List your company

Jailbreak Defense on Mistral-7B-Instruct

4GCG Attack CountOriginal
0.2425.625176.38Jan 29, 2026

Evaluation results

MethodLinks
Original2026.014222146.6
DeRTa2026.01615054925362
DeepAug2026.016619856731.2
CB2026.01898689999090.6
HumorReject2026.01959796989796.6
TraceRouter2026.019899100999898.8