framework
active
framework:superposition-hypothesis

Superposition Hypothesis

Core theoretical framework: neural networks represent more features than neurons by encoding features as directions in superposition

Neighborhood — ranked by edge-count

Thinkers (2)

thinker

Methods (3)

method
  • Sparse dictionary learning variant that reads and writes to multiple layers or models, yielding shared features.
  • General method for finding overcomplete sparse decompositions; the paper uses sparse autoencoders as an approximation
  • Measure of superposition in toy models.

Concepts (9)

concept
  • Polysemanticity
    associated_with
    Neurons that respond to multiple unrelated concepts, limiting interpretability.
  • The idea that features are encoded as directions in activation space.
  • Representation of features spread across multiple layers, complicating dictionary learning.
  • When non-orthogonal features cause logistic regression to identify a suboptimal probe direction
  • Property that features activate on only a small fraction of inputs; enables compressed sensing and is what allows superposition to work
  • Mathematical object defining J-space as union of cones.
  • Specific phrases or sequences memorized via binary features in superposition, enabling narrow pattern matching despite few neurons
  • Mechanism by which superposition works: small neural networks exploit sparsity to approximately simulate much larger sparse networks
  • Overcomplete Basis
    associated_with
    A set of feature directions that is larger than the dimensionality of the space, enabling superposition

Claims (2)

claim

Frameworks (3)

framework
  • Primary method introduced: trains a one-hidden-layer MLP with L1 sparsity penalty to decompose model activations into overcomplete feature dictionaries
  • Compressed Sensing
    associated_with
    Mathematical framework enabling recovery of high-dimensional sparse vectors from low-dimensional projections; theoretically underpins sparse autoencoder approach
  • Disentanglement
    contradicts
    Related research agenda seeking representations that separate conceptually distinct factors; contrasted with superposition approach

Findings (1)

finding

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.