This model was discussed recently: https://news.ycombinator.com/item?id=38522428 https://news.ycombinator.com/item?id=38522428 It's a new kind of ML model archi

This model was discussed recently: https://news.ycombinator.com/item?id=38522428 https://news.ycombinator.com/item?id=38522428 It's a new kind of ML model architecture that can be used instead of a transformer in LLMs. See also the original repo from the paper: https://github.com/state-spaces/mamba https://github.com/state-spaces/mamba