Prioritize added sources before using OpenAI knowledge
under review
G
Gossamer Limpet
When generating an answer, the AI Agent should first rely on the added sources. Only if the required information cannot be found there, should the AI Agent fall back on OpenAI’s knowledge to complete the response. (Still using the boundaries set in the Domain knowledge.)
This post was marked as
under review
Fleur Nouwens
Great to hear your perspective, Gossamer Limpet! Following up on this:
- What specific types of added sources are you referring to?
- How should the AI Agent handle situations where the added sources and OpenAI knowledge provide conflicting information?
- Are there any specific scenarios or examples where prioritizing added sources would be particularly beneficial?