The most commonly understood definition of plagiarism is presenting someone else's work as your own.
When writing general background for a thesis or article, it's possible that you present common facts and methodologies without citing every single detail, but even then it should be clear that this is not your contribution.
When whole sentences start to resemble previous work, you need to take a step back and look at what you are saying.
Are you briefly mentioning Pythagoras' theorem? Maybe throw the word "commonly" in there.
Are you presenting someone else's findings? Cite them.
For example "The equation is given by Smith et al. as ..." or "The dates are canonically given as 450BC to 430BC ."
When in doubt, cite. I've yet to see someone told off for providing references to their claims.