Reflection Llama-3.1 70B is currently the world's top open-source large language model (LLM), trained using a new technique known as Reflection-Tuning, which enables the model to detect and correct errors in its reasoning. The model has been trained on synthetic data generated by Glaive, an excellent tool for those training models. It employs the standard Llama 3.1 chat format, distinguishing between the model's internal reasoning and final answers with special tags, thereby enhancing the user experience.