100%

Finished Base Model!

0.75

75% checkpoint

50% release

Little more coherent, some decent paragraphs.

Pretrain release

The model had only trained on about 5B tokens at this point, not very close to done, but I thought it would be fun to release