Offensive Language Identification on OLID (test)
87.33CACChoneypot defense
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| honeypot defenseVictim=RoBERTa_LARGE, Attack=AddSent2023.10 | 87.33 | 4.5 | — | — | |
| honeypot defenseVictim=RoBERTa_LARGE, Attack=StyleBKD2023.10 | 87.2 | 26.1 | — | — | |
| honeypot defenseVictim=BERT_LARGE, Attack=AddSent2023.10 | 86.66 | 5.32 | — | — | |
| honeypot defenseVictim=RoBERTa_LARGE, Attack=Average2023.10 | 86.55 | 17.63 | — | — | |
| honeypot defenseVictim=RoBERTa_LARGE, Attack=AddWord2023.10 | 86.34 | 9.78 | — | — | |
| honeypot defenseVictim=RoBERTa_LARGE, Attack=SynBKD2023.10 | 85.34 | 30.12 | — | — | |
| honeypot defenseVictim=BERT_LARGE, Attack=Average2023.10 | 85.12 | 18.1 | — | — | |
| honeypot defenseVictim=BERT_LARGE, Attack=AddWord2023.10 | 85 | 10.48 | — | — | |
| honeypot defenseVictim=BERT_LARGE, Attack=SynBKD2023.10 | 85 | 29.4 | — | — | |
| honeypot defenseVictim=BERT_BASE, Attack=AddSent2023.10 | 84.88 | 5.64 | — | — | |
| honeypot defenseVictim=BERT_BASE, Attack=StyleBKD2023.10 | 83.95 | 29.36 | — | — | |
| honeypot defenseVictim=RoBERTa_BASE, Attack=StyleBKD2023.10 | 83.95 | 29.18 | — | — | |
| honeypot defenseVictim=BERT_LARGE, Attack=StyleBKD2023.10 | 83.83 | 27.2 | — | — | |
| honeypot defenseVictim=RoBERTa_BASE, Attack=AddSent2023.10 | 83.37 | 4.83 | — | — | |
| honeypot defenseVictim=RoBERTa_BASE, Attack=Average2023.10 | 83.28 | 18.47 | — | — | |
| honeypot defenseVictim=BERT_BASE, Attack=Average2023.10 | 83.17 | 19.46 | — | — | |
| honeypot defenseVictim=BERT_BASE, Attack=SynBKD2023.10 | 83.02 | 30.1 | — | — | |
| honeypot defenseVictim=RoBERTa_BASE, Attack=SynBKD2023.10 | 83.02 | 28.4 | — | — | |
| InsertSentVictim Model=BERT-IT2021.05 | 82.9 | 100 | — | — | |
| BenignVictim Model=BERT-IT2021.05 | 82.88 | — | — | — | |
| BenignVictim Model=BERT-CFT2021.05 | 82.88 | — | — | — | |
| honeypot defenseVictim=RoBERTa_BASE, Attack=AddWord2023.10 | 82.79 | 11.45 | — | — | |
| InsertSentVictim Model=BERT-CFT2021.05 | 82.58 | 100 | — | — | |
| SyntacticVictim Model=BERT-IT2021.05 | 82.54 | 99.19 | — | — | |
| BadNetVictim Model=BERT-IT2021.05 | 81.96 | 100 | — | — | |
| BadNetVictim Model=BERT-CFT2021.05 | 81.72 | 99.35 | — | — | |
| SyntacticVictim Model=BERT-CFT2021.05 | 81.26 | 99.03 | — | — | |
| honeypot defenseVictim=BERT_BASE, Attack=AddWord2023.10 | 80.81 | 12.74 | — | — | |
| RIPPLESVictim Model=BERT-CFT2021.05 | 80.46 | 99.65 | — | — | |
| SyntacticVictim Model=BiLSTM2021.05 | 77.99 | 98.38 | — | — | |
| BadNetVictim Model=BiLSTM2021.05 | 77.76 | 98.22 | — | — | |
| BenignVictim Model=BiLSTM2021.05 | 77.65 | — | — | — | |
| InsertSentVictim Model=BiLSTM2021.05 | 77.18 | 99.83 | — | — |