Larger S_max correlates with smaller θ50 across backbones in E3 (negative association consistent across pooling and metric choices)

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

• AUS_N is a weaker correlate of θ50 than S_max across E3 backbonesfinding0.842
  E3 finding distinguishing the two geometry summaries; breadth less predictive than peak height
• Larger Sbmax associated with smaller θ50 in E3 sweepfinding0.830
  Geometry-to-behavior correlate within E3.
• 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^-105finding0.775
  Validates robustness of alignment metric choice
• Peak anchoring Sbmax and normalized area AUSN correlate with per-item success and internal shot midpoints θ50, providing a geometry-to-behavior bridge.claim0.774
  Main interpretation of E3.
• Li et al. 2024: larger LLMs outperform smaller ones at distinguishing self-related from non-self-related properties on self-awareness benchmarksfinding0.770
  Prior finding showing scale-dependent self-awareness, consistent with the scale effect observed in the paper's Experiment 1
• Larger models should amplify bias less than smaller models, with model biases more accurately reflecting data biases rather than exacerbating themclaim0.768
  Implication of PRH for AI fairness and bias
• Bigger models are more likely to converge to a shared representation than smaller modelshypothesis0.765
  Selective pressure toward convergence via model capacity
• 82% of features in 1M SAE had maximum Pearson correlation ≤0.3 with any MLP neuron, and manual inspection showed no semantic resemblance.finding0.765
  SAE features are not simply mirroring individual neurons.