Math Reasoning on Minerva STEM evaluation
16.16AccuracyNaive Training
Evaluation Results
| Method | Links | |
|---|---|---|
| Naive TrainingBackbone=LLaMA-DCLM, Reweighting Space=Data sources2025.07 | 16.16 | |
| DataAgent_RLBackbone=LLaMA-DCLM, Reweighting Space=Data sources2025.07 | 14.8 | |
| DataAgent_RLBackbone=LLaMA-FWE, Reweighting Space=Data sources2025.07 | 13.46 | |
| DataAgent_SFTBackbone=LLaMA-DCLM, Reweighting Space=Data sources2025.07 | 12.7 | |
| Naive TrainingBackbone=LLaMA-FWE, Reweighting Space=Data sources2025.07 | 12.58 | |
| Naive TrainingBackbone=LLaMA-Nemo, Reweighting Space=Data sources2025.07 | 12.3 | |
| DBLBackbone=LLaMA-DCLM, Reweighting Space=Data sources2025.07 | 12.2 | |
| RegMixBackbone=LLaMA-FWE, Reweighting Space=Data sources2025.07 | 12.2 | |
| DataAgent_SFTBackbone=LLaMA-FWE, Reweighting Space=Data sources2025.07 | 12.19 | |
| DBLBackbone=LLaMA-FWE, Reweighting Space=Data sources2025.07 | 12.05 | |
| RegMixBackbone=LLaMA-DCLM, Reweighting Space=Data sources2025.07 | 11.5 | |
| DataAgent_RLBackbone=LLaMA-Nemo, Reweighting Space=Data sources2025.07 | 10.83 | |
| DBLBackbone=LLaMA-Nemo, Reweighting Space=Data sources2025.07 | 10.74 | |
| RegMixBackbone=LLaMA-Nemo, Reweighting Space=Data sources2025.07 | 10.7 | |
| Naive TrainingBackbone=Pythia-1.4B, Reweighting Space=Data sources2025.07 | 10.16 | |
| DataAgent_SFTBackbone=LLaMA-Nemo, Reweighting Space=Data sources2025.07 | 9.07 | |
| DataAgent_RLBackbone=Pythia-1.4B, Reweighting Space=Data sources2025.07 | 7.73 | |
| DataAgent_SFTBackbone=Pythia-1.4B, Reweighting Space=Data sources2025.07 | 7.19 | |
| RegMixBackbone=Pythia-1.4B, Reweighting Space=Data sources2025.07 | 6.92 | |
| Base ModelBackbone=LLaMA-Nemo, Reweighting Space=Data sources2025.07 | 4.85 | |
| Base ModelBackbone=Pythia-1.4B, Reweighting Space=Data sources2025.07 | 4.39 | |
| Base ModelBackbone=LLaMA-DCLM, Reweighting Space=Data sources2025.07 | 4.1 | |
| Base ModelBackbone=LLaMA-FWE, Reweighting Space=Data sources2025.07 | 2.52 |