Rank-one subcomponents

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

• Parameter subcomponentconcept0.799
  One of the simple rank-one matrices resulting from VPD that sums with others to reconstruct the original model weights and has a specific functional role.
• Rank-one matrixconcept0.771
  The simple matrix form into which VPD constrains subcomponents to enforce mechanistic simplicity.
• Rank-one matrix decompositionmethod0.745
  Constraint in VPD where each parameter subcomponent is constrained to be a rank-one matrix for simplicity.
• Adversarial search for causally unimportant subcomponentsmethod0.743
  Procedure in VPD that actively searches for combinations that break the prediction of which subcomponents are unimportant, stress-testing the decomposition.
• Emoticon eye subcomponentconcept0.743
  The part of the emoticon subcomponent responsible for recognizing the 'eyes' of emoticons like ';', ':' or '=', which was edited in the demo.
• Subordinationmethod0.738
  Attribute: spatial positioning that signals inferiority, using lower positioning or smaller size.
• class-one structureconcept0.727
  A structure generated by a living process, with deep interlocking and wholeness, as described in earlier volumes.
• VPD subcomponents are sparse, interpretable, and avoid feature splitting.claim0.720
  Assertion about the qualitative advantages of VPD's rank-one decomposition.