The PILE dataset is full of racist content and thus GPT-J produces racist thinking.
azeemh opened this issue · comments
azeemh commented
Any way to train on a different dataset?
Joseff Betancourt commented
I too have a question on how to train the dataset, specifically can we set GPT-J to learn from it's interactions?
havietisov commented
Model card mentions this :
https://huggingface.co/EleutherAI/gpt-j-6b#out-of-scope-use