Hallucination detection on Dolly AC (test)
81.59AUCTPA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TPAModel=LLaMA2-13B2025.12 | 81.59 | 97.41 | 80.75 | |
| TPAModel=LLaMA3-8B2025.12 | 76.08 | 65.61 | 75.29 | |
| TSVModel=LLaMA2-13B2025.12 | 75.52 | 59.52 | 60.43 | |
| TSVModel=LLaMA2-7B2025.12 | 74.54 | 87.28 | 76.84 | |
| NovoModel=LLaMA3-8B2025.12 | 74.18 | 58.54 | 63.16 | |
| TSVModel=LLaMA3-8B2025.12 | 73.47 | 64.67 | 66.95 | |
| TPAModel=LLaMA2-7B2025.12 | 71.34 | 78.97 | 75.27 | |
| NovoModel=LLaMA2-13B2025.12 | 69.09 | 72.22 | 69.03 | |
| NovoModel=LLaMA2-7B2025.12 | 64.23 | 80.7 | 72.44 | |
| ReDeEPModel=LLaMA2-13B2025.12 | 58.42 | 85.18 | 76.03 | |
| ReDeEPModel=LLaMA2-7B2025.12 | 51.36 | 82.45 | 78.33 | |
| ChainPollModel=LLaMA2-13B2025.12 | 47.58 | 43.64 | 60 | |
| ReDeEPModel=LLaMA3-8B2025.12 | 36.52 | 83.92 | 71 | |
| RAGASModel=LLaMA3-8B2025.12 | 36.28 | 80 | 52.46 | |
| ChainPollModel=LLaMA2-7B2025.12 | 35.02 | 41.38 | 55.81 | |
| TrulensModel=LLaMA3-8B2025.12 | 33.52 | 36.59 | 51.72 | |
| TrulensModel=LLaMA2-7B2025.12 | 31.98 | 55.17 | 66.67 | |
| EigenScoreModel=LLaMA2-13B2025.12 | 29.48 | 81.81 | 72 | |
| RAGASModel=LLaMA2-7B2025.12 | 28.77 | 53.45 | 63.92 | |
| RefCheckModel=LLaMA2-13B2025.12 | 28.69 | 25.45 | 39.44 | |
| RAGASModel=LLaMA2-13B2025.12 | 28.4 | 41.82 | 54.76 | |
| SEPModel=LLaMA2-13B2025.12 | 28.23 | 65.45 | 69.23 | |
| ChainPollModel=LLaMA3-8B2025.12 | 26.91 | 34.15 | 45.16 | |
| SEPModel=LLaMA2-7B2025.12 | 26.05 | 62.16 | 70.23 | |
| TrulensModel=LLaMA2-13B2025.12 | 25.65 | 38.18 | 49.41 | |
| RefCheckModel=LLaMA2-7B2025.12 | 24.94 | 39.66 | 54.12 | |
| EigenScoreModel=LLaMA2-7B2025.12 | 24.28 | 75 | 72.41 | |
| EigenScoreModel=LLaMA3-8B2025.12 | 20.65 | 71.42 | 59.52 | |
| ITIModel=LLaMA2-13B2025.12 | 6.46 | 53.85 | 67.12 | |
| SEPModel=LLaMA3-8B2025.12 | 6.39 | 68.29 | 68.29 | |
| ITIModel=LLaMA2-7B2025.12 | 4.42 | 58.16 | 62.81 | |
| ITIModel=LLaMA3-8B2025.12 | 0.24 | 30.91 | 42.5 | |
| RefCheckModel=LLaMA3-8B2025.12 | -0.89 | 19.51 | 27.59 |