Learn Before
The scaled dot-product attention formula, Softmax((QK^T / sqrt(d)) + Mask)V, is used when an entire input sequence is available for simultaneous processing. Which specific operation within this formula directly represents the parallel computation of interaction scores between every possible pair of tokens in the sequence, a step that is only feasible because the entire input is present at once?
0
1
Tags
Ch.5 Inference - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Analysis in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
The scaled dot-product attention formula,
Softmax((QK^T / sqrt(d)) + Mask)V, is used when an entire input sequence is available for simultaneous processing. Which specific operation within this formula directly represents the parallel computation of interaction scores between every possible pair of tokens in the sequence, a step that is only feasible because the entire input is present at once?Optimizing Prefilling Phase Performance
Consequences of Removing the Causal Mask