General Question: Which framework do you use to evaluate MiniCPM3-4B?

[xunkai55]

Hi there,

I tried to evaluate MiniCPM3-4B by my self (deployed locally with OpenBMB/vllm) with OpenAI/simple-evals but got weirdly low scores, i.e. MMLU = 59.8. Before digging it deeper, I'd like to know which evaluation framework do you use.

Thanks a lot!