Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Hallucination monitoring on New York City domain
Loading...
89
AUROC
HalMit
52.6
62.05
71.5
80.95
Jul 21, 2025
AUROC
AUC-PR
F1 Score
Accuracy
Updated 19d ago
Evaluation Results
Method
Method
Links
AUROC
AUC-PR
F1 Score
Accuracy
HalMit
Backbone=Llama3.1
2025.07
89
82
88
84
HalMit
Backbone=Llama2
2025.07
88
77
75
89
SCG
Backbone=Llama3.1
2025.07
86
84
87
86
SCG
Backbone=Llama2
2025.07
82
72
85
82
PP
Backbone=Llama3.1
2025.07
60
58
52
53
ICL
Backbone=Llama2
2025.07
59
22
32
75
ICL
Backbone=Llama3.1
2025.07
58
64
63
45
PP
Backbone=Llama2
2025.07
54
12
21
74
Feedback
Search any
task
Search any
task