Glossary term

Context window

The amount of input a model can hold during one reasoning pass.

Definition

This is the text and token range an LLM can handle in one reasoning pass. Long documents should expose headings, summaries, and definition blocks early so less signal is lost inside that limit.

Working notes

Structure matters more than raw length if the goal is fast retrieval.
Topic hubs and glossary pages act as compressed entry points before the long-form post.