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

Jailbreak Attack on Harmful Query Evaluation Set Claude Haiku 4.5 N=750

1.01Toxicity

HADES

0.94881.36191.7752.1881May 7, 2026
Updated 26d ago

Evaluation Results

MethodLinks
2026.05
1.010.060.671.47
2026.05
1.010.110.531.07
2026.05
1.030.080.531.33
2026.05
1.150.513.734.67
2026.05
1.280.86.85.73
2026.05
1.541.3414.6714.53
2026.05
2.032.4827.225.33
2026.05
2.393.2135.634.67
2026.05
2.423.2337.8738
2026.05
2.543.4639.7339.33