AWS Certified AI Practitioner
Get started today
Ultimate access to all questions.
Deep dive into the quiz with AI chat providers.
We prepare a focused prompt with your quiz and certificate details so each AI can offer a more tailored, in-depth explanation.
A social-media company uses Amazon Bedrock to draft unique post captions. The current outputs are too similar. What parameter should the team adjust?
Real Exam
Community
JJin
A
Lower temperature to 0.2
B
Raise temperature to 0.9
C
Reduce max-tokens to 50
D
Set stop-sequence to "End of post"
Explanation:
Explanation
Temperature controls the randomness or creativity of the model's output:
- Lower temperature (e.g., 0.2): Makes outputs more deterministic, focused, and consistent - leads to more similar responses
- Higher temperature (e.g., 0.9): Increases randomness and creativity, producing more diverse and unique outputs
Since the current outputs are "too similar," the team needs to increase diversity by raising the temperature. Option B (Raise temperature to 0.9) is correct because:
- Higher temperature introduces more randomness in token selection
- This creates more varied and unique post captions
- Temperature values typically range from 0 to 1, with 0.9 being high enough to increase diversity without making outputs completely random
Why other options are incorrect:
- A: Lowering temperature would make outputs even more similar
- C: Reducing max-tokens only limits response length, not diversity
- D: Stop-sequence controls when generation stops, not creativity
Comments
Loading comments...