This model is a model that always answers in the same language as the questioner’s language. You can think of it as a version with reduced language matching errors in ChatGPT because it does not apply any rules related to conversation topics.
The reason I made it
As a result of using ChatGPT and Bing, I found that no matter how many questions I asked in Korean, I often received answers in English. So, I wrote a prompt at the bottom of the question that said, “Please answer in Korean,” but even so, if the conversation continued for a long time, I ended up answering in English again.
I think this is because the learning rate for English materials is very high. To correct this problem, we created a model called Language Mirror. Only language matching rules are applied and topic-related rules are not applied, so I personally created it thinking it was a modified version of chatgpt’s language matching.
Preview
Results from my own experience
Personally, I think I will continue to use this model instead of chatgpt. It is much more convenient because language matching is good. While making 4 versions of mygpts, I finally remembered “Jarvis from the movie Iron Man.”
