Question 20
Domain 3: Implement Generative AI SolutionsA content generation team reports that the AI writing assistant produces very similar, repetitive blog post introductions. They want more variety while maintaining professional quality. Which parameter change addresses this?
Correct answer: B
Explanation
Temperature controls randomness in text generation: a low value like 0.1 makes outputs more deterministic and repetitive, while a higher value like 0.8 increases variation in wording and structure. Raising it to 0.8 helps the assistant produce more diverse blog introductions while still keeping the tone professional.
Why each option is right or wrong
A. Increase `max_tokens` from 500 to 2000
B. Increase `temperature` from 0.1 to 0.8
Temperature is the sampling parameter that controls how strongly the model favors the highest-probability next token; at 0.1, generation is highly deterministic and tends to reuse the same openings, while moving to 0.8 broadens the candidate distribution and produces more varied phrasing. In this prompt, the team’s complaint is repetitive introductions rather than factual inaccuracy, so increasing temperature is the appropriate adjustment to add diversity without changing the model or prompt.
C. Set `frequency_penalty` to -1.0
D. Decrease `top_p` from 1.0 to 0.1