How does the choice of embedding model affect retrieval quality in a RAG system?

It determines the language of the retrieved documents

It determines the speed of the retriever's search algorithm

It determines how well semantic similarity is captured in the vector space

It determines the maximum number of tokens the LLM can generate

A RAG system for test analysis sometimes produces answers that are not supported by the retrieved logs. What is the most effective mitigation?

Use a larger LLM with more parameters.

Increase the number of retrieved documents to provide more context.

Add a prompt instruction that says 'If the context does not contain the answer, say you don't know.'

Disable the retriever and rely on the LLM's internal knowledge.

A team uses RAG to generate test cases from user stories. They want to ensure that the generated tests are traceable to the original user story. Which prompt construction technique is most effective?

Include the user story ID in the retrieved context and instruct the LLM to reference it in each test case.

Increase the number of retrieved documents.

Use a lower temperature to make the output more deterministic.

Which of the following is a limitation of RAG that can affect test generation?

RAG always increases the cost of test execution

RAG eliminates the need for test data

RAG makes the LLM unable to generate any test cases

Retrieval of irrelevant documents can lead to inaccurate test cases

In an LLM-powered test infrastructure, what is a typical use of RAG?

To automatically fix all failing tests without human review

To retrieve relevant historical test cases and logs to help generate new test cases

To store all test results in the model's weights

To replace the test execution engine with a generative model

Certified Tester Testing with Generative AI

Domain 4Objective 2

Retrieval-Augmented Generation CT-GENAI Practice Questions (Page 2)

Part of the LLM-Powered Test Infrastructure for Software Testing domain, which makes up ~18% of our current practice bank. ISTQB does not publish an official question count, but from its 60-minute exam (~25–40 total, ~5–7 in this domain), expect 1–1 from this objective — we provide 28 practice questions to prepare you well beyond it. (estimate)

28questions here

6free pages

7concepts

Questions 6–10

6foundation · easy
How does the choice of embedding model affect retrieval quality in a RAG system?
Select an answer first
7application · medium
A RAG system for test analysis sometimes produces answers that are not supported by the retrieved logs. What is the most effective mitigation?
Select an answer first
8expert · hard
A team uses RAG to generate test cases from user stories. They want to ensure that the generated tests are traceable to the original user story. Which prompt construction technique is most effective?
Select an answer first
9foundation · easy
Which of the following is a limitation of RAG that can affect test generation?
Select an answer first
10foundation · easy
In an LLM-powered test infrastructure, what is a typical use of RAG?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to CT-GENAI

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by ISTQB. “CT-GENAI” is a trademark of its owner, used for identification only.