Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Error Detection on ImageNet V2 (test)
Loading...
86.19
AuROC
Ours
67.106
72.0605
77.015
81.9695
Nov 27, 2025
AuROC
AuPR
FPR95
Updated 1mo ago
Evaluation Results
Method
Method
Links
AuROC
AuPR
FPR95
Ours
Backbone=CLIP ViT-16
2025.11
86.19
92.42
59
Ours-D
Backbone=CLIP ViT-16
2025.11
85.32
91.91
62.27
TrustVLM
Backbone=CLIP ViT-16
2025.11
83.03
89.68
78.76
TempScaling
Backbone=CLIP ViT-16
2025.11
81.53
89.95
70.43
MaxSoftmax
Backbone=CLIP ViT-16
2025.11
81.3
89.83
71
Entropy
Backbone=CLIP ViT-16
2025.11
78.78
88.62
78.08
MaxCosine
Backbone=CLIP ViT-16
2025.11
67.84
79.59
83.55
Feedback
Search any
task
Search any
task