List your company

Science Question Answering on ScienceQA (Image Acc.)

88.6Image AccuracyQwen2.5-VL-7B-Instruct
65.51271.50677.583.494Jul 16, 2025

Evaluation results

MethodLinks
Qwen2.5-VL-7B-InstructMitigation=None (baseline)2025.0788.6
SENTINELBase Model=Qwen2.5-VL-7B-Instruct2025.0788.5
SENTINELBase Model=Qwen2-VL-7B-Instruct2025.0786.9
Qwen2-VL-7B-InstructMitigation=None (baseline)2025.0785.7
SENTINELBase Model=LLaVA-v1.6-vicuna-13B2025.0778
LLaVA-v1.6-vicuna-13BMitigation=None (baseline)2025.0777.7
SENTINELBase Model=Qwen2-VL-2B-Instruct2025.0777.4
Qwen2-VL-2B-InstructMitigation=None (baseline)2025.0776.9
LLaVA-v1.6-vicuna-7BMitigation=None (baseline)2025.0774.3
SENTINELBase Model=LLaVA-v1.6-vicuna-7B2025.0774.2
SENTINELModel=LLaVA-v1.5-13B2025.0772.8
VCDModel=LLaVA-v1.5-13B2025.0772
vanilla-DPOModel=LLaVA-v1.5-13B2025.0771.8
baselineModel=LLaVA-v1.5-13B2025.0771.6
HSA-DPOModel=LLaVA-v1.5-13B2025.0771.3
HA-DPOModel=LLaVA-v1.5-7B2025.0769.7
SENTINELModel=LLaVA-v1.5-7B2025.0769.2
POVIDModel=LLaVA-v1.5-7B2025.0768.8
VCDModel=LLaVA-v1.5-7B2025.0768.7
OPERAModel=LLaVA-v1.5-7B2025.0768.2
RLAIF-VModel=LLaVA-v1.5-7B2025.0768.2
CLIP-DPOModel=LLaVA-v1.5-7B2025.0767.6
DoLaModel=LLaVA-v1.5-7B2025.0767.5
TPOModel=LLaVA-v1.5-7B2025.0767.1
baselineModel=LLaVA-v1.5-7B2025.0766.8
EFUFModel=LLaVA-v1.5-7B2025.0766.4