HallusionBench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
HallusionBench
76.6Accuracy
156
HallusionBench
82.02Accuracy
153
HallusionBench
81.8Accuracy (aACC)
61
HallusionBench
73.2Accuracy
43
HallusionBench
0.729Accuracy
42
HallusionBench
71.6Answer Accuracy (aAcc)
39
HallusionBench
77.2Accuracy
37
HallusionBench
58.4Score
37
HallusionBench
69.2HallusionBench Score
26
HallusionBench
71.08Accuracy
24
HallusionBench
70.7Hallucination Score
22
HallusionBench visual-dependent setting (test)
33.21qAcc (Question Accuracy)
21
HallusionBench
70Accuracy
21
HallusionBench
67.05Hallusion Score
20
HallusionBench
73.48Accuracy
20
HallusionBench
75.2Simple Accuracy
19
HallusionBench
74Pass@1
16
HallusionBench
96LI Score
15
HallusionBench
59.5Score
15
HallusionBench
69.9Pass@1
14
HallusionBench 2024
52.2Score
13
HallusionBench (HallB)
41.7HallB Score
13
HallusionBench
67.9Score
12
HallusionBench
22.51Accuracy
12
HallusionBench GPT4-assisted (All)
49.94Accuracy (All)
11