concept
active
concept:feature-absorption

Feature Absorption

Phenomenon where sparsity yields gerrymandered latents for hierarchical concepts.

Neighborhood — ranked by edge-count

Methods (1)

method

Concepts (2)

concept

Questions (1)

question

Related by similarity (8)

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.

  • Synthesizing inputs maximizing component activation.
  • Featuresconcept0.773
    Interpretable concept representations hypothesized as basic units of computation.
  • Functionally significant directions in the residual stream that organize activations in mechanistic interpretability
  • When non-orthogonal features cause logistic regression to identify a suboptimal probe direction
  • Shared featuresconcept0.763
    Features with similar decoder magnitude across both models.
  • Error featuresconcept0.762
    Extra feature representing unexplained error to make isomorphism exact.
  • Domain of techniques for constructing informative features from raw data; covariance pooling is a feature engineering method for token sequences.
  • Feature utilityconcept0.761
    Frequency-weighted spectral value πkλkj determining which features are retained.