Bug Report:
Steps to reproduce:
- Initiate a conversation with GPT4, and add a number of attachments in the initial message
- Send multiple messages to Chat GPT4, say a dozen or so
- Upload a new document, say a PDF, and ask a question or offer a prompt which requires that GPT4 engage with the newly uploaded document
Expected result:
Chat GPT4 should first read, and then engage with the document by responding to the prompt as asked.
Actual result:
Chat GPT4 replies with the message verbatim or similar "I cannot directly access or view files, including the one you've just uploaded...." and continues onto a response related to the query, but without engaging with the newly added document.
Additional information
Browser: Google Chrome Version 121.0.6167.160 (Official Build) (64-bit)
OS: Windows 11 Home