What does a larger vocabulary enable in Llama 3? #llama3

The smallest Llama 3 model increased to 8B parameters vs the 7B parameters with Llama 2. Aston Zhang, author of D2L and researcher on the Llama team at Meta shared a little bit more on one of the factors that led to this and how it's enabling stronger performance on some tasks with Llama 3.

--

Subscribe: youtube.com/facebookai?sub_confirmation=1

Learn more about our work: ai.meta.com/

Follow us on Twitter: twitter.com/aiatmeta
Follow us on Facebook: www.facebook.com/aiatmeta
Connect with us on LinkedIn: www.linkedin.com/showcase/aiatmeta/

Meta focuses on bringing the world together by advancing AI, powering meaningful and safe experiences, and conducting open research.