Question Answering on HotpotQA (Cover Exact Match)
52.4Cover EMPAGER
Evaluation Results
| Method | Links | |
|---|---|---|
| PAGERBackbone=Llama3.1-70B-Instruct2026.01 | 52.4 | |
| DeepNoteBackbone=Llama3.1-70B-Instruct2026.01 | 51.7 | |
| PAGERBackbone=Qwen3-32B2026.01 | 50.6 | |
| Search-o1Backbone=Llama3.1-70B-Instruct2026.01 | 50.2 | |
| StructRAGBackbone=Llama3.1-70B-Instruct2026.01 | 49.6 | |
| RATBackbone=Llama3.1-70B-Instruct2026.01 | 48.9 | |
| DeepNoteBackbone=Qwen3-32B2026.01 | 48.4 | |
| Vanilla RAGBackbone=Llama3.1-70B-Instruct2026.01 | 48.2 | |
| Search-o1Backbone=Qwen3-32B2026.01 | 47.6 | |
| IRCoTBackbone=Qwen3-32B2026.01 | 45.1 | |
| RATBackbone=Qwen3-32B2026.01 | 43.7 | |
| Iter-RetGenBackbone=Qwen3-32B2026.01 | 43.2 | |
| Vanilla RAGBackbone=Qwen3-32B2026.01 | 42.9 | |
| StructRAGBackbone=Qwen3-32B2026.01 | 42.5 | |
| Iter-RetGenBackbone=Llama3.1-70B-Instruct2026.01 | 41.1 | |
| Vanilla LLMBackbone=Llama3.1-70B-Instruct2026.01 | 37.7 | |
| IRCoTBackbone=Llama3.1-70B-Instruct2026.01 | 35.2 | |
| Vanilla LLMBackbone=Qwen3-32B2026.01 | 28.6 |