M1-T1-MLC3 - What Is an LLM? How Language Models Generate Text – and Why They Hallucinate

Chatbots such as ChatGPT, Claude or Gemini answer in fluent, confident language – even when they are wrong. The aim of this material is to explain, with concrete examples, what a large language model (LLM) is, how it produces text-based answers, and why it can generate inaccurate, invented or misleading information. By the end of the learning content, teachers will be able to define the concept of “token”, to explain the principle of the next-token prediction, recognise the typical forms and warning signs of hallucination, and use practical strategies and tools for validating AI outputs..