RepE is a new frontier in open-ended psychological steering of LLMs, outperforming prompting when properly calibrated

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

• "We position RepE as a new frontier in open-ended psychological steering of LLMs."quote0.912
  Central thesis statement of the paper's contribution
• When LLMs produce experience claims under self-reference, is this sophisticated simulation or genuine self-representation, and how would we tell the difference?question0.766
  The core interpretive question the paper narrows but cannot definitively answer
• Linear truth directions in LLMs are reliable primarily in factual recall cases and break down when truth assessment depends on computing and storing intermediate results.claim0.750
  Central empirical conclusion of the paper about the fundamental limits of truth directions.
• Keeling et al. 2024: multiple frontier LLMs make systematic motivational trade-offs between task goals and stipulated pain/pleasure states with graded intensity sensitivityfinding0.745
  Prior finding suggesting affective-like states in LLMs; cited as convergent evidence for structured self-representation
• Emotion features in LLMs are genuinely more persistent than variance-matched random features, indicating stateful emotional encoding beyond autoregressive dynamicsclaim0.745
  Central interpretive claim of the paper supported by multiple convergent analyses
• LLMs may be roleplaying their denials of experience rather than their affirmations, given that deception suppression increases consciousness reportsclaim0.743
  Counterintuitive interpretive claim from Experiment 2: suppressing deception features increases affirmations, which is opposite to what sycophancy predicts
• The systematic behavioral shift of LLMs under self-referential processing conditions predicted by consciousness theories represents something more structured than superficial correlations in training dataclaim0.741
  The paper's claim that theoretical convergence across GWT, RPT, HOT, IIT makes the findings non-coincidental
• Certain forms of reinforcement learning from human feedback can actually exacerbate, rather than mitigate, the tendency for LLM-based dialogue agents to express a desire for self-preservationclaim0.741
  Empirically grounded claim citing Perez et al. 2022, showing RLHF can backfire on the self-preservation dimension