Aggregate Evaluation on Multi-domain Aggregate Suite
57.1General AverageBase
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BaseBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Pre-trained2026.05 | 57.1 | 23.8 | 46.6 | |
| Anchored LearningBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 1: iGSM2026.05 | 53.3 | 45.9 | 45.5 | |
| Low-SFTBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 3: iGSM -> MedCalc -> IFEval2026.05 | 52.3 | 33.1 | 45.9 | |
| Anchored LearningBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 3: iGSM -> MedCalc -> IFEval2026.05 | 51.8 | 50.5 | 49.2 | |
| Low-SFTBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 1: iGSM2026.05 | 49.6 | 44.1 | 42.3 | |
| Low-SFTBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 2: iGSM -> MedCalc2026.05 | 46.2 | 47 | 44.8 | |
| Anchored LearningBackbone=Qwen2.5-3B-Instruct, Incremental Learning Stage=Stage 2: iGSM -> MedCalc2026.05 | 44.8 | 57.7 | 45.7 |