method
active
method:compressed-sensing-ideal-circuit-library-visualizationCompressed-sensing ideal-circuit library visualization
Maps transcoder features onto a library of ideal circuits via compressed sensing, with residual.
Neighborhood — ranked by edge-count
Papers (1)
paper
- A Toy Model of Mechanistic (Un)Faithfulnessintroducesmentions
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.
- Mathematical framework enabling recovery of high-dimensional sparse vectors from low-dimensional projections; theoretically underpins sparse autoencoder approach
- Interactive tool for visualizing and inspecting learned binary logic circuits using modified DigitalJS library
- Lists issues: concept-based tasks, flawed decomposition, low faithfulness.
- A goal in mechanistic interpretability to identify sparse computational subgraphs; VPD promotes sparse parameter circuits.
- Used to visually inspect separation of truth-related directions in model activation space across layers
- Mechanism for how the model modulates representation strength.
- Central thesis of the paper unifying cognitive phenomena under one objective function
- Will circuit discovery provide insights into arbitrary tasks, or only where tasks can be crisply defined?question0.685Circuit discovery scope.