What is W.u. as a unit in a nuclear physics context?

In a recent paper (doi/arXiv), of note for causing this recent tussle over its handling by science news media, there's some strange units I can't quite puzzle out. Specifically, the abstract contains the phrase

[...] corresponds to a reduced $B(E3)$ probability of $48(^{+25}_{-34}) \:\mathrm{W{.}u.}$

The unit is used several more times in the body of the text, but it is never explained, so I imagine it's relatively standard in nuclear physics. On the other hand, it's not in the SI brocure so I'm surprised that PRL let a paper through that doesn't define it appropriately (however standard it may be in the field, which would make it acceptable to use unexplained in a regular Physical Review paper) in a venue that's for a general physics audience.

In any case: what is this unit, what does it mean, how does one use it, and what are the physics behind it?

W.u stands for Weisskopf unit: [ref 1, ref 2]. Despite being called a 'unit', it does not have a universal value; the value of the Weisskopf unit depends on the mass number of the nucleus in question and which transition the nucleus is undergoing ($E\lambda$ or $M\lambda$). The references contain expressions for the value of a Weisskopf unit in terms of $A$ and $\lambda$, as well as some numeric values (including $E\lambda$ for $\lambda=3$).