Semantic-anchoring assumption

Assumption that structured inputs act as anchors biasing activation toward target patterns PT.

Neighborhood — ranked by edge-count

concept

semantic anchoring
related_to
The central idea that external structure binds latent patterns to desired targets.

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.

anchorconcept0.745
External structure (in-context examples, retrieval, tuning) that biases latent pattern activation.
Pattern-repository assumptionconcept0.733
Assumption that pretraining populates a repository of unlabeled latent patterns Pprior.
Threshold-like activation assumptionconcept0.719
Assumption that small anchor changes can produce sharp performance shifts when conditions are favorable.
Implication (A → B)concept0.713
Steering Language Models With Activation Engineering (Turner et al., 2023)concept0.713
Foundational paper introducing activation steering methodology used in this work
NLA explanations can contain claims about the target model's input context that are verifiably false, but are typically thematically faithful to the context.claim0.713
Key limitation identified: NLAs hallucinate specific details while preserving thematic accuracy; informs practical usage.
In-Context Learning of Representations (Park et al. 2025)framework0.711
Reports phase-like breakpoints and geometry changes as context scales; UCCT provides measurable predictor
Cross-domain anchoringconcept0.709
Observation that anchoring effects appear across text and vision modalities.