Fact-checking on HOVER
71.82Macro F1 (2-hop)FOLK
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FOLKMethod Category=LLM-based reasoning methods, Source=Reproduction, Backbone Model=DeepSeek-Coder, k=10, threshold=0.92026.01 | 71.82 | 61.3 | 54.89 | — | — | |
| ProgramFCMethod Category=LLM-based reasoning methods, Source=Reproduction, Backbone Model=DeepSeek-Coder, k=10, threshold=0.92026.01 | 69.78 | 60.27 | 56.41 | — | — | |
| RRCMethod Category=LLM-based reasoning methods, Backbone Model=DeepSeek-Coder, k=10, threshold=0.92026.01 | 69.33 | 63.98 | 60.95 | — | — | |
| FLAN-T5Method Category=LLM-based ICL methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 69.02 | 60.23 | 55.42 | — | — | |
| DeBERTaV3-NLIMethod Category=Finetuned methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 68.72 | 60.76 | 56 | — | — | |
| BootstrappingMethod Category=LLM-based reasoning methods, Source=Reproduction, Backbone Model=DeepSeek-Coder, k=10, threshold=0.92026.01 | 68.45 | 61.29 | 59.23 | — | — | |
| BiDevMethod Category=LLM-based reasoning methods, Source=Reproduction, Backbone Model=DeepSeek-Coder, k=10, threshold=0.92026.01 | 67.85 | 62.03 | 58.87 | — | — | |
| CodexMethod Category=LLM-based ICL methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 65.07 | 56.63 | 57.27 | — | — | |
| RoBERTa-NLIMethod Category=Finetuned methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 63.62 | 53.99 | 52.4 | — | — | |
| MULTIVERSMethod Category=Finetuned methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 60.17 | 52.55 | 51.86 | — | — | |
| LisT5Method Category=Pretrained methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 52.56 | 51.89 | 50.46 | — | — | |
| BERT-FCMethod Category=Pretrained methods, Source=Reported in [9], k=10, threshold=0.92026.01 | 50.68 | 49.86 | 48.57 | — | — | |
| Factcheck-GPTLLM=GPT-4o2026.01 | — | — | — | 42 | 58 | |
| Factcheck-GPTLLM=GPT-4o-mini2026.01 | — | — | — | 22 | 52 | |
| FactoolLLM=GPT-4o2026.01 | — | — | — | 44 | 56 | |
| FactoolLLM=GPT-4o-mini2026.01 | — | — | — | 41 | 51 | |
| FIRELLM=GPT-4o2026.01 | — | — | — | 39 | 66 | |
| FIRELLM=GPT-4o-mini2026.01 | — | — | — | 30 | 70 | |
| PCCLLM=GPT-4o2026.01 | — | — | — | 52 | 76 | |
| PCCLLM=GPT-4o-mini2026.01 | — | — | — | 45 | 71 | |
| SAFELLM=GPT-4o2026.01 | — | — | — | 44 | 68 | |
| SAFELLM=GPT-4o-mini2026.01 | — | — | — | 28 | 68 |