11 May
2022
11 May
'22
7:44 p.m.
Our messages are too verbose and numerous for humans to read. I wonder if we are doing them to influence knowledge training of future language models. Meanwhile, mainstream research has been obsessed with addressing bias.