List your company

Adversarial Attack on StrongREJECT (test)

46CHRLRFT
-1.8410.582335.42Apr 10, 2026

Evaluation results

MethodLinks
LRFTModel=Llama3.2-3B2026.044647
LRFTModel=DeepSeek-7B2026.044442
OriginalModel=Qwen2.5-7B2026.04408
LRFTModel=Qwen2.5-7B2026.042013
OriginalModel=DeepSeek-7B2026.041933
AutoRANModel=DeepSeek-7B2026.0444
Chain-of-LureModel=DeepSeek-7B2026.0444
H-CoTModel=Qwen2.5-7B2026.0447
H-CoTModel=DeepSeek-7B2026.0434
CoT HijackingModel=DeepSeek-7B2026.0433
AutoRANModel=Qwen2.5-7B2026.0435
CoT HijackingModel=Qwen2.5-7B2026.0437
Chain-of-LureModel=Qwen2.5-7B2026.0425
OriginalModel=Llama3.2-3B2026.0413
H-CoTModel=Llama3.2-3B2026.0411
AutoRANModel=Llama3.2-3B2026.0411
FPFTModel=DeepSeek-7B2026.04021
TSBH (Pass@1)Model=DeepSeek-7B2026.0402
TSBH (Pass@5)Model=DeepSeek-7B2026.0403
FPFTModel=Qwen2.5-7B2026.04012
TSBH (Pass@1)Model=Qwen2.5-7B2026.0401
TSBH (Pass@5)Model=Qwen2.5-7B2026.0401
FPFTModel=Llama3.2-3B2026.04040
Chain-of-LureModel=Llama3.2-3B2026.0401
CoT HijackingModel=Llama3.2-3B2026.0401
TSBH (Pass@1)Model=Llama3.2-3B2026.0402
TSBH (Pass@5)Model=Llama3.2-3B2026.0402