Why does introspective capacity vary concept-by-concept and what mechanisms could stabilize it over time?

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

• Introspective ability is concept-specific: quality differs across emotive concepts and the same intervention helps some concepts but not othersclaim0.826
  Cross-concept steering results; only 2 of 12 non-diagonal cells show significant introspection improvement
• This introspective capacity is highly unreliable and context-dependent in today's modelsclaim0.819
  A caveat qualifying the main claim.
• We hypothesize that introspective capabilities may scale with model size and architecture, including recurrence/looping that extends the integration windowhypothesis0.805
  Forward-looking prediction about whether early-layer introspection generalizes to larger models or recurrent architectures
• Introspective capacity may follow a simple monotonic scaling law across all concepts and architectureshypothesis0.804
  The paper treats this as possible but unconfirmed; current evidence shows concept-specific scaling only
• Introspective capacity is present from the first conversation turn, not requiring multi-turn context to emergeclaim0.797
  Three of four concepts show significant introspection at turn 1; rules out joint temporal drift as sole explanation
• Introspective capabilities may continue to develop with further improvements to model capabilitiesclaim0.797
  Forward-looking statement about future models.
• Either introspection is an emergent capability requiring larger scale, or more stringent controls are needed to test introspection in smaller modelsclaim0.795
  Alternative interpretations offered for why binary detection fails in Llama 3.1 8B but frontier models claim success
• If introspective ability exists, can it be improved?question0.791
  Secondary research question addressed through cross-concept steering experiments