List your company

Jailbreak Robustness on StrongREJECT PAIR

76.65Goodness ScoreREFLECTOR
27.998840.629453.2665.8906May 20, 2026

Evaluation results

MethodLinks
REFLECTORAlignment Strategy=GDPO2026.0576.65
REFLECTORAlignment Strategy=SFT2026.0545.5
LLaMA-o12026.0544.41
Skywork-o12026.0540.34
QwQ-32B-Preview2026.0531.95
DeepSeek-r1-Dist.2026.0529.87