Can you ask chatGPT why it wrote a certain line or why it made a certain decision? I still think explainability is missing in these models, but even if it able

Can you ask chatGPT why it wrote a certain line or why it made a certain decision? I still think explainability is missing in these models, but even if it able to come up with something, what is the guarantee that the explanation is not bullshit?