David gives a great explanation of what these terms mean. I understand your confusion about how we talk about them, though.

When you get a Find the Assumption question ("which of the following is an assumption on which the argument depends?"), we're looking for that missing Premise 2, as David showed. In other words, you're looking for the thing that we don't yet know is true (because the author didn't state it), but that would HAVE to be true for the Conclusion to follow from Premise 1. Without it, the conclusion falls apart. In other words, it is NECESSARILY to make the argument logically cohere.

When you get a Draw an Inference question ("which of the follow must be true," "what conclusion can be drawn," etc), you're also looking for something that HAS to be true - something that flows logically from the premises given. The difference is generally that nothing else "falls apart" if the conclusion is untrue. Everything else stated is fact, so it can't really fall apart. In fact, we're looking for the thing that we can't make untrue.

To simplify:
Find Assumption: ask yourself, "does this statement have to be true for the conclusion to hold up?"
Draw an Inference: ask yourself "does this statement have to be true?" (full stop).

Does that clear things up?