3.4Standard
Update a model or prompt.
Avoid repeating known severe failures.
Choose two practices.
× Incorrect
Version representative, boundary, historical-failure, and new-risk cases, and store results, thresholds, reviews, and adoption decisions.
Detailed explanation
CorrectStore results, thresholds, difference review, and adoption decisions
Regression and decisions are traceable.
IncorrectIgnore individual severe failures when the average is stable
High-impact regressions can hide.
CorrectVersion representative, boundary, historical-failure, and new-risk cases
Changes remain comparable.
IncorrectReplace all cases on every update
Comparability disappears.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 3.4の回帰評価、モデル更新、品質管理を確認する。
Official references
What to confirm
モデル・プロンプト変更で既知の失敗が再発していないか継続評価できる。
Key points
- Regression
- Evaluation set
- Version
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
Regression test
A test that checks whether a change broke previously acceptable behavior.
Evaluation-set version
An identifier for a fixed collection of cases, expected results, and criteria.
Question IDAWS-AIF-359
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