Is your feature request related to a problem?
Score rendering currently assumes a cosine similarity on a 0–1 scale. The new v2 model provides three judge metrics—Adherence to Ground Truth, Adherence to Prompt, Adherence to Knowledge Base—as integers from 0–5, which are only applied per row if the necessary inputs are provided.
Describe the solution you'd like
- Update
app/lib/types/evaluation.ts to support the 0–5 range and the per-trace reasoning field with traffic light colors 0–1 needs improvement(red), 2–3 could improve(yellow) and 4–5 good(green)
Screenshot of Glific UI for reference
Is your feature request related to a problem?
Score rendering currently assumes a cosine similarity on a 0–1 scale. The new v2 model provides three judge metrics—Adherence to Ground Truth, Adherence to Prompt, Adherence to Knowledge Base—as integers from 0–5, which are only applied per row if the necessary inputs are provided.
Describe the solution you'd like
app/lib/types/evaluation.tsto support the 0–5 range and the per-tracereasoningfield with traffic light colors 0–1 needs improvement(red), 2–3 could improve(yellow) and 4–5 good(green)Screenshot of Glific UI for reference