g1: Using Llama-3.1 70b on Groq to create o1-like reasoning chains - View it on GitHub