Rank-one matrix decomposition

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Rank-one matrix decomposition constraint enforcing mechanistic simplicityclaim0.813
  Core design principle of VPD: each parameter subcomponent is constrained to be a simple rank-one matrix to enable isolated understanding and combination.
• Weight matrix decompositionconcept0.805
  The core idea of decomposing weight matrices into components for interpretability.
• What matrix decomposition or dimensionality reduction best summarizes the enormous low-rank OV and QK matrices?question0.769
  Open methodological question about converting the 50k x 50k expanded matrices into human-graspable summaries
• Rank-one subcomponentsconcept0.745
  The atomic pieces into which weight matrices are decomposed by VPD; each rank-one component is interpretable.
• Singular Value Decompositionmethod0.725
  Used to summarize principal patterns of internal functional states.
• Copying Matrixconcept0.720
  An OV circuit matrix that maps tokens to increasing the logit of those same tokens; detectable via positive eigenvalues
• Activation decompositionconcept0.715
  The conventional approach (e.g., SAEs, transcoders) of decomposing activations into interpretable features.
• Transition Matrix (B matrix)concept0.708
  Prescribes transitions among hidden state factors under action; encodes policy-dependent dynamics