I am using the gpt-4o-mini as the model for summarization of text. The prompt is as follows:
Provide a concise summary in less than 400 characters of the following text. Keep the output language the same as the input language.: ${text}.
Sometimes, the model seems to recognize the language wrong. While the input text is clearly in English, the summary turns out in Spanish or French. This happens only rarely, but often enough to be a problem.
The other issue is that when summarizing Kannada text, sometimes, it returns a summary predominantly in Kannada by with some of the words or parts of words in English.
What can I do to eliminate these problems? Any help or suggestions would be greatly appreciated.
Thanks