Softmax
Softmax maps a vector of real scores to positive components that sum to one. Subtracting the maximum score before exponentiation preserves the result while improving numerical stability.
\operatorname{softmax}(z_i)=\frac{e^{z_i}}{\sum_j e^{z_j}}
Overview. Softmax maps a vector of real scores to positive components that sum to one.
Context. Subtracting the maximum score before exponentiation preserves the result while improving numerical stability.
Claims and evidence
Claim ledger, 2Softmax maps a vector of real scores to positive components that sum to one.
Subtracting the maximum score before exponentiation preserves the result while improving numerical stability.