List your company

Reasoning on GPQA Diamond (test)

34.85AccuracyLIFT
26.446828.628430.8132.9916Jun 1, 2025

Evaluation results

MethodLinks
LIFTModel=Qwen2.5-3B, Rank (r)=2562025.0634.85
Full FTModel=Qwen2.5-3B2025.0633.33
LIFTModel=Qwen2.5-1.5B, Rank (r)=1282025.0628.79
Full FTModel=Qwen2.5-1.5B2025.0626.77