question
active
question:does-the-apparent-phase-change-we-observe-in-features-neurons-have-any-connection-to-phase-changes-in-compressed-sensingDoes the apparent phase change we observe in features/neurons have any connection to phase changes in compressed sensing?
Open question.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Toy Models of Superpositionmentions
Frameworks (1)
framework
- Compressed SensingcitesMathematical framework enabling recovery of high-dimensional sparse vectors from low-dimensional projections; theoretically underpins sparse autoencoder approach
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.
- Motivating question.
- All neural codes, not just communication ones, could plausibly be compressed to some degreeclaim0.738Neuroscience implication.
- The phase diagram explains why neurons are sometimes monosemantic and sometimes polysemanticclaim0.737Theory contribution.
- Ensuing need for active, intelligent decoding on the expansion side.
- Falsifiability test built into the PC analysis design
- Central claim of the paper, supported by detailed feature analysis, human evaluation, automated interpretability of activations, and automated interpretability of logit weights
- Perturbing a single feature.
- Highlights the non-genetic control of large-scale anatomy.