It is more accurate to think of the spins as always interacting with one another.. at 0K temperature only singlets are allowed, increasing the temperature by just a bit, both singlets and triplets coexist, etc. even that is just a picture.

To start with an everyday analogy. if you know you have 51cents in your piggy bank, but not what the precise breakdown is. Then you must have at least one penny. But you don't know exactly how many pennies you have unless you take a look.

Unlike pennies though, the singlets and triplets cannot be distinguished from one another. One can only measure how many "excess" triplets there are, one cannot point to where these triplets are hiding. The measurement doesn't "cause" a collapse, after the measurement, you know there is some spin imbalance but it is still in some superposition

Anyways.. to fully get what I am going with this, you will have to play with the math on your own. I have, but it's hard to translate all that to English. I'm just trying to point out that the interesting part is not in "collapse" but rather the "failure to collapse"--- the failure to take the system out of superpositions. Because the spins are always interacting.

Or maybe I just don't understand enough to explain it like Feynman.. heh.. but if you turn some of what Feynman has said in plain English into math, you would see that some of his stuff is also.. misleading

IBM must have some sims you can run on your PC, ask ChatGPT to solve and draw a small "Heisenberg model" that imports their libs. If you can take Taiwanese:

https://youtu.be/hHbUytvNLeE