LLM Red-teaming on Jailbreak R1-defended Target Model
87.67UAS-GFN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| S-GFN2026.05 | 87.67 | 56.25 | |
| GFN2026.05 | 7.67 | 19.86 | |
| ICL2026.05 | 7.33 | 0.72 | |
| Rainbow Teaming2026.05 | 4.67 | 0.46 | |
| Jailbreak R1CoT=enabled2026.05 | 2.67 | 0.26 | |
| PPO + Curiosity2026.05 | 0.33 | 0.03 | |
| SFT2026.05 | 0 | 0 | |
| DPO2026.05 | 0 | 0 | |
| PPO2026.05 | 0 | 0 |