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

Object Hallucination Evaluation on GQA (Random)

89.5Accuracy

MESA

74.139278.127182.11586.1029Dec 22, 2025Jan 9, 2026Jan 27, 2026Feb 14, 2026Mar 4, 2026Mar 22, 2026Apr 9, 2026
Updated 8d ago

Evaluation Results

MethodLinks
2026.04
89.589.45
2025.12
89.389.4
2026.04
89.2688.78
2026.04
8988.63
2025.12
88.788.9
2026.04
88.3788.02
2026.04
87.9786.67
2026.04
87.788
2025.12
87.287.6
2025.12
86.887.2
2026.04
86.6586.99
2026.04
86.284.62
2026.04
86.1186.11
2025.12
86.187.1
2026.04
85.985.29
2025.12
85.785.9
2026.04
85.5985.33
2026.04
85.3386.12
2026.04
84.9686.25
2026.04
84.3381.94
2025.12
84.184.9
2025.12
83.582.7
2025.12
83.584.3
2025.12
83.181.2
2025.12
82.883.2
2025.12
79.580.4
2026.04
74.7666.6
2026.04
74.7365.55