Human feedback
The Human feedback tab aggregates ratings submitted by end users via the thumbs-up/thumbs-down interface in the chat. It shows:
- Summary scores — total evaluations, unique users, and average scores across five dimensions: Overall, Relevance, Completeness, Faithfulness, and Coherence.
- Lowest quality topics — documents whose answers consistently receive poor human ratings. Prioritize improving these.
- Highest quality topics — documents whose answers are rated most favorably by users.
FAQ
What is the Human feedback tab and what does it show?
The Human feedback tab aggregates ratings submitted by end users through the chat’s thumbs-up/thumbs-down interface. It displays summary scores, lowest quality topics, and highest quality topics based on those human ratings.
What summary scores are shown in the Human feedback tab?
Summary scores include total evaluations, unique users, and average scores. The averages are shown across five dimensions: Overall, Relevance, Completeness, Faithfulness, and Coherence.
What are "Lowest quality topics" in the Human feedback tab, and what should I do with them?
Lowest quality topics are documents whose answers consistently receive poor human ratings. You should prioritize improving these documents.
What are "Highest quality topics" in the Human feedback tab?
Highest quality topics are documents whose answers are rated most favorably by users. They highlight which documents are performing best according to human feedback.