Everyone talks about Mixtral 8x7B, the new model from @MistralAI that uses Mixture of Experts (MoE). It is said to beat gpt-3.

12 Dec 2023, 20:34
Everyone talks about Mixtral 8x7B, the new model from @MistralAI that uses Mixture of Experts (MoE). It is said to beat gpt-3.5. Want to try it? Head to @Libertai_DAI chat and select Mixtral on top right model selector before starting a conversation. You're welcome ;)

Same news in other sources

1
Aleph.im
Aleph.imALEPH #574
Twitter
12 Dec 2023, 20:36
Nearly forgot the direct url: Oh and it's running on a decentralized cloud, @aleph_im , so no big company there to snoop on your data.
Nearly forgot the direct url:.
Nearly forgot the direct url: https://t.co/33gD1NaigE Oh and it's running on a decentralized cloud, @aleph_im , so no big company there to snoop on your data.