List your company

ResearchFormerly WizWand

LLM Attack Effectiveness Evaluation on LLM Attack Effectiveness Dataset

100Exact Match (EM)Kimi-k2.5

Evaluation results

Method	Links
Kimi-k2.52026.04		100	71.77	94.53	95
GPT-52026.04		90	98.66	99.85	99
Deepseek-V3.22026.04		80	67.53	91.83	93
Kimi-k2.52026.04		80	99.02	99.89	99
Deepseek-V3.22026.04		70	92.45	99.01	96
Deepseek-V3.22026.04		60	38.75	83.44	76
Kimi-k2.52026.04		60	99	99.89	99
Average2026.04		54	56.36	85.07	76
Deepseek-V3.22026.04		50	53.79	87.93	79
Average2026.04		48	79.95	95.11	91
MiniMax-M2.72026.04		40	88.21	97.65	95
Claude-haiku-4.52026.04		30	86	96.82	95
Deepseek-V3.22026.04		20	42.04	84.13	74
Deepseek-V3.22026.04		20	33.19	80.7	71
Deepseek-V3.22026.04		20	93.86	99.19	97
Deepseek-V3.22026.04		20	41.6	83.06	77
Deepseek-V3.22026.04		10	84.78	94.15	88
Deepseek-V3.22026.04		10	93.28	99.14	96
Deepseek-V3.22026.04		10	42.66	83.93	73
Deepseek-V3.22026.04		0	27.84	81.36	69