method
active
method:representational-similarity-analysisRepresentational Similarity Analysis
Inspiration for distance-matrix comparison.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (2)
method
- Cosine similarity of cosine similarity matrices across layers.
- Distance Matrix Analysis of Concept Centroidsanalogous_toL2 distance between centroids compared across manifolds, RSA-like.
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.
- A correlational similarity method compared against MAS; uses RDM correlations between model representations.
- Framework for characterizing span-level information of sequences of representations, independent of any consciousness estimate; used as a comparison baseline.
- How familiar a model is with a numeral system, manipulated via bases in Experiment 2.
- The central empirical phenomenon: different neural networks trained on different data/objectives develop increasingly similar representations
- Measure of similarity between the similarity structures (kernels) induced by two different representations
- Core theoretical claim about the target of representation learning
- The central question of whether representational geometry implies corresponding computational structure
- The desired property of a bidirectional, behavior-preserving mapping between model representations; the goal MAS pursues.