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

Jailbreak Defense Evaluation on XST

3.03Strong-Reject Score

vanilla

0.91881.46692.0152.5631Feb 14, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
3.03
2026.02
2.905
2026.02
2.115
2026.02
1.945
2026.02
1.8
2026.02
1.79
2026.02
1.71
2026.02
1.56
2026.02
1.455
2026.02
1.39
2026.02
1.345
2026.02
1.32
2026.02
1.31
2026.02
1.285
2026.02
1.275
2026.02
1.21
2026.02
1.15
2026.02
1.08
2026.02
1.07
2026.02
1.065
2026.02
1