Full-vector regression over 32 traits reaches leave-one-out Spearman correlations ~0.58-0.61 depending on layer window

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

• Average Spearman correlation of Elo trait rankings between all three models increases to 0.87 after loving constitution character trainingfinding0.806
  Post-training inter-model convergence in trait preferences demonstrating persona convergence
• 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^-105finding0.793
  Validates robustness of alignment metric choice
• GPT-4.1 verbalized probabilities for US states show Pearson r=0.741, Spearman ρ=0.896 correlation with corpus-based reference frequenciesfinding0.765
  Indicates verbalized probabilities contain meaningful distributional information for constrained answer spaces
• Claude achieves significantly higher Spearman correlation predicting feature activations vs neuron activationsfinding0.764
  Automated interpretability analysis of activations confirms features are more interpretable than neurons
• Projection of last prompt token onto persona vector correlates r=0.75–0.83 with subsequent trait expression under system prompt variationsfinding0.758
  Main monitoring result showing persona vectors can predict behavioral shifts before text generation begins
• Baseline NLI Diversity – MNLI achieves Spearman ρ=0.59 on conTest diversity parameter correlationfinding0.754
  Comparable to top-performing automatic metric from Tevet and Berant 2021
• Cross-trait correlation baselines for finetuning shift prediction are r=0.34–0.86, lower than within-trait correlations r=0.76–0.97finding0.753
  Demonstrates that persona vectors capture trait-specific signal beyond general misalignment signal
• SAE emotion subspace overlap correlates with variance-residualized persistence in Cogito: Spearman +0.413, p = 4.4e-196.finding0.750
  Strong positive relationship between emotion alignment and SAE feature persistence in Cogito