I have one with a cracked open case that I’ve been meaning to probe out, I guess I should do that. It has been on my list to build a replacement PCB that has 32k SRAM chip and ties A15 low to override the EI memory because I’ve kinda lost patience with the whole DIN plug mess.
There's pictures floating of the NewBuff which was a replica of the buffered EI cable but it seems the site that lists it is currently down. I think the photos still work though.
I just popped the lid off my original and did a little probing. Looks to me like, no, they're not, the D* signals pass straight through from one side to the other. Looks like the three '244s are entirely just addresses and control signals.
That's what I figured. If there was a bi-directional buffer such as a 74LS245 to handle the data lines, it would need an enable line derived from the address decoding inside the E.I. unit to prevent read contentions with the keyboard unit.