How can I see that for a scalar field
$$\phi(x)=e^{i\hat{p}\cdot x}\phi(0)e^{-i\hat{p}\cdot x}$$
if we have translation invariance?
How can I see that for a scalar field
$$\phi(x)=e^{i\hat{p}\cdot x}\phi(0)e^{-i\hat{p}\cdot x}$$
if we have translation invariance?
This is a straightforward application of Lagrange's shift operator, namely the formal rewriting of Taylor's expansion around 0. ($\phi$ being translationally invariant means its form does not change under translations--only its argument.) However, you must be very careful with the labelling of your double-duty variables, since the momentum operator acts to shift the implicit "x" variable of the field, starting from 0, by an amount x. To avoid confusion, and mistakes, we rename that shift to y. Work in one dimension, as extension to more dimensions is self-evident.
Now, when you are working in the x-representation, the arguments of the field, you have $\hat p=-i\hbar \partial_x$, hence Lagrange's/Taylor's shift is just $$ e^{y \cdot \partial_x }\phi(x)= \phi(x+y) e^{y \cdot \partial_x } ~~~\leadsto \\ e^{y \cdot \partial_x }\phi(x) e^{-y \cdot \partial_x } = \phi(x+ y) $$ for an arbitrary differentiable function $\phi(x)$.
Now, take x=0, $$ e^{iy \cdot \hat p/\hbar }\phi(0) e^{-iy \cdot \hat p /\hbar } = \phi( y),$$ and rename $ y\to x$, to get your dimensionally consistent expression. (Instead, you appear to have set $\hbar=1$.)