WingNews logo WingNews
top | new | best | ask | show | jobs
top | item 46455302

The Bayesian Geometry of Transformer Attention

4 points| samwillis | 1 month ago |arxiv.org

1 comment

order

samwillis|1 month ago

Higher level overview and links to the other related papers: https://medium.com/@vishalmisra/attention-is-bayesian-infere...
powered by hn/api // news.ycombinator.com