Follow-up Question Answering on MathChat-FQA 3rd turn
43.4AccuracyLEMMA
Evaluation Results
| Method | Links | |
|---|---|---|
| LEMMABackbone=LLaMA3-8B2025.03 | 43.4 | |
| MetaMathBackbone=LLaMA3-8B2025.03 | 40.1 | |
| ISCBackbone=LLaMA3-8B2025.03 | 39.3 | |
| RefAug-90kBackbone=LLaMA3-8B2025.03 | 38.6 | |
| GPTAugBackbone=LLaMA3-8B2025.03 | 37.9 | |
| DART-Math-U.Backbone=LLaMA3-8B2025.03 | 37.1 | |
| RefAugBackbone=LLaMA3-8B2025.03 | 36.5 | |
| DART-Math-H.Backbone=LLaMA3-8B2025.03 | 35.1 | |
| RFTBackbone=LLaMA3-8B2025.03 | 29.1 | |
| SFTBackbone=LLaMA3-8B2025.03 | 28.3 |