Dec 1, 2023
Selective state-space models process sequences with an architecture that does not use Transformer attention.
A prominent alternative explores efficient long-sequence modeling and linear scaling in sequence length.
Explore on Humanity Unfolds
