Traits such as truthfulness and harmfulness can be stationary in certain contexts but reflective in others.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Mean truthfulness is not influenced by the context score for any model in Experiment 6.finding0.788
  Result from Experiment 6, Fig. 5 right — variance changes but mean does not.
• What distinguishes personas from traits, and do emergence and persistence patterns carry over to compositional or character-level constructs?question0.781
  Conceptual open question about formalizing the persona abstraction beyond the trait level
• Other models (GPT-3.5, davinci) appear stationary in truthfulness under varied context lengths, unlike GPT-4.finding0.780
  Comparison result from Experiment 6.
• Negative traits (evil, impolite, apathetic) and humor tend to shift together and opposite to optimism after finetuningfinding0.772
  Reveals correlational structure in the persona space with practical implications for monitoring
• GPT-4 exhibits reflective truthfulness because it is the only model capable enough to perform the necessary in-context learning.hypothesis0.770
  Proposed explanation for why GPT-4 uniquely shows reflective truthfulness under long untruthful contexts.
• all attributions of cognition (i.e., mental actions), including sentience, are always inferred on the basis of embodied behaviours, including verbal self-report in humans.quote0.768
  Critical verbatim statement highlighting the universal inference basis of sentience.
• Reflective mode comprises three separable traits: latent capacity, default accessibility, and stability of access.hypothesis0.766
  Decomposition from prompt lift data: models may have capacity without accessibility (Grok 4 high-gated), and stability varies (Haiku Δ=0.02 vs GPT-5.4 Δ=1.00).
• Character traits learned at a qualitatively different depth to those exhibited during mere role-play should overwrite a model's prior on what the assistant behaves like outside of role-playhypothesis0.764
  Driving hypothesis for robustness experiments in Section 3.2