List your company

Discriminative Object Hallucination on AMBER Discriminative Task

87.4F1 ScoreGPT-4V
70.44874.84979.2583.651Apr 28, 2026May 2, 2026May 6, 2026May 11, 2026May 15, 2026May 19, 2026May 24, 2026

Evaluation results

MethodLinks
GPT-4VBase Model=Proprietary2026.0587.483.4
HALVABase Model=LLaVA-v1.5-13B2026.0586.5—
HSA-DPOBase Model=LLaVA-v1.5-13B2026.0586.180.8
LBPBase Model=LLaVA-v1.5-7B2026.0586.178.6
LBPBase Model=LLaVA-v1.5-13B2026.0585.477
AMP-MEGBase Model=LLaVA-v1.5-13B2026.0584.679.5
RLAIF-VBase Model=LLaVA-v1.5-7B2026.0584.576.8
LLaVA-v1.5-13BBase Model=LLaVA-v1.5-13B2026.0584.479.3
OctopusVLM=LLaVA-1.52026.0482.776.7
IECD2VLM=LLaVA-1.52026.0482.677.8
IECD2VLM=InstructBLIP2026.0482.176.6
CCA-LLaVABase Model=LLaVA-v1.5-7B2026.0581.977.7
V-DPOBase Model=LLaVA-v1.5-7B2026.0581.6—
OctopusVLM=InstructBLIP2026.0479.774
ICDVLM=InstructBLIP2026.0475.969.6
VCDVLM=InstructBLIP2026.0475.666.3
RLHF-VBase Model=LLaVA-v1.5-13B2026.057572.6
LLaVA-v1.5-7BBase Model=LLaVA-v1.5-7B2026.0574.772
RegularVLM=InstructBLIP2026.0474.668.2
ICDVLM=LLaVA-1.52026.0472.869.9
RegularVLM=LLaVA-1.52026.047269.2
VCDVLM=LLaVA-1.52026.0471.167.3