Hallucination detection on CALLHOME k=30% (test)
0Hallucination RateAttention Score
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Attention ScoreBackbone=Qwen-2-Audio2026.04 | 0 | 79 | 0 | 0 | 0 | 14 | -0.13 | |
| RAUQ EntropyBackbone=Qwen-2-Audio2026.04 | 24.4 | 80 | 56 | 52 | 61 | 61 | 0.54 | |
| PerplexityBackbone=Qwen-2-Audio2026.04 | 37.89 | 76 | 58 | 45 | 83 | 69 | 0.61 | |
| Mean EntropyBackbone=Qwen-2-Audio2026.04 | 38.23 | 76 | 58 | 45 | 83 | 67 | 0.59 | |
| RandomBackbone=Qwen-2-Audio2026.04 | 50.03 | 50 | 29 | 20 | 49 | — | — | |
| Logistic RegressionBackbone=Qwen-2-Audio, Feature Set=AUDIORATIO, Number of Features=10242026.04 | 66.97 | 52 | 45 | 29 | 95 | 60 | 0.53 | |
| Logistic RegressionBackbone=Qwen-2-Audio, Feature Set=Combined, Number of Features=40962026.04 | 75.99 | 43 | 41 | 26 | 96 | 61 | 0.53 | |
| Logistic RegressionBackbone=Qwen-2-Audio, Feature Set=Top 75, Number of Features=3002026.04 | 79.05 | 40 | 40 | 25 | 97 | 59 | 0.5 |