Your work has left a profound impression on us and provided us with great inspiration.
I would like to compare the ability of Chinese LLM. JudgeLM demonstrates a good understanding of Chinese from LLM output, but the explanations(reason) it provides are in English, even when our questions and the answers from both models are in Chinese.
Is there a simple way to make the model's output in Chinese? Alternatively, could retraining the model using translated data from the Judge105K dataset achieve this goal?
Thank you very much.
Your work has left a profound impression on us and provided us with great inspiration.
I would like to compare the ability of Chinese LLM. JudgeLM demonstrates a good understanding of Chinese from LLM output, but the explanations(reason) it provides are in English, even when our questions and the answers from both models are in Chinese.
Is there a simple way to make the model's output in Chinese? Alternatively, could retraining the model using translated data from the Judge105K dataset achieve this goal?
Thank you very much.