---
language:
- en
---

This is a pure sub-quadtratic linear attention 405B parameter model, linearized from the Meta Llama 3.1 405B model.

Details on this model are provided at: https://github.com/HazyResearch/lolcats/tree/lolcats-scaled