What are supervised, unsupervised, and reinforcement learning?

Answer preview

LLM training uses all three: Supervised (text prediction), unsupervised (next-token), reinforcement (RLHF — human feedback on quality).…