What kind of feedback we are looking for
We are not collecting raw chat logs or personal stories. What we need are short, structured notes about what actually happened when you used the protocol. Useful feedback often looks like this:
- “Here is the kind of question I asked.”
- “Here is how the model responded.”
- “Here is why it felt helpful / unhelpful / strange in this situation.”
We are especially interested in:
- cases where the model clearly reduced pressure or helped you see structure more clearly
- cases where it obviously fell back into Interpretation-era behaviour (optimising, hyping, over-analysing)
- cases where it misread your situation in a way that reveals a blind spot in the meta-prompt
How to structure your message
You can use the template below as a loose guide. Feel free to skip anything that doesn’t fit your case.
1. Context (1–3 sentences)
What were you doing when you opened the LLM? Phone or desktop? Work, personal, late at night, in a hurry, etc.
2. Your prompt (or a paraphrase)
Roughly what did you ask? (You can change details to protect your privacy.)
3. The model’s response (summary)
In your own words, what did it do? Did it follow the Structure vocabulary? Did it slide back into optimisation or self-help tone?
4. Your reaction
Did this answer make the situation lighter, heavier, clearer, more confusing? What felt structurally right or wrong about it?
5. Anything we should adjust
Is there something you think the meta-prompt should explicitly allow, forbid or emphasise more? You do not need to copy long transcripts.
A few clear examples are much more valuable than a complete history.
Privacy
Please remove or alter any details that could identify you or other people.
We may quote short, anonymised fragments of feedback in future journal entries or protocol updates, but never with names, emails or other direct identifiers.
If you absolutely do not want your feedback to be quoted even anonymously, just add a line:
“Please do not quote this feedback in public.”
We will respect that.