I don’t understand it. In the playground, I try to add a new .txt file with some tags separated by “,”. I always get a failure message. When I upload the whole thing normally in GPT in the chat, I always get the message that an unknown error has occurred. I’ve tried everything, encoding the file in different ways: UFT8 ANSI, UFT8 with BOM, etc. Nothing works. It’s a simple .txt file with one line like “green, red, blue, orange” and nothing else, and it can’t upload it. I keep getting the same error.
I created a new file on my phone and tried to upload it, but it doesn’t work. On another computer with a different internet connection, it doesn’t work; other browsers don’t work.
The API disallows CSV files from being included in file search’s vector stores. What you’ve provided would be chunked into 800 token parts to run a search on.
Were it CSV, say with those being the title row, the vector store would give extremely poor results, and probably will anyway. Thus, file inspection disallows certain data types.
If it is information that must be seen in full or is significantly smaller than the AI model’s context window, I would just program a client-side file extractor that places text information as a user message “I have this text file for you to answer about:” and then the latest user message.
It’s not a .csv file, but a simple .txt file.
Since I don’t want to change the prompt every time, I have several such smaller files sets that I adjust as needed and that the prompt accesses. It worked perfectly a few months ago, but now updating this file no longer works cause the upload ends in an error.
Just wanted to chip in that I’m suffering the same error, but not on the playground but just in a ChatGPT project. Other browsers don’t work, other devices don’t work. Other text files don’t work. What does work is I can zip the text file, then ask chatgpt to unzip it to /usr/mnt. However that doesn’t make it available to other chats in the project and the file will be deleted after a while. I have been having this problem since yesterday.
@_j Thanks for your help, but honestly, I don’t understand why this shouldn’t work, since it’s barely a large amount of data. And somehow that doesn’t make sense for me, especially since there are no concrete error messages in the playground. And I only get the info that it can’t be uploaded and in the GPT chat I always only get the message that an unknown error has occurred. I mean its a normal textfile with some words and some commas.
In any case, I wanted to make progress, so I simply converted one file to .html and the others to .pdf. It works, and the token usage isn’t particularly high.