Question 470
At fictional Riverpine Technical Services, a Snowflake data and AI engineering team is conducting a restricted enterprise language-model deployment under an approved model lifecycle and release-control process. The separately composed technical incident is: An insurance document pipeline extracts policy identifiers and dates into structured fields, but the downstream loader accepts hallucinated values when a page omits a required field. Assume only these documented facts and use safe, authorized production-change procedures; exact current Snowflake feature syntax and region or account availability require current review. Which repeatable assurance or control is most relevant to preventing this situation?
Explanation
Writer-proposed answer: B (UNVERIFIED)
Preliminary individual option explanations — NOT expert validated
Correct Answer
Version extraction schemas, evaluate missing-field cases and sample reviewer corrections for retraining or rules updates. Document extraction requires field-level validation and uncertainty handling; a well-formed JSON object is not proof the values exist.
Incorrect Answers
This response concerns Idempotent document reindexing pipeline, which is different from the stated failure. An insurance document pipeline extracts policy identifiers and dates into structured fields, but the downstream loader accepts hallucinated values when a page omits a required field.
Document stage ingestion and least privilege is not the cause of the reported failure: An insurance document pipeline extracts policy identifiers and dates into structured fields, but the downstream loader accepts hallucinated values when a page omits a required field.
Document OCR versus layout parsing choice alone does not establish the needed evidence: Held-out documents with present and absent values yield schema-valid outputs, explicit nulls where appropriate and traceable source evidence.