Is there a way to update llama_index query function to use GPT4? I'm assuming so - just combing through the code now, figured I'd ask in case anyone knows off the bat 🙂