How does GPT-3 spend its 175B parameters? - LessWrong - discu.eu

Linking pages

Training Deep Networks with Data Parallelism in Jax https://www.mishalaskin.com/posts/data_parallel 37 comments
Knowing Enough About MoE to Explain Dropped Tokens in GPT-4 - 152334H https://152334h.github.io/blog/knowing-enough-about-moe/ 1 comment

Related searches:

Search whole site: site:www.lesswrong.com

Search title: How does GPT-3 spend its 175B parameters? - LessWrong

See how to search.

Submit link to: