Mathematical Reasoning on Nemotron Math (test)
59.2AccuracyMoLEM
Evaluation results
| Method | Links | |
|---|---|---|
| MoLEMBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 3 Math (Nemotron Math)2026.05 | 59.2 | |
| MemGenBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 3 Math (Nemotron Math)2026.05 | 58.4 | |
| SFTBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 3 Math (Nemotron Math)2026.05 | 54.8 | |
| SFTBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 1 Code (KodCode)2026.05 | 50.6 | |
| MemGenBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 2 Science (Nemotron Science)2026.05 | 35.4 | |
| MemGenBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 1 Code (KodCode)2026.05 | 33.2 | |
| ExpeLBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 2 Science (Nemotron Science)2026.05 | 31.4 | |
| ExpeLBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 1 Code (KodCode)2026.05 | 30.8 | |
| MoLEMBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 2 Science (Nemotron Science)2026.05 | 30.8 | |
| VanillaBase Model=Qwen3-4B-Instruct-2507, Training Stage=Vanilla2026.05 | 30.4 | |
| MoLEMBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 1 Code (KodCode)2026.05 | 30.4 | |
| ExpeLBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 3 Math (Nemotron Math)2026.05 | 29.6 | |
| SFTBase Model=Qwen3-4B-Instruct-2507, Training Stage=Stage 2 Science (Nemotron Science)2026.05 | 27 |