Question Answering on Bamboogle (Cover Exact Match)
62.4Cover Exact MatchPAGER
Evaluation Results
| Method | Links | |
|---|---|---|
| PAGERBackbone=Qwen3-32B2026.01 | 62.4 | |
| PAGERBackbone=Llama3.1-70B-Instruct2026.01 | 62.4 | |
| RATBackbone=Llama3.1-70B-Instruct2026.01 | 58.4 | |
| Search-o1Backbone=Llama3.1-70B-Instruct2026.01 | 58.4 | |
| DeepNoteBackbone=Llama3.1-70B-Instruct2026.01 | 58.4 | |
| Vanilla LLMBackbone=Llama3.1-70B-Instruct2026.01 | 57.6 | |
| StructRAGBackbone=Llama3.1-70B-Instruct2026.01 | 57.6 | |
| Vanilla RAGBackbone=Llama3.1-70B-Instruct2026.01 | 56.8 | |
| RATBackbone=Qwen3-32B2026.01 | 55.2 | |
| Vanilla RAGBackbone=Qwen3-32B2026.01 | 52 | |
| StructRAGBackbone=Qwen3-32B2026.01 | 44.8 | |
| Vanilla LLMBackbone=Qwen3-32B2026.01 | 44 | |
| DeepNoteBackbone=Qwen3-32B2026.01 | 39.2 | |
| Iter-RetGenBackbone=Qwen3-32B2026.01 | 37.6 | |
| Search-o1Backbone=Qwen3-32B2026.01 | 33.6 | |
| Iter-RetGenBackbone=Llama3.1-70B-Instruct2026.01 | 32.8 | |
| IRCoTBackbone=Qwen3-32B2026.01 | 32 | |
| IRCoTBackbone=Llama3.1-70B-Instruct2026.01 | 11.2 |