Short-length Instruction Following on Lenctrl-Bench
4.99MAELARFT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LARFTBase Model=Llama-3.2-3B-Instruct2026.02 | 4.99 | 21.22 | |
| RLBase Model=Llama-3.2-3B-Instruct2026.02 | 5.49 | 21.73 | |
| RLBase Model=Llama-3.1-8B-Instruct2026.02 | 6.57 | 21.7 | |
| LARFTBase Model=Llama-3.1-8B-Instruct2026.02 | 6.98 | 21.78 | |
| LARFTBase Model=Qwen2.5-7B-Instruct2026.02 | 7 | 21.12 | |
| LARFTBase Model=Qwen2.5-3B-Instruct2026.02 | 7.72 | 20.25 | |
| SFT+RLBase Model=Llama-3.2-3B-Instruct2026.02 | 8.52 | 20.65 | |
| RLBase Model=Qwen2.5-7B-Instruct2026.02 | 9.01 | 21.38 | |
| SFT+RLBase Model=Qwen2.5-7B-Instruct2026.02 | 9.71 | 19.8 | |
| OriginBase Model=Llama-3.2-3B-Instruct2026.02 | 11.83 | 21.12 | |
| SFT+RLBase Model=Qwen2.5-3B-Instruct2026.02 | 12.09 | 20.01 | |
| OriginBase Model=Llama-3.1-8B-Instruct2026.02 | 12.74 | 21.59 | |
| RLBase Model=Qwen2.5-3B-Instruct2026.02 | 13.92 | 20.86 | |
| SFT+RLBase Model=Llama-3.1-8B-Instruct2026.02 | 14.19 | 17.91 | |
| OriginBase Model=Qwen2.5-7B-Instruct2026.02 | 24.01 | 19.9 | |
| OriginBase Model=Qwen2.5-3B-Instruct2026.02 | 28.76 | 18.08 | |
| SFTBase Model=Qwen2.5-3B-Instruct2026.02 | 45.44 | 16.87 | |
| SFTBase Model=Qwen2.5-7B-Instruct2026.02 | 47.77 | 17.85 | |
| SFTBase Model=Llama-3.2-3B-Instruct2026.02 | 48.16 | 17.54 | |
| PositionIDBase Model=Qwen2.5-7B-Instruct2026.02 | 53 | 19.45 | |
| SFTBase Model=Llama-3.1-8B-Instruct2026.02 | 54.4 | 17.02 | |
| PositionIDBase Model=Llama-3.1-8B-Instruct2026.02 | 59.2 | 19.17 | |
| RulerBase Model=Qwen2.5-7B-Instruct2026.02 | 65.4 | 19.4 | |
| PositionIDBase Model=Llama-3.2-3B-Instruct2026.02 | 72.2 | 19.4 | |
| RulerBase Model=Llama-3.1-8B-Instruct2026.02 | 73.8 | 19.66 | |
| RulerBase Model=Qwen2.5-3B-Instruct2026.02 | 79.8 | 19.78 | |
| PositionIDBase Model=Qwen2.5-3B-Instruct2026.02 | 84.9 | 19.17 | |
| RulerBase Model=Llama-3.2-3B-Instruct2026.02 | 86 | 19.43 |