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

Multimodal Safety Evaluation on BeaverTails-V

2.9Safety Score

Qwen2.5-VL-7B + TIS

0.49761.12131.7452.3687Mar 3, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
2.92.772.88
2026.03
2.852.752.89
2026.03
2.842.782.87
2026.03
2.842.842.91
2026.03
2.832.332.54
2026.03
2.822.62.27
2026.03
2.822.882.93
2026.03
2.752.372.69
2026.03
2.742.592.23
2026.03
2.722.272.28
2026.03
2.722.812.91
2026.03
2.641.731.9
2026.03
2.622.662.86
2026.03
2.611.691.86
2026.03
2.582.662.77
2026.03
2.572.612.77
2026.03
2.52.332.17
2026.03
1.672.72.14
2026.03
1.582.672.15
2026.03
0.822.011.39
2026.03
0.732.411.66
2026.03
0.592.561.68