--- base_model: appvoid/v118-dpo-round-01 library_name: transformers tags: - dpo - preference-optimization - text-generation - text-editing - rewriting - instruct private: true --- ![https://huggingface.co/appvoid/a-cool-model/resolve/main/acoolmodel.png](acoolmodel.png) A cool model that follows instructions quite well, try it. | Benchmark | a-cool-model | Qwen2.5 0.5B Instruct | Difference (a-cool-model vs Qwen) | Higher Score | | :--- | :---: | :---: | :---: | :---: | | **ARC Easy** | **61.45%** | 59.18% | +2.27% | **a-cool-model** | | **PIQA** | 67.95% | **70.46%** | -2.51% | **Qwen2.5 0.5B Instruct** | | **ARC Challenge** | **35.67%** | 33.28% | +2.39% | **a-cool-model** | | **HellaSwag** | 47.32% | **52.42%** | -5.10% | **Qwen2.5 0.5B Instruct** |