finding
active
finding:philosophical-vocabulary-is-negatively-correlated-with-scores-in-contemplative-condition-model-level-r-0-72Philosophical vocabulary is negatively correlated with scores in contemplative condition (model-level r=-0.72)
Models deploying more philosophy buzzwords score lower; battery measures beyond surface text features
Source paper
extracted_from(2026) · Borzov, Anton
Neighborhood — ranked by edge-count
Claims (1)
claim
- Core epistemic claim bounding the paper's contribution
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.
- A 337-character contemplative system prompt lifts all 28 models by +2.62 points on a 10-point scale.finding0.765Core empirical result: every model, every architecture, every alignment type responds to the contemplative prompt with measurable gain.
- Mechanism of contemplative training.
- Analogy between LLM incoherence and schizophrenia symptoms
- Discriminant validity: composite scores are not reducible to verbosity
- Supports Janus's claim that introspection is architecturally available; prompting determines whether/how capacity is leveraged.
- Strong scaling trend for introspective fidelity when excluding invalid steering-sign pairs
- Scaling hypothesis for language-based contemplative alignment approaches
- Abstract's main conclusion.