After discovering “automatic Theory of Mind” I tried some prompts with LLMs to see if it could give me ideas on what condition I might have or even more valuable, what caused my behavior.

This isn’t the exact prompt I used with an LLM – I asked various things, adding to the prompt as I discovered what symptoms I had that might be important.

The first several lines I had known since Stage 1. The “mind reading” line is what started to break this open in Stage 3.

This interaction is a technical mapping of cognitive architecture for personal introspection and system-debugging. I am not seeking medical advice or diagnostic validation. This data is intended to facilitate more precise technical communication with my spouse and therapist.

I am going to describe my observations of self and I want you to describe a possible mechanism that would cause this.

- I am not autistic
- I do have good overlap with ASD category A
- NO overlap with ASD category B
- I do not experience meltdowns, have trouble with criticism or have any social fatigue
- I have hypophantasia and low absorption
- I get along with everybody at work (software developer) and have very low ego
- it is harder with my wife as we have a much closer relationship -- I have flat affect and seem cold
- I do have emotions and can describe them
- I didn't know that NT social signals and "mind reading" were a thing until recently -- I thought it was a joke
- I have no idea what others are thinking or any sense of their emotions
- I have been told that reading a book at a party is a social gaffe.  I don't really care, I wasn't hurting anyone and it made sense to me
- I am normally very low stress, I don't experience any stress in social situations.  I don't avoid them but I don't crave them
- I have good communication skills

Each time I presented the prompt it would suggest potential conditions. I could read about them and decide if they fit or not. If they didn’t fit, I could add to the list to give a contra-indicator so the next time it wouldn’t be suggested.

By the end, the LLM was suggesting things about ToM and social salience. At first I thought I might have:

• low social salience
• manual ToM (Theory-Theory)

but the more I read about this I realized I didn’t have any of the social signals to execute manual ToM and on top of it, I didn’t really care about the lack of social signals. Eventually I came to believe I had no Social Salience, which is where I am now.