Singular Value Decomposition

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

• Weight matrix decompositionconcept0.737
  The core idea of decomposing weight matrices into components for interpretability.
• Rank-one matrix decompositionmethod0.725
  Constraint in VPD where each parameter subcomponent is constrained to be a rank-one matrix for simplicity.
• Activation decompositionconcept0.716
  The conventional approach (e.g., SAEs, transcoders) of decomposing activations into interpretable features.
• valueconcept0.709
  Probability of sensory input expected by an agent, aligning value maximization with surprise minimization.
• Value extractionconcept0.708
  The initial stage of uncertainty metabolization, pulling usable value from sensations.
• Partial Information Decomposition (PID)framework0.700
  Williams and Beer's decomposition of joint mutual information into unique, redundant, and synergistic components.
• Helmholtz Decompositionmethod0.700
  Mathematical technique decomposing flow into curl and divergence-free components; enables derivation of free energy principle.
• structured dataconcept0.691
  Training data with inherent geometric or relational structure, which induces geometric organization in model internals.