method
active
method:umap-visualization-for-features

UMAP visualization for features

Dimensionality reduction of SAE decoder vectors to create interactive feature maps.

Neighborhood — ranked by edge-count

Methods (1)

method
  • 2D embedding of feature direction vectors used to visualize feature clusters and splitting geometry

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.

  • Method of optimizing input to cause a neuron to fire maximally, used to characterize what a neuron detects; establishes causal link
  • Used to visualize LLM true/false representations, revealing clear linear structure separating true from false statements
  • Used to visually inspect separation of truth-related directions in model activation space across layers
  • Metaphor treating each system feature or function as a separate application that can be independently loaded and managed.
  • The central object of study — the idea that a concept like truth is encoded as a direction in the LLM's latent space
  • Context Featureconcept0.711
    Feature that activates across all tokens within a specific context (e.g., DNA sequences, base64 strings)
  • Action Featuresconcept0.710
    Dual interpretation of features: in addition to responding to inputs, features also act to increase probability of specific output tokens
  • Pure Featureconcept0.709
    A feature that responds to only a single latent variable, contrasted with polysemantic features