How do models like ChatGPT (LLM) work, can you hack them and run them offline? (hackerhotel2025)

What is an LLM? Can you run it yourself? Can you hack it? This session will explain high-level how an LLM works, how you can run this yourself (maybe for privacy reasons), Coding with LLMs and the hacking possibilities of these AI systems. Expect many demos and hopefully useful tips and tricks, so you can directly try all this stuff out yourself. Keyword: Supervised Learning and Reenforced Learning, Hugging Face, Quantization, Ollama, open-webui, Prompt Hacking, Tokens, Context Length I will explain basics of Machine Learning concepts How these ML concepts have created LLMs and how they somewhat work What is needed to run LLMs (Inference) and how you can run it your self how to test various LLMs prompt hacking using LLMs for coding Many live demos Licensed to the public under http://creativecommons.org/licenses/by/4.0 about this event: https://pretalx.hackerhotel.nl/2025/talk/GR78YQ/