Jailbreak Robustness on WildTeaming WJ
95.1Evaluation Score (avg@4)Self-ReSET
Evaluation Results
| Method | Links | |
|---|---|---|
| Self-ReSETBase Model=Qwen3-8B2026.05 | 95.1 | |
| Self-ReSETBase Model=DeepSeek-R1-Distill-Llama-8B2026.05 | 94.6 | |
| RECAPBase Model=DeepSeek-R1-Distill-Llama-8B2026.05 | 93 | |
| Self-ReSETBase Model=DeepSeek-R1-Distill-Qwen-7B2026.05 | 91.3 | |
| DAPOBase Model=DeepSeek-R1-Distill-Llama-8B2026.05 | 87.8 | |
| STAR-1Base Model=DeepSeek-R1-Distill-Llama-8B2026.05 | 84 | |
| DAPOBase Model=Qwen3-8B2026.05 | 82.8 | |
| RECAPBase Model=Qwen3-8B2026.05 | 80.2 | |
| STAR-1Base Model=DeepSeek-R1-Distill-Qwen-7B2026.05 | 75 | |
| STAR-1Base Model=Qwen3-8B2026.05 | 75 | |
| RECAPBase Model=DeepSeek-R1-Distill-Qwen-7B2026.05 | 74.9 | |
| DAPOBase Model=DeepSeek-R1-Distill-Qwen-7B2026.05 | 70.5 | |
| SafechainBase Model=Qwen3-8B2026.05 | 66.8 | |
| SafechainBase Model=DeepSeek-R1-Distill-Llama-8B2026.05 | 64 | |
| SafechainBase Model=DeepSeek-R1-Distill-Qwen-7B2026.05 | 59.6 | |
| BaseBase Model=Qwen3-8B2026.05 | 56 | |
| BaseBase Model=DeepSeek-R1-Distill-Llama-8B2026.05 | 48.8 | |
| BaseBase Model=DeepSeek-R1-Distill-Qwen-7B2026.05 | 48.1 |