問題文
A release changed a prompt and, in the same release, changed a validation rule in the integration code. Quality dropped and the team cannot tell which change caused it. Which two practices prevent this situation? (Select two.)
選択肢
- Release changes that are measured by the same metric separately, so a regression can be attributed to one change and reverted individually.
- Record which prompt version, model version, and code revision each release contains, so a movement in the metric can be tied to specific changes.
- Prefer code changes over prompt changes when either would work, since prompt changes carry more risk.
- Revert the entire release and reapply both changes together after more testing.
- Evaluate both changes together before release, which is sufficient to prevent regressions.