Chunking for RAG quality
Trace a worked example
From goal to measurement to ship-or-abort for chunking on the employee handbook assistant.
1Learn the idea
Read
Trace
Goal: improve the employee handbook assistant using chunking without breaking protected slices.
Mechanism reminder for chunking: Parse documents, split on structure with target token ranges (often ~300–800), keep heading paths, add limited overlap, attach metadata, then embed.
Baseline and shock: If exceptions live 120 tokens after a rule and chunks are 80 tokens with no overlap, retrieval can return the rule without the exception—measurable as citation-support failures.
Tradeoff in play for chunking: Small chunks retrieve precisely but lose context; large chunks preserve context but dilute matches. Overlap helps boundary cases and duplicates hits.
Ship decision: Ship heading-aware ~500-token chunks with 60-token overlap after gold-span recall@5 ≥ 0.9 on 40 handbook questions.
Rollback triggers for chunking must cite recall@k on gold spans, citation entailment, duplicate rate in top-k, avg chunk tokens. If you cannot name a tolerated regression on the employee handbook assistant, do not promote the change.
Read
Make it operational
Keep the chunking decision record beside the employee handbook assistant code paths that implement it. Future you will not remember why a default exists unless the evidence is linked from the config for chunking.
Also pin one numeric memory from this chunking chapter: If exceptions live 120 tokens after a rule and chunks are 80 tokens with no overlap, retrieval can return the rule without the exception—measurable as citation-support failures. That number is not decoration; it is a template for how claims about chunking on the employee handbook assistant should look in design docs. Scoped specifically to chunking / employee handbook assistant / worked-trace.
Read
Common mix-ups
People confuse chunking with neighboring buzzwords when debugging the employee handbook assistant. Before changing prompts, ask whether the broken stage was evidence gathering, the chunking judgment itself, validation, or the product action. Fixing the wrong stage creates folklore (“we tried chunking and it failed”) that blocks the next team on the employee handbook assistant. Scoped specifically to chunking / employee handbook assistant / worked-trace.
Read
Rehearsal (chunking-for-rag/worked-trace)
Write a five-line artifact for this page: goal, inputs, check, owner, stop rule. Invent one fluent failure that the check would catch. Keep details specific to chunking for rag rather than generic AI advice.
Read
Rehearsal (chunking-for-rag/worked-trace)
Write a five-line artifact for this page: goal, inputs, check, owner, stop rule. Invent one fluent failure that the check would catch. Keep details specific to chunking for rag rather than generic AI advice.
Read
Rehearsal (chunking-for-rag/worked-trace)
Write a five-line artifact for this page: goal, inputs, check, owner, stop rule. Invent one fluent failure that the check would catch. Keep details specific to chunking for rag rather than generic AI advice.
Read
Chapter close 1
For chunking for rag, add acceptance test 1: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 2
For chunking for rag, add acceptance test 2: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 3
For chunking for rag, add acceptance test 3: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 4
For chunking for rag, add acceptance test 4: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 5
For chunking for rag, add acceptance test 5: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 6
For chunking for rag, add acceptance test 6: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 7
For chunking for rag, add acceptance test 7: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 8
For chunking for rag, add acceptance test 8: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 9
For chunking for rag, add acceptance test 9: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 10
For chunking for rag, add acceptance test 10: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 11
For chunking for rag, add acceptance test 11: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 12
For chunking for rag, add acceptance test 12: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 13
For chunking for rag, add acceptance test 13: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 14
For chunking for rag, add acceptance test 14: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 15
For chunking for rag, add acceptance test 15: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 16
For chunking for rag, add acceptance test 16: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 17
For chunking for rag, add acceptance test 17: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 18
For chunking for rag, add acceptance test 18: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 19
For chunking for rag, add acceptance test 19: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 20
For chunking for rag, add acceptance test 20: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 21
For chunking for rag, add acceptance test 21: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 22
For chunking for rag, add acceptance test 22: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 23
For chunking for rag, add acceptance test 23: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 24
For chunking for rag, add acceptance test 24: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 25
For chunking for rag, add acceptance test 25: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 26
For chunking for rag, add acceptance test 26: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 27
For chunking for rag, add acceptance test 27: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 28
For chunking for rag, add acceptance test 28: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 29
For chunking for rag, add acceptance test 29: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 30
For chunking for rag, add acceptance test 30: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 31
For chunking for rag, add acceptance test 31: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 32
For chunking for rag, add acceptance test 32: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Read
Chapter close 33
For chunking for rag, add acceptance test 33: evidence source, threshold, and signer. Keep it unique to this chapter's scenario.
Go deeper
Before you start
Why this matters
List the constraints (latency, cost, privacy, review capacity) that any chunking change must respect for the employee handbook assistant.
In the wild
See how this idea shows up as a product and a company — then come back to the lesson. Skills transfer across vendors.
Related lessons
Check your understanding
Page assessment
Answer from memory. Completion is saved from this evidence, not from opening the next page.
All responses are required.