Pangram Labs' _ueaj posts a parody paper mocking machine learning research that claims to simplify or generalize attention mechanisms

From: Pangram Labs' _ueaj posts a parody paper mocking machine learning research that claims to simplify or generalize attention mechanisms

ueaj@_ueaj·13hOriginal post

"Attention is just a special case of <abstract math thing> so we generalized it by <neglecting the other 30 abstractions and conditions required for frontier architecture> and we found it performed <p hacking> compared to <naive baseline>"

View on

Pangram Labs' _ueaj posts a parody paper mocking machine learning research that claims to simplify or generalize attention mechanisms · Digg