Got told my open-source model experiments are too scattered. I'm organizing a journal to provide clarity before structuring the first git release. Is this readable for ML folks who aren’t in mech interp? Open to ANY feedback [D] — PLINKFEED