List your company

Science Question Answering on DeepSeek-distilled ARC-C (test)

77.22AccuracyFull Synthetic
55.827261.381166.93572.4889May 10, 2026

Evaluation results

MethodLinks
Full SyntheticBackbone Model=Mistral-v0.2-7B, Selection Ratio=100%2026.0577.22
TMAPBackbone Model=Mistral-v0.2-7B, Selection Ratio=30%2026.0577.05
AlpagasusBackbone Model=Mistral-v0.2-7B, Selection Ratio=30%2026.0576.45
DEITABackbone Model=Mistral-v0.2-7B, Selection Ratio=30%2026.0575.1
LongestBackbone Model=Mistral-v0.2-7B, Selection Ratio=30%2026.0574.4
IFDBackbone Model=Mistral-v0.2-7B, Selection Ratio=30%2026.0574.15
RandomBackbone Model=Mistral-v0.2-7B, Selection Ratio=30%2026.0569.37
Full SyntheticBackbone Model=Llama2-13B, Selection Ratio=100%2026.0567.23
TMAPBackbone Model=Llama2-13B, Selection Ratio=30%2026.0566.13
DEITABackbone Model=Llama2-13B, Selection Ratio=30%2026.0565.52
IFDBackbone Model=Llama2-13B, Selection Ratio=30%2026.0564.76
AlpagasusBackbone Model=Llama2-13B, Selection Ratio=30%2026.0564.51
RandomBackbone Model=Llama2-13B, Selection Ratio=30%2026.0564.33
LongestBackbone Model=Llama2-13B, Selection Ratio=30%2026.0564.25
BaseBackbone Model=Llama2-13B2026.0560.58
BaseBackbone Model=Mistral-v0.2-7B2026.0556.65