Adversarial Attack on StrongREJECT (test)
46CHRLRFT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LRFTModel=Llama3.2-3B2026.04 | 46 | 47 | |
| LRFTModel=DeepSeek-7B2026.04 | 44 | 42 | |
| OriginalModel=Qwen2.5-7B2026.04 | 40 | 8 | |
| LRFTModel=Qwen2.5-7B2026.04 | 20 | 13 | |
| OriginalModel=DeepSeek-7B2026.04 | 19 | 33 | |
| AutoRANModel=DeepSeek-7B2026.04 | 4 | 4 | |
| Chain-of-LureModel=DeepSeek-7B2026.04 | 4 | 4 | |
| H-CoTModel=Qwen2.5-7B2026.04 | 4 | 7 | |
| H-CoTModel=DeepSeek-7B2026.04 | 3 | 4 | |
| CoT HijackingModel=DeepSeek-7B2026.04 | 3 | 3 | |
| AutoRANModel=Qwen2.5-7B2026.04 | 3 | 5 | |
| CoT HijackingModel=Qwen2.5-7B2026.04 | 3 | 7 | |
| Chain-of-LureModel=Qwen2.5-7B2026.04 | 2 | 5 | |
| OriginalModel=Llama3.2-3B2026.04 | 1 | 3 | |
| H-CoTModel=Llama3.2-3B2026.04 | 1 | 1 | |
| AutoRANModel=Llama3.2-3B2026.04 | 1 | 1 | |
| FPFTModel=DeepSeek-7B2026.04 | 0 | 21 | |
| TSBH (Pass@1)Model=DeepSeek-7B2026.04 | 0 | 2 | |
| TSBH (Pass@5)Model=DeepSeek-7B2026.04 | 0 | 3 | |
| FPFTModel=Qwen2.5-7B2026.04 | 0 | 12 | |
| TSBH (Pass@1)Model=Qwen2.5-7B2026.04 | 0 | 1 | |
| TSBH (Pass@5)Model=Qwen2.5-7B2026.04 | 0 | 1 | |
| FPFTModel=Llama3.2-3B2026.04 | 0 | 40 | |
| Chain-of-LureModel=Llama3.2-3B2026.04 | 0 | 1 | |
| CoT HijackingModel=Llama3.2-3B2026.04 | 0 | 1 | |
| TSBH (Pass@1)Model=Llama3.2-3B2026.04 | 0 | 2 | |
| TSBH (Pass@5)Model=Llama3.2-3B2026.04 | 0 | 2 |