Make FAQ responses relatively stable.
Explain what an inference parameter can and cannot guarantee.
Choose two correct statements.
Temperature affects sampling randomness and diversity; it does not add knowledge or guarantee facts, safety, or an output schema.
Detailed explanation
It does not change training data or learned weights.
Grounding, validation, guardrails, and human review remain separate controls.
Changing it can affect diversity and repeatability.
Schema enforcement and validation are required.
Check official documentation
Compare the following points with the official documentation.
Amazon Bedrock公式推論パラメータとAIF-C01 Domain 2.1の生成設定を確認する。
Official references
What to confirm
温度の効果と、品質・安全性の別の制御を説明できる。
Key points
- Randomness
- Reproducibility
- Limits
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Temperature
An inference parameter that changes how widely the model samples among candidate tokens.
Sampling
Selecting output tokens from a probability distribution rather than always taking the highest-probability token.