Updated ITk ambi scoring tool to use 5 GeV cluster cut
With !52633 (merged), the ITk config for the AmbiScoringTool (unintentionally?) diverged from the Run-3 one, using a 0 GeV cut for the EM RoI clusters. The 21.9 implementation of the InDetAmbiScoringTool.cxx used instead a 5 GeV cut, consistent with the current Run-3 configuration InDetTrackScoringToolsConfig.py. This MR restores this behaviour. Run-4 outputs are expected to change. Might be related to https://its.cern.ch/jira/browse/ATLIDTRKCP-428