Talk for HackSoc Nottingham, Feb 2025

"How do LLMs Actually Work?"

Summary: Have you ever used ChatGPT, Claude, or DeepSeek and wondered how they worked? In this talk we will be going on a detailed deep dive of the structure of Generative Pre-trained Transformer Models and how they work, as well as how to run them locally on your computer for personal use, or even Hacks! We'll also investigate some of the recent developments of these models, and what the whole Deepseek drama was about. No previous ML experience needed, Attention is all you need!