Sections of styled text (italic, bold, etc) which contain multiple textRuns render incorrectly

[ChrisRimmer]

When rendering something like the image caption "Feature visualisation [...]" in this doc the text is broken into three pieces, each of which gets a set of asterisks around it to make it bold. This results in *Feature visualization of a neuron that corresponds to dog-like features (**[image source](https://distill.pub/2020/circuits/zoom-in/)**)* which is understandably interpreted by the markdown renderer as a bold segment inside an italic one, rather than as three italic segments back to back.

Seems super minor so I'm going to call it low priority for now, but it seems good to have a record of it

[LeMurphant]

I've also seen traces of this happening in #69

e.g. **A****lignment is hard like cryptographic security is hard: you** **have to plan for the possibility of intelligent** **adversaries.**

I think it's minor enough as well, especially if we can use blockquotes, but worth keeping in mind.