I am working from page $8$ of this paper.
The generator for the coset representatives of $C^{\perp}$ in $C$ for the $[6,5,2]$ classical code is given by $$G_{C/C^{\perp}} = \begin{bmatrix} 1 & 1 & 0 & 0 & 0 & 0 \\ 1 & 0 & 1 & 0 & 0 & 0 \\ 1 & 0 & 0 & 1 & 0 & 0 \\ 1 & 0 & 0 & 0 & 1 & 0 \end{bmatrix}$$
The rows of this matrix, $\underline{g_{i}}, i \in 1,2,3,4$, generate all coset representatives for $C^{\perp}$ in $C$.
CSS construction provides a $[[6,4]]$ stabilizer code $Q$ spanned by the set of basis vectors {$| \psi_{x} \rangle | \underline{x} \in \mathbb{F}_{2}^{4}$}, where $\underline{x} = [x_{1}, x_{2}, x_{3}, x_{4}]$ and $$|\psi_{x} \rangle = \frac{1}{\sqrt{2}}| (000000) + \sum_{j=1}^{4}x_{j}\underline{g_{j}} \rangle + \frac{1}{\sqrt{2}}| (111111) + \sum_{j=1}^{4}x_{j}\underline{g_{j}} \rangle$$
However, I am very confused by this. If $\underline{x} \in \mathbb{F}_{2}^{4}$, then how are we multiplying it by $\underline{g_{j}}$, which has length $5$ (if it is referring to the rows of the generator of $C/C^{\perp}$, however, if it is referring to the columns then $\underline{g_{j}}$ has length $4$, but then it doesn't make sense to add the resulting vector $x_{j}\underline{g_{j}}$ of length $4$ to the vectors $(000000)$ and $(111111)$ of length $6$.