Mathematical Reasoning on MetaMathQA (test)
53.37AccuracyDA
Evaluation results
| Method | Links | |
|---|---|---|
| DAsource model=qwen3-32b, train dataset=gsm8k2026.02 | 53.37 | |
| DAsource model=qwen3-8b, train dataset=gsm8k2026.02 | 51.78 | |
| RSFTsource model=gemma3-27b-it, train dataset=gsm8k2026.02 | 49.73 | |
| DAsource model=gemma3-27b-it, train dataset=gsm8k2026.02 | 49.2 | |
| RSFTsource model=qwen3-32b, train dataset=gsm8k2026.02 | 48.14 | |
| LoRARank=16, Base Model=LLaMA2-7B, Fine-tuning=true2024.02 | 47.76 | |
| RSFTsource model=qwen3-8b, train dataset=gsm8k2026.02 | 45.03 | |
| ResLoRAbsBase Model=LLaMA2-7B, Fine-tuning=true2024.02 | 43.9 | |
| PADsource model=gemma3-27b-it, train dataset=gsm8k2026.02 | 43.72 | |
| ResLoRAisBase Model=LLaMA2-7B, Fine-tuning=true2024.02 | 43.37 | |
| PADsource model=qwen3-32b, train dataset=gsm8k2026.02 | 42.53 | |
| PADsource model=qwen3-8b, train dataset=gsm8k2026.02 | 42.51 | |
| LoRARank=4, Base Model=LLaMA2-7B, Fine-tuning=true2024.02 | 42.23 | |
| DAsource model=gemma3-27b-it, train dataset=medmcqa2026.02 | 35.43 | |
| PADsource model=qwen3-32b, train dataset=medmcqa2026.02 | 34.54 | |
| RSFTsource model=qwen3-32b, train dataset=medmcqa2026.02 | 34.4 | |
| RSFTsource model=gemma3-27b-it, train dataset=medmcqa2026.02 | 34.4 | |
| DAsource model=qwen3-32b, train dataset=medmcqa2026.02 | 34.38 | |
| PADsource model=gemma3-27b-it, train dataset=medmcqa2026.02 | 33.27 | |
| ResLoRAmsBase Model=LLaMA2-7B, Fine-tuning=true2024.02 | 33.13 | |
| PADsource model=qwen3-8b, train dataset=medmcqa2026.02 | 32.38 | |
| DAsource model=qwen3-8b, train dataset=medmcqa2026.02 | 30.27 | |
| LoHABase Model=LLaMA2-7B, Fine-tuning=true2024.02 | 30.17 | |
| RSFTsource model=qwen3-8b, train dataset=medmcqa2026.02 | 28.19 | |
| AdaLoRABase Model=LLaMA2-7B, Fine-tuning=true2024.02 | 25.17 | |
| LoKrBase Model=LLaMA2-7B, Fine-tuning=true2024.02 | 21.76 |