Mathematical Reasoning on GSM8k (test) (ACC, ΔACC, AGR, KLD)
47.8AccuracyQwen 2.5 7B Instruct
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen 2.5 7B Instructk=42025.09 | 47.8 | 2.5 | 60.3 | 0.86 | |
| Llama 3.1 8B Instructk=42025.09 | 32.6 | 4.1 | 53.7 | 0.35 |
Loading the SOTA2 catalog…