AI and LLMs · Middle · Question 2 of 75
Why is chunk overlap often used when splitting documents for RAG?
- 1So the vector index uses less storage
- 2So every chunk contains the whole document
- 3So facts on chunk borders aren't cut in half
Show the answer
Correct answer: So facts on chunk borders aren't cut in half
Explanation
A sentence split between two chunks may never be retrieved whole. A small overlap keeps context together at the cost of some duplication.