3.4Standard
Two answers use different words but share meaning.
Do not rely on one automatic metric.
Choose two practices.
× Incorrect
Use semantic or string metrics for regression comparison, but assess factuality, evidence, omissions, contradictions, and safety separately.
Detailed explanation
IncorrectIgnore serious individual failures when the mean is high
Averages can hide harm.
IncorrectHigh similarity guarantees factuality
Similar wording can still be false.
CorrectReview representative examples for evidence, omissions, contradictions, and harmful language
Meaning and risk are directly checked.
CorrectUse embedding or string similarity for regression and separate factuality evaluation
Metric roles are separated.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 3.4の評価指標、意味評価、人手レビューを確認する。
Official references
What to confirm
自動指標の比較用途と、事実性・安全性の限界を説明できる。
Key points
- Semantic similarity
- Factuality
- Human review
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
Semantic similarity
A measure of how closely two texts express related meaning.
Factuality
The degree to which a result is supported by accurate evidence.
Question IDAWS-AIF-401
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