Verifiable Instruction Following on IFEval
84.3IFEval ScoreREDIPO
Evaluation Results
| Method | Links | |
|---|---|---|
| REDIPOModel=OLMo-3-7B2026.05 | 84.3 | |
| DivPOModel=Qwen3-4B2026.05 | 83.2 | |
| DivPOModel=OLMo-3-7B2026.05 | 82.8 | |
| InstructModel=OLMo-3-7B2026.05 | 82.3 | |
| DPOModel=OLMo-3-7B2026.05 | 81.9 | |
| InstructModel=Qwen3-4B2026.05 | 81.7 | |
| DPOModel=Qwen3-4B2026.05 | 81.3 | |
| REDIPOModel=Qwen3-4B2026.05 | 81.2 | |
| InstructModel=LLaMA-3.1-8B2026.05 | 74.3 | |
| REDIPOModel=LLaMA-3.1-8B2026.05 | 68.4 | |
| DPOModel=LLaMA-3.1-8B2026.05 | 64.5 | |
| DivPOModel=LLaMA-3.1-8B2026.05 | 60.4 | |
| BaseModel=Qwen3-4B2026.05 | 34.9 | |
| BaseModel=OLMo-3-7B2026.05 | 22.9 | |
| BaseModel=LLaMA-3.1-8B2026.05 | 8.9 |