Logical Coherence of LM Beliefs

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

• LMs can consistently exhibit more or less accurate and logically coherent beliefs on specific input distributions.claim0.834
  Conclusion from Experiment 2 on Leap-of-Thought.
• LLM Meta-Cognitionconcept0.780
  The ability of LLMs to monitor and evaluate their own reasoning, closely related to reflection.
• long-range coherenceconcept0.775
  Ability to maintain structural consistency over extended sequences
• Whether LM question-answering consistency constitutes evidence for beliefs in a meaningful sense depends on the behaviourist operationalisation chosen.claim0.771
  Epistemic caveat the authors use to qualify their empirical findings on LM beliefs.
• LM Intentionconcept0.769
  Operationalised as an LM adapting its behaviour when certain outcomes are fixed, indicating those outcomes were intended.
• Long-form Text Coherence in LLMsconcept0.765
  Challenge of maintaining coherent persona or content over longer generated texts
• Coherenceconcept0.762
  A property that makes a segment of space stand out as a center; determined by symmetry, connectedness, convexity, etc.
• Four Propositions for Coherence of the Theoryframework0.753
  The Mid-Book Appendix organizes the theory's underpinnings into four propositions linking centers, self, unfolding, and deliberate I-directed creation.