Textual understanding on SedarEval
57.7Overall AccuracySimPO
Evaluation Results
| Method | Links | |
|---|---|---|
| SimPOModel=Gemma-2-9B-IT2025.06 | 57.7 | |
| Uni-DPOModel=Gemma-2-9B-IT2025.06 | 57.5 | |
| Uni-DPOModel=Qwen2.5-7B2025.06 | 49.9 | |
| SimPOModel=Qwen2.5-7B2025.06 | 48.2 | |
| Uni-DPO v0.2Model=Llama3-8B Instruct2025.06 | 46.5 | |
| Uni-DPOModel=Llama3-8B Instruct2025.06 | 46.32 | |
| SimPOModel=Llama3-8B Instruct2025.06 | 44.81 | |
| SimPO v0.2Model=Llama3-8B Instruct2025.06 | 42.37 | |
| SFTModel=Llama3-8B Instruct2025.06 | 40.93 | |
| Uni-DPO v0.2Model=Llama3-8B Base2025.06 | 39.62 | |
| Uni-DPOModel=Llama3-8B Base2025.06 | 38.49 | |
| SimPO v0.2Model=Llama3-8B Base2025.06 | 36.23 | |
| SimPOModel=Llama3-8B Base2025.06 | 32.43 | |
| DPOModel=Llama3-8B Base2025.06 | 31.8 | |
| SFTModel=Llama3-8B Base2025.06 | 20.85 |