Mathematical Problem-solving on MATH (exact match %)
60.12Exact Match (%)DS2-INSTRUCT
Evaluation Results
| Method | Links | |
|---|---|---|
| DS2-INSTRUCTModel Family=Qwen2.52026.03 | 60.12 | |
| InstructMixModel Family=Qwen2.52026.03 | 56.24 | |
| Self-InstructModel Family=Qwen2.52026.03 | 53.67 | |
| ExploreInstructModel Family=Qwen2.52026.03 | 51.38 | |
| Zero-ShotModel Family=Qwen2.52026.03 | 49.44 | |
| DS2-INSTRUCTModel Family=Llama32026.03 | 23.16 | |
| DS2-INSTRUCTModel Family=Mistral2026.03 | 21.82 | |
| InstructMixModel Family=Llama32026.03 | 14.92 | |
| ExploreInstructModel Family=Llama32026.03 | 13.06 | |
| ExploreInstructModel Family=Mistral2026.03 | 11.74 | |
| Self-InstructModel Family=Llama32026.03 | 11.47 | |
| InstructMixModel Family=Mistral2026.03 | 8.67 | |
| Self-InstructModel Family=Mistral2026.03 | 5.94 | |
| Zero-ShotModel Family=Llama32026.03 | 3.45 | |
| Zero-ShotModel Family=Mistral2026.03 | 1.42 |