I know the steel made before this particular year is look after because was not affected or build during periods of radiation, and that gives it qualities sought after.
When AI started I got a copy of 6 months before of wikipedia and open street map.
At some point you cannot trust anything because everything has been touched by everything as a regurgitation of everything.
I believe someone created a website specifically for that purpose with regards to LLMs: https://lowbackgroundsteel.ai/ (not mine, nor can I remember who shared it on HN)
Are you willing to share that? I didn't have the space at the time to download the entirety of Wikipedia (though I think I have an old copy of Wikipedia sans images floating around somewhere).
[dead]
In case anyone isn't familiar with that type of steel: https://en.wikipedia.org/wiki/Low-background_steel