Cross-lingual Question Answering on MKQA Thai
27.83fEMCroSearch-R1
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CroSearch-R1Backbone=Qwen2.5-7B-Instruct, Paradigm=Post-training2026.04 | 27.83 | 45.98 | |
| Search-R1Backbone=Qwen2.5-7B-Instruct, Paradigm=Post-training2026.04 | 25.29 | 37.13 | |
| CroSearch-R1Backbone=Qwen2.5-3B-Instruct, Paradigm=Post-training2026.04 | 24.41 | 42.26 | |
| SFTBackbone=Qwen2.5-7B-Instruct, Paradigm=Post-training2026.04 | 23.26 | 34.69 | |
| Search-R1Backbone=Qwen2.5-3B-Instruct, Paradigm=Post-training2026.04 | 22.36 | 33.95 | |
| Search-o1Backbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.04 | 18.54 | 23.65 | |
| IRCoTBackbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.04 | 18.06 | 23.65 | |
| SFTBackbone=Qwen2.5-3B-Instruct, Paradigm=Post-training2026.04 | 16.91 | 29.06 | |
| RAGBackbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.04 | 16.76 | 22.55 | |
| IRCoTBackbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.04 | 15.46 | 21.13 | |
| Search-o1Backbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.04 | 15.25 | 20.35 | |
| RAGBackbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.04 | 14.02 | 19.04 | |
| Direct InferenceBackbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.04 | 14.02 | 20.75 | |
| Direct InferenceBackbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.04 | 11.43 | 18.75 |