← Back to Glossary
Chunk Overlap
Duplicating a small portion of text between adjacent chunks to prevent sentences or context from getting cut in half.
Why it exists
Arbitrary document splitting can cut a sentence in half, destroying key context across chunk boundaries.
How it works
If chunk 1 covers tokens 0–500, chunk 2 spans tokens 450–950, preserving a 50-token shared overlap region.
Related terms
Where you'll use this
roadmap