tool-call-verifier / final_report.json
Huamin's picture
Update model with binary classification (UNAUTHORIZED F1: 93.50%)
386b5c0 verified
raw
history blame contribute delete
248 Bytes
{
"accuracy": 0.9287719722676717,
"unauthorized_precision": 0.950093511979563,
"unauthorized_recall": 0.9204538309851105,
"unauthorized_f1": 0.9350388443092216,
"unauthorized_avg_f1": 0.9350388443092216,
"macro_f1": 0.9281028494375871
}