---
license: apache-2.0
---
Finetune DPO viethq188/LeoScorpius-7B with ultrafeedback_binarized dataset https://huggingface.co/datasets/HuggingFaceH4/ultrafeedback_binarized


You can use alpaca template.
```
template_format = """{system}
### Instruction:
{prompt}

### Response:
"""
```