Hallucination detection on SVAMP (test)
53.3ANLLGemma2
Evaluation Results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Gemma2Model=Gemma22025.12 | 53.3 | 47.5 | 69.1 | 75.3 | 83.2 | 83.8 | 84.9 | — | 82.8 | 82.9 | 83.7 | 85 | |
| Llama3.1Model=Llama3.12025.12 | 57.9 | 57.9 | 68.9 | 84.3 | 86.7 | 87.8 | 92.4 | 64.4 | 90.4 | 90.2 | 90.6 | 91.4 | |
| Llama3.2Model=Llama3.22025.12 | 62.1 | 62.1 | 57.7 | 80.3 | 80 | 79.3 | 86.2 | 60.8 | 84.8 | 84.4 | 84.6 | 86.4 | |
| Falcon3Model=Falcon32025.12 | 91.7 | 91.7 | 92.3 | 89.2 | 90.9 | 93 | 94.4 | 66.6 | 93.4 | 94 | 94.7 | 95.1 |