concept
active
concept:internal-threshold-50internal threshold θ50
Few-shot midpoint in E3's geometric analysis.
Neighborhood — ranked by edge-count
Methods (1)
method
- Quantitative study correlating layer-wise anchoring geometry (S_max, AUS_N) with behavioral thresholds θ50
Concepts (4)
concept
- normalized area under S-curve AUSNassociated_withAverage S(ℓ) across layers; another geometry summary.
- normalized area AUSNassociated_withAverage of S(ℓ) across layers.
- shot midpoint k50associated_withNumber of in-context exemplars to reach 50% accuracy in E2.
- peak anchoring Sbmaxassociated_withMaximum layer-wise anchoring score across layers.
Artifacts (1)
artifact
- Main paper presenting UCCT and semantic anchoring framework.
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.
- Shot count needed to reach 50% accuracy; reflects when anchoring strength crosses critical value.
- Criterion requiring that causal influence of internal state on description be internal, not routed through sampled outputs; rules out pseudo-introspection via self-observation.
- Per-latent calibrated steering strength that yields average first-attempt judge score of 30/100
- Value of S above which performance sharply increases; varies by model, layer, and task.
- Assumption that small anchor changes can produce sharp performance shifts when conditions are favorable.
- Task-dependent value of S at which performance flips.
- The possibility of a stably encoded, causally active emotional state within LLMs, as distinct from token-by-token semantic content