← BACK TO AI LAB

AI LAB / RAG / Chunking

Chunk Size vs Answer Quality

Testing how document chunk size changes retrieval recall, context precision, and citation quality.

RAGChunkingEvaluation
EXPERIMENT
EXP-019
STATUS
complete
DATE
Jul 30, 2026
TAGS
RAG · Chunking · Evaluation

EXPERIMENT / HYPOTHESIS

Structure-aware chunks will beat a fixed token window across mixed technical documents.

  1. 01Create fixed chunks at three sizes
  2. 02Create heading-aware semantic chunks
  3. 03Evaluate retrieval and citation completeness
Best recall
94.1%
Citation precision
90.7%
Documents
180

CONCLUSIONStructure-aware chunks had the best balance. Very small chunks improved retrieval precision but often lost the context needed for complete answers.

01

Variants

The test compared 256, 512, and 1,024-token windows with a structure-aware parser preserving headings and lists.

  • Same embeddings
  • Same retriever
  • Same answer model
02

Finding

Chunking quality depended on document structure more than one globally optimal token count.

  • Preserve semantic boundaries
  • Include parent headings
  • Measure citation completeness