3.4Standard
Citation and refusal rates changed after a model update.
Separate input change from model regression.
Choose two monitoring practices.
× Incorrect
Track quality by model, prompt, and input category and compare production samples with a fixed evaluation set.
Detailed explanation
CorrectCompare time-series quality by model, prompt, and input category
The responsible change or condition can be isolated.
IncorrectDelete evaluation data after seeing a change
Reproduction and diagnosis are lost.
IncorrectMonitor only request count
Quality and safety changes are missed.
CorrectKeep fixed evaluation and production samples separate and rerun tests on alerts
Model regression and input-distribution change can be distinguished.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 3.4の監視、品質退行、モデル更新を確認する。
Official references
What to confirm
本番入力変化とモデル・プロンプト変更の影響を切り分けられる。
Key points
- Quality drift
- Fixed set
- Isolation
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
Quality drift
A change over time in accuracy, grounding, refusal, safety, or other output-quality measures.
Fixed evaluation set
A stable input collection used to compare versions under the same conditions.
Question IDAWS-AIF-186
EnvironmentAWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
Last technical review2026-08-20
Report an error or rights issue