5-fold Cross-Validated Logistic Regression AUC

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

• Logistic regression correctness probemethod0.737
  Logistic regression trained on GSM8k training set to predict answer correctness from projection features along reflection direction
• Logistic Regression Probemethod0.724
  Standard linear probing technique; compared to mass-mean probing for classification accuracy and causal implication
• The logistic fit for threshold behavior is a phenomenological surrogate for interpretability, not a mechanistic derivationclaim0.699
  Authors' explicit epistemic limitation on the threshold model
• logistic fitting for shot thresholdsmethod0.689
  Fit a sigmoid to accuracy vs. k to estimate k50 and phase width.
• A/5 autoencoder (131,072 features) recovers 94.5% of MLP log-likelihood loss reductionfinding0.681
  Shows that loss recovery increases with autoencoder size
• Gene Ontology prediction: +8.4% AUC improvement with unsupervised autoencoder and covariance pooling embeddingsfinding0.680
  Empirical result: covariance pooling combined with unsupervised autoencoder embeddings improves Gene Ontology prediction AUC by 8.4% over mean pooling.
• AIC/BIC Model Selection Criteriaframework0.678
  Used as theoretical motivation for UCCT's log k budget term as complexity penalty mirroring model selection
• Criterion 3: IIT estimates must achieve higher mean AUC than Span Representation for ToM score classification.concept0.677
  Third of three operational criteria; distinguishes consciousness from inherent LLM representational separations.