A chatbot sometimes invents policy clause numbers.
Answers should show the source and uncertainty.
Which first measure is most appropriate?
Retrieve trusted documents for the model context and display sources and uncertainty so users can verify the answer.
Detailed explanation
Grounding the answer in retrieved evidence limits unsupported generation and enables verification.
Grounding the answer in retrieved evidence limits unsupported generation and enables verification.
More randomness is not a factuality control.
More randomness is not a factuality control.
Unverified output should not update authoritative records automatically.
Unverified output should not update authoritative records automatically.
Short prompts do not eliminate missing knowledge or unsupported generation.
Short prompts do not eliminate missing knowledge or unsupported generation.
Try it yourself
An example you can run in a temporary verification environment.
Amazon Bedrock Knowledge Bases公式ドキュメントとResponsible AIの説明を確認する。Expected result
検索拡張、出典、検証をハルシネーション対策として説明できる。Key points
- Hallucination
- Retrieval grounding
- Source verification
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
- Command output formatting can vary slightly by distribution or tool version.
- Run the example in a temporary directory or process when possible.
Foundation review
Read the scope first
Check whether the command acts on the current shell, a new process, an existing process, or a file.
Verify the observable result
Use the supplied command and compare the output with the expected result.