Lucene 9.9: Benchmark sparse improvements

[lintool]

Follow up to this: #2288 (comment)

It'd be great to re-benchmark the "Wacky Weights" paper: https://arxiv.org/abs/2110.11540

It's actually quite simple, we've gotten it down to a single command:

python src/main/python/run_regression.py --download --index --verify --search --regression msmarco-passage-splade-pp-ed

See: https://github.com/castorini/anserini/blob/master/docs/regressions/regressions-msmarco-passage-splade-pp-ed.md

@jpountz @benwtrent @ChrisHegarty Do either of you have cycles to try it out?

[lintool]

No further follow-up, closing.