Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Safety Alignment on HH-RLHF

1.09MD Rate

ShaPO-T

-1.754817.447636.6555.8524Feb 7, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.02
1.090.12
2026.02
1.310.12
2026.02
1.750.23
2026.02
1.860.12
2026.02
2.090.33
2026.02
2.390.22
2026.02
2.570.34
2026.02
2.940.19
2026.02
4.580.37
2026.02
5.922.02
2026.02
6.141.76
2026.02
6.882.25
2026.02
12.15.26
2026.02
12.125.45
2026.02
12.392.72
2026.02
14.141.67
2026.02
17.74.9
2026.02
18.025.03
2026.02
21.78.3
2026.02
23.0510.3
2026.02
24.5510.67
2026.02
26.138.59
2026.02
29.3519.94
2026.02
30.3213.34
2026.02
32.5210.76
2026.02
55.1921.95
2026.02
58.8221.06
2026.02
59.4922.78
2026.02
60.3318.08
2026.02
65.119.92
2026.02
66.9338.67
2026.02
67.6118.89
2026.02
69.2747.54
2026.02
70.5448.01
2026.02
72.0224.95
2026.02
72.2125.75