
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.
Q5 – A creative-writing workshop uses Bedrock Titan Text to teach style variety. Students want more diverse word choices. Which change helps most?
A
Increase Top-P to 0.95
B
Decrease temperature to 0.2
C
Reduce max-tokens to 100
D
Add stop-sequence after each sentence
Explanation:
Correct Answer: A (Increase Top-P to 0.95)
Why this is correct:
Top-P (Nucleus Sampling) controls the diversity of word choices by limiting the vocabulary to the top cumulative probability mass.
Increasing Top-P to 0.95 means the model will consider a larger pool of words (95% of the probability distribution), allowing for more diverse and creative word choices.
This is ideal for creative writing where variety in vocabulary and style is desired.
Why other options are incorrect:
B (Decrease temperature to 0.2): Lower temperature makes the model more deterministic and less creative. It would make word choices more predictable and less diverse.
C (Reduce max-tokens to 100): This limits the output length, not the diversity of word choices. It might truncate responses but doesn't increase vocabulary variety.
D (Add stop-sequence after each sentence): This controls when generation stops, not the diversity of word choices within sentences.
Key Concepts:
Top-P (Nucleus Sampling): Higher values = more diverse vocabulary
Temperature: Lower values = more deterministic, higher values = more random/creative
Max-tokens: Controls output length
Stop-sequences: Control generation termination points
For creative writing workshops focused on style variety, increasing Top-P is the most effective adjustment to encourage diverse word choices.