framework
active
framework:topk-saeTopK SAE
Evaluated architecture (Gao et al.).
Neighborhood — ranked by edge-count
Papers (5)
paper
- Do Sparse Autoencoders Capture Concept Manifolds?citesmentions
Frameworks (2)
framework
- BatchTopK SAEextendsEvaluated architecture.
- ReLU SAEextendsBaseline SAE architecture.
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.
- Sparse dictionary learning method used to extract interpretable features from EEG transformer embeddings.
- Hierarchical architecture that excels on disentanglement and scales positively.
- Hybrid SAE jointly training turn-averaged inner and per-token outer partitions.
- The individual, supposedly monosemantic directions learned by SAEs; argued here to fragment manifolds into disconnected pieces.
- Internal concepts a model uses, extracted as human-interpretable components.
- Architecture compared.
- Standard baseline SAE operating on individual token activations.
- Hybrid with separate turn-averaged and residual per-token SAEs.