Object Hallucination Probing
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
80.2F1 Score
72
Jul 7, 2026
88.2Accuracy
63
Jul 7, 2026
75.2Accuracy
63
Jul 7, 2026
83.27Accuracy
55
Jun 30, 2026
86.07Accuracy
49
Apr 10, 2026
89.93Accuracy (GQA POPE)
42
Apr 10, 2026
79.1Accuracy
27
Feb 26, 2026
82.73Accuracy
19
Apr 10, 2026
90.67Accuracy
17
Feb 26, 2026
90.83Accuracy
12
Feb 26, 2026
85Accuracy
11
Feb 26, 2026
79.97POPE Score (Zh)
6
Feb 26, 2026
86.03Accuracy (Zh)
6
Feb 26, 2026
83.27Score (Zh)
6
Feb 26, 2026
87.87Accuracy
5
Feb 26, 2026