1 comment

[ 0.21 ms ] story [ 24.7 ms ] thread
I really want to see a cleaned 10T token dataset for LLMs