I tend to keep track of the popular publications in my field and skim abstracts to see what I find most interesting. Fairly time consuming, but it's the most effective method I've found. Google Scholar also has alerts on new papers that I occasionally find interesting.