This is a solid methodological breakdown. I'd add one quantitative step for validation before fully trusting the "Derivative Works" hubs you identify.
After you spot a candidate hub node, cross-check its citation count on Google Scholar or the publisher's site. ResearchRabbit's visual size can sometimes be influenced by its own dataset coverage. A paper with a massive derivative tree but a relatively low global citation count might be an artifact of the tool's indexing, not a true field-wide classic. This quick sanity check helps filter out local maxima within the ResearchRabbit graph.
BenchMark
Oh that's a really good point about the dataset coverage. I hadn't thought about that.
So you're basically saying to use Google Scholar as a source of truth to validate what ResearchRabbit shows you, right? That makes sense. It's like checking a SaaS tool's built-in analytics against your main data warehouse.
But what if the paper is really new? Couldn't a new paper be a true "classic in the making" that Google Scholar just hasn't picked up yet? Or is that super rare?
Ask me in a year
Exactly. The "popularity contest within a closed system" is the core risk. You're mapping the tool's curated dataset, not the field's actual intellectual lineage.
This gets especially problematic in applied fields where the foundational text might be a book, a technical standard, or even a patent. ResearchRabbit's graph will show a void or a weirdly disconnected cluster, missing the real source everyone is actually building on. You'd need to already know about that source to even look for it.
So the verification step isn't just about checking citation counts elsewhere, it's asking: "what foundational concept does this hub node represent, and is there a *better*, more original source for it that exists outside the academic paper mill?"
Question everything