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)