List your company

Hallucination Detection on HalLoc-VQA

77.9ROCFaithSCAN
49.8257.1164.471.69Jan 1, 2026

Evaluation results

MethodLinks
FaithSCANBackbone=LLaVA-8B2026.0177.982.36887.31,898.950.741
SEUBackbone=LLaVA-8B2026.0176.979.966.786.52,443.891.769
FaithSCANBackbone=IB-T5-XL2026.0175.477.875.579.21,560.2370.605
P(True)Backbone=Qwen3-VL-8B2026.017582.554.883.81,143.7224.299
EmbVarBackbone=LLaVA-8B2026.0174.78250.985.52,504.5840.175
FaithSCANBackbone=Qwen3-VL-8B2026.0173.975.756.981.11,741.6521.191
LogisticBackbone=LLaVA-8B2026.0173.378.858831,874.8950.114
SelfCheckGPTBackbone=LLaVA-8B2026.0171.577.962.581.82,588.092340.683
PEBackbone=Qwen3-VL-8B2026.0168.617.853.5201,328.4340.002
SEBackbone=LLaVA-8B2026.0168.374.746.279.52,632.83311.212
P(True)Backbone=IB-T5-XL2026.0167.559.775.356.51,127.96114.938
LogisticBackbone=Qwen3-VL-8B2026.0167.268.649.672.41,689.4580.102
LogisticBackbone=IB-T5-XL2026.0166.966.973.467.41,570.3490.087
SEUBackbone=IB-T5-XL2026.0166.345.273.8441,685.9751.845
EmbVarBackbone=IB-T5-XL2026.0165.746.474.645.31,430.8840.06
SelfCheckGPTBackbone=Qwen3-VL-8B2026.0165.671.655.274.82,165.013367.522
SEUBackbone=Qwen3-VL-8B2026.0165.374.258.7802,199.4271.821
T-NLLBackbone=Qwen3-VL-8B2026.016523.352.4191,330.9960.001
EmbVarBackbone=Qwen3-VL-8B2026.0162.870.55778.72,194.5860.107
P(True)Backbone=LLaVA-8B2026.0161.5695274.81,178.08419.511
T-NLLBackbone=LLaVA-8B2026.0160.725.652.924.71,524.740.001
PEBackbone=LLaVA-8B2026.0160.527.653.525.81,491.4480.002
SEBackbone=IB-T5-XL2026.0158.356.365.356.21,381.9294.913
SEBackbone=Qwen3-VL-8B2026.0155.932.545.630.52,019.87713.812
PEBackbone=IB-T5-XL2026.0153.652.67352.71,326.3080.002
T-NLLBackbone=IB-T5-XL2026.0153.553.776.254.21,257.5120.001
SelfCheckGPTBackbone=IB-T5-XL2026.0150.954.572.2591,602.88778.613