I have just started experimenting with the 7B Q4 model by running "ollama run wizardlm2" and so far it seems to be very good at arithmetic word problems. I usu

I have just started experimenting with the 7B Q4 model by running "ollama run wizardlm2" and so far it seems to be very good at arithmetic word problems. I usually use Mistral 7B-Q4 with Ollama's REST interface. I am curious about and looking forward to slightly adjust the prompts I use for wizardlm2 and see how it does.