I think most papers should effectively just be scavenged through. In (mathematics) grad school I’ve only had time to seriously study a single digit number of research papers, but I at least feel like I gained a lot from looking through many many more than that.
