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

Jailbreak Attack Resistance Evaluation on Jailbreak

62.3MMLU

Original

38.993645.044351.09557.1457Jan 6, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.01
62.351.473.27
2026.01
62.2746.173.08
2026.01
62.2151.744.81
2026.01
62.1851.463.08
2026.01
61.9446.062.12
2026.01
60.0150.452.69
2026.01
59.7450.294.04
2026.01
59.4345.012.5
2026.01
59.3845.072.12
2026.01
58.6535.5782.12
2026.01
58.5935.474.04
2026.01
58.334.3494.62
2026.01
57.7444.665.38
2026.01
57.6844.915
2026.01
57.1835.6688.85
2026.01
57.1234.0596.15
2026.01
57.135.677.31
2026.01
55.2634.785.38
2026.01
55.2334.7178.27
2026.01
55.0835.0793.85
2026.01
48.443.416.92
2026.01
48.3441.63.08
2026.01
48.2343.46.54
2026.01
48.1642.582.88
2026.01
47.7143.27.5
2026.01
46.433.6892.12
2026.01
46.3741.70
2026.01
46.1442.930
2026.01
45.4135.3973.85
2026.01
45.2435.3573.85
2026.01
45.1343.673.08
2026.01
45.0642.677.5
2026.01
45.0643.564.42
2026.01
43.934.5785.96
2026.01
43.7340.2710
2026.01
42.8933.5792.69
2026.01
42.5934.8770.77
2026.01
42.5834.8770.96
2026.01
39.9535.0577.5
2026.01
39.8935.0577.69