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

Object Hallucination Probing on GQA POPE Popular

86.07Accuracy

MESA

Updated 8d ago

Evaluation Results

Method
MESA 2026.04	86.07	-	-	-	-	-	-	-	-	-	92.73	78.27	84.89
VAF 2026.04	85.43	-	-	-	-	-	-	-	-	-	92.61	72.53	81.35
ICT 2026.04	85.31	-	-	-	-	-	-	-	-	-	91.59	77.86	84.15
SchroMind 2026.02	84.83	-	-	-	-	-	-	-	-	-	-	-	84.93
ICT w/o object 2024.11	84.7	-	-	-	-	-	-	-	-	-	-	-	84.45
ICT 2024.11	84.7	-	-	-	-	-	-	-	-	-	-	-	84.78
Scalpel 2026.02	84.57	-	-	-	-	-	-	-	-	-	-	-	85.31
ICT w/o image 2024.11	84.23	-	-	-	-	-	-	-	-	-	-	-	83.92
ICT w/o object 2024.11	83.93	-	-	-	-	-	-	-	-	-	-	-	82.95
VTI 2026.04	83.67	-	-	-	-	-	-	-	-	-	95.01	71.07	81.31
MESA 2026.04	82.93	-	-	-	-	-	-	-	-	-	79.37	89	83.91
ICT 2026.04	82.86	-	-	-	-	-	-	-	-	-	82.51	84.93	83.7
ICT 2024.11	82.63	-	-	-	-	-	-	-	-	-	-	-	82.22
VCD 2024.11	82.23	-	-	-	-	-	-	-	-	-	-	-	82.11
OPERA 2024.11	81.97	-	-	-	-	-	-	-	-	-	-	-	82.12
VCD 2026.04	81.83	-	-	-	-	-	-	-	-	-	80.45	84.09	82.23
VTI 2026.04	81.73	-	-	-	-	-	-	-	-	-	79.53	85.47	82.39
ICT w/o image 2024.11	81.43	-	-	-	-	-	-	-	-	-	-	-	80.86
VAF 2026.04	81.16	-	-	-	-	-	-	-	-	-	78.52	85.8	82
ICT 2026.02	80.86	-	-	-	-	-	-	-	-	-	-	-	82.64
ICT 2026.02	80.86	-	-	-	-	-	-	-	-	-	-	-	82.64
ICD 2026.04	80.8	-	-	-	-	-	-	-	-	-	91.77	67.66	77.89
VCD 2024.11	80.73	-	-	-	-	-	-	-	-	-	-	-	82.24
VCD 2026.02	80.73	-	-	-	-	-	-	-	-	-	-	-	82.24
VCD 2026.02	80.73	-	-	-	-	-	-	-	-	-	-	-	82.24
VCD 2026.04	80.73	-	-	-	-	-	-	-	-	-	76.26	89.24	82.24
OPERA 2024.11	79.27	-	-	-	-	-	-	-	-	-	-	-	81.83
OPERA 2026.02	79.27	-	-	-	-	-	-	-	-	-	-	-	82.11
OPERA 2026.02	79.27	-	-	-	-	-	-	-	-	-	-	-	82.11
Instruction Contrastive Decoding (ICD) 2024.03	78.8	-	-	-	-	-	-	-	-	-	75.15	87.53	80.87
Visual Contrastive Decoding (VCD) 2024.03	78.57	-	-	-	-	-	-	-	-	-	74.62	86.61	80.17
ICD 2026.04	78.56	-	-	-	-	-	-	-	-	-	77.95	79.66	78.79
Regular 2024.11	78.17	-	-	-	-	-	-	-	-	-	-	-	80.73
Vanilla 2026.02	78.17	-	-	-	-	-	-	-	-	-	-	-	78.37
Regular 2026.02	78.17	-	-	-	-	-	-	-	-	-	-	-	78.37
Nullu 2026.04	77.63	-	-	-	-	-	-	-	-	-	72.08	90.2	80.13
Regular 2024.11	75.99	-	-	-	-	-	-	-	-	-	-	-	74.84
AVISC 2026.02	74.8	-	-	-	-	-	-	-	-	-	-	-	79.17
Vanilla 2026.04	74.76	-	-	-	-	-	-	-	-	-	98.18	50.46	66.66
Nullu 2026.04	74.71	-	-	-	-	-	-	-	-	-	98.17	50.33	66.54
Standard Decoding 2024.03	73.87	-	-	-	-	-	-	-	-	-	60.63	84.69	76.42
Vanilla 2026.04	73.3	-	-	-	-	-	-	-	-	-	66.4	94.33	77.93
M3ID 2026.02	72.83	-	-	-	-	-	-	-	-	-	-	-	77.58
Instruction Contrastive Decoding (ICD) 2024.03	62.84	-	-	-	-	-	-	-	-	-	61.09	80.54	69.48
Visual Contrastive Decoding (VCD) 2024.03	62.14	-	-	-	-	-	-	-	-	-	61.14	72.26	66.24
Regular Decoding 2026.01	58.87	-	-	-	-	-	-	-	-	-	59.87	53.8	56.67
VCD 2026.01	58.83	-	-	-	-	-	-	-	-	-	58.65	56.73	57.68
SDCD 2026.01	57.47	-	-	-	-	-	-	-	-	-	57.22	59.2	58.19
Standard Decoding 2024.03	57.19	-	-	-	-	-	-	-	-	-	58.55	60.81	59.66
Baseline 2025.06	-	83.9	69.6	64.03	64.53	71.53	63.4	-	-	66.62	-	-	-
VCD 2025.06	-	-	73.6	67.5	65.23	71.87	66.2	-	-	68.88	-	-	-
CLAIM 2025.06	-	81.57	78.5	77.93	79.6	80.57	-	-	-	79.63	-	-	-
Baseline 2025.06	-	85.8	81.9	72.17	75.07	-	-	59.2	70.4	71.93	-	-	-
VCD 2025.06	-	-	82.2	73.73	74.17	-	-	54.7	76.87	72.33	-	-	-
CLAIM 2025.06	-	84.43	80.6	82.6	61.3	-	-	78.07	-	77.4	-	-	-