List your company

Hallucination detection on VOXPOPULI (test)

2.03Hallucination Rate (%)RAUQ Entropy
0.100413.125226.1539.1748Apr 21, 2026

Evaluation results

MethodLinks
RAUQ EntropyModel=Voxtral-3B, Category=Baseline2026.042.03963547283243
Mean EntropyModel=Voxtral-3B, Category=Baseline2026.042.1974255344443
Logistic RegressionModel=Voxtral-3B, Feature Configuration=Top 75, Number of features=3002026.042.99986266586855
Logistic RegressionModel=Voxtral-3B, Feature Configuration=AUDIORATIO, Number of features=9602026.043.04986468616754
PerplexityModel=Voxtral-3B, Category=Baseline2026.043.12964042394140
Logistic RegressionModel=Voxtral-3B, Feature Configuration=Combined, Number of features=38402026.043.23986465626956
Attention ScoreModel=Voxtral-3B, Category=Baseline2026.047.1291171226910
RandomModel=Voxtral-3B, Category=Baseline2026.0450.27506349——