Maximum Margin Separator

The direction logistic regression converges to on linearly separable data; shown to be suboptimal for identifying truth direction

Neighborhood — ranked by edge-count

method

Logistic Regression Probe
implements
Standard linear probing technique; compared to mass-mean probing for classification accuracy and causal implication

concept

Feature Interference
associated_with
When non-orthogonal features cause logistic regression to identify a suboptimal probe direction

cosine ≥ 0.65 · no typed edge

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

marginconcept0.741
The outer edge space that protects the text block but can be invaded by marginalia.
Linearly Separable Functionconcept0.705
A function separable by a single linear boundary; allows independent contributions of inputs to output; computable by Perceptron
Maximal Marginal Relevancemethod0.702
Diversity-based reranking approach from Carbonell and Goldstein 1998 for document summarization
maximum-norm gradient normalizationmethod0.693
Training-free technique normalizing all task gradients to the maximum gradient norm magnitude
Margins are highly charged spacesclaim0.689
Margins function actively, not neutrally; they define boundaries and create the distinction between text-space and world-space.
Design and construction separationconcept0.688
The 20th-century divorce of design from making, which prevents ornament from arising naturally.
separation from realityconcept0.677
A state caused by disorderly, dead environments where people lose connection to everyday human events and freedom.
Feature splittingconcept0.677
Phenomenon where a feature in a small SAE splits into multiple finer features in a larger SAE.