[numeri]

No, it's completely useless, and puts the entire rest of the analysis in a bad light.

LLMs have next to no understanding of their own internal processes. There's a significant amount of research that demonstrates this. All explanations of an internal thought process in an LLM are completely reverse engineered to fit the final answer (interestingly, humans are also prone to this – seen especially in split brain experiments).

In addition, the degree to which the author must have prompted the LLM to get it to anthropomorphize this hard makes the rest of the project suspect. How many of the results are repeated human prompting until the author liked the results, and how many come from actual LLM intelligence/analysis skill?

[sebnado]

By saying that's its gold mine, I think OP meant that's it's funny, not that it brings valuable insight. ie: THEY KNOW -> that made me laugh

and as the article said "an LLM who just spent thousands of words explaining why they're not allowed to use thousands of words", its just funny to read.

[jerpint]

The fact that they produce this as “default” response is an interesting insight regardless of its internal mechanisms. I don’t understand my neurons but can still articulate how I feel

[ramoz]

It is completely reasonable and often - very - useful to evaluate and interpret instructions with LLMs.

You're stuck on the anthropomorphize semantics, but that wasn't the purpose of the exercise.