List your company

Reasoning and Knowledge Assessment on ARC bo

53.67AccuracyOurs-MoE-SFT
27.46234.26641.0747.874Jul 12, 2025

Evaluation results

MethodLinks
Ours-MoE-SFTModel=Ours-MoE-SFT2025.0753.67
Ours-SFTModel=Ours-SFT2025.0748.39
Ours-MoE-Base-8kModel=Ours-MoE-Base-8k2025.0745.79
Ours-MoE-BaseModel=Ours-MoE-Base2025.0744.96
Ours-Base-32kModel=Ours-Base-32k2025.0744.62
Ours-BaseModel=Ours-Base2025.0744.32
DeepSeek-R1-Distill-Llama-8BModel=DeepSeek-R1-Distill-Llama-8B2025.0730.38
Qwen2.5-7B-baseModel=Qwen2.5-7B-base2025.0729.6
LLaMA3.1-8B-InstructModel=LLaMA3.1-8B-Instruct2025.0729.31
Qwen3-8BModel=Qwen3-8B2025.0729.16
Qwen2.5-7B-InstructModel=Qwen2.5-7B-Instruct2025.0728.47