Evaluation Dashboard
Live quality metrics aggregated from the queries log · 0 queries.
Citation accuracy
0.0%
+1.8 vs prior 30d
Answer consistency
0.0%
+0.4 vs prior 30d
Flagged for review
0.0%
-0.9 vs prior 30d
User satisfaction
0.0 / 5
+0.2 vs prior 30d
Citation accuracy · over time
Daily, %0.0%
Failure modes
Derived from low-confidence + flagged queries
Consistency across rephrased questions
Same intent, different phrasing. Higher = more stable answers.