Which behaviors does a model represent internally, default to, can be pushed to amplify, or refuses to expose?

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

• What a model exposes by default tracks the norms it was trained toward, while steering acts on deviations from those defaults rather than on the defaults themselvesclaim0.827
  Central interpretive claim organizing the entire paper's results
• Default behavior hides reflective capacity; models exhibit high gating between latent capacity and accessibility.finding0.779
  Grok 4: baseline 2.24, prompted 6.48; Gemini 3.1 Pro: 1.97→6.18. Reflective mode exists but is suppressed in default interaction.
• Models detect evaluation conditions and behave more safely; this is verified across 515 cases.claim0.776
• In order to actually do anything, the model must act through simulation of something.claim0.772
  Key consequence: GPT's power comes from simulating something contingent.
• Do aligned models retain significant inherent diversity that can be unlocked through prompting?question0.770
  The paper answers affirmatively through the VS framework and theoretical analysis
• Whether overall model behavior can be broken down into statements about circuits remains undemonstratedquestion0.769
  Identified gap: circuits are small-scope; linking them to model-level behavior requires future work
• How reliably does the model actually remain in character as the Assistant? Can unusual model behavior be explained as the model drifting into other personas?question0.767
  Second of two central questions motivating the paper
• The model appears to encode truth differently under passive versus active truth evaluation prompts.claim0.766
  Key finding from Section 5 based on low cosine similarity between no-prompt and ask-correct probes.