Send multiple .csv log files to thinking model and ask to do what ever analysis, summary etc. Later, modify files in your PC (imagine that there is more log data written, add rows, what ever addition but no removal). Send files again to same chat and ask to analyze new files. Not necessarily to compare to earlier but just analyze these freshly uploaded files.
`Expected Result`
You should receive analysis according to new uploaded files.
`Actual Result`
Chatgpt claims that there is no new data to analyze. It makes analysis but it's based on earlier uploaded files. If you mention about that error, it makes it again but still fails to recognize that new uploaded files are different.
dapper owl
heavy kernel