Yannic Explains: How are Large Language Models Trained?

In this video, Yannic Kilcher, PhD, co-founder and CTO at DeepJudge, breaks down the surprisingly simple way that large language models (LLMs) are trained. He shows step-by-step how language models learn by reading trillions of pages of text predicting missing words, and adjusting through feedback. You’ll learn why LLMs can seem so powerful, but also where they have limits (like hallucinations and knowledge cutoffs).