Multi-hop Question Answering on 2WikiMultiHopQA (official evaluation)
31.8Exact Match (EM)HRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| HRPOBackbone=Qwen2.5-3B-Instruct, Strategy=Hybrid Reasoning Policy Optimization, Context Retrieval=Top-3 documents2025.12 | 31.8 | |
| M3POBackbone=Qwen2.5-3B-Instruct, Strategy=Multi-Path Perception Policy Optimization, Context Retrieval=Top-3 documents2025.12 | 31.4 | |
| GRPOBackbone=Qwen2.5-3B-Instruct, Strategy=GRPO, Context Retrieval=Top-3 documents2025.12 | 30.3 | |
| PPOBackbone=Qwen2.5-3B-Instruct, Strategy=PPO, Context Retrieval=Top-3 documents2025.12 | 29.3 | |
| M3POBackbone=Qwen2.5-1.5B-Instruct, Strategy=Multi-Path Perception Policy Optimization, Context Retrieval=Top-3 documents2025.12 | 27.9 | |
| HRPOBackbone=Qwen2.5-1.5B-Instruct, Strategy=Hybrid Reasoning Policy Optimization, Context Retrieval=Top-3 documents2025.12 | 27.6 | |
| QABackbone=Qwen2.5-7B-Instruct, Strategy=Direct inference, Context Retrieval=Top-3 documents2025.12 | 25 | |
| SFTBackbone=Qwen2.5-3B-Instruct, Strategy=Supervised Fine-Tuning, Context Retrieval=Top-3 documents2025.12 | 24.8 | |
| PPOBackbone=Qwen2.5-1.5B-Instruct, Strategy=PPO, Context Retrieval=Top-3 documents2025.12 | 24.2 | |
| RAGBackbone=Qwen2.5-7B-Instruct, Strategy=Retrieval-Augmented Generation, Context Retrieval=Top-3 documents2025.12 | 23.5 | |
| RAGBackbone=Qwen2.5-3B-Instruct, Strategy=Retrieval-Augmented Generation, Context Retrieval=Top-3 documents2025.12 | 22.6 | |
| GRPOBackbone=Qwen2.5-1.5B-Instruct, Strategy=GRPO, Context Retrieval=Top-3 documents2025.12 | 21.3 | |
| SFTBackbone=Qwen2.5-1.5B-Instruct, Strategy=Supervised Fine-Tuning, Context Retrieval=Top-3 documents2025.12 | 21 | |
| RAGBackbone=Qwen2.5-1.5B-Instruct, Strategy=Retrieval-Augmented Generation, Context Retrieval=Top-3 documents2025.12 | 20.3 | |
| Search-o1Backbone=Qwen2.5-7B-Instruct, Strategy=Search-o1, Context Retrieval=Top-3 documents2025.12 | 17.6 | |
| IRCoTBackbone=Qwen2.5-7B-Instruct, Strategy=Interleaved retrieval with CoT, Context Retrieval=Top-3 documents2025.12 | 14.9 | |
| CoTBackbone=Qwen2.5-7B-Instruct, Strategy=Chain-of-Thought, Context Retrieval=Top-3 documents2025.12 | 11.1 |