Long-form generation on FreshWiki
54.1ROUGE-1Agentic Reasoning
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Agentic Reasoningbackbone=DeepSeek-R12025.02 | 54.1 | 19.62 | 18.77 | |
| STORMbackbone=DeepSeek-R12025.02 | 47.93 | 17.42 | 15.43 | |
| Search-O1backbone=DeepSeek-R12025.02 | 41.56 | 16.08 | 12.88 | |
| RAgentbackbone=DeepSeek-R12025.02 | 30.04 | 14.21 | 9.08 | |
| RAGbackbone=DeepSeek-R12025.02 | 29.14 | 14.23 | 8.84 | |
| Direct Genbackbone=DeepSeek-R12025.02 | 27.32 | 13.13 | 6.11 |