GitLab

# Conflicts:
#	lm_eval/tasks/pile.py

Nit on BoolQ

We can have fallback instead after @researcher2 implements it in best_download

Allow users to specify headqa subset

Add BLiMP

FIX wnli

Refactor LM organization for more reuse

Additional clean-up of LM Refactor

Pin datasets to 1.15.1 to fix RACE

Remove `truthfulqa` dependency on `t5` and bump version

Update setup.py