method
active
method:l0-active-feature-countL0 active-feature count
Measure of features activated by clean vs adversarial inputs.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Adversarial Examples Are Not Bugs, They Are Superpositionintroducesmentions
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.
- Average number of nonzero feature entries per input; primary measure of activation sparsity in the autoencoder
- Using Claude to search for features activating on specific concepts and automated labeling.
- Metric of features discovered.
- Sparsity trade-off.
- Dual interpretation of features: in addition to responding to inputs, features also act to increase probability of specific output tokens
- A feature that responds to only a single latent variable, contrasted with polysemantic features
- Feature represents the 'addition' function abstractly.
- Warm-up derivation.