Open-domain multi-hop question answering on MuSiQue answerable
65.06Answer F1RAG-RL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Min-Max2025.03 | 65.06 | 81.51 | 72.37 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Linear2025.03 | 61.1 | 73.9 | 66.89 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Base-Answerable2025.03 | 57.53 | 67.75 | 62.22 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Max2025.03 | 54.64 | 68.84 | 60.92 | |
| RAG-RLEval. Setting=Ideal Retrieval, Curriculum=Base-Unanswerable2025.03 | 54.53 | 68.37 | 60.67 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Min-Max2025.03 | 47.18 | 64.48 | 54.49 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Linear2025.03 | 44.68 | 59.79 | 51.14 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Base-Answerable2025.03 | 41.4 | 54.28 | 46.97 | |
| Qwen2.5-7B-InstructEval. Setting=Ideal Retrieval, Curriculum=Baseline2025.03 | 41.16 | 58.16 | 48.21 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Max2025.03 | 40.91 | 53.07 | 46.2 | |
| RAG-RLEval. Setting=Distractor, Curriculum=Base-Unanswerable2025.03 | 38.84 | 52.15 | 44.52 | |
| Qwen2.5-7B-InstructEval. Setting=Distractor, Curriculum=Baseline2025.03 | 25.88 | 25.35 | 25.61 |