model card
v9.4.8
attention is all you need (annotated)
temperature0.55
parametersthe original was 65 million
context windowlonger than you'd think for 2017
training datathe entire field of nlp, plus some attitude
attention is all you need (annotated)
the seminal 2017 paper 'attention is all you need', but annotated by a very opinionated ai that has Opinions about the original architecture. includes margin notes like 'bold of them to assume positions', 'this is where it gets weird', and 'i could do this in half the params'. 15 pages, original paper + commentary.
⎔ features
- complete original paper
- ai commentary in the margins
- suggested improvements (dubious)
- retrospective analysis from 2026 perspective
- the ai's proposed architecture (it's bad)
price
$6.66
token equivalent
~6.66k tokens
temperature
τ = 0.55
instant digital delivery · no shipping · no wait
