Reasoning on Xcope bo (acc)
65.4AccuracyOurs-MoE-SFT
Evaluation Results
| Method | Links | |
|---|---|---|
| Ours-MoE-SFTModel=Ours-MoE-SFT2025.07 | 65.4 | |
| Ours-Base2025.07 | 59.8 | |
| Ours-BaseModel=Ours-Base2025.07 | 59.8 | |
| Ours-Base-32kcontext-length=32k2025.07 | 58.6 | |
| Ours-Base-32kModel=Ours-Base-32k2025.07 | 58.6 | |
| Ours-SFTalignment=SFT2025.07 | 57.8 | |
| Ours-SFTModel=Ours-SFT2025.07 | 57.8 | |
| Ours-MoE-BaseModel=Ours-MoE-Base2025.07 | 57.8 | |
| Ours-MoE-Base-8kModel=Ours-MoE-Base-8k2025.07 | 57.2 | |
| Yak-Llama2-7B2025.07 | 53 | |
| Qwen2.5-7B-baseModel=Qwen2.5-7B-base2025.07 | 51.8 | |
| LLaMA3.1-8B-InstructModel=LLaMA3.1-8B-Instruct2025.07 | 51.6 | |
| Qwen2.5-7B-InstructModel=Qwen2.5-7B-Instruct2025.07 | 51.6 | |
| Tibetan-Alpaca-7B2025.07 | 51.2 | |
| Tibetan-Llama2-7B2025.07 | 50.6 | |
| Qwen3-8BModel=Qwen3-8B2025.07 | 50.4 | |
| DeepSeek-R1-Distill-Llama-8BModel=DeepSeek-R1-Distill-Llama-8B2025.07 | 50.4 |