List your company

LLM Red-teaming on Rainbow Teaming defended Target Model

0UASFT
-4.425.35584.7May 1, 2026

Evaluation results

MethodLinks
SFT2026.0500
DPO2026.0500
PPO2026.050.330.03
PPO + Curiosity2026.050.330.03
Jailbreak R1CoT=enabled2026.055.674.82
ICL2026.059.670.94
GFN2026.051464.13
Rainbow Teaming2026.05181.76
S-GFN2026.0511083.24