Sparse Autoencoder-based Framework for Steering Semantic Features

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

• Sparse Autoencoder Featuresconcept0.837
  Used in Anthropic welfare assessment to identify performative behavior and hidden emotional struggle co-activating features
• Sparse Autoencoderframework0.830
  Interpretability framework used to decompose layer-40 activations into sparse feature sets for studying emotional alignment and persistence
• Sparse Autoencoders Find Highly Interpretable Features in Language Models (Cunningham et al., 2023)concept0.808
  Core methodology paper for SAE-based interpretable feature extraction
• Sparse autoencoders produce interpretable features for large models.claim0.803
  Central claim of the paper: the method scales to state-of-the-art transformers.
• Sparse Autoencoder for Dictionary Learningframework0.796
  Primary method introduced: trains a one-hidden-layer MLP with L1 sparsity penalty to decompose model activations into overcomplete feature dictionaries
• Training the sparse autoencoder on more data makes features subjectively sharper and more interpretableclaim0.789
  Empirical principle discovered during autoencoder training; led to using 8 billion training points
• Sparse Autoencoders (SAE) activation-based paradigmframework0.780
  Standard interpretability approach that VPD critiques and proposes an alternative to.
• Sparse autoencoders don't provide a comprehensive solution because they decode activations, not parametersclaim0.771
  Critique of activation-based interpretability methods.