Sparse Crosscoders

Extension of SAEs that jointly learns latents across representations from different models; proposed as alternative for extended fine-tuning

Neighborhood — ranked by edge-count

paper

framework

Sparse Autoencoder
related_to
Interpretability framework used to decompose layer-40 activations into sparse feature sets for studying emotional alignment and persistence

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Sparse Autoencoder Featuresconcept0.827
Used in Anthropic welfare assessment to identify performative behavior and hidden emotional struggle co-activating features
Sparse Autoencoder for Dictionary Learningframework0.810
Primary method introduced: trains a one-hidden-layer MLP with L1 sparsity penalty to decompose model activations into overcomplete feature dictionaries
Sparse and smooth codingconcept0.796
Coding scheme where qualities are represented by few neurons with continuous similarity relations.
Sparse Autoencoders (SAE)method0.795
Interpretability method criticized in this paper for shattering manifolds into atomic pieces, obscuring overarching semantic structure.
TopK Sparse Autoencodersframework0.785
The central mechanistic interpretability tool applied across all three EEG transformers to extract sparse feature dictionaries
Sparse autoencoders don't provide a comprehensive solution because they decode activations, not parametersclaim0.774
Critique of activation-based interpretability methods.
Sparse Coding Hypothesisconcept0.771
The hypothesis that semantic concepts in neural representations can be captured using sparsity priors
Sparse circuitsconcept0.766
A goal in mechanistic interpretability to identify sparse computational subgraphs; VPD promotes sparse parameter circuits.