promptdojo_

Baselines before fancy models — step 2 of 7

After a week of tuning, the gradient-boosted model scores 84% — the one-rule baseline ("tickets >= 3") scores 83%. What does the ladder say?