About LLM 3D Academy

Making Large Language Model education accessible to everyone through free, interactive 3D visualizations. No math PhD required.

Our Mission

LLM 3D Academy exists to bridge the gap between cutting-edge AI research and public understanding. We believe that understanding how AI works should not require a graduate degree in computer science.

Through interactive 3D visualizations, we transform abstract concepts like attention mechanisms, tokenization, and model training into tangible, explorable experiences. Click, rotate, and learn at your own pace.

What We Offer

Nine free interactive modules covering the complete LLM pipeline, from how text becomes tokens to how models generate grounded answers:

🏗️

Transformer Architecture

Explore the complete architecture in 3D

✂️

Tokenization

How text becomes tokens (BPE)

📐

Embedding

Words as vectors in high-dimensional space

🔍

Attention Variants

MHA, MQA, GQA, and Flash Attention

⚡

KV Cache

Accelerating autoregressive inference

🏋️

Training Pipeline

Pre-training, fine-tuning, loss landscapes

🎯

LoRA

Efficient parameter fine-tuning

🤝

RLHF

Aligning models with human preferences

📚

RAG

Retrieval-Augmented Generation

Who Is This For?

Students — Learn AI fundamentals interactively, not from dry textbooks
Developers — Build intuition for the models you work with daily
AI Enthusiasts — Satisfy your curiosity about how ChatGPT actually works
Educators — Use our visualizations as teaching aids in your classroom

Technology

Built with a minimalist philosophy — no frameworks, no build tools, no dependencies to install. Just HTML, CSS, JavaScript, and Three.js for stunning 3D visualizations.

Three.js WebGL Vanilla JS CSS Custom Properties Zero Dependencies

The entire site is static HTML — no server, no database, no tracking beyond basic analytics. Fast, accessible, and available to anyone with a modern browser.

Project

Created with ❤️ for the AI community. LLM 3D Academy is an open-source project dedicated to making AI education freely available to everyone, everywhere.

Have feedback, suggestions, or want to contribute? We'd love to hear from you!

Get in touch →