concept
active
concept:prediction-orthogonality-thesisPrediction orthogonality thesis
A model optimized for prediction can simulate agents with any objectives and any degree of optimality.
Neighborhood — ranked by edge-count
Frameworks (1)
framework
- Orthogonality thesisextendsClassical thesis that any level of intelligence can be combined with any goal.
Claims (1)
claim
- Prediction orthogonality thesis.
Artifacts (1)
artifact
- Simulators (LessWrong post)introducesThe paper being extracted.
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 signed directional computation of how actual outcomes compare to goal-specified targets; identified with valence
- Hypothesis that information may be encoded in arbitrary non-linear subspaces of a neural network
- The hypothesis that models internalize concepts as approximately linear directions in representation space; used to interpret MDS injection behavior
- Theory that perception and cognition involve predictive coding; listed with indicators in Butlin et al. 2023.
- If simulators are not inner aligned, then many important properties like prediction orthogonality may not hold.hypothesis0.696Conditional importance of inner alignment.
- In active inference, predictions about what should be, held to motivate action; overflow leads to tanha.
- Key mathematical object in DAS; transforms neural representations to alternative bases to reveal distributed structure.
- The theory of welfare that there is a plurality of welfare goods, some independent of pro-attitudes.