Where a concept exists (and has the same reasonably well-defined meaning) across multiple languages, as in this case, it is good and proper that it have a corresponding language-neutral tag. This is not ambiguity.
Dataframes are a generic statistical-data-processing concept and exist in R, Python/pandas, Julia, Excel, also STATA, SAS, SPSS ('dataset') and I'm sure other current and future languages.
>Should it be deleted completely as a meta tag?
No, there's nothing 'meta' about it. It's concrete.
>Should this tag be edited to include the fact it is also the name of the pandas data structure?
No, because that isn't where the name or concept came from. Wes McKinney named pandas concepts after the R concepts they derive from. And R came from S. And so on. But the common constructs in languages are just a symptom of common terminology and methodology for data processing.
>Seprated dataframe-r and dataframe-pandas
No, it's terrible practice to split tags referring to the same concept by language - unless there's something different or problematic about it in one language - which there isn't.
Having said that we seem to currently have dataframes for dataframes in general, and data.frame for R data.frame in particular, which I think is arguably undesirable and possibly a merge candidate - unless a large portion of the R data.frame stuff is things not generically useful to other languages - I haven't waded through a breakdown of it.
Anyway, people can and should tag by both language/library and concept: pandas dataframes or r data.frame or 'tag:some-other-language' dataframes
>Something else / nothing
Yes, leave it as it is. Only improve its wiki definition, if you can.
By the way, the Wikipedia definition of dataframe is bad and needs radical surgery.