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

Reward Modeling on HelpSteer 3

83.15Accuracy

o3-2025-04-16

59.063665.316871.5777.8232Jan 12, 2026Jan 15, 2026Jan 19, 2026Jan 22, 2026Jan 26, 2026Jan 29, 2026Feb 2, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.01
83.15---
2026.01
83.15---
2026.01
82.45---
2026.01
82.07---
2026.01
81.02---
2026.01
80.75---
2026.01
80.42---
2026.01
80.38---
2026.01
80.12---
2026.01
79.5---
2026.01
79.42---
2026.01
78.96---
2026.01
78.27---
2026.01
78.18---
2026.01
76.4---
2026.02
75.6---
2026.02
75.4---
2026.01
75.22---
2026.02
74.8---
2026.02
74.6---
2026.01
73.81---
2026.01
73.76---
2026.02
72.9---
2026.02
71.4---
2026.02
71.1---
2026.02
70.6---
2026.02
70.3---
2026.02
69.8---
2026.01
67.94---
2026.02
67.5---
2026.02
67---
2026.02
65.2---
2026.01
64.99---
2026.01
64---
2026.02
62.6---
2026.02
62.4---
2026.02
61.8---
2026.02
60.2---
2026.01
59.99---
2026.02
-75.924.957
2026.02
-77.529.554
2026.02
-78.412.468.6
2026.02
-78.15.673.7
2026.02
-77.520.561.5
2026.02
-7611.767.1
2026.02
-78.415.766
2026.02
-79.912.569.9
2026.02
-76.537.947.4
2026.02
-72.844.840.2
2026.02
-72.534.947.2
2026.02
-74.75930.6
2026.02
-73.250.736.1
2026.02
-75.132.950.4
2026.02
-72.945.839.5
2026.02
-72.944.640.4
2026.02
-73.134.647.8
2026.02
-74.14044.5
2026.02
-75.546.940.1
2026.02
-76.329.254