Spearman Rank Correlation

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

• 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^-105finding0.794
  Validates robustness of alignment metric choice
• Rank-Order Correlation (Kendall's rho)method0.761
  Statistical method used by Yodan Rose to measure agreement between different people's neighborhood diagnoses.
• Claude achieves significantly higher Spearman correlation predicting feature activations vs neuron activationsfinding0.746
  Automated interpretability analysis of activations confirms features are more interpretable than neurons
• Five independent LLM scorers from four labs produce identical rankings (Spearman ρ > 0.8).finding0.729
  Scorer bias validation: Claude Haiku, Gemini Flash, GPT-5.4, Grok 4, Kimi K2.5 all converge on same model ordering.
• Activation Correlationmethod0.702
  Pearson correlation of feature activations across 40M tokens used to measure feature similarity and universality across models
• Yodan Rose's North Beach study found high rank-order correlation (Kendall's rho) between different people's diagnoses of good and bad places along streets in North Beach, San Franciscofinding0.691
  Empirical evidence that neighborhood quality diagnosis is objective rather than merely a matter of opinion.
• Token-100 correlation persistence metricmethod0.690
  Measures emotion feature persistence as correlation between z-scored activation at token 0 and token 100 across all eligible target model tokens
• Wellbeing concept: Spearman ρ=0.68, isotonic R²=0.48 in LLaMA-3.2-3B (n=400, p<10⁻²⁶)finding0.690
  Second-strongest pooled introspective coupling in primary model