Chunking
Also: Chunks · Textsegmentierung
Splitting documents into smaller sections so that they can be searched for RAG.
Long documents are split into “chunks” of a few paragraphs, each stored as an embedding. If the sections are too small, context is missing; if they are too large, search becomes imprecise.
Good chunking follows the structure: headings, chapters, tables. It is one of the underrated levers for the quality of RAG systems.
General information, not legal or professional advice.