List your company
ResearchFormerly WizWand

Hallucination Detection on RAGTruth MS MARCO (subsample)

0.77AUROCINTRYGUEmin-max
0.44760.53130.6150.6987Mar 23, 2026
Updated 7mo ago

Evaluation results

MethodLinks
INTRYGUEmin-maxModel=Mistral-7B2026.030.77
MaxProbModel=Mistral-7B2026.030.75
INTRYGUEmin-maxModel=Llama-2-13B2026.030.73
INTRYGUEmin-maxModel=Llama-2-7B2026.030.72
ReDeEPModel=Mistral-7B2026.030.72
INTRYGUEmeanModel=Mistral-7B2026.030.71
MaxProbModel=Llama-2-7B2026.030.69
LSCModel=Mistral-7B2026.030.68
INTRYGUEmeanModel=Llama-2-7B2026.030.67
INTRYGUEmeanModel=Llama-2-13B2026.030.67
FocusModel=Mistral-7B2026.030.67
MaxEntropyModel=Mistral-7B2026.030.67
MaxProbModel=Llama-2-13B2026.030.67
SARModel=Mistral-7B2026.030.67
ReDeEPModel=Llama-2-13B2026.030.66
LN-EntropyModel=Mistral-7B2026.030.65
MaxEntropyModel=Llama-2-7B2026.030.65
PerplexityModel=Mistral-7B2026.030.64
FocusModel=Llama-2-7B2026.030.63
LN-EntropyModel=Llama-2-7B2026.030.63
ReDeEPModel=Llama-2-7B2026.030.63
LN-EntropyModel=Llama-2-13B2026.030.62
MaxEntropyModel=Llama-2-13B2026.030.62
Semantic EntropyModel=Llama-2-13B2026.030.62
FocusModel=Llama-2-13B2026.030.61
EigenScoreModel=Mistral-7B2026.030.6
Semantic EntropyModel=Mistral-7B2026.030.6
EigenScoreModel=Llama-2-7B2026.030.58
EigenScoreModel=Llama-2-13B2026.030.58
LSCModel=Llama-2-7B2026.030.57
LSCModel=Llama-2-13B2026.030.57
RAUQModel=Mistral-7B2026.030.56
SARModel=Llama-2-7B2026.030.55
Semantic DensityModel=Mistral-7B2026.030.55
Semantic DensityModel=Llama-2-7B2026.030.55
RAUQModel=Llama-2-13B2026.030.54
RAUQModel=Llama-2-7B2026.030.53
Semantic DensityModel=Llama-2-13B2026.030.53
Semantic EntropyModel=Llama-2-7B2026.030.53
Attention ScoreModel=Llama-2-13B2026.030.51
Attention ScoreModel=Mistral-7B2026.030.49
PerplexityModel=Llama-2-7B2026.030.48
SARModel=Llama-2-13B2026.030.48
Attention ScoreModel=Llama-2-7B2026.030.47
PerplexityModel=Llama-2-13B2026.030.46