The conversation was conducted in Chinese, but in the end, the memory list retrieved from mem0 showed English
Issue with current documentation:
The conversation was conducted in Chinese, but in the end, the memory list retrieved from mem0 showed English
What's the reason for this? Can it be controlled?
Is it like adding a sentence 'Please answer in Chinese' directly in the content? But if it is stored in the content, then the sentence 'please answer in Chinese' will also be stored as memory in mem0, and adding 'please answer in Chinese' to each content will also be troublesome
I also encountered the same problem.
I know how to do it now. The source code can provide a place to add system prompts
https://docs.mem0.ai/open-source/features/custom-fact-extraction-prompt#prompt-blueprint
建议从源代码提示词哪里改