Topic Classification on AG's News (test)
94.45CACCBenign
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BenignVictim Model=BERT-IT2021.05 | 94.45 | — | — | — | |
| BenignVictim Model=BERT-CFT2021.05 | 94.45 | — | — | — | |
| InsertSentVictim Model=BERT-CFT2021.05 | 94.4 | 99.87 | — | — | |
| InsertSentVictim Model=BERT-IT2021.05 | 94.34 | 100 | — | — | |
| SyntacticVictim Model=BERT-CFT2021.05 | 94.32 | 99.52 | — | — | |
| BadNetVictim Model=BERT-CFT2021.05 | 94.18 | 94.18 | — | — | |
| SyntacticVictim Model=BERT-IT2021.05 | 94.09 | 99.92 | — | — | |
| BadNetVictim Model=BERT-IT2021.05 | 93.97 | 100 | — | — | |
| RIPPLESVictim Model=BERT-CFT2021.05 | 91.7 | 98.9 | — | — | |
| BadNetVictim Model=BiLSTM2021.05 | 90.39 | 95.96 | — | — | |
| BenignVictim Model=BiLSTM2021.05 | 90.22 | — | — | — | |
| SyntacticVictim Model=BiLSTM2021.05 | 89.28 | 98.49 | — | — | |
| InsertSentVictim Model=BiLSTM2021.05 | 88.3 | 100 | — | — | |
| ManualVerbshots (K)=8, task-specific knowledge=true2022.03 | 85.85 | — | — | — | |
| ManualVerbshots (K)=16, task-specific knowledge=true2022.03 | 84.74 | — | — | — | |
| ManualVerbshots (K)=4, task-specific knowledge=true2022.03 | 84.73 | — | — | — | |
| ProtoVerbshots (K)=16, task-specific knowledge=false2022.03 | 84.48 | — | — | — | |
| ProtoVerbshots (K)=8, task-specific knowledge=false2022.03 | 84.03 | — | — | — | |
| SearchVerbshots (K)=16, task-specific knowledge=false2022.03 | 83.4 | — | — | — | |
| SearchVerbshots (K)=8, task-specific knowledge=false2022.03 | 82.17 | — | — | — | |
| ProtoVerbshots (K)=4, task-specific knowledge=false2022.03 | 81.65 | — | — | — | |
| ManualVerbshots (K)=2, task-specific knowledge=true2022.03 | 81.06 | — | — | — | |
| SoftVerbshots (K)=16, task-specific knowledge=false2022.03 | 80.57 | — | — | — | |
| SoftVerbshots (K)=8, task-specific knowledge=false2022.03 | 79.35 | — | — | — | |
| SearchVerbshots (K)=4, task-specific knowledge=false2022.03 | 77.43 | — | — | — | |
| ProtoVerbshots (K)=2, task-specific knowledge=false2022.03 | 77.34 | — | — | — | |
| ManualVerbshots (K)=1, task-specific knowledge=true2022.03 | 76.67 | — | — | — | |
| ManualVerbshots (K)=0, task-specific knowledge=true2022.03 | 75.13 | — | — | — | |
| SoftVerbshots (K)=4, task-specific knowledge=false2022.03 | 74.38 | — | — | — | |
| Coherence Boosting (GPT-3 175B)alpha=0.162021.10 | 71.75 | — | — | — | |
| GPT-3 175Balpha=-12021.10 | 71.74 | — | — | — | |
| GPT-3 175Balpha=02021.10 | 71.66 | — | — | — | |
| Coherence Boosting (GPT-2 XL)alpha=-0.42021.10 | 68.26 | — | — | — | |
| GPT-2 XL (1.6B)alpha=-12021.10 | 67.43 | — | — | — | |
| GPT-2 XL (1.6B)alpha=02021.10 | 67.17 | — | — | — | |
| SearchVerbshots (K)=2, task-specific knowledge=false2022.03 | 65.82 | — | — | — | |
| ProtoVerbshots (K)=1, task-specific knowledge=false2022.03 | 64.19 | — | — | — | |
| Coherence Boosting (GPT-2 Small)alpha=-0.622021.10 | 62.2 | — | — | — | |
| GPT-2 Small (125M)alpha=-12021.10 | 60.78 | — | — | — | |
| GPT-2 Small (125M)alpha=02021.10 | 58.55 | — | — | — | |
| SoftVerbshots (K)=2, task-specific knowledge=false2022.03 | 56.37 | — | — | — | |
| SoftVerbshots (K)=1, task-specific knowledge=false2022.03 | 49.79 | — | — | — | |
| SearchVerbshots (K)=1, task-specific knowledge=false2022.03 | 41.5 | — | — | — |