concept
active
concept:semantic-structuresemantic structure
The meaningful organization of concepts in a model's representation space, claimed to be better captured by manifolds than by SAEs.
Neighborhood — ranked by edge-count
Claims (2)
claim
- Core critique of sparse autoencoders: they break the geometric structure of representations, making it harder to see the big picture.
- Positive claim that geometric descriptions retain the conceptual coherence lost in atomized feature decompositions.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Conventional programming constructs like variables, arrays; claimed unnecessary for Elephant programs.
- Meaning that arises from relations within the graphical system, not inherent in elements.
- The actual computational operations a model performs, which the paper argues need not mirror representational structure
- The central question of whether representational geometry implies corresponding computational structure
- Central mechanism in denotational design: precise mathematical meaning assigned to types and expressions, independent of implementation.
- Central claim that meaning emerges from spatial positioning and relational organization of elements on a page.
- Question about the inner structure of the plenum, leading to the answer that it is pure unity without structure.