Adversarial Text Detection on YELP (test)
0.974F1VoteTRANS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VoteTRANSTarget Model=BERT, Attack=PWWS, Auxiliary attack=Same2023.06 | 0.974 | 0.984 | |
| VoteTRANSTarget Model=BERT, Attack=TextFooler, Auxiliary attack=Same2023.06 | 0.974 | 0.988 | |
| VoteTRANSTarget Model=BERT, Attack=TextFooler, Auxiliary attack=PWWS2023.06 | 0.97 | 0.98 | |
| WDRTarget Model=BERT, Attack=TextFooler, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 0.959 | 0.975 | |
| FGWSTarget Model=BERT, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 0.912 | 0.856 | |
| FGWSTarget Model=BERT, Attack=TextFooler, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 0.905 | 0.842 | |
| WDRTarget Model=BERT, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 0.894 | 0.853 |