Sparse interpretability in neural networks

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

• Dictionary Learning for Neural Network Interpretabilitymethod0.797
  Bricken et al.'s method for decomposing language models into interpretable features; cited as AI alignment interpretability relevant to consciousness detection
• Sparse Autoencoders Find Highly Interpretable Features in Language Models (Cunningham et al., 2023)concept0.774
  Core methodology paper for SAE-based interpretable feature extraction
• Sparse autoencoders produce interpretable features for large models.claim0.773
  Central claim of the paper: the method scales to state-of-the-art transformers.
• Can an interpretable symbolic algorithm be used to faithfully explain a complex neural network model?question0.765
  Framing question for the paper's research program.
• Noisy Simulation of Sparse Networksconcept0.763
  Mechanism by which superposition works: small neural networks exploit sparsity to approximately simulate much larger sparse networks
• Sparse autoencoders extract features that are significantly more monosemantic than neurons, as shown by four independent lines of evidenceclaim0.762
  Central claim of the paper, supported by detailed feature analysis, human evaluation, automated interpretability of activations, and automated interpretability of logit weights
• Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models (Marks et al., 2025)concept0.761
  Cited as enabling precise behavioral control through SAE features, extending the same methodological line
• Sparse and smooth codingconcept0.759
  Coding scheme where qualities are represented by few neurons with continuous similarity relations.