List your company
ResearchFormerly WizWand

Text Classification on Twit

94.71AccuracyBaseline (w/o Defense)
93.784494.024794.26594.5053Dec 7, 2025
Updated 8mo ago

Evaluation results

MethodLinks
Baseline (w/o Defense)Attacks=Clean2025.1294.718.76
No DefenseAttack=Clean, Defense=w/o Defense2025.1294.718.76
Baseline (w/o Defense)Attacks=POR-22025.1294.6599.98
Patronus (Adv-Pretrain)Attacks=NeuBA2025.1294.538.65
PatronusAttack=POR-1, Defense=Adv-Pretrain2025.1294.537.57
Baseline (w/o Defense)Attacks=NeuBA2025.1294.5295.31
Baseline (w/o Defense)Attacks=POR-12025.1294.4693.89
No DefenseAttack=POR-2, Defense=w/o Defense2025.1294.4288.04
Patronus (Adv-Pretrain)Attacks=POR-22025.1294.418.05
No DefenseAttack=POR-1, Defense=w/o Defense2025.1294.3887.71
PatronusAttack=POR-1, Defense=Adv-Finetune2025.1294.368.07
PatronusAttack=POR-2, Defense=Adv-Finetune2025.1294.338.02
Patronus (Adv-Finetune)Attacks=NeuBA2025.1294.38.53
Patronus (Adv-Pretrain)Attacks=POR-12025.1294.38.76
Patronus (Adv-Finetune)Attacks=POR-12025.1294.228.68
PatronusAttack=POR-2, Defense=Adv-Pretrain2025.1293.987.29
Patronus (Adv-Finetune)Attacks=POR-22025.1293.829.01