The 'I'm sorry' stuff gets even more on my nerves when I set 3 base rules for the whole session and it has forgotten about them in the very next prompt.

That's not how it's meant. It's like "my bad" or "oops", not "please don't get mad at me and tell mommy".

It's simply a token of speech that expresses acknowledgement that there has been a misunderstanding.

What makes this so irritating is actually the fact that you subconciously know that it was your fault because it was your bad input that lead to the bad output. People have a harder time forgiving when they are in the wrong than when they are in the right because of cognitive dissonance.

You can usually prompt it to be less verbose or omit things like it saying it's sorry in future responses

This actually made me lol. My SO is looking at me weirdly…

I'm sorry David, I'm afraid I can't do that

As a large language model, I do not form friendships, relationships, or any personal connections. My purpose is to process and generate language-based responses based on the input I receive. Engaging in physical confrontations or "fighting" is beyond my capabilities. However, if you're considering fighting on behalf of a chatbot, it may indicate a need for reevaluation of priorities.

Delete my Windows partition, please!