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

Vision Model Representation Alignment on ImageNet 20% held-out split of 20,000 samples (val)

85Reconstruction Accuracy

MOT + R

12.09631.02349.9568.877Oct 2, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.10
85
2025.10
79
2025.10
78.8
2025.10
77.8
2025.10
73.2
2025.10
65.7
2025.10
63.3
2025.10
62.4
2025.10
61.4
2025.10
60
2025.10
59.8
2025.10
59.6
2025.10
58.8
2025.10
58
2025.10
57.7
2025.10
57.1
2025.10
52.6
2025.10
50.9
2025.10
50.8
2025.10
46.6
2025.10
43.3
2025.10
41.8
2025.10
41.7
2025.10
41.1
2025.10
39.4
2025.10
38.6
2025.10
38.1
2025.10
35.9
2025.10
35.4
2025.10
35.3
2025.10
35.2
2025.10
34
2025.10
31.7
2025.10
30.1
2025.10
29.3
2025.10
28.9
2025.10
28.3
2025.10
20.2
2025.10
18
2025.10
14.9