hypothesis
active
hypothesis:if-bold-is-split-into-context-specific-features-performance-is-capped-unless-all-such-features-are-aggregatedIf bold is split into context-specific features, performance is capped unless all such features are aggregated
Feature-splitting hypothesis.
Source paper
extracted_from(2026) · Leon Bergen · Usha Bhalla · Sidharth Baskaran · Max Loeffler +13
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Hypothesized explanation.
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.
- Limitation of single-prompt selection.
- Main proposal.
- Motivation for using sparsity-based dictionary learning on language models
- Motivates nested SAE.
- Discovery story.
- Supports monosemantic reading.
- Rotation-invariance argument.
- Figure 5.