평가 점수 언어 기준은 영어인가요?
한국어로도 같은 점수가 나오는지 궁금하네요.
I just evaluated with open-llm-leaderboard data, they use their dataset; You can check if you are interested in { https://huggingface.co/docs/leaderboards/open_llm_leaderboard/about }. You can also check about these datasets from this model's result on related datasets. In my personal opinion, this model can infer multilingual such as English, Russian, Chinese, ALSO Korean. Throughout it can not be perfect frankly speaking it speaks mix-language; it likes '나는 이런 feelings를 느껴요.' It quite often prints that way in my experience. But still, it can speak multi-language when you order; using prompts correctly. Plus I tested the actual model in Q4_K_M.gguf , it might be more greater result I hoped. Hope this answer will be what you want and thank you for your interest in my merge.
Oh, I thought you were Korean. Anyway thanks for your answer :)