The Neural Net

28.02.26 > LOADING... > TOKENIZE... > DEREFERENCE > GLITCH > RESTART_

Ever wondered if AI is actually smart… or just really good at guessing?

BREAKING AN LLM is a fun, beginner-friendly workshop where you get to peek inside the mind of a language model and see what's really going on. Instead of only chatting with AI, you'll observe its behavior, why it predicts certain words, why it repeats patterns, and why it sometimes gives confident but wrong answers.

We'll walk you through how an LLM reads words, remembers context, and predicts the next token, and you'll quickly notice something surprising: many times it isn't “thinking”, it's following patterns. For example, if the model keeps seeing a sequence like “2, 4, 6, 8…”, it may continue it automatically even when the question changes. That's your first clue... AI behaves more like a pattern machine than a brain.