concept
active
concept:linear-directionlinear direction
A straight vector in activation space, traditionally used for concept manipulation; claimed to be insufficient when true concept geometry is curved.
Neighborhood — ranked by edge-count
Claims (1)
claim
- Proposes that nonlinear geometric structure is superior to linear feature spaces for capturing semantic content.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- The sequential, continuous order of text, often challenged by diagrammatic branching.
- Typical approach that adds a scaled steering vector to representations; the paper argues this is mismatched with actual representation geometry.
- Semantic domain for linear transformations; denotation as actual linear function; Category instance generated from homomorphism principle.
- Manipulation of activations along a straight line; shown to fail when it crosses voids, in contrast to manifold-following interventions.
- Alignment map ϕ(h)=W_orth*h using orthogonal matrix; assumes linear representation hypothesis
- Correlative technique measuring the type of information encoded in distributed representations via linear predictability.
- A direction in the model's representation space that governs self-reflection behavior, computed as mean difference between reflection and non-reflection embeddings