The link function provides the relationship between the linear predictor and the mean of the distribution function.
In the context of a generalized linear model (GLM)
$\operatorname {E} (\mathbf {Y} )={\boldsymbol {\mu }}=g^{-1}(\mathbf {X} {\boldsymbol {\beta }})$
log linear regression,
$\operatorname {log} (\mathbf {Y} ) = (\mathbf {X} {\boldsymbol {\beta }})$
the link function looks like
g(·) = log(·)
what does the link function of a normal distribution look like?
For a normal distribution, the link function is the identity link
$ g(\mathbf{\mu}) = \mathbf{\mu} = \mathbf{X}\mathbf{\beta}$.
Reference: https://newonlinecourses.science.psu.edu/stat501/node/378/