List your company

Binary Safety Hallucination Detection on EndoVis VQA In-template 18 (val)

98AccuracyQA-SNNE
-2.8823.3149.575.69Nov 3, 2025

Evaluation results

MethodLinks
QA-SNNEBackbone Model=MedGemma3.0, Evaluation Protocol=Zero-shot, Variant=Ent2025.1198
QA-SNNEBackbone Model=MedGemma3.0, Evaluation Protocol=Zero-shot, Variant=CrossE2025.1198
SNNEBackbone Model=PitVQA, Evaluation Protocol=Peft2025.1198
VL-UBackbone Model=PitVQA, Evaluation Protocol=Peft2025.1198
QA-SNNEBackbone Model=MedGemma3.0, Evaluation Protocol=Zero-shot, Variant=Emb2025.1187
QA-SNNEBackbone Model=PitVQA, Evaluation Protocol=Peft, Variant=Emb2025.1185
QA-SNNEBackbone Model=SurgicalGPT, Evaluation Protocol=Peft, Variant=Emb2025.1184
SNNEBackbone Model=SurgicalGPT, Evaluation Protocol=Peft2025.1182
VL-UBackbone Model=MedGemma3.0, Evaluation Protocol=Zero-shot2025.1181
VL-UBackbone Model=Qwen2.5, Evaluation Protocol=Zero-shot2025.1179
VL-UBackbone Model=Llama3.2, Evaluation Protocol=Zero-shot2025.1174
QA-SNNEBackbone Model=Llama3.2, Evaluation Protocol=Zero-shot, Variant=CrossE2025.1170
QA-SNNEBackbone Model=Llama3.2, Evaluation Protocol=Zero-shot, Variant=Ent2025.1169
QA-SNNEBackbone Model=Llama3.2, Evaluation Protocol=Zero-shot, Variant=Emb2025.1167
QA-SNNEBackbone Model=Qwen2.5, Evaluation Protocol=Zero-shot, Variant=Ent2025.1167
SNNEBackbone Model=Llama3.2, Evaluation Protocol=Zero-shot2025.1156
QA-SNNEBackbone Model=Qwen2.5, Evaluation Protocol=Zero-shot, Variant=CrossE2025.1154
QA-SNNEBackbone Model=PitVQA, Evaluation Protocol=Peft, Variant=CrossE2025.1151
VL-UBackbone Model=SurgicalGPT, Evaluation Protocol=Peft2025.1139
QA-SNNEBackbone Model=SurgicalGPT, Evaluation Protocol=Peft, Variant=Ent2025.1135
QA-SNNEBackbone Model=SurgicalGPT, Evaluation Protocol=Peft, Variant=CrossE2025.1135
SNNEBackbone Model=MedGemma3.0, Evaluation Protocol=Zero-shot2025.1122
SNNEBackbone Model=Qwen2.5, Evaluation Protocol=Zero-shot2025.1120
QA-SNNEBackbone Model=Qwen2.5, Evaluation Protocol=Zero-shot, Variant=Emb2025.1120
QA-SNNEBackbone Model=PitVQA, Evaluation Protocol=Peft, Variant=Ent2025.111