The degree of $\mathbb{Q}(\alpha^8)$ over $\mathbb{Q}$ would be 2, because the minimum polynomial of $\alpha^8$ is $x^2-3$ which has degree 2. To compute $[\mathbb{Q}(\alpha^4):\mathbb{Q}(\alpha^8)]$ we first compute $[\mathbb{Q}(\alpha^4):\mathbb{Q}]$. This is 4 because the minimal polynomial is $x^4-3$. So $[\mathbb{Q}(\alpha^4):\mathbb{Q}(\alpha^8)]$ is 2, by tower law. Similarly, we get that the degrees of the other intermediate extensions are 2. To check that these types of polynomials are irreducible, we just use eisenstein's criterion with $p = 3$.
Now for the second part of the question I'm not completely sure. An intermediate field of $F$ must be generated by some $\alpha^n$ since if it doesn't contain any, then it can only be $\mathbb{Q}$. If $n$ does not divide 16, then the minimal polynomial has degree greater than 16(I think), which means it can't be a subfield. But how does this argument require using the constant term?
Source: https://dornsife.usc.edu/assets/sites/363/docs/F17_510ab.pdf

As mentioned in the comments the subfields aren't neccesarily of the form $\mathbb{Q}(\alpha^{\frac 1n})$. Nevertheless the easiest way would be to consider the Galois closure of the field and use Galois correspondence.
It's not hard to compute that $\mathbb{Q}(\alpha,\zeta_{16})$, where $\zeta_{16}$ is the principal 16-th root of unity, is the Galois closure of $\mathbb{Q}(\alpha)$. Also we have that $x^{16}-3$ is the minimal polynomial of $\alpha$ over $\mathbb{Q}(\zeta_{16})$. This follows as $3$ is a prime in the ring of integers $\mathbb{Z}[\zeta_{16}]$ and thus we can use the generalized Eisenstein Criterion. Now it's not hard to see that the Galois group is generated by:
$$\tau: \begin{array}{lr} \alpha \to \zeta_{16}\alpha\\ \zeta_{16} \to \zeta_{16} \end{array} \quad \quad \sigma: \begin{array}{lr} \alpha\to \alpha\\ \zeta_{16} \to \zeta_{16}^3 \end{array} \quad \quad \pi: \begin{array}{lr} \alpha\to \alpha\\ \zeta_{16} \to \zeta_{16}^7 \end{array}$$
In fact we have that the Galois group is isomorphic to the semidirect product $\mathbb{Z}_{16}\rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right)$. As $\{e\} \rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right)$ is the subgroup fixing $\mathbb{Q}(\alpha)$ you need to find all subgroups of $\mathbb{Z}_{16}\rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right)$ containing it. They are exactly:
$$\langle e \rangle \rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right), \langle 8 \rangle \rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right),\langle 4 \rangle \rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right),\langle 2 \rangle \rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right),\langle 1 \rangle \rtimes \left(\mathbb{Z}_2 \times \mathbb{Z}_4\right)$$
Thus by Galois correspondence $\mathbb{Q} \subset \mathbb{Q}(\alpha)$ has exactly $5$ intermediate subfields and they must be the one already mentioned.
Hence the proof.