For positive integers $m$ and $n$, let $D_m(n)$ denote the determinant $\det[(j-k)^m+\delta_{jk}]_{1\le j,k\le n}$, where the Kronecker delta $\delta_{jk}$ is $1$ or $0$ according as $j=k$ or not.
Recently, joint with my graduate student Han Wang, we proved in arXiv:2206.12317 the identity $$D_1(n)=1+\frac{n^2(n^2-1)}{12}\tag{1}$$ via eigenvalues. (See also A079034 at OEIS.) Based on my numerical computations, I conjectured that $$\begin{align}D_2(n)&=1+\frac{n^2(n^2-1)}{1080}(n^5-5n^3-36n^2+4n+54) \\&=\frac{(n^2-4)(n^2+2n+3)(n^5-2n^4-n^3-28n^2+60n-90)}{1080}\end{align}\tag{2}$$ and $$\begin{align}D_3(n)=&1+\frac{n^2(n^2-1)}{672000} \\&\times(n^{12}-19n^{10}+123n^8-337n^6+12376n^4-44144n^2+40000)\end{align}\tag{3}$$ (cf. A355175 and A355326 at OEIS). Note that the equation $$x^5-2x^4-x^3-28x^2+60x-90=0$$ over $\mathbb Q$ is not solvable by radicals.
In view of $(1)-(3)$, I have the following general conjecture.
Conjecture. Let $m$ be any positive integer. Then $D_m(n)$ has the form $1+n^2(n^2-1)P_m(n)$, where $P_m(n)$ is a polynomial in $n$ with rational number coefficients whose degree is $(m+1)^2-4$.
QUESTION. How to prove the formulas $(2)$ and $(3)$? Is the above general conjecture true?
Your comments are welcome!