If we assume that the sample is statistically significant (if total reviews are in single digits it doesn't matter), then this indicates that the company's hiring process "works as intended".
When you interview, if you don't end up being an employee, you will have a bad experience. But if you do end up working there, you will not see the interview in a negative light*, on the contrary, you will be very happy with the working conditions. This filters perfectly good matches for the company culture.
Of course one can argue that disgruntling other potential hires is a failure, but, still the hiring process filters out completely anyone who would not be 100% happy working there, so it sounds like a resounding success.
All of that, is of course assuming, that the reviews are legitimate. Negative reviews are usually legitimate, but they can be very biased, by simply having a much greater chance of being posted (people are incentivized to share negative experiences rather than neutral ones). As for the positive employee reviews, these are far more questionable. Your best course of action, is to find people working there currently (through eg Linkedin) and ask them to give you their unofficial opinion, why they would (or not) recommend a friend to work there, and what was their worst day on the job.
Personally if I knew that if I were to become an employee I would have an insane chance of liking it, and otherwise I would refuse the offer or not get one, I'd be happy with the review process, as it would filter me out of an environment I wouldn't want to be part of (no matter whose fault it is, theirs or mine).
*Assuming that none of the interview reviewers accepted a job offer (if they were offered but did not accept, it still was a good filter, by presenting them why they would not want to work there)