Query/Key/Value
Core terms in the Transformer attention mechanism. Query, Key, and Value are all vectors obtained by applying different linear transformations to the input embedding; attention weights are computed from the dot product of Query and Key passed through softmax, and the output is a weighted sum of Value.
Core Facts
Timeline (last 90 days)
Query/Key/Value are standard terms for the Transformer attention mechanism, originally derived from an information-retrieval analogy.
All Facts (2)
The self-attention mechanism maps each token in the input sequence to three vectors—Query, Key, and Value—and computes attention weights by taking the dot product of Query and Key followed by Softmax normalization.
80%UnverifiedQuery/Key/Value are standard terms for the Transformer attention mechanism, originally derived from an information-retrieval analogy.
50%