Interesting analysis. I wonder if the Chinese skills are undercounted somewhat- there could be more repositories the researchers didn’t have access to.

Would love to know how changing the language impacts the output quality. Do the models using these non-English skills still reason in English? Do different languages make it easier to express certain ideas more clearly? Considering how linguistic relativity might apply here.