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

Human Preference Evaluation on HPD v2 (test)

85.4Preference Accuracy

HPSv3 - 7B

64.18469.69275.280.708May 23, 2024Sep 22, 2024Jan 22, 2025May 24, 2025Sep 23, 2025Jan 23, 2026May 25, 2026
Updated 7d ago

Evaluation Results

MethodLinks
2026.05
85.4
2026.02
85.36
2026.02
84.31
2024.05
83.5
2026.05
83.5
2024.05
83.3
2026.05
83.3
2026.02
83.27
2026.02
83.1
2026.02
82.7
2026.02
82.58
2026.05
82.2
2026.02
82.13
2026.05
80.8
2026.05
80.1
2024.05
79.8
2026.05
79.8
2026.02
79.44
2026.05
77.6
76.8
74
2026.02
73.95
2024.05
73.1
2024.05
72.6
2026.02
72.39
2024.05
71.2
2026.02
71.19
2024.05
70.6
2026.05
70.1
2026.05
68.5
65.1
2026.05
65