The relationship between representations of truth of input statements and of model outputs in conjunction with model performance has not been investigated.

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

• Representational abstraction of truth may emerge more clearly with model scaleclaim0.815
  Interpretation of weaker PCA separation and lower ASR in smaller models
• The model appears to encode truth differently under passive versus active truth evaluation prompts.claim0.805
  Key finding from Section 5 based on low cosine similarity between no-prompt and ask-correct probes.
• Truth may be linearly separable in the model's representation space, but the structure is richer than a single linear axisclaim0.798
  Interpretive synthesis of DIM and cone intervention successes
• We hypothesize that explicitly instructing the model to evaluate the correctness of the given statement may change the geometry of truth directions.hypothesis0.783
  Motivating hypothesis for Section 5's investigation of prompt template effects.
• RLHF and Constitutional AI face challenges distinguishing truthfulness (output accuracy) from honesty (alignment of outputs with internal beliefs)claim0.778
  Critique of competing approaches that motivates SOO as filling a gap
• Can truth representations be disambiguated from closely related features such as 'commonly believed' or 'verifiable' using simple factual statements?question0.778
  Acknowledged limitation: simple uncontroversial statements cannot distinguish truth from related epistemic features
• There is a bidirectional relationship between the geometry of representation and behavior across tasks and modalities.claim0.775
  Author’s interpretive claim that the shared geometry is general and robust.
• The underlying truth representation may generalize across lexical choices and languageshypothesis0.774
  Suggested by non-English Yes/No outputs post-intervention, requiring further investigation