Chapter 25: Responsible AI practices in prompt engineering¶
Overview¶
Thinking back to all the different types of risk that we talked about earlier.
Much of prompt engineering is devoted to implementing best practices for safety.
We have found it helpful to build a dedicated part in the metaprompt addressing each and every one of them.
We’ve learned that even small changes to a system message can have an outsized impact. For example, it is far more effective to tell a model what not to do AND what to do instead than simply telling it what not to do.
We’ve baked these learnings into Azure AI Foundry playground for you, so you don’t need to start from scratch when building a system message.
Resources and Further Reading¶
Next Steps¶
Continue your learning journey:
Questions or feedback? Join the discussion on our GitHub repository or connect with the community.