Large Open-Weight MoEs Are Converging on Lower Expert Activation Ratios(twitter.com)
twitter.com
Large Open-Weight MoEs Are Converging on Lower Expert Activation Ratios
https://twitter.com/ayushpai0/status/2077817066619351456
0 comments
—
https://twitter.com/ayushpai0/status/2077817066619351456
—