reducing the size of large datasets

[KennethEnevoldsen]

If it is possible to reduce the size of some datasets without changing the performance too much it would be great to ensure that the benchmark runs faster.

I am especially thinking of ScaLA, Da Political comments, as well as massive intent and massive scenario.

@x-tabdeveloping would you think this is reasonable as well?

[x-tabdeveloping]

Hmm yeah it would be nice if we could make it faster somehow, especially if we're planning on bootstrapping stuff, then it's a really good idea.

[KennethEnevoldsen]

Well only bootstrapping the evaluation (not the encoding) - but agree

[KennethEnevoldsen]

Notably better at the fixes in #130 (doesn't actually change the size)