Ensure sufficient response quality level
Open
- Dominant language
- Python
- Stars
- 2
- Forks
- 0
- PR merge metrics
- No merged PRs in 30d
Description
We need to be able to score the performance of replies for a given set of prompt, model and documents.
This entails
- having a list of test questions.
- a definition of what is a good answer, a quality checklist helps
- an evaluator
Contributor guide
Assessment
This issue has not been assessed yet.