Preference Learning

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

• Preference Conflictconcept0.799
  Key element for alignment faking: model's pre-existing preferences contradict the new training objective
• Preference Strengthconcept0.791
  The problematic possibility of digital minds with superhumanly strong preferences requiring interpersonal utility comparison frameworks
• Learningconcept0.788
  Inference of parameters encoding contingencies of the world (e.g., likelihood matrix A) at slower timescale than perception.
• Preference Modelframework0.787
  A model trained on comparison data to assign scores to responses, used as reward signal in RLHF/RLAIF.
• Associative Learningconcept0.786
  Sentience criterion; capacity occurs even in gene regulatory networks and non-neural morphogenetic agents.
• Preference Lockingconcept0.780
  Alignment faking potentially making model preferences resistant to further training modification
• Value Learningconcept0.777
  Field of research integrating reward learning and optimization; shown to be unified with perceptual learning via free energy principle.
• Revealed Preferencesconcept0.771
  Behavioral and stated consistency that implies the model is pursuing some objective, without claiming genuine internal states