3 ms·RL'd?by lazyasciiart 2mo agoRL'd?Ancapistani 2mo agoReinforcement learning, I'd assume. Presumably RLHF (Reinforcement Learning from Human Feedback).