Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Hallucination Evaluation on POPE Popular MSCOCO, A-OKVQA, GQA average

88.48Accuracy

Vanilla

80.99282.93684.8886.824Feb 14, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2026.02
88.4886.6191.3588.8
2026.02
88.4585.7892.688.92
2026.02
88.3586.6191.0488.64
2026.02
87.4684.9791.2987.9
2026.02
87.1483.4593.1187.88
2026.02
86.0181.8393.8287.15
2026.02
86.0181.8593.887.14
2026.02
84.6478.8695.8986.33
2026.02
84.0978.8395.1385.88
2026.02
83.4578.9493.2285.12
2026.02
83.3177.3194.585
2026.02
82.8577.692.5984.35
2026.02
82.8476.4595.1784.74
2026.02
82.6476.3294.9184.55
2026.02
81.2875.1893.5383.31