Adds DPO fine-tuned Llama-3.1-8B-Instruct model

initial commit