At the moment, Datasets Viewer sometimes displays non-ASCII symbols (including but not limited to Cyrillic letters) in a way that’s basically unreadable. For example,
Гамбурге may be displayed as
\u0413\u0430\u043c\u0431\u0443\u0440\u0433\u0435. This, of course, makes it quite difficult to explore a given dataset.
The problem occurs only under specific conditions: the list view should be unchecked and the dataset feature causing this should be a sequence of strings.
I made a small PR fixing that some time ago. Basically, the cause of this issue is that by default
json.dumps that is used for table view escapes non-ASCII symbols. So you just need to override that behavior.