Mathematical Reasoning on GSM8K-Aug, ASDiv-Aug, and DU Average
77.08AccuracyLTF
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LTFModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 77.08 | 1.94 | |
| ReGuLaRModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 74.4 | 2.13 | |
| CoLaRModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 70.97 | 5.1 | |
| LTFModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 68.85 | 1.95 | |
| CoconutModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 66.98 | 6 | |
| ReGuLaRModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 65.12 | 2.1 | |
| iCoTModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 62.18 | 0 | |
| LTFModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 62.16 | 1.88 | |
| SoftCoT++Model=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 59.95 | 120.48 | |
| LTFModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 59.68 | 1.91 | |
| CoLaRModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 58.97 | 4.93 | |
| Assist-CoTModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 58.41 | 120.98 | |
| ReGuLaRModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 58.06 | 1.98 | |
| CoTModel=LLaMA-3.1 Instruct 8B, Finetuning Setting=true2026.06 | 58.02 | 121.42 | |
| ReGuLaRModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 55.56 | 2.04 | |
| CoconutModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 52.23 | 6 | |
| CoLaRModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 49.03 | 5.04 | |
| CoLaRModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 47.97 | 5 | |
| iCoTModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 44.73 | 0 | |
| SoftCoT++Model=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 42.76 | 120.2 | |
| CoconutModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 42.74 | 6 | |
| Assist-CoTModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 40.2 | 120.38 | |
| CoconutModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 39.22 | 6 | |
| CoTModel=LLaMA-3.2 Instruct 3B, Finetuning Setting=true2026.06 | 39.03 | 121.6 | |
| iCoTModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 36.86 | 0 | |
| SoftCoT++Model=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 36.58 | 120.13 | |
| iCoTModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 35.84 | 0 | |
| Assist-CoTModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 34.15 | 120.1 | |
| SoftCoT++Model=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 33.65 | 119.64 | |
| CoTModel=DeepSeek-R1-Distill-Qwen-1.5B, Finetuning Setting=true2026.06 | 32.89 | 120.79 | |
| CODIModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 31.76 | 6 | |
| Assist-CoTModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 31.27 | 119.65 | |
| CoTModel=LLaMA-3.2 Instruct 1B, Finetuning Setting=true2026.06 | 29.91 | 120.37 |