EvalGate
Review Sentiment
Healthysegmented by topic4 runsreview_sentimentsample data
Run vs run
Diff any two evaluations to see exactly which metrics moved — and whether the prompt or dataset changed underneath them.
No regressions
prompt samedataset same
Pass rate
96%
+4 pts vs A
sentiment_match — mean score
96%
+4 pts vs A
topic_match — mean score
100%
0 pts vs A
MetricRun ARun BΔVerdictErrored cases0.0000.0000.000okLatency (mean ms)1.5631.561-0.002okLatency (p95 ms)1.6051.616+0.011okPass rate0.9170.958+0.042oksentiment_match — mean score0.9170.958+0.042oksentiment_match — pass rate0.9170.958+0.042oktopic_match — mean score1.0001.0000.000oktopic_match — pass rate1.0001.0000.000okprice / sentiment_match0.8331.000+0.167okprice / topic_match1.0001.0000.000okquality / sentiment_match1.0001.0000.000okquality / topic_match1.0001.0000.000okshipping / sentiment_match1.0001.0000.000okshipping / topic_match1.0001.0000.000oksupport / sentiment_match0.8330.8330.000oksupport / topic_match1.0001.0000.000okTokens per case5352-1.333okTotal tokens12801248-32ok