List your company

Medical on MedCalc

57.4AccuracySFT
5.29618.82332.3545.877May 6, 2026

Evaluation results

MethodLinks
SFTBackbone=Qwen2.5-3B-Instruct2026.0557.4
Anchored LearningBackbone=Llama-3.2-3B-Instruct2026.0556.4
Anchored LearningBackbone=Qwen2.5-3B-Instruct2026.0556.3
SFTBackbone=Llama-3.2-3B-Instruct2026.0554.7
Low-SFTBackbone=Qwen2.5-3B-Instruct2026.0554.4
Low-SFTBackbone=Llama-3.2-3B-Instruct2026.0551.3
Anchored LearningBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 2: iGSM -> MedCalc2026.0551.2
Low-SFTBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 2: iGSM -> MedCalc2026.0550.5
KL-SFTBackbone=Qwen2.5-3B-Instruct2026.0535.6
KL-SFT 0.2Backbone=Llama-3.2-3B-Instruct2026.0535.6
Anchored LearningBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 3: iGSM -> MedCalc -> IFEval2026.0529.7
Iter-SFTBackbone=Qwen2.5-3B-Instruct2026.0529
Iter-SFTBackbone=Llama-3.2-3B-Instruct2026.0523.8
Self-sftBackbone=Qwen2.5-3B-Instruct2026.0521.6
Self-SFTBackbone=Llama-3.2-3B-Instruct2026.0517.3
Low-SFTBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 3: iGSM -> MedCalc -> IFEval2026.0515.7
STMBackbone=Qwen2.5-3B-Instruct2026.0514.7
DFTBackbone=Qwen2.5-3B-Instruct2026.0514.5
DFTBackbone=Llama-3.2-3B-Instruct2026.0514.5
BaseBackbone=Qwen2.5-3B-Instruct2026.0513.3
BaseBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Pre-trained2026.0513.3
Low-SFTBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 1: iGSM2026.0510
STMBackbone=Llama-3.2-3B-Instruct2026.058.5
BaseBackbone=Llama-3.2-3B-Instruct2026.057.8
Anchored LearningBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 1: iGSM2026.057.3