claim
active
claim:fuzzy-induction-can-be-understood-as-performing-induction-in-the-feature-basisFuzzy induction can be understood as performing induction in the feature basis
Central interpretation of induction case study.
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (3)
finding
- Fuzziness lives in QK.
- Layer-dependent induction on Haiku 3.5.
- Fuzzy induction heads match words across languages, antonyms, structural positions and kinship termssupportsVariety of fuzzy matching found.
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.
- Induction performed in the feature basis rather than token space.
- Fuzzy poetry induction head OV circuit is identical to literal induction (newline value to 'say \n' output)finding0.777Copying behavior unchanged.
- Interpretation of attributions.
- Emerging framework that seeks invariants between evolution and learning; cited as future direction.
- The mechanistic explanation of how induction heads are implemented in two-layer models
- Central empirical claim of the paper; induction heads are shown to be the mechanism for powerful in-context learning
- Quantitative relationship between concept frequency and feature presence.