hypothesis
active
hypothesis:if-curvature-measures-memorization-eigenvector-bands-will-activate-differently-on-memorized-vs-clean-dataIf curvature measures memorization, eigenvector bands will activate differently on memorized vs clean data
Section 4 hypothesis.
Source paper
extracted_from(2025) · Merullo, Jack · Vatsavaya, Srihita · Bushnaq, Lucius · Lewis, Owen
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (3)
finding
- At OLMo-7B layer 22 MLP input, bottom 50% K-FAC eigenvectors activate 23.1% more on memorized data; top 10% activate 26% more on cleanassociated_withsupportsDisentanglement evidence.
- App. H.1.
- Vision disentanglement.
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.
- Section 4.
- Consistent with prior work.
- K-FAC eigenvector bands show sorted activation strength top10%>10-25%>25-50%>bottom50%, unlike SVDfinding0.768Supports curvature basis interpretability.
- Section 6.
- Basis for editing.
- Key quote connecting path redundancy to interferometric information encoding.
- Ringing evidence.
- Motivates Jacobian connection.