First, you should always try to find answers on details and your questions on a paper in the references of that paper. It's also easy to overlook the supplementary material of a paper, when available. If your questions/details are very fundamental and basic, it's likely these issues are resolved in those sources, otherwise it's unlikely it would have passed the review process of a good journal.
On average (even on researchgate (imo horrible to search this website)) you will not find many readers interested in a particular article, as not many read an individual article, and the likelihood that you meet at the same time/spot in the WWW is low. If the article has an higher citation count, your chances are better to discuss via googlegroups/stackexchange/boards a paper with other being interested, but also the chances that you missed or misunderstood something reading the paper.
Another option is looking into papers citing the particular paper, via google scholar you can even search only within such papers and using the right keywords.
One imho very unused option is looking into a related thesis to that paper. In depth explanations, technical setups, formulas are often described there in detail and often the thesis is freely available in the www or the server of the university. In Germany all PhD thesis are freely downloadable.
If your confusion and questions are not answered by those methods, you should contact the corresponding author of a paper via the given mail adress. But keep the length ouf your mail short and the questions clear, especially and even if you think some results are "suspicious/unfair". Here you can also ask if a thesis exists that gives more details than the paper. It's only your suspicion and maybe wrong. Rather ask how than why questions, especially under such a suspicion.
Personally, if I ask a corresponding author something I'm missing or not understanding after applying above methods before, it's how they did something (technical measurement, simulation details). Why questions are mostly all resolved by authors, journal editor and reviewers during planning, writing and reviewing a paper. The details of how are quite often missing, due to limited space/interest or to keep an advance. In the latter case, if no other researcher/group can reproduce specific results in a paper (this happens quite often), at some point the authors will/have to reveal some secrets, before being suspected of pseudo-science and I have experienced also several times that results/methods of paper have been doubted directly after presentation in the questions time at conferences by the audience. If you can meet an author of a paper which is important for your research on a conference, take the opportunity, asking someone direct critical questions in a talk or poster presentation will often reveal much faster, if something is made-up or the well thought out.