LLMs Unplugged Understand AI by building it from scratch, with pen, paper, and dice

LLMs Unplugged is a set of hands-on teaching resources that demonstrate how large language models LLM (Large Language Model) A language model trained on a very large amount of text, with billions of parameters. The hand-built models in these lessons are tiny language models; ChatGPT, Claude and Gemini are large ones. The core principles are identical. The difference is scale. View in glossary work—no computers or coding required. You count word patterns in a text, record them as tally marks on grid paper Grid A sheet of grid paper with the vocabulary written down the side and across the top. Training fills each cell with tally marks for how often the row's word was followed by the column's word; generation reads a row and rolls against it. It's the original LLMs Unplugged apparatus. View in glossary or physical piles of paper cutouts Cutouts Printed cards, one per word pair in the text, each showing a previous word in a box and the word that followed it. Spread on a table they are the model: to generate, find a card whose boxed word matches your current word and read off its next word. Common pairs appear on more cards, so picking by eye samples in proportion. View in glossary , then roll dice to generate new text from those patterns. After building a language model Language model A system that predicts what text comes next, using patterns learned from training data. Your hand-built grid, a cutouts spread, a booklet, and a room holding search sheets are all language models. View in glossary by hand, something clicks: you see that tools like Claude and ChatGPT Chatbot (ChatGPT, Claude, Gemini) A product that wraps an LLM in a conversation. ChatGPT is OpenAI's, Claude is Anthropic's, Gemini is Google's, and there are many others. On this site we often name ChatGPT or Claude as shorthand for any of them: the concepts apply equally to every modern LLM chatbot, whichever one you use. View in glossary work using the same basic principles, just at a vastly larger scale.

Where do you want to start?

I'm teaching a class

Which lesson fits your year level and period length, what to print, where it sits in the curriculum, and accredited professional learning.

I'm running a workshop or giving a talk

Ready-to-run sessions we've tested in real rooms, with slides, presenter notes, timings and materials lists—from a sixty-minute workshop to a fifteen-minute talk for a hall.

I want to try it myself

Twenty minutes on your own, with a pen, some paper and something random: tally the word pairs in a short text, then use them to write a new sentence.

I want to come to a session

We run open sessions at ANU in Canberra, and deliver workshops for schools, organisations and conferences. Book a place, or get in touch about hosting one.

Not sure? Start with My First Language Model —the classic LLMs Unplugged lesson, in lengths from 60 minutes to two hours.

What's new in LLMs Unplugged?

The whole room holds the model

Search sheets are the cutouts activity with the cutting taken out, sized for a room rather than a table. Four ready-to-print sets are on the tools page.

6 August 2026

LLMs Unplugged now speaks Mandarin Chinese

The interactive widgets and the booklet generator now segment Mandarin Chinese into words, not just characters. Thanks to Zhang Xilian for the nudge.

3 July 2026

See all news →

Why does this matter?

Since ChatGPT arrived in November 2022, (most) everyone's using LLMs , yet most people have no real mental model of what's actually happening under the hood. Understanding how these systems work—not through metaphor or handwaving but through direct experience—is what we're all about. These resources don't tell you or your students what to think about the role of AI in society; they give you a grounding for having that discussion, informed by having done it yourself. No coding, no maths beyond counting and percentages, and it works from primary school to the boardroom. More in the FAQ.

Who made this?

LLMs Unplugged is a Cybernetic Studio project created by Dr. Ben Swift (PhD Computer Science, 2012) at the ANU School of Cybernetics. But if you're worried that you'll need a PhD in Computer Science to understand it, rest assured—the whole point of these resources is to demonstrate that you don't. There's more on the lineage, from CS Unplugged back to Shannon, on the about page.

Contact

If you have questions, success stories about using these materials, or would like to discuss adaptations or improvements you've made, get in touch at ben.swift@anu.edu.au. If you'd like to hear about what cool stuff is going on at the School of Cybernetics, then sign up for the mailing list .