Instruction Following on IF-Eval
84.66AccuracyGeneral Teacher
Evaluation Results
| Method | Links | |
|---|---|---|
| General Teacher2026.05 | 84.66 | |
| Basebackbone=DS-8B2025.09 | 63.7 | |
| CaMOPD2026.05 | 59.89 | |
| IPObackbone=DS-8B2025.09 | 56.2 | |
| RealSafebackbone=DS-8B, alignment=SFT-based2025.09 | 54.7 | |
| SelecTKD2026.05 | 49.17 | |
| Medical Teacher2026.05 | 48.98 | |
| Relaxed OPD2026.05 | 48.61 | |
| Vanilla MOPD2026.05 | 48.06 | |
| Base ModelBackbone=Meta-Llama-3-8B-Instruct2026.02 | 40.48 | |
| NumericalBackbone=Meta-Llama-3-8B-Instruct2026.02 | 39.93 | |
| RandomBackbone=Meta-Llama-3-8B-Instruct2026.02 | 39.74 | |
| WIM Fixed JudgeBackbone=Meta-Llama-3-8B-Instruct, Judge Strategy=Fixed2026.02 | 39.56 | |
| WIM Changing JudgeBackbone=Meta-Llama-3-8B-Instruct, Judge Strategy=Changing2026.02 | 39.19 |