Skip to course content
Free RAG fundamentals course

Introduction to RAG and Knowledge Assistants

Module 02 Knowledge Check

5 questions. Pass mark 4 out of 5. Answer every question before checking the answer key below, then retry after reading the feedback.

1. Why does chunk size matter so much?

2. A chunk that begins mid-sentence with 'This must be approved by the regional head' is dangerous because…

3. What is metadata used for in a retrieval system?

4. Permissions must be enforced at retrieval time because…

5. Overlapping chunks are commonly used to…

---

Answer Key and Explanations

Check these only after attempting every question.

1. B - Too small loses context; too large dilutes the match and wastes the context window

Chunking is a trade-off between precision of match and sufficiency of context. Both failure directions produce wrong answers.

2. B - 'This' has no referent - the chunk is unusable without its context

Dangling references are a classic chunking bug: the text retrieves well but means something different in isolation.

3. B - Filtering, permissioning, freshness checks, and citation back to the source

Without metadata you cannot restrict by audience, prefer recent documents, or tell the user where an answer came from.

4. B - If a restricted chunk enters the context, the model may quote it regardless of instructions

A prompt instruction is not an access control. Filter before retrieval, not after generation.

5. B - Reduce the chance that an answer is split across a chunk boundary

Overlap trades storage for a lower risk of losing the exact sentence that answers the question.

Practical Check

Apply this module to your own work: complete the module activity for *Documents, Chunks, Metadata, and Permissions*, then write one sentence naming what your result shows and one naming what it does not.

Strong Answer Pattern

A strong answer names the task, the evidence used, the check performed, and the remaining limitation. It avoids "proved", "guaranteed", or "always" unless the evidence genuinely supports it.