Fact-checking on AggreFact Xsum
76.4Balanced AccuracyGPT-4o
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4oScale=>300B2025.02 | 76.4 | |
| Claude 3.5-SonnetScale=>300B2025.02 | 75.7 | |
| GPT-4Scale=>300B2025.02 | 75.4 | |
| OpenAI o1Scale=>300B2025.02 | 74.8 | |
| DeepSeek-V3 671BScale=>300B2025.02 | 74.6 | |
| GraphCheckBackbone=Llama3.3 70B2025.02 | 72.9 | |
| GraphCheckBackbone=Qwen 72B2025.02 | 72.1 | |
| MiniCheck2025.02 | 70.8 | |
| AlignScore2025.02 | 68 | |
| GraphEval2025.02 | 67.6 | |
| Llama3.3 70B2025.02 | 60.1 | |
| Qwen2.5 72B2025.02 | 55.6 | |
| ACUEval2025.02 | 55.5 | |
| Llama3 8B2025.02 | 53.4 | |
| Qwen2.5 7B2025.02 | 53.2 |