cohesion ρd

Within-cluster tightness of target pattern representations.

Neighborhood — ranked by edge-count

method

Whitening of span embeddings
associated_with
Preprocessing step that uses dev-set covariance to standardize embedding scales before computing ρd and dr.

concept

anchoring strength S
associated_with
Composite score S = ρd − dr − log k predicting anchoring success.
Anchoring strength S = ρd - dr - log k
associated_with
The calibrated score measuring how effectively anchors bind target patterns; a predictive correlate of success.
shot midpoint k50
associated_with
Number of in-context exemplars to reach 50% accuracy in E2.

artifact

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.

Target Pattern Cohesion (ρd)concept0.847
Measures how tightly the target pattern PT clusters in representation space; one of three components of S
Social cohesionconcept0.794
Second key outcome variable hypothesized to vary with cultural wealth-sharing institutions.
Coherenceconcept0.750
A property that makes a segment of space stand out as a center; determined by symmetry, connectedness, convexity, etc.
Cooperationconcept0.744
Competing or cooperating with neighbours; a plant behaviour.
complianceconcept0.731
The model's tendency to comply with harmful requests, the opposite of refusal.
the relative harmony or 'life' of a given building may be understood directly from the internal cohesion of the structure... without reference to opinion, prejudice or philosophy, merely as a consequence of the wholeness which exists.quote0.724
Strong claim that life/beauty is an objective property of the wholeness structure.
Corrigibilityframework0.715
The property of an AI being safe to shut down or modify; discussed in context of GPT.
Alignmentconcept0.714
The goal of making model behavior match human values and intentions, often addressed during post-training.