Memorization in Superposition

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

• Superpositionconcept0.842
  Phenomenon where models represent more features than dimensions via almost-orthogonal directions.
• Superposition in Neural Networksconcept0.817
  Theoretical model of how neural networks encode more features than dimensions, informing linear representation work.
• Superposition of Simulacraconcept0.800
  The state in which a dialogue agent maintains multiple possible characters simultaneously, refined as the conversation proceeds
• Simulacra in Superposition Frameworkframework0.772
  The more nuanced second metaphor: LLM as simulator maintaining a superposition of possible simulacra across a multiverse of characters
• Superposition is in some sense deliberate: the model converts pure neurons into polysemantic neurons to store more features in fewer neurons.claim0.770
  Interpretation of the cars-in-superposition circuit finding as an intentional representational strategy
• Superposition of Sparse Featuresconcept0.763
  Mechanistic finding by Bricken et al. 2023 about how LLMs store features; cited as operational justification for pattern-repository assumption
• Superposition in Residual Streamconcept0.762
  The phenomenon where the residual stream communicates many more features than its dimensionality by encoding information across overlapping subspaces
• Cross-layer superposition is a fundamental challenge for dictionary learning.claim0.749
  Features smeared across layers cannot be fully disentangled by SAE on a single residual stream.