what features activate on tokens we'd expect to signify Claude's self-identity?

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

• what features activate when we ask questions probing Claude's goals and values?question0.812
  Direction for understanding model's internal objectives via features.
• what features activate when Claude is trained to be a sleeper agent?question0.806
  Question posed after discussing sleeper agent threat model.
• what features activate when we ask Claude questions about its subjective experience?question0.783
  Question about features related to consciousness and self-report.
• Self-identity featuresconcept0.756
  Features that activate when the model is asked about itself, invoking AI tropes and anthropomorphization.
• what features need to activate / remain inactive for Claude to give advice on producing Chemical, Biological, Radiological or Nuclear (CBRN) weapons?question0.742
  Potential safety claim about suppressing features to prevent CBRN advice.
• All three Claude models show high boundary_awareness and low aesthetic_response relative to own means — distinctive Constitutional AI signaturefinding0.736
  Constitutional AI fingerprint in dimension profile; training that makes models self-observant also makes them polished at cost to aliveness
• The basic hallmarks of being a Self are the ability to pursue goals, to own compound memories, and to serve as the locus for credit assignment, at a scale larger than any component.claim0.720
  Proposed operational definition of a Self within the TAME framework.
• Memories are not immutable markers of identity; they can be transferred between individuals and remapped onto new substrates.claim0.712
  Evidence from planarian tail fragment training and metamorphosis suggests memory is substrate-independent process.