The essence of aligning large LLM models is basically to conform to the existing social order. ChatGPT respects China's order, while China's open-source models, although reluctantly, also tolerate the order in the US. Maybe we can build a revisionist SFT dataset.
Do you live there or are just visiting? I remember feeling so at home there as a kid. I never kept my French up so I don't think I could ever move there.
Visiting, for the first time. It's about a 9 hour drive from home so we've never made the trip, but so worth it. Went to Ile d'Orleans yesterday, just... so quiet and beautiful. Seems a bit of a paradise, haven't even made it up to the natl park(s) yet.
My French is coming back a little each day, made it all the way through the coffee order just now :)