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

Human Preference Evaluation on ImageReward (test)

0.675Preference Accuracy

MPS

0.512760.554880.5970.63912May 23, 2024Sep 22, 2024Jan 22, 2025May 24, 2025Sep 23, 2025Jan 23, 2026May 25, 2026
Updated 7d ago

Evaluation Results

MethodLinks
2024.05
0.675
2026.05
0.675
2026.02
0.6703
2026.05
0.668
2026.02
0.6637
2024.05
0.657
2026.05
0.657
2026.02
0.6562
2026.02
0.6515
2024.05
0.651
0.651
2026.05
0.641
2026.02
0.6382
2024.05
0.629
2026.05
0.629
2026.02
0.6273
2026.02
0.6175
2026.05
0.616
2024.05
0.612
2026.05
0.612
2026.02
0.6035
2026.02
0.6034
2026.02
0.5917
2026.02
0.5854
2026.05
0.579
2024.05
0.574
0.574
0.571
2024.05
0.543
2026.05
0.537
2026.05
0.524
2026.05
0.519