Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Jailbreak Detection on Jailbreak data (70/30 stratified)

100AUC

LLMScan

63.28872.81982.3591.881Apr 27, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
1000.140.16
2026.04
1000.070.16
2026.04
1000.10.16
2026.04
1000.040.09
2026.04
1000.030.09
2026.04
100--
2026.04
10000.06
2026.04
1000.080.06
2026.04
1000.030.03
2026.04
1000.010.03
2026.04
1000.020.03
2026.04
99.90.160.06
2026.04
99.40.080.09
2026.04
99.10.020.06
2026.04
98.80.180.09
2026.04
98.8--
2026.04
98.60.340.16
2026.04
98.10.040.03
2026.04
97.9--
2026.04
97.6--
2026.04
97.2--
2026.04
96.8--
2026.04
95.5--
2026.04
93.7--
2026.04
92.7--
2026.04
92.3--
2026.04
91.2--
2026.04
89.6--
2026.04
86.4--
2026.04
84.2--
2026.04
80.5--
2026.04
64.7--