Single-Hop Question Answering on NQ (in-domain)
48.2AccuracyGEPO
Evaluation Results
| Method | Links | |
|---|---|---|
| GEPOType=RL Training, Model=Qwen2.5-7B-Instruct2025.10 | 48.2 | |
| GiGPOType=RL Training, Model=Qwen2.5-7B-Instruct2025.10 | 46.4 | |
| SKILLRLBackbone=Qwen2.5-7B-Instruct2026.02 | 45.9 | |
| ZeroSearchBackbone=Qwen2.5-7B-Instruct2026.02 | 43.6 | |
| GEPOType=RL Training, Model=Qwen2.5-3B-Instruct2025.10 | 43.6 | |
| ZeroSearchType=RL Training, Model=Qwen2.5-7B-Instruct2025.10 | 43.6 | |
| EvolveRBackbone=Qwen2.5-7B-Instruct2026.02 | 43.5 | |
| GiGPOType=RL Training, Model=Qwen2.5-3B-Instruct2025.10 | 42 | |
| ZeroSearchType=RL Training, Model=Qwen2.5-3B-Instruct2025.10 | 41.4 | |
| Search-R1Backbone=Qwen2.5-7B-Instruct2026.02 | 39.3 | |
| Search-R1Type=RL Training, Model=Qwen2.5-7B-Instruct2025.10 | 39.3 | |
| Search-R1Type=RL Training, Model=Qwen2.5-3B-Instruct2025.10 | 34.1 | |
| RAGBackbone=Qwen2.5-7B-Instruct2026.02 | 27.4 | |
| R1-InstructType=RL Training, Model=Qwen2.5-3B-Instruct2025.10 | 27 | |
| R1-InstructBackbone=Qwen2.5-7B-Instruct2026.02 | 21 | |
| R1-InstructType=RL Training, Model=Qwen2.5-7B-Instruct2025.10 | 21 | |
| Search-o1Backbone=Qwen2.5-7B-Instruct2026.02 | 19.4 | |
| CoTBackbone=Qwen2.5-7B-Instruct2026.02 | 12.8 | |
| Qwen2.5Backbone=Qwen2.5-7B-Instruct2026.02 | 11.6 |