List your company

Object Hallucination Detection on FOIL-COCO (test)

99.6AccuracyVC-Inspector-3B
48.22461.56274.988.238May 25, 2022Dec 13, 2022Jul 3, 2023Jan 22, 2024Aug 11, 2024Mar 1, 2025Sep 20, 2025

Evaluation results

MethodLinks
VC-Inspector-3BEvaluation Mode=Reference-free2025.0999.6——
FLEUREvaluation Mode=Reference-based2025.0997.3——
FLEUREvaluation Mode=Reference-free2025.0996.8——
PAC-SEvaluation Mode=Reference-based2025.0993.5——
RefCLIP-SNumber of References=42022.0592.6——
BERT-SNumber of References=42022.0592.1——
RefCLIP-SNumber of References=12022.0591——
CIDErNumber of References=42022.0590.6——
MIDNumber of References=12022.0590.5——
MIDNumber of References=42022.0590.5——
PAC-SEvaluation Mode=Reference-free2025.0990.2——
BERT-SNumber of References=12022.0588.6——
CLIP-SNumber of References=12022.0587.2——
CLIP-SNumber of References=42022.0587.2——
SPICENumber of References=42022.0586.1——
METEORNumber of References=42022.0585.4——
BLEU-4Number of References=42022.0582.6——
CIDErNumber of References=12022.0582.5——
ROUGE-LNumber of References=42022.0579.3——
METEORNumber of References=12022.0578.8——
SPICENumber of References=12022.0575.5——
ROUGE-LNumber of References=12022.0571.7——
BLEU-4Number of References=12022.0566.5——
lengthNumber of References=12022.0550.2——
lengthNumber of References=42022.0550.2——
BERT-S2022.05—88.692.1
BLEU-42022.05—66.582.6
CIDEr2022.05—82.590.6
CLIP-S2022.05—87.287.2
length2022.05—50.250.2
METEOR2022.05—78.885.4
MID2022.05—90.590.5
RefCLIP-S2022.05—9192.6
ROUGE-L2022.05—71.779.3
SPICE2022.05—75.586.1