best : Beyond self-attention: How a small language model predicts the next token

Beyond self-attention: How a small language model predicts the next token
463 by tplrbv | 85 comments on Hacker News.