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

Error Detection on ImageNet

88.57AuROC

Ours

Updated 3d ago

Evaluation Results

Method	Links
Ours 2025.11		88.57	95.85	48.87
Ours 2025.11		88.01	93.23	54.1
Ours 2025.11		87.52	91.61	55.26
Ours-D 2025.11		86.8	92.57	56.57
Ours-D 2025.11		86.17	94.8	56.88
Ours-D 2025.11		85.8	90.53	59.02
MaxSoftmax 2025.11		84.21	94.1	65.3
MaxSoftmax 2025.11		84.21	94.1	65.3
TempScaling 2025.11		84.12	94.06	66.06
TracIn 2025.06		83.58	-	-
TRAK 2025.06		83.48	-	-
Entropy 2025.11		83.46	93.87	68.64
Entropy 2025.11		83.46	93.87	68.64
TrustVLM* 2025.11		83.39	90.43	75.67
TrustVLM* 2025.11		81.66	90.85	81.65
TempScaling 2025.11		81.28	89.87	72.16
MaxSoftmax 2025.11		80.71	87.43	71.38
TempScaling 2025.11		80.63	87.37	71.64
TRAK 2025.06		80.08	-	-
Zero 2025.11		79.43	93.55	68.65
BayesVLM 2025.11		79.41	86.05	72.83
Zero 2025.11		78.87	90.84	73.95
Entropy 2025.11		78.27	86.01	76.75
PCME++ 2025.11		77.31	90.68	71.82
TracIn 2025.06		71.27	-	-
ProLIP 2025.11		70.57	85.77	81.23
MaxCosine 2025.11		70.2	87.04	82.31
MaxCosine 2025.11		70.2	87.04	82.31
BayesVLM 2025.11		70.06	87.84	83.01
TRAK 2025.06		67.72	-	-
MaxCosine 2025.11		65.79	74.58	86.4
TracIn 2025.06		61.46	-	-
TRAK 2025.06		59.66	-	-
TracIn 2025.06		54.4	-	-
ProbVLM 2025.11		50.18	61.79	95.31