Top all-time · #24855
Amirhossein Kazemnejad
Personal blog · kazemnejad.com/blogaiprogramming
Amirhossein Kazemnejad's personal blog at kazemnejad.com covers ai and programming. Since 2023, 3 different Hacker News readers have linked it 3 times, most often “Transformer Architecture: The Positional Encoding”. The last post was in 2020, but the archive is still widely read.
Too few sampled comments (1) to say how often HN recommends it.?
Why HN links it
HN readers link this post to explain transformer positional encoding, citing its list of criteria a positional encoding should satisfy. The one comment cites it as a source and does not judge it.
Summarised by Claude from a sample of 1 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
I'm an M.Sc. student at Mila and McGill University, and currently, I work at RL Lab as a research assistant under the supervision of Siva Reddy. My interest ...
Most-linked posts?
- Transformer Architecture: The Positional Encoding3 people · 3 mentions · first linked 2023-04-26
What Hacker News says?
According to https://kazemnejad.com/blog/transformer_architecture_positio... they explain it as a clever way to satisfy the following criteria: - It should output a unique encoding for each time-step (word’s position in a sentence) - Distance between any two time-steps should be consistent across sentences with different lengths. - Our model should generalize to longer sentences without any efforts. Its values shoul…
Comment linking /blog/transformer_architecture_positional_encoding/
on “Transformers from Scratch”Submitted Transformer Architecture: The Positional Encoding as a story · 2 points
HN discussionSubmitted The Positional Encoding (2019) as a story · 1 points
HN discussion