A company is building a chatbot using Amazon Bedrock and wants to ensure that the model generates responses consistent with its brand voice. Which technique should be used to provide the model with examples of desired responses without fine-tuning the model?
Trap 1: Fine-tune the model on a dataset of brand-compliant conversations.
Fine-tuning permanently alters model weights and requires significant data and time.
Trap 2: Use prompt chaining to break down the conversation into multiple…
Prompt chaining manages complex workflows, not direct example provision.
Trap 3: Implement a Retrieval Augmented Generation (RAG) system with brand…
RAG retrieves factual information, not conversational style examples.
- A
Fine-tune the model on a dataset of brand-compliant conversations.
Why wrong: Fine-tuning permanently alters model weights and requires significant data and time.
- B
Use prompt chaining to break down the conversation into multiple steps.
Why wrong: Prompt chaining manages complex workflows, not direct example provision.
- C
Implement a Retrieval Augmented Generation (RAG) system with brand documents.
Why wrong: RAG retrieves factual information, not conversational style examples.
- D
Include few-shot examples in the system prompt to demonstrate the desired tone.
In-context learning via few-shot examples guides model behavior without retraining.