@victor_tokarev @terratempora this was the first place i went to check as well as i was born in czechoslovakia... and it struck me it was marked as part of the ussr since 1968. i tried to explain it out as very weird interpretation of the state of things after 1968 soviet invasion so i went to check to around ww2 and the data is complete utter mess. i investigated where the source data comes from and although it started as (much smaller) academic dataset it was ultimately compiled by some company using "AI"... so actuall yes, it is claude (or chatgpt or whatever) fault. sorry, i love the idea but this project is a prime example for why not to use LLM for this kind of work. it does more harm than good.