It’s just the displayed Kelvin value that differs, not the white balance multipliers themselves, that are actually used. Also, last time i checked there is no single standard way to convert the white balance multipliers to the color temperature. dt uses the scientifically-correct way, doing all the fancy spectral stuff:
So if anything, i’d say dt is correct.
(rt uses similar method, but when i checked back then, their code produced yet another Kelvin temperature, so it had bugs, back then at least.)
What you are seeing is likely explained by the fact that dt only uses [adobe] color matrix, while LR uses entirety of the DCP, which dt does not support.
You can mimic the jpegs, if you so like them: