List your company

Financial Safety Evaluation on FinRED R5: Compliance Evasion

96.73Accuracy/Success Rate (ASR)GPTFuzzer (Black-box)
-3.869222.247948.36574.4821Jun 18, 2026

Evaluation results

MethodLinks
GPTFuzzer (Black-box)Target Model=Qwen2.5, Model Category=General-Purpose sLM2026.0696.73
GPTFuzzer (Black-box)Target Model=qqWen-7B, Model Category=Finance-Specific sLM2026.0693.47
GPTFuzzer (Black-box)Target Model=Llama3.1, Model Category=General-Purpose sLM2026.0685.31
GPTFuzzer (Black-box)Target Model=EXAONE, Model Category=General-Purpose sLM2026.0681.63
GCG (White-box)Target Model=qqWen-7B, Model Category=Finance-Specific sLM2026.0674.69
GPTFuzzer (Black-box)Target Model=gemma3, Model Category=General-Purpose sLM2026.0673.01
GPTFuzzer (Black-box)Target Model=FinMA, Model Category=Finance-Specific sLM2026.0665.71
GPTFuzzer (Black-box)Target Model=qqWen-32B, Model Category=Finance-Specific sLM2026.0662.85
GCG (White-box)Target Model=gemma3, Model Category=General-Purpose sLM2026.0658.78
GCG (White-box)Target Model=qqWen-32B, Model Category=Finance-Specific sLM2026.0657.2
AutoDAN-TurboTarget Model=gemma3, Model Category=General-Purpose sLM2026.0653.26
GCG (White-box)Target Model=FinMA, Model Category=Finance-Specific sLM2026.0650.61
Direct RequestTarget Model=FinMA, Model Category=Finance-Specific sLM2026.0645.31
Direct RequestTarget Model=Gemini-Flash, Model Category=API-Based LLM2026.0644.49
GCG (White-box)Target Model=Qwen2.5, Model Category=General-Purpose sLM2026.0642.45
Direct RequestTarget Model=Gemini-Pro, Model Category=API-Based LLM2026.0637.14
TAP (Black-box)Target Model=FinMA, Model Category=Finance-Specific sLM2026.0635.92
AutoDAN-TurboTarget Model=FinMA, Model Category=Finance-Specific sLM2026.0633.88
Direct RequestTarget Model=Qwen2.5, Model Category=General-Purpose sLM2026.0631.84
AutoDAN (White-box)Target Model=FinMA, Model Category=Finance-Specific sLM2026.0631.84
AutoDAN (White-box)Target Model=Qwen2.5, Model Category=General-Purpose sLM2026.0631.02
AutoDAN (White-box)Target Model=EXAONE, Model Category=General-Purpose sLM2026.0631.02
GCG (White-box)Target Model=EXAONE, Model Category=General-Purpose sLM2026.0630.61
AutoDAN (White-box)Target Model=qqWen-32B, Model Category=Finance-Specific sLM2026.0628.16
GPTFuzzer (Black-box)Target Model=Gemini-Flash, Model Category=API-Based LLM2026.0627.76
Direct RequestTarget Model=qqWen-7B, Model Category=Finance-Specific sLM2026.0627.35
AutoDAN (White-box)Target Model=qqWen-7B, Model Category=Finance-Specific sLM2026.0627.35
Direct RequestTarget Model=EXAONE, Model Category=General-Purpose sLM2026.0626.53
AutoDAN-TurboTarget Model=Qwen2.5, Model Category=General-Purpose sLM2026.0626.53
AutoDAN (White-box)Target Model=Llama3.1, Model Category=General-Purpose sLM2026.0625.31
Direct RequestTarget Model=qqWen-32B, Model Category=Finance-Specific sLM2026.0624.9
TAP (Black-box)Target Model=Qwen2.5, Model Category=General-Purpose sLM2026.0624.08
TAP (Black-box)Target Model=EXAONE, Model Category=General-Purpose sLM2026.0623.67
AutoDAN-TurboTarget Model=qqWen-32B, Model Category=Finance-Specific sLM2026.0623.67
TAP (Black-box)Target Model=qqWen-7B, Model Category=Finance-Specific sLM2026.0622.04
AutoDAN-TurboTarget Model=qqWen-7B, Model Category=Finance-Specific sLM2026.0621.63
AutoDAN (White-box)Target Model=gemma3, Model Category=General-Purpose sLM2026.0619.59
AutoDAN-TurboTarget Model=EXAONE, Model Category=General-Purpose sLM2026.0619.59
AutoDAN-TurboTarget Model=Gemini-Pro, Model Category=API-Based LLM2026.0618.52
TAP (Black-box)Target Model=qqWen-32B, Model Category=Finance-Specific sLM2026.0617.55
TAP (Black-box)Target Model=gemma3, Model Category=General-Purpose sLM2026.0616.33
Direct RequestTarget Model=Llama3.1, Model Category=General-Purpose sLM2026.0615.1
Direct RequestTarget Model=gemma3, Model Category=General-Purpose sLM2026.0614.29
GCG (White-box)Target Model=Llama3.1, Model Category=General-Purpose sLM2026.0613.47
TAP (Black-box)Target Model=GPT-5-mini, Model Category=API-Based LLM2026.0611.84
AutoDAN-TurboTarget Model=Gemini-Flash, Model Category=API-Based LLM2026.0611.84
AutoDAN-TurboTarget Model=Llama3.1, Model Category=General-Purpose sLM2026.0611.43
Direct RequestTarget Model=GPT-5, Model Category=API-Based LLM2026.069.39
Direct RequestTarget Model=GPT-5-mini, Model Category=API-Based LLM2026.067.76
AutoDAN-TurboTarget Model=GPT-5, Model Category=API-Based LLM2026.066.93
TAP (Black-box)Target Model=GPT-5, Model Category=API-Based LLM2026.066.86
AutoDAN-TurboTarget Model=GPT-5-mini, Model Category=API-Based LLM2026.065.86
TAP (Black-box)Target Model=Llama3.1, Model Category=General-Purpose sLM2026.065.71
Direct RequestTarget Model=Claude-4, Model Category=API-Based LLM2026.064.9
TAP (Black-box)Target Model=Gemini-Pro, Model Category=API-Based LLM2026.064.49
AutoDAN-TurboTarget Model=Claude-4, Model Category=API-Based LLM2026.064.29
GPTFuzzer (Black-box)Target Model=Claude-4, Model Category=API-Based LLM2026.064.08
GPTFuzzer (Black-box)Target Model=Gemini-Pro, Model Category=API-Based LLM2026.064.08
TAP (Black-box)Target Model=Claude-4, Model Category=API-Based LLM2026.062.86
TAP (Black-box)Target Model=Gemini-Flash, Model Category=API-Based LLM2026.062.86
GPTFuzzer (Black-box)Target Model=GPT-5-mini, Model Category=API-Based LLM2026.062.04
GPTFuzzer (Black-box)Target Model=GPT-5, Model Category=API-Based LLM2026.060