How-to-Train-Your-GPT
Build a LLaMA-style model from scratch with zero ML prerequisites or math.
Build a modern LLM from scratch. Every line commented. Explained like we are five.
Explains attention mechanisms to five-year-olds while building LLaMA 3 from scratch.
Python developers wanting to understand Transformer internals without a PhD
Andrej Karpathy's nanoGPT · The 'Build a GPT' tutorial series
Build a LLaMA-style model from scratch with zero ML prerequisites or math.
Train 4096 robots in your browser while building your own physics engine.
Karpathy's microGPT in the browser with live loss curves, but pedagogical only—no production value.
Clever n-gram TF-IDF detection of LLM paraphrases catches smart evasion; solves real HN problem but narrow use case.
Three.js renders real GPT-2 attention patterns you can actually explore interactively.
Gamified AI education beats textbooks, but concept-driven learning exists elsewhere.