Follow-up Question Answering on MathChat-FQA 1st turn
83.4AccuracyLEMMA
Evaluation Results
| Method | Links | |
|---|---|---|
| LEMMABackbone=LLaMA3-8B2025.03 | 83.4 | |
| DART-Math-U.Backbone=LLaMA3-8B2025.03 | 82.4 | |
| DART-Math-H.Backbone=LLaMA3-8B2025.03 | 81.7 | |
| MetaMathBackbone=LLaMA3-8B2025.03 | 80.3 | |
| GPTAugBackbone=LLaMA3-8B2025.03 | 78.8 | |
| ISCBackbone=LLaMA3-8B2025.03 | 78.4 | |
| RefAug-90kBackbone=LLaMA3-8B2025.03 | 71.7 | |
| RefAugBackbone=LLaMA3-8B2025.03 | 69.3 | |
| RFTBackbone=LLaMA3-8B2025.03 | 64 | |
| SFTBackbone=LLaMA3-8B2025.03 | 63.2 |