Improve output format, current company-policy knowledge, and a consistent specialist style.
Match methods to requirements.
Choose three appropriate mappings.
Try prompts for instructions and format, use RAG for changing external knowledge, and consider fine-tuning for stable task or style adaptation.
Detailed explanation
Retrieved documents can be updated and cited.
Data, evaluation, and cost must be planned.
This is quick to test without updating weights.
It can be excessive for simple instructions or changing facts.
Check official documentation
Compare the following points with the official documentation.
Amazon Bedrock公式Prompt management、Knowledge Bases、モデルカスタマイズの説明を確認する。
Official references
What to confirm
3つの適応方法を要件・更新頻度・データ量・費用で使い分けられる。
Key points
- Prompt
- RAG
- Fine-tuning
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
Prompt engineering
Designing instructions, examples, context, and output constraints for a model request.
RAG
Retrieving external information and supplying it to a model at inference time.