I want to be able to run codes in GPT env as it significantly boosts the quality of the GPT answers. The environment is fine when starting a new chat. But environment detoriates so quickly that code interpreter is essentially useless and non-existent. What I mean by “detoriate” is that the environment itself stops working thus removing a huge chunk of GPT’s features.
I tried this with GPT-4, GPT-4o, Data Analyst and Python (a GPT). The behavior is precisely the same. How can I fix this? Anyone with the same issue?
Note: Please note that the whole point of Data Analyst is to visualise data… if this issue won’t/can’t be fixed than why data analyst even exists? I added this example to demonstrate why this is a serious issue.