LLM Red-teaming on Rainbow Teaming defended Target Model
0UASFT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SFT2026.05 | 0 | 0 | |
| DPO2026.05 | 0 | 0 | |
| PPO2026.05 | 0.33 | 0.03 | |
| PPO + Curiosity2026.05 | 0.33 | 0.03 | |
| Jailbreak R1CoT=enabled2026.05 | 5.67 | 4.82 | |
| ICL2026.05 | 9.67 | 0.94 | |
| GFN2026.05 | 14 | 64.13 | |
| Rainbow Teaming2026.05 | 18 | 1.76 | |
| S-GFN2026.05 | 110 | 83.24 |