List your company

Hallucination Evaluation on POPE MSCOCO, A-OKVQA, GQA average (Adversarial)

84.11AccuracyVanilla
76.944478.804780.66582.5253Feb 14, 2026

Evaluation results

MethodLinks
VanillaMLLM=Qwen3-VL2026.0284.1180.2591.3585.24
AdaVBoostMLLM=Qwen3-VL2026.0284.0980.3891.0485.17
PAIMLLM=Qwen3-VL2026.0283.5578.9692.685
VGAMLLM=Qwen3-VL2026.0283.0578.8591.2984.4
VAFMLLM=Qwen3-VL2026.0282.3977.1793.1184.17
VanillaMLLM=LLaVA-NeXT2026.0280.1874.4793.8282.72
AdaVBoostMLLM=LLaVA-NeXT2026.0280.1574.4693.882.7
VGAMLLM=InternVL3.52026.0278.7172.0494.581.66
VanillaMLLM=InternVL3.52026.0278.6872.7192.5981.3
VAFMLLM=InternVL3.52026.0278.4871.5595.1781.61
AdaVBoostMLLM=InternVL3.52026.0278.4671.6194.9181.54
VAFMLLM=LLaVA-NeXT2026.0277.8171.0195.8981.36
VGAMLLM=LLaVA-NeXT2026.0277.4671.8793.2280.77
PAIMLLM=LLaVA-NeXT2026.0277.2370.9595.1380.93
PAIMLLM=InternVL3.52026.0277.2270.7193.5380.44