How Do LLMs Learn?

Large Language Models learn by predicting the next token in billions of text samples, gradually capturing patterns, facts, and reasoning abilities from data.