concept
active
concept:truesightTruesight
A base model's superhuman ability to pick up precise statistical signatures of its own autoregression or the user's hidden aspects.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Anima Labs Phenomenology Pt1implements
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- The deep, unpretentious ease that arises in places where people can be themselves, supported by the subtle adaptation of the physical environment.
- The quality of being in touch with the I; beauty so deep that it makes God visible and which 20th-century design systematically avoided.
- A hypothesized direction in LLM activation space that encodes the truth or falsehood of factual statements
- Institutional affiliation of Owain Evans
- Administered the AI Safety Grants program that supported this research
- Belonging that arises from living processes and unique adaptation, not from wealth or ostentation
- The multi-dimensional activation subspace whose directions causally mediate truthful behavior in LLMs
- Wertheimer's protagonist's quest.