Correct number of parameters

README.md

@@ -32,7 +32,7 @@ language:
 
 On April 10th, [@MistralAI](https://huggingface.co/mistralai) released a model named "Mixtral 8x22B," an 176B MoE via magnet link (torrent):
 
-- 176B MoE with ~40B active
+- 141B MoE with ~35B active
 - Context length of 65k tokens
 - The base model can be fine-tuned
 - Requires ~260GB VRAM in fp16, 73GB in int4