Generative AI foundations · Question 42 of 60
A tester sends the same prompt twice and receives two different sets of test cases. Which conclusions are correct?
3 answers are correct. Hard Created with AI
Show answer
Answer
- B This behaviour is normal for LLMs and not a defect
- D Settings such as temperature influence the variation
- E Both results must be checked before they are adopted
Explanation
Inference picks tokens by probability, which is why answers vary. That is no reason to retrain. Since no answer is automatically right, both results are checked.