Inexpensive Preferences

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

• Revealed Preferencesconcept0.770
  Behavioral and stated consistency that implies the model is pursuing some objective, without claiming genuine internal states
• Prior Preferencesconcept0.759
  Target distribution over states or outcomes encoded in the generative model; goal states.
• Preference Learningconcept0.756
  The ability of active inference agents to learn their own prior preferences over outcomes by accumulating Dirichlet parameters from experience.
• Direct Preference Optimizationframework0.751
  Post-training alignment method during which undesirable behaviors emerged in the studied model.
• Preference Conflictconcept0.747
  Key element for alignment faking: model's pre-existing preferences contradict the new training objective
• Where Do The Goals Preferences And Attachments Ofquestion0.737
• Preference Strengthconcept0.734
  The problematic possibility of digital minds with superhumanly strong preferences requiring interpersonal utility comparison frameworks
• Soft preference labelsmethod0.733
  Using normalized log-probabilities from the feedback model as soft targets for preference model training.