- Alternative sparsity notion proposed for predictive component.
- The extracted set of sparse interpretable features from model embeddings via SAEs
- Classical literature on sparsity priors over admissible supports.
- Proposed variation designating ~10k of ~250k features as shared with penalty 0.1-0.2x baseline to soak up shared variance.
- Interpretable concept representations hypothesized as basic units of computation.
- Datapoint features form only if there is a repeated datapoint in training and sparsity regularization is appliedhypothesis0.793Two necessary conditions.
- Synthesizing inputs maximizing component activation.