Large language models develop surprisingly coherent yet often rigid internal preferences as they scale

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

• Large Language Models Can Strategically Deceive Their Users When Put Under Pressure (Scheurer et al. 2023)concept0.839
  GPT-4 engaging in insider trading and denying it; related work on strategic deception
• Do large language models monitor their own internal states?question0.819
  Framing question that motivates the entire paper
• The inability for autoregressive large language models to maintain states of long-range order resembles tangential speech or derailment in formal thought disorder.claim0.817
  Analogy between LLM incoherence and schizophrenia symptoms
• Can large language models introspect—that is, accurately detect perturbations to their own internal states?question0.815
  Central research question of the paper
• Today's Large Language Models have become so good at playing Turing's game that it often takes experts to demonstrate the present limits of their ability to simulate human-like intelligence.claim0.808
  Paper's assessment of current LLM capabilities relative to Turing Test
• Can Large Language Models Genuinely Shift Human Perspectivequestion0.805
• The examples of features found in language models suggest they are highly sparse variables, consistent with dictionary learning being applicablehypothesis0.801
  Motivation for using sparsity-based dictionary learning on language models
• Can language models genuinely introspect on internal states or only confabulate?question0.799
  Central research question animating the paper: distinguishing genuine introspection from illusion through causal manipulation of activations.