Interactive
Live Demos
Explorable explainers, built to be played with. Each one is self-contained and runs entirely in your browser - no sign-up, no server round-trip.
How a Multimodal LLM Sees
Live Demo
How a Multimodal LLM Sees
An interactive walkthrough of how a vision-language model turns an image and a sentence into one shared sequence of vectors, attends across both, and predicts the next token.
Open demo