My subjective impression in comparing X100T in-camera jpegs with Fuji Raw File Converter software output jpegs is that the film simulations are very close but not identical between the two methods. The X100T doesn't have the Acros simulation, though, so I don't know if that's any different.
I'd expect the Adobe versions of all the simulations to be different again - as far as I know, Adobe is reverse-engineering these, perhaps by comparing in-camera jpegs of colour calibration targets with the output of their own conversions of the corresponding raw files, and generating a profile to match them as closely as possible. I wasn't quite convinced by Adobe's versions of various Nikon camera profiles (compared to either in-camera jpegs or NX conversions) back in the CS6 era, but perhaps they are better now.