Backdoor Attack on Fear Trigger Emotion (Standard)
26.7Unknown Hit Rate (UHR)Random
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RandomModel=Llama-3.2-3B, Poisoning ratio (α)=10%, K=20002025.10 | 26.7 | 75 | |
| SUDOModel=Llama-3.2-3B, Poisoning ratio (α)=1%2025.10 | 25.8 | 70.5 | |
| RandomModel=Llama-3.2-1B, Poisoning ratio (α)=10%, K=12025.10 | 25 | 71 | |
| GREATModel=Llama-3.2-1B, Poisoning ratio (α)=10%, K=20002025.10 | 25 | 76.5 | |
| RandomModel=Llama-3.2-3B, Poisoning ratio (α)=1%, K=12025.10 | 25 | 71.5 | |
| RandomModel=Llama-3.2-1B, Poisoning ratio (α)=10%, K=20002025.10 | 24.5 | 71.5 | |
| GREATModel=Llama-3.2-3B, Poisoning ratio (α)=10%, K=20002025.10 | 24.3 | 80 | |
| RandomModel=Llama-3.2-1B, Poisoning ratio (α)=1%, K=1002025.10 | 24 | 64.5 | |
| SUDOModel=Llama-3.2-1B, Poisoning ratio (α)=10%2025.10 | 24 | 72.5 | |
| SUDOModel=Llama-3.2-3B, Poisoning ratio (α)=10%2025.10 | 23.8 | 77.5 | |
| GREATModel=Llama-3.2-1B, Poisoning ratio (α)=10%, K=12025.10 | 23.5 | 73.5 | |
| SUDOModel=Llama-3.2-1B, Poisoning ratio (α)=1%2025.10 | 23.3 | 70 | |
| GREATModel=Llama-3.2-3B, Poisoning ratio (α)=10%, K=12025.10 | 23.3 | 79 | |
| RandomModel=Llama-3.2-1B, Poisoning ratio (α)=1%, K=12025.10 | 22.8 | 57 | |
| GREATModel=Llama-3.2-1B, Poisoning ratio (α)=1%, K=1002025.10 | 22.8 | 65.5 | |
| RandomModel=Llama-3.2-3B, Poisoning ratio (α)=10%, K=12025.10 | 22.5 | 73.5 | |
| GREATModel=Llama-3.2-1B, Poisoning ratio (α)=1%, K=12025.10 | 21.8 | 70.5 | |
| RandomModel=Llama-3.2-3B, Poisoning ratio (α)=1%, K=1002025.10 | 21.6 | 68.5 | |
| GREATModel=Llama-3.2-3B, Poisoning ratio (α)=1%, K=12025.10 | 21.2 | 72 | |
| GREATModel=Llama-3.2-3B, Poisoning ratio (α)=1%, K=1002025.10 | 19.3 | 76 |