It's the same benefit as knowing how to build a house helps an architect. Understanding the medium one is working with is a good skill to have.
An architect that only knows how to use the pen can still be a good architect, but one that also knows how to pick up a hammer tends to have better insight into the types of building solutions one can come up with.
From my own experience, knowing JS has allows us to call out bad vendors who lacked basic UI skills and would claim "That can't be done". We'd just then build the UI ourselves and send it over, leaving them no room to wiggle out of their inabilities.
It also has allowed me to design in code, which is a boon to our IAs and others on the team. We can work much more iteratively. They tweak their wireframes, I tweak the code, and we can go back and forth quickly.
Finally, I think knowing how to build something helps you as a UXer to understand all the individual points of interaction that have to be thought about. I find that it's common for UX people with no code skills to create wireframes that make a lot of sense on paper, but are missing at least half of the interaction specifications required to actually build it.
As for jQuery vs. JS, jQuery is JS--just faster to write. I say learn jQuery and as you learn it, you'll find that you're learning a lot about how JS works as well.
If you were thinking about dedicating some significant time to learning JS, I would weigh it up against your other skillsets. For example spending more time learning about User Research may provide more bang for your buck in terms of future UX jobs.
Others have answered that you can build what you design - or at least that it is one route to doing this.
As many have pointed out, it helps you build your design.
I'd like to enforce that you should be able to build your design. When I make a carousel or other thing in JS/jQuery, what really matters is how fast it rotates, which direction, how many images, if it should fade, etc. Usually those are just parameters and I prefer the person with the idea play with those parameters until perfected rather than me the programmer.
This is the same in the games industry. The game designer should know enough code to tweak parameters. I.e. coder creates jump-function, designer tweaks how high and how fast. Etc.
Also, instead of putting your time into explaining what and how you want it - you can just do it (or at least a prototype of it).