Conditioning of a Triangular Weight Matrix

A small linear layer uses the weight matrix

$W = \begin{bmatrix} 3 & 4 \\ 0 & 5 \end{bmatrix}$

Because $W$ is upper triangular, its eigenvalues can be read straight off the diagonal: $3$ and $5$.

Two standard facts you may use:

• The singular values of $W$ are the square roots of the eigenvalues of $W^{\mathsf{T}}W$, ordered $\sigma_1 \ge \sigma_2 \ge 0$.
• The best rank-$1$ approximation $W_1$ leaves Frobenius error $\|W - W_1\|_F = \sigma_2$, and the $2$-norm condition number is $\kappa_2(W) = \sigma_1/\sigma_2$.

Which pair of values is correct?