This is what we need more of.

Harvard is publishing a collection of around 1m books, which is 394m pages and 242bn tokens. Soon all your base LLMs will be pre-trained on it.

The beautiful and enlightened future is incoming.

This is what we need more of.

Harvard is publishing a collection of around 1m books, which is 394m pages and 242bn tokens. Soon all your base LLMs will be pre-trained on it.

The beautiful and enlightened future is incoming.

https://apnews.com/article/ai-chatbot-training-data-libraries-idi-e096a81a4fceb2951f232a33ac767f53