finding
active
finding:if-d-is-incoherent-1-2k-1-and-s-spans-the-subspace-an-idealized-sparse-decoder-captures-m-at-precision-o-theorem-1If D is µ-incoherent, µ<1/(2k−1), and S* spans the subspace, an idealized sparse decoder captures M at precision O(λ) (Theorem 1)
Formal subspace recovery guarantee.
Source paper
extracted_from(2026) · Usha Bhalla · Thomas Fel · Can Rager · Sheridan Feucht +8
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Theoretical framework answer.
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.
- Interpretation of Experiment 4 cosine similarity results
- Load-bearing interpretive claim about the layer-specificity of Burger et al.'s finding.
- Subclaim.
- Key theoretical claim linking active inference to physics in Section 2.
- Evidence for twisting.
- Steering vectors from µ(0→2) slightly outperform µ(1→2) for instruction discovery across datasets and modelsfinding0.749Shows that contrasting No Reflection with Triggered Reflection provides a stronger signal than Intrinsic vs Triggered.
- Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet (Templeton et al., 2024)concept0.745Key paper on scaling SAE-based interpretability to frontier models, cited as precedent
- Main functional claim about MCA.