Comment by jangletown

"51% fewer false positives", how were you measuring? is this an internal or benchmarking dataset?