The simplicity of the term "literature review" is a little deceptive. This is because the term can refer to many things; its meaning is different for some fields, professionals, and situations. Sometimes "literature review" refers to a section of an essay or report. Other times it is an entire article in itself. Most confusingly, the term can refer not to a thing at all (section, paper, etc.) but can refer to a process. (In a thesis, for example, the "literature review" may be spread throughout the entire work, bolstering certain sections with evidence.) So then, how can we make sense of it? Are there commonalities among all of these use cases?
A literature review is the action (or the product of the action) of aggregating and analyzing evidence on whatever you're studying. Simply put, literature reviews give context to your work.
How do you know what you know about your topic? What have other researchers said on the subject? What conclusions or gaps can you draw by looking at the scholarship about your topic? These are all questions that literature reviews generally try to answer.
Similar to a lot of science, the most crucial feature of any literature review is minimizing bias. There is enough evidence in the world that no matter how farfetched a hypothesis may be, you can probably find something that aligns with your take. However, just because a piece of evidence aligns with your idea does not mean it is proven. What if two articles have opposite conclusions? Do they simply cancel each other out? What if one paper reports findings after one test while the other ran the same test one hundred times? Do their findings have equal weight?
To minimize bias in a literature review, you must aggregate your evidence dispassionately. Simply put, do not search for evidence that proves your theory. Search for whatever evidence is out there, and build your theory with it.