Exams / AIF-C01 / RAG, knowledge bases and vector stores / Question 4
AIF-C01 practice question: RAG, knowledge bases and vector stores (4 of 4)
Applications of Foundation Models A free question from our original AIF-C01 practice exam. Try it, then open the answer to see why each option is right or wrong.
A university is loading 2,000 long research reports into a knowledge base. Why are the reports split into smaller chunks before they are converted to embeddings?
Choose one answer.
- AA. So that each report can be stored in a different AWS Region
- BB. So that the model is fine-tuned on each chunk separately
- CC. So that relevant passages can be found and fit in the prompt
- DD. So that the reports are encrypted one chunk at a time
Show the answer and explanation
Answer: C
- A. Wrong: Chunking has nothing to do with where data is stored. It is about how text is retrieved.
- B. Wrong: A knowledge base does not fine-tune the model. Chunks are retrieved and added to prompts at query time.
- C. Right: Correct. Smaller chunks give more precise matches, and only the useful passages, not whole reports, are added to the prompt.
- D. Wrong: Encryption protects data at rest and in transit. It is not the purpose of chunking.
Explanation
Chunking splits documents into manageable pieces before embedding. Each chunk gets its own vector, so a search can return the specific passages that answer a question. This improves relevance and keeps the augmented prompt within the model's context window and token budget.
Study
Keep practising
Answer questions with a score at the end, or practise just this topic with the explanation after each one. No sign-up needed.