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

Image Captioning Hallucination Detection on FOIL (test)

95.4Accuracy

Polos

65.34473.14780.9588.753Feb 28, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.02
95.4
2024.02
94.9
2024.02
93.7
2024.02
93.3
2024.02
92.6
2024.02
92.1
2024.02
91.1
2024.02
91
2024.02
90.6
2024.02
90.5
2024.02
90.5
2024.02
89.9
2024.02
89.9
2024.02
88.6
2024.02
88.4
2024.02
88.4
2024.02
87.2
2024.02
87.2
2024.02
86.1
2024.02
85.3
2024.02
82.6
2024.02
82.6
2024.02
82.5
2024.02
79.3
2024.02
78.8
2024.02
75.5
2024.02
71.7
2024.02
66.5