claim
active
claim:observed-neural-networks-are-noisily-simulating-larger-highly-sparse-networksObserved neural networks are noisily simulating larger, highly sparse networks
Hypothesis about real networks.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Toy Models of Superpositionmentions
Hypotheses (1)
hypothesis
- Central hypothesis.
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.
- The paper's central thesis statement, presented prominently after the abstract
- Mechanism by which superposition works: small neural networks exploit sparsity to approximately simulate much larger sparse networks
- The central hypothesis of the paper; the platonic representation hypothesis itself
- Universality hypothesis to be tested across families.
- Extends convergence argument to brain-machine alignment
- Vision statement in the conclusion.
- Superposition hypothesis: neural networks represent more features than dimensions using almost-orthogonal directions.hypothesis0.759Explanation for why dictionary learning can recover many more features than dimensions.