Mathematical Reasoning on GSM8K (Accuracy, Length)
95.1AccuracyEvo-L2S (Accuracy)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Evo-L2S (Accuracy)Model Scale=14B, Merging Strategy=Single-Objective Evolutionary Merging2026.04 | 95.1 | 945.7 | |
| Evo-L2S-02Model Scale=14B, Merging Strategy=Multi-Objective Evolutionary Merging2026.04 | 94.8 | 763.6 | |
| Task ArithmeticModel Scale=14B, Merging Strategy=Arithmetic Merging2026.04 | 94.5 | 865.8 | |
| Evo-L2S-01Model Scale=14B, Merging Strategy=Multi-Objective Evolutionary Merging2026.04 | 94.3 | 858.5 | |
| Evo-L2S-04Model Scale=14B, Merging Strategy=Multi-Objective Evolutionary Merging2026.04 | 94.2 | 273.9 | |
| DeepSeek-R1-Distill-Qwen-14BModel Scale=14B, Merging Strategy=Distilled Model2026.04 | 93.4 | 1,061.1 | |
| Evo-L2S-03Model Scale=14B, Merging Strategy=Multi-Objective Evolutionary Merging2026.04 | 93.3 | 593.4 | |
| ACM-TAModel Scale=14B, Merging Strategy=Activation-informed Merging (ACM)2026.04 | 92.6 | 1,545.6 | |
| Qwen2.5-14BModel Scale=14B, Merging Strategy=Base Model2026.04 | 91.7 | 202.4 | |
| Evo-L2S-05Model Scale=14B, Merging Strategy=Multi-Objective Evolutionary Merging2026.04 | 91.5 | 245.2 | |
| ACM-TIESModel Scale=14B, Merging Strategy=Activation-informed Merging (ACM)2026.04 | 91.2 | 640.6 | |
| ACM-AverageModel Scale=14B, Merging Strategy=Activation-informed Merging (ACM)2026.04 | 90.7 | 742.2 | |
| Average MergingModel Scale=14B, Merging Strategy=Arithmetic Merging2026.04 | 89.8 | 781.8 | |
| TIES-MergingModel Scale=14B, Merging Strategy=Arithmetic Merging2026.04 | 89.3 | 712.5 | |
| Evo-L2S (Length)Model Scale=14B, Merging Strategy=Single-Objective Evolutionary Merging2026.04 | 89.1 | 367.5 |