Localist Representations

Prior assumption that high-level variables align with disjoint groups of neurons in standard basis; contrasted with distributed representations.

Neighborhood — ranked by edge-count

concept

Distributed Neural Representations
contradicts
Representations where individual neurons play multiple conceptual roles; patterns consisting of linear combinations of unit vectors.

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.

Local vs Compositional Representationsframework0.814
Theoretical distinction between representing token-context pairs as individual features (local) vs combining independent features (compositional)
Localist Alignment Baselinemethod0.759
Baseline that finds the axis-aligned orthogonal matrix closest to the learned distributed rotation, assuming disjoint neuron groups.
Natural Distribution of Representationsconcept0.751
The distribution of latent representations produced by the model under unperturbed inputs
Higher-Order Representationsconcept0.750
Representations of one's own mental states; associated with consciousness in higher-order theories.
Linear representationconcept0.750
The idea that features are encoded as directions in activation space.
Deep Representationsconcept0.748
Hierarchical representations in neural networks that allow compression and coordinated behaviour while retaining sensitivity to input changes.
Structure in representationsconcept0.738
The central question of whether representational geometry implies corresponding computational structure
Local Symmetries Transformationconcept0.737
A structure-preserving transformation: introducing symmetries within a center, not globally, to strengthen local order.