Dataset
#2
by
mdegans
- opened
How exactly was the dataset filtered? Are you just removing "As an AI Language Model" or also words like "LGBT", "feminism", etc. as Eric Hartford is prone to do?
Edit: the answer is (via discord) that Eric is off the team. The dataset is filtered for refusals, but does not target minorites, unlike Eric Hartford.
imone
changed discussion status to
closed