It seems that GPT sometimes misses certain details when given non-English documents to read. However, it appears to respond more effectively when the text is directly pasted into it. Has anyone else experienced a similar situation?
Side note: How do you feed PDF to GPT?
I’m using another platform that offers free access to ChatGPT-3.5 because I can’t log in openai chatgpt. It has a plugin that allows me to add PDFs for it to read. I believe it functions in a similar manner?
What’s the name of the tool you are using to read PDFs?
I utilize Forefront AI by uploading a PDF document and then posing a question to the system.
It looks like the PDF text extractor of the site you are using is the one that has a problem since you can get exactly what you want when you feed the text directly by yourself.
Perhaps I will try using another PDF extraction tool to see if it can resolve the issue. Thank you.