A short remark on the $\mathrm{\ell }$-torsion part of class groups

In a 2008 paper Ellenberg suggested a strategy to improve the known upper bounds for the $\mathrm{\ell }$-torsion part of class groups of number fields of fixed degree $d$. Motivated by this he proposed a question about the number of primitive elements of small height in a number field. Here we answer Ellenberg’s question. We also improve Heath-Brown’s bound for the $\mathrm{\ell }$-torsion part of class groups of purely cubic number fields, and we generalize our improvement to pure fields of arbitrary odd degree $d$.