| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| 1D Convection-Diffusion-Reaction (CDR) Equation (test) | MODE | Test Loss2.5 | 14 | 2mo ago | |
| 1D Convection-Diffusion-Reaction (CDR) Equation (train) | MODE | Train Loss2.34 | 14 | 2mo ago | |
| MMLU (val) | FOAM | MMLU STEM Accuracy69.12 | 14 | 3mo ago | |
| GSM8K (test) | GaLore-AF | Test Perplexity2.4887 | 12 | 27d ago | |
| GSM8K (train) | AdamW16 | Training Loss0.3347 | 12 | 27d ago | |
| Alpaca (test) | Adam-mini | Test Perplexity7.3804 | 12 | 27d ago | |
| Alpaca (train) | AdamW16 | Loss0.5622 | 12 | 27d ago | |
| Target domain | LA-LoRA | Learning Score0.5 | 8 | 3mo ago | |
| WizardLM | RefLoRA | Evaluation Loss0.661 | 7 | 2d ago | |
| OpenOrca | RefLoRA | Evaluation Loss0.773 | 7 | 2d ago | |
| OpenHermes | RefLoRA | Evaluation Loss0.706 | 7 | 2d ago | |
| CodeFeedback | LoRA | Evaluation Loss0.638 | 7 | 2d ago | |
| Alpaca | RefLoRA | Evaluation Loss1.35 | 7 | 2d ago | |
| LLaMA-2-7B | QFT | Weights Memory Usage7.42 | 5 | 2mo ago | |
| Open-Platypus | SL-SAM | Max Memory (MB)47,388 | 4 | 3mo ago |
