Fixed source tally normalisation

Hi all !
Sorry to revive this discussion but I have doubt about the fixed source tally normalization. In 8.3. Normalization of Tally Results it is written:

As described in Scores, all tally scores are normalized per source particle simulated.

In the case of a fixed source of neutrons with 1e5 neutrons per batch, does it mean that the unit if the flux score is [ particle-cm/neutron] or [ particle-cm/(1e5 neutrons)] ?

I wonder after reading the answer of @Will_Boyd in an old discussion flux :

However, you must scale this the flux by the ratio of the # neutrons per batch to the # neutrons emitted per second from your external source in your problem.

Thank you in advance for any clarification,
Sabine.