ATTENTION
Visualize how transformers decide which words to focus on when processing text
Built for curious minds.
Eight interactive visualizations. Each one dismantles a core concept of how language models work. Click, explore, understand.
Watch how conversation history accumulates and fills the context window with each exchange
See text break apart into tokens in real-time and understand how models read language
Visualize how transformers decide which words to focus on when processing text
Explore the high-dimensional space where words become mathematical vectors
Walk through the full Retrieval-Augmented Generation pipeline: ingest, chunk, embed, retrieve, and generate
See what really happens when a model 'uses a tool' — it's just structured text generation
The USB port for AI — see how models connect to the outside world through the Model Context Protocol
Peek into the model's mind and see the probability of every possible next word