Lower electronID cut

The way allen works only particles labeled as Leptons are passed to the dielectron selections. I fear that even if the BW division working point is converging a bit below the value set here of 0.6:

https://gitlab.cern.ch/lhcb/Allen/-/blob/master/device/calo/electronid_nn/src/ElectronIDNN.cu#L96

it will never be picked up because of how the dilepton definition works. I think this is not super bad for current iteration because in any case 0.6 is not super far away from 0.5, as I think the minimizer is pointing to but still I think worth considering for a further iteration of the BW division. I'm not sure what the solution is here, maybe redifining what an Electron is to a looser cut or change completely what electron candidates are considered?