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

Hallucination Detection on VisionHall

45.15F1 Score

ZINA

Updated 1mo ago

Evaluation Results

Method	Links
ZINA 2025.06		45.15
GPT-4o 2025.06		29.37
GPT-4o (w/o images) 2025.06		27.02
LLaVA-OV-Qwen2-72B 2025.06		25.7
Qwen2.5-VL-72B-Instruct 2025.06		21.31
LLaVA-NeXT-Qwen-32B 2025.06		19.09
Llama-3.2-90B-Vision-Instruct 2025.06		16.92
LLaVA-1.5-13B 2025.06		4.73
LLaVA-OV-Qwen2-7B 2025.06		3.39
Qwen2-VL-7B 2025.06		3.36
LLaVA-1.5-7B 2025.06		0.82