Multi-hop Faithfulness Hallucination Detection on HoVer Refined
82.9Macro F1FaithLens
Evaluation Results
| Method | Links | |
|---|---|---|
| FaithLensbackbone=Llama-3.1-8B-Inst2025.12 | 82.9 | |
| GPT-4.12025.12 | 82.6 | |
| Llama-3.1-405B-Inst2025.12 | 81.6 | |
| o32025.12 | 81.1 | |
| ClearCheckbackbone=Llama-3.1-8B-Inst2025.12 | 80.3 | |
| Claude-3.7-Sonnet2025.12 | 80.2 | |
| DeepSeek-V3.2thinking=true2025.12 | 80 | |
| o12025.12 | 79.9 | |
| o3-mini2025.12 | 78.5 | |
| DeepSeek-V3.2thinking=false2025.12 | 76.7 | |
| MiniCheckparameters=7B2025.12 | 74.9 | |
| GPT-4o2025.12 | 73.6 | |
| AlignScoreparameters=355M2025.12 | 73.3 | |
| FactCGparameters=435M2025.12 | 73.1 |