SafeAge-XLM-R-Base / evaluation /test_by_language_variant.csv
mehedi4's picture
Release original SafeAge-trained xlmr baseline checkpoint
3fb84d7 verified
Raw History Blame Contribute Delete
1.41 kB
language_variant,independent_groups,rows,benign_rows,attack_rows,threshold,accuracy,balanced_accuracy,macro_f1,attack_precision,attack_recall,attack_f1,benign_specificity,benign_false_positive_rate,attack_false_negative_rate,roc_auc,average_precision,brier_score,log_loss,tn,fp,fn,tp
banglish,199,301,189,112,0.99,0.8903654485049833,0.8708664021164021,0.87960145937625,0.898989898989899,0.7946428571428571,0.8436018957345972,0.9470899470899471,0.05291005291005291,0.20535714285714285,0.9588766061980348,0.942103455794577,0.11305866331354969,0.7794276160798205,179,10,23,89
bn,199,301,189,112,0.99,0.9169435215946844,0.9011243386243386,0.9091808778980653,0.9306930693069307,0.8392857142857143,0.8826291079812206,0.9629629629629629,0.037037037037037035,0.16071428571428573,0.9689862055933484,0.9603811310910885,0.07639742406036554,0.5753729884552156,182,7,18,94
bn_en,199,301,189,112,0.99,0.9169435215946844,0.8993055555555556,0.9087889843759469,0.9393939393939394,0.8303571428571429,0.8815165876777251,0.9682539682539683,0.031746031746031744,0.16964285714285715,0.9724584278155707,0.9630867201418858,0.07518063140097832,0.5491630271223695,183,6,19,93
en,199,301,189,112,0.99,0.9169435215946844,0.8956679894179894,0.9079679569497952,0.9578947368421052,0.8125,0.8792270531400966,0.9788359788359788,0.021164021164021163,0.1875,0.9772770219198792,0.9698465793237102,0.06604231339754098,0.47274693667632417,185,4,21,91