In short, and I think most SEOs understand this already, is Penguin looks at the page linking to your site. If that page is really bad, it can be labeled as a Penguin link.
Some SEOs may have confused links as just the anchor text or the URL. But when I pried on the concept that Google calls Penguin a "web spam algorithm" and specifically not a "link spam algorithm" - Gary said this is why - because it looks at the source the link is coming from.
Google Penguin looks at "the source site, not on the target site" he said in the interview. So it is not about the destination of the link but rather the source of the link.
An example he gave of a link source that might be slapped with a Penguin label is a source page that is a low quality forum profile page that links to you. He said "empty profile pages, forum profile pages" might be one example.
Here is the audio embed:
The transcript is over here.
Glenn Gabe nicely summed it up in this tweet as well:
Google: "Webspam" with Penguin relates to the SOURCE site, not DESTINATION. Many were looking at the wrong side :) https://t.co/PNQ2QSiyFu— Glenn Gabe (@glenngabe) October 11, 2016
Again, most of you all probably knew this but I don't think we had Google provide such transparency on the concept that Penguin is based on the link source versus link destination or anchor text or something else.
Forum discussion at Twitter.