Practice
Learn
← Quit
0 / 0 correct
LLMs & AI
Medium
1 of 20
What does the feed-forward (MLP) sub-layer in a transformer block do?
A
Processes each token's vector independently with two linear layers and a nonlinearity, holding much of the model's knowledge
B
Mixes information between tokens
C
Tokenizes text
D
Computes the loss