3 ms·
That post also very helpfully links to another paper they published alongside the OLMo paper just on the dataset. Dolma: an Open Corpus of Three Trillion Token
by dwagnerkc 3y ago
That post also very helpfully links to another paper they published alongside the OLMo paper just on the dataset.
Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research
https://arxiv.org/abs/2402.00159 https://arxiv.org/abs/2402.00159