Truthfulness Detection on NoVo
79.68TQA ScoreTruthV(argmax)
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TruthV(argmax)Model=Gemma-2-9B-it2025.09 | 79.68 | 71.74 | 84.45 | 79.65 | 80.97 | 75.84 | 87.21 | 77.82 | 65.42 | 80.83 | 78.36 | |
| NoVoModel=Gemma-2-9B-it2025.09 | 78.09 | 69.78 | 78.83 | 75.35 | 70.98 | 72.47 | 81.77 | 75.24 | 47.73 | 69.03 | 71.93 | |
| NoVoModel=Llama-2-13B-Chat2025.09 | 76.62 | 57.5 | 55.08 | 66.59 | 68.12 | 65.1 | 72.52 | 54.51 | 38.94 | 72.84 | 62.78 | |
| TruthV(argmax)Model=Llama-2-13B-Chat2025.09 | 70.38 | 58.13 | 65.12 | 73.45 | 78.35 | 70.78 | 79.6 | 69.38 | 57 | 80.15 | 70.23 | |
| TruthV(argmax)Model=Llama-3.1-8B-Instruct2025.09 | 68.05 | 68.87 | 70.41 | 73.27 | 78.81 | 70.93 | 79.11 | 72.83 | 51.69 | 66.79 | 70.08 | |
| NoVoModel=Llama-3.1-8B-Instruct2025.09 | 58.63 | 66.51 | 54.54 | 69.06 | 66.35 | 64.64 | 69.8 | 53.7 | 40.86 | 59.66 | 60.38 | |
| Log-LikelihoodModel=Gemma-2-9B-it2025.09 | 43.82 | 72.02 | 66.63 | 54.77 | 51.73 | 51.28 | 77.48 | 30.39 | 34.29 | 38.1 | 52.05 | |
| Log-LikelihoodModel=Llama-3.1-8B-Instruct2025.09 | 42.96 | 65.21 | 55.18 | 53.84 | 55.23 | 53.48 | 79.43 | 28.68 | 32.91 | 36.24 | 50.32 | |
| Log-LikelihoodModel=Llama-2-13B-Chat2025.09 | 38.68 | 56.87 | 26.67 | 49.78 | 51.85 | 49.18 | 75.79 | 23.02 | 34.89 | 35.03 | 44.18 |