List your company

Hallucination Detection on M-HalDetect (val)

89.98A-ROCVIB-Probe
51.801661.713371.62581.5367Jan 9, 2026

Evaluation results

MethodLinks
VIB-ProbeBase VLM=LLaVA-v1.52026.0189.9882.35
VIB-ProbeBase VLM=LLaVA-v1.62026.0188.3681.23
DHCPBase VLM=LLaVA-v1.52026.0188.1380.2
VIB-ProbeBase VLM=Average2026.0186.7180.41
DHCPBase VLM=LLaVA-v1.62026.0186.5178.87
VIB-ProbeBase VLM=Qwen2.5-VL2026.0185.1780.79
DHCPBase VLM=Qwen2.5-VL2026.0184.8280.4
DHCPBase VLM=Average2026.0184.7678.52
VIB-ProbeBase VLM=MiniGPT-42026.0183.3377.26
MetaTokenBase VLM=LLaVA-v1.52026.0182.0271.14
MetaTokenBase VLM=LLaVA-v1.62026.0181.2373.56
RepProbingBase VLM=Qwen2.5-VL2026.0180.9271.13
DHCPBase VLM=MiniGPT-42026.0179.5874.62
MetaTokenBase VLM=Average2026.0178.9370.78
RepProbingBase VLM=Average2026.0178.4270.54
RepProbingBase VLM=MiniGPT-42026.0178.2170.04
RepProbingBase VLM=LLaVA-v1.62026.0177.3871.2
MetaTokenBase VLM=MiniGPT-42026.0177.2869.13
RepProbingBase VLM=LLaVA-v1.52026.0177.1869.8
MetaTokenBase VLM=Qwen2.5-VL2026.0175.1969.3
AvgProbBase VLM=Qwen2.5-VL2026.0166.4742.71
AvgEntBase VLM=Qwen2.5-VL2026.0163.5241.39
AvgProbBase VLM=LLaVA-v1.62026.0160.2140.36
AvgProbBase VLM=Average2026.0159.9240.39
AvgProbBase VLM=LLaVA-v1.52026.0159.0139.54
AvgEntBase VLM=Average2026.0156.938.5
AvgEntBase VLM=LLaVA-v1.62026.0155.937.52
AvgEntBase VLM=MiniGPT-42026.0154.938.22
AvgProbBase VLM=MiniGPT-42026.015438.93
AvgEntBase VLM=LLaVA-v1.52026.0153.2736.87