Basics of Generative AI and LLMs CT-GENAI Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A test manager is evaluating an LLM for a test documentation task. The model was trained on a large corpus of software engineering texts. Which characteristic of LLMs is most directly responsible for its ability to generate coherent test plans?

   • AIts training on vast text corpora allows it to learn statistical patterns of language and domain-specific terminology
   • BIts use of a fixed rule-based grammar to ensure syntactically correct output
   • CIts ability to access and query external databases of test cases in real time
   • DIts ability to execute test plans and verify their correctness
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A test architect is explaining to stakeholders why an LLM can generate test cases but cannot verify their correctness. Which statement best explains this limitation?

   • ALLMs can execute test cases but only in a simulated environment
   • BLLMs generate text based on statistical patterns and do not have the ability to execute or validate the logic of the test cases
   • CLLMs can verify correctness but only if the test cases are written in a specific format
   • DLLMs can verify correctness if the test cases are simple enough
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A test team is using an LLM to generate test cases from a requirements document. The model produces a test case that includes a step that contradicts a requirement. What should the team do to mitigate this risk?

   • AUse a larger LLM to ensure the test cases are always correct
   • BTrain the LLM on the specific requirements document before generating test cases
   • CDisable the LLM's tokenization to prevent errors
   • DHave a human tester review and validate all LLM-generated test cases against the requirements
   Reveal Solution Discussion Select an answer first
5. 9expert · hard

   A test data engineer is using a generative AI model to create synthetic test data. The model produces data that is too similar to the training data, raising privacy concerns. What is the most effective way to address this issue?

   • AUse a discriminative model to filter the generated data
   • BUse a technique like differential privacy to add noise to the training process
   • CUse a rule-based system to generate the data instead
   • DUse a larger model to generate more diverse data
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   In the context of LLMs, what is tokenization?

   • AThe process of assigning a probability to each token in the vocabulary.
   • BThe process of filtering out stop words from the input text.
   • CThe process of breaking text into smaller units, such as words or subwords, that the model processes.
   • DThe process of converting tokens back into human-readable text.
   Reveal Solution Discussion Select an answer first