News for hackers. Stuff that matters.

A comment on Quadrillion-parameter transformers and how to train them. Parent comment

Re:Quadrillion-parameter transformers and how to train them by derac on Sunday November 08, 2026 @09:37AM (#50413734)
The data is massively conpressed and pruned. Where are the input tokens in an LLM?
[ Parent | on HN ]

There are no comments on this comment yet.