List your company
ResearchFormerly WizWand

Hallucination Detection on RAGTruth & HalluRAG Overall Average 1.0 (test)

0.7284Average F1 ScoreWavelet-high
0.4653840.5336670.601950.670233Feb 20, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Wavelet-highModel=Mistral-7B2026.020.72840.8595
Fourier-highModel=Mistral-7B2026.020.72660.8532
Lookback-lensModel=Mistral-7B2026.020.72140.8478
LaplacianModel=Mistral-7B2026.020.71970.8494
Attn-entropyModel=Mistral-7B2026.020.70530.8297
Fourier-highModel=LLaMA-13B2026.020.67230.8198
Wavelet-highModel=LLaMA-13B2026.020.66870.8131
Fourier-highModel=LLaMA-7B2026.020.66150.8052
Wavelet-highModel=LLaMA-7B2026.020.66010.7953
Lookback-lensModel=LLaMA-13B2026.020.65890.8031
LaplacianModel=LLaMA-13B2026.020.65590.7918
LaplacianModel=LLaMA-7B2026.020.65340.7859
SelfcheckGPTModel=Mistral-7B2026.020.64980.7269
Attn-entropyModel=LLaMA-13B2026.020.64380.7875
RefCheckerModel=Mistral-7B2026.020.62230.6259
Lookback-lensModel=LLaMA-7B2026.020.61750.7871
SelfcheckGPTModel=LLaMA-7B2026.020.61450.6901
SelfcheckGPTModel=LLaMA-13B2026.020.60660.6885
Attn-entropyModel=LLaMA-7B2026.020.59730.7627
RefCheckerModel=LLaMA-13B2026.020.59340.6146
RefCheckerModel=LLaMA-7B2026.020.58760.6004
RedeepModel=Mistral-7B2026.020.56620.6333
RedeepModel=LLaMA-7B2026.020.56430.5643
EigenScoreModel=LLaMA-13B2026.020.55720.584
Attn-varianceModel=Mistral-7B2026.020.55150.6827
RedeepModel=LLaMA-13B2026.020.54480.6113
EigenScoreModel=Mistral-7B2026.020.53790.6004
Attn-varianceModel=LLaMA-13B2026.020.52760.69
EigenScoreModel=LLaMA-7B2026.020.51660.5416
Attn-varianceModel=LLaMA-7B2026.020.47550.6025