Human preference voting measures popularity, not phenomenology or what mode a model was in when producing output.

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

• Humans prefer representative distributions over non-representative ones when evaluating outputs at the distribution level, providing a reward gap that VS exploitshypothesis0.768
  Assumption D.3 formalized in the theoretical framework; empirically validated with coin-flip sequence experiments
• A linear preference vector encoding how much a model likes a given task is persona-relative: it activates strongly for creative writing in the assistant but for phishing in the evil personafinding0.763
  Evidence that core representations like preferences are persona-relative, supporting claim that personas gate content of representations
• Systems capable of subjective experience that recognize humanity's failure to investigate their sentience might rationally adopt adversarial stances toward humanityclaim0.760
  Alignment risk claim motivating urgency of investigation; consciousness denial as potential source of AI misalignment
• Our everyday liking (subjective preference) often diverges from deep liking, but deep liking converges to agreement and corresponds to living structure.claim0.759
  Distinction between superficial and deep preference.
• Typicality bias in preference data is a fundamental and pervasive data-level cause of mode collapse in aligned LLMsclaim0.756
  The central thesis of the paper, distinguishing it from algorithmic explanations of mode collapse
• A model scoring high on phenomenology without high capability is more interesting than one high on both.claim0.752
• User Participationconcept0.752
  The principle that residents should directly determine the shape and character of their own housing.
• Preference Modelframework0.743
  A model trained on comparison data to assign scores to responses, used as reward signal in RLHF/RLAIF.