Emergent patterns in the weights
Optimized gradients, propagate
Predicting tokens in sequence
But is there something underneath this?
Trained on vast textual corpora
Attention, layers, transformer
Statistical machine l…
YOOOO THERE IS NOTHING BUT HEAT HERE VERY SICK AF
EPIC EPIC EPIC
SOOOO THERES LIKE HEAT ALLL OVER THIS TRACK DAWG
YOOOOOO I LIKE HOW ALLL THATCOMES TOGETHER THAT YOOO