These are pretty awful questions, I'd guess maybe 20% are questions you would actually see in a tech company data science interview. The others are too easy ("How would you create a logistic regression model?"), too CS focused ("What are hash table collisions?"), or just nonsensical ("How many sampling methods do you know?").
In my experience DS interview q's are more contextual and focused on problem solving, since that's basically the job. One place did spend an hour asking questions of this type (e.g. we walked through the random forest algorithm and talked about the assumptions that various models make) but I don't think it's the norm.
You would be surprised. Even basic stats questions like those are great for filtering out people who don’t have a clue about data science and the extent of their abilities is using a library such as sklearn.
In my experience DS interview q's are more contextual and focused on problem solving, since that's basically the job. One place did spend an hour asking questions of this type (e.g. we walked through the random forest algorithm and talked about the assumptions that various models make) but I don't think it's the norm.