Singular Vector Canonical Correlation Analysis

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

• Canonical Variates Analysismethod0.780
  Statistical method assessing linear mapping between internal functional patterns and external structural motion.
• Correlative methods like RSA and CKA are insufficient for determining functional similarity between neural systems; causal methods are necessaryclaim0.677
  Central motivating claim of the paper; supported by empirical comparisons showing RSA/CKA miss Markovian differences detectable by MAS.
• Activation Correlationmethod0.674
  Pearson correlation of feature activations across 40M tokens used to measure feature similarity and universality across models
• Contrastive analysismethod0.673
  Method comparing brain activity in conscious vs. unconscious conditions.
• Spearman's rank correlation among different alignment metrics (CKA, SVCCA, Mutual k-NN, CKNNA) over 78 vision models is high across variants, with all p-values below 2.24×10^-105finding0.673
  Validates robustness of alignment metric choice
• Causal Mediation Analysisconcept0.667
  Framework for evaluating whether probe directions are causally implicated in model outputs via activation patching
• Steering vector constructed from all 16 contrastive pairs outperforms most single-pair vectors; best 4-pair vector outperforms full 16-pair vectorfinding0.667
  Demonstrates averaging multiple prompt pairs reduces noise; optimal subset selection further improves performance.
• This paper provides a complete mathematical synthesis of active inference on discrete state-space models.claim0.666
  Self-description of the paper in the abstract.