various prompts from lmsys and wildchat that were filtered to be (theoretically) at least somewhat distinct from one another
also tried to remove all the prompts with slurs in them
this is not a good dataset and i do not recommend you use it (like other recent stuff this is just public for storage reasons)