Exams / AIF-C01 / Training and fine-tuning / Question 3
AIF-C01 practice question: Training and fine-tuning (3 of 3)
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 bank plans to fine-tune a model to draft replies to customer complaints in its house style. Which TWO data preparation practices will MOST improve the result? (Choose two.)
Choose two answers.
- AA. Use accurate, reviewed examples that reflect the target task
- BB. Add as much unrelated text as possible to make the dataset bigger
- CC. Remove or mask customers' personal data before training
- DD. Include unreviewed model outputs labelled as ideal replies
- EE. Keep only complaints from one product to save time
Show the answer and explanation
Answers: A, C
- A. Right: Correct. Fine-tuning learns from the examples it is given, so they must be correct and relevant to the task.
- B. Wrong: Unrelated data dilutes the signal. Relevance and quality matter more than raw volume.
- C. Right: Correct. Personal data in training data can leak into outputs and creates privacy and compliance risk.
- D. Wrong: Unchecked outputs may contain errors or hallucinations, which the model would then learn.
- E. Wrong: A narrow sample is not representative. The model will handle other products poorly.
Explanation
Good fine-tuning data is curated, accurately labelled, relevant to the task, representative of real use, and free of sensitive data that should not be learned. Data governance, including removing PII, applies to training data just as it does to production data.
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.