1 result for “context-window”
The maximum number of tokens — including the prompt, prior conversation, retrieved documents, and the model's own output — that a large language model can process in a single forward pass.