Open-domain multi-hop question answering on 2Wiki
75.06Answer F1RAG-RL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Min-Max2025.03 | 75.06 | 80.37 | 77.63 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Linear2025.03 | 74.31 | 88.82 | 80.92 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Base-Answerable2025.03 | 72.61 | 79.98 | 76.11 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Max2025.03 | 71.82 | 78.53 | 75.02 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Base-Unanswerable2025.03 | 71.04 | 74.3 | 72.63 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Min-Max2025.03 | 70.77 | 76.37 | 73.46 | |
| Qwen2.5-7B-InstructEval. Setting=Ideal Retrieval, Curriculum=Baseline2025.03 | 70.29 | 54.49 | 61.39 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Linear2025.03 | 68.92 | 83.23 | 75.4 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Base-Answerable2025.03 | 68.56 | 76.36 | 72.25 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Max2025.03 | 67.99 | 77.08 | 72.25 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Base-Unanswerable2025.03 | 66.59 | 72.04 | 69.21 | |
| Qwen2.5-7B-InstructEval. Setting=Distractor, Curriculum=Baseline2025.03 | 48.71 | 40.18 | 44.03 |