List your company

Discriminative Hallucination Evaluation on AMBER-d

89.2AccuracyQwen2VL-7B + ACT
68.08873.56979.0584.531Apr 1, 2026

Evaluation results

MethodLinks
Qwen2VL-7B + ACTArchitecture=Qwen2VL-7B, ACT=Enabled2026.0489.2
Qwen2VL-7BArchitecture=Qwen2VL-7B, ACT=Disabled2026.0486.6
InternVL2.5-8B + ACTArchitecture=InternVL2.5-8B, ACT=Enabled2026.0484.8
InternVL2.5-8BArchitecture=InternVL2.5-8B, ACT=Disabled2026.0483.4
InternVL2-8B + ACTArchitecture=InternVL2-8B, ACT=Enabled2026.0481.8
Qwen2.5VL-7B + ACTArchitecture=Qwen2.5VL-7B, ACT=Enabled2026.0478.9
Qwen2.5VL-7BArchitecture=Qwen2.5VL-7B, ACT=Disabled2026.0477.8
LLaVA-1.5-7B + ACTArchitecture=LLaVA-1.5-7B, ACT=Enabled2026.0477
LLaVA-1.5-13B + ACTArchitecture=LLaVA-1.5-13B, ACT=Enabled2026.0474
LLaVA-1.5-7BArchitecture=LLaVA-1.5-7B, ACT=Disabled2026.0471.6
InternVL2-8BArchitecture=InternVL2-8B, ACT=Disabled2026.0471.2
LLaVA-1.5-13BArchitecture=LLaVA-1.5-13B, ACT=Disabled2026.0468.9