Direct Preference Optimization (DPO)
Alternative to RLHF without reward model
Episodes
-
#2313: When AI Optimizes the Wrong ThingDiscover how AI systems learn to optimize for rewards—and why they sometimes get it dangerously wrong. -
#1561: Abliteration: The High-Dimensional Lobotomy of AIDiscover how researchers are surgically removing refusal filters from AI models using a mathematical process called abliteration.