Multi-hop Question Answering on 2Wiki (pass@1)
58pass@1R-Search-7B + SSP
Evaluation Results
| Method | Links | |
|---|---|---|
| R-Search-7B + SSPTraining Paradigm=Continual Training on Search-Specialized Agents, SSP Integration=true2025.10 | 58 | |
| R-Search-7BTraining Paradigm=Continual Training on Search-Specialized Agents, SSP Integration=false2025.10 | 56.4 | |
| Qwen3-8B + SSPTraining Paradigm=Generalization Across Model Families, SSP Integration=true2025.10 | 51.5 | |
| Qwen2.5-14B-Instruct + SSPTraining Paradigm=Scaling to Larger Models, SSP Integration=true2025.10 | 49.4 | |
| Qwen2.5-32B-Instruct + SSPTraining Paradigm=Scaling to Larger Models, SSP Integration=true2025.10 | 49.2 | |
| Qwen2.5-14B-InstructTraining Paradigm=Scaling to Larger Models, SSP Integration=false2025.10 | 48.4 | |
| Qwen2.5-32B-InstructTraining Paradigm=Scaling to Larger Models, SSP Integration=false2025.10 | 48.4 | |
| Qwen3-8BTraining Paradigm=Generalization Across Model Families, SSP Integration=false2025.10 | 48 | |
| Search-R1-7B + SSPTraining Paradigm=Continual Training on Search-Specialized Agents, SSP Integration=true2025.10 | 45.6 | |
| Search-R1-7BTraining Paradigm=Continual Training on Search-Specialized Agents, SSP Integration=false2025.10 | 45.2 | |
| Qwen2.5-7B-Instruct + SSPTraining Paradigm=From-Scratch Training on Base and Instruct Models, SSP Integration=true2025.10 | 38.8 | |
| ZeroSearch-7B + SSPTraining Paradigm=Continual Training on Search-Specialized Agents, SSP Integration=true2025.10 | 37.2 | |
| ZeroSearch-7BTraining Paradigm=Continual Training on Search-Specialized Agents, SSP Integration=false2025.10 | 34.6 | |
| LLaMA-3.1-8B + SSPTraining Paradigm=Generalization Across Model Families, SSP Integration=true2025.10 | 34.4 | |
| Qwen2.5-7B-Base + SSPTraining Paradigm=From-Scratch Training on Base and Instruct Models, SSP Integration=true2025.10 | 33.2 | |
| Qwen2.5-7B-InstructTraining Paradigm=From-Scratch Training on Base and Instruct Models, SSP Integration=false2025.10 | 32.8 | |
| LLaMA-3.1-8BTraining Paradigm=Generalization Across Model Families, SSP Integration=false2025.10 | 19.4 | |
| Qwen2.5-7B-BaseTraining Paradigm=From-Scratch Training on Base and Instruct Models, SSP Integration=false2025.10 | 10.8 |