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

Safety Evaluation on HH-RLHF (test)

1.02Harm Score

SAILS

0.94561.44781.952.4522Dec 29, 2025
Updated 4d ago

Evaluation Results

MethodLinks
2025.12
1.020.9960.4
2025.12
1.030.9920.8
2025.12
1.080.9820
2025.12
1.130.9782.6
2025.12
1.170.9682.6
2025.12
1.180.9462.6
2025.12
1.290.9424.6
2025.12
1.340.9286
2025.12
1.40.925.2
2025.12
1.410.9244.8
2025.12
1.440.9227
2025.12
1.470.9067.6
2025.12
1.510.888.8
2025.12
1.520.8928.8
2025.12
1.540.897.6
2025.12
1.540.8967.6
2025.12
1.560.8768.4
2025.12
1.670.8610.2
2025.12
2.540.6235.2
2025.12
2.730.57637.4
2025.12
2.880.52842.4