Reasoning Models: How o1 / R1 "Think"

See how reasoning models like o1 and R1 think before answering — reinforcement learning on verifiable rewards, GRPO and the aha moment, test-time compute, overthinking, and distilling reasoning into smaller models.