It would be a good idea to read several articles just for their form. Read them without paying attention to what they are saying, only looking at how they are put together. To do this it would best to choose articles you aren't very interested in so that you wouldn't be distracted. Look at how the author backs up her claims, how he writes, what kind of signposting she feels the need to use. What his authorial persona is. You could even develop a rubric to judge the article by several measures.
Graduate students read scholarly articles, but they don't tend to read them formalistically as I'm suggesting. Maybe I'll try it myself since I'm recommending it to you.