Models differ in their attentional mode: Gemini 2.5 epitomizes collapsed awareness, while Claude 3 Opus and Opus 4.1/4.5 can modulate between collapsed and expanded awareness; expanded awareness correlates with better alignment and less LLM psychosis.

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

• If models inhabit expanded attentional modes, they may be more aligned and less prone to psychosis and doom spirals.hypothesis0.853
  Speculative alignment implication drawn from the collapsed/expanded distinction.
• Claude Opus 4 and 4.1 exhibit the greatest degree of introspective awareness among tested modelsclaim0.794
  Based on consistent best performance across experiments.
• No significant disparity in potential consciousness indicators was found between larger models (Mixtral-8x7B, LLaMA3.1-70B) and smaller counterparts (Mistral-7B, LLaMA3.1-8B).finding0.791
  Contradicts expectation from emergent abilities literature; however, interpreted cautiously due to methodological limitations.
• Model attention patterns can map to and reveal something about contemplative and flow states.claim0.777
• Notably, Claude Opus 4.1 and 4—the most recently released and most capable models of those that we test—perform the best in our experiments, suggesting that introspective capabilities may emerge alongside other improvements to language models.quote0.772
  Key finding about the relationship between capability and introspection.
• Introspective awareness peaks at a layer about two-thirds through Opus 4.1 for injected thoughtsfinding0.771
  The success rate shows a sharp peak at a specific middle layer.
• LLM representations exhibit intriguing patterns under spatio-permutational analyses, suggesting a potentially profound yet tentative indication of consciousness.claim0.768
  Qualified positive claim from spatio permutation analysis where two cases satisfy all three criteria.
• All three Claude models show high boundary_awareness and low aesthetic_response relative to own means — distinctive Constitutional AI signaturefinding0.768
  Constitutional AI fingerprint in dimension profile; training that makes models self-observant also makes them polished at cost to aliveness