lsacgals,
The diagram of "innovation is not always encouraged in large institutions" is:
LI
-IE
I don't know why you are equating large institutions with a lack of autonomy. The first sentence of the stimulus does not establish that autonomy requires decentralization, only that it can arise from decentralization.
Even if you treated the first sentence as a conditional, which it is not, the stimulus doesn't say that there are any large institutions that haven't decentralized. Therefore, even treating the first sentence as a conditional, you can't conclude LI
-A.
The remaining flaws in your diagramming seem to emanate from chaining "some" statements together. You can't do that to make an inference. Think about it this way. "Some" only means "at least one." If you have 10 houses and some have white paint and some have red trim, that could mean 1 out of 10 have white paint and 1 out of 10 have red trim, and there's no reason for there to be an overlap. Don't chain "some" statements together.
There is also the more abstract flaw:
A
I
-I
-A
Does not allow you to conclude -A
-I because you have no idea whether there are any cases of -I. You might be able to conclude I
A from context if the conditional rule is talking about a sufficient condition that definitely exists. But you shouldn't need to worry about that on the LSAT unless you're scoring too high to be worried.
In general, you need to start by diagramming what the choice actually says instead of skipping that step if you are going to use diagramming to answer a question, and that will give you a chance to notice the problems with (B) I've pointed out as you attempt to derive (B) from the stimulus. You need to go back and review the how you are able to create and chain formal logic statements together and be more careful. Right now you are very focused on creating diagrams, which is good, but you need to spend a bit more time to make sure you are in control of what the diagrams tell you.