Quiet Nodes
← Wiki Quiet Nodes

Vocabulary

Context window

How much text a model can keep in working memory for one reply: your prompt, history, documents, and its answer share the same budget.

Also called: context · long context

The context window is the model’s short-term memory for a single generation. Everything you send and everything it writes competes for the same limited space, measured in tokens.

“Long context” means that budget got bigger. It does not mean the model reads a novel with perfect attention all the way through. Useful rule of thumb: more room helps, but relevance and structure still beat dumping the attic.

Cups that use this term