List your company

Skill Learning on Medical and Multi-Task Suite (Hellaswag, Humaneval, IFeval, MMLU, TruthfulQA, Winogrande)

40.2Medical ScoreSDFT
29.69632.42335.1537.877Jan 27, 2026

Evaluation results

MethodLinks
SDFT2026.0140.261.467.772.371.547.371.965.4
DFT2026.0136.261.964.674.671.640.171.364
SFT + re-invoke2026.0135.661.563.167.67042.371.462.6
SFT2026.0135.559.562.156.670.539.872.960.2
BaseBackbone=Qwen2.5-7B2026.0130.16265.874.371.747.971.165.5