I have a large subject-based TM consisting of about 254k segments which occupies 1GB. Lately it has been slowing down, so I upgraded it again today. 

I am using the default settings and I can have more than one translation for the same source segment, with the Merge option enabled. This is okay as it accounts for stylistic preferences (like the use of the capitalized or lower case formal 'you', or of some specific terminology for a given client). However, the translation of some repetitive segments, such as those from disclaimers which I translate at least three times per week, have multiplied over the years as I improved my translation.

Now I would like to tidy my TM keeping a reasonable number of options (3-4) for the same source segments. Is there a way to do so?

For example, if I could apply a filter so that I can see my TM contents sorted by number of occurrences in the TM, i.e. if a segments has 10 possible translations (10 TUs with the same source), it should be displayed before a segment which has only 1 translation (1 TU). I do not pretend to review all 254k segments, just the most popular ones.

Also, similar segments (within a certain margin of fuzziness, like 95%) because maybe there is just a tag of difference or a typo in the source and I don't want to miss those, either.

Any other tip is welcome, also. Thank you in advance.