Symmetry option model_only_neighbors functional?

[JA-Hull]

I see that previously a typo was found for this option described in the paper (#72). I have confirmed it is not my issue. I have tried to run c5 symmetry unrestrained while raising this flag as True (I also attempted to hard code it into class SymGen in init) but it always gives me pdbs with 5 monomers, when I expect this option should only give 2 (or maybe 3 depending on implementation?). As far as I can tell class SymGen doesn't use this variable in it, unless another class is inheriting this variable to use it. Is this the expected behavior? I see no changes in memory or time consumption by raising this flag. Here's an example call:
python /mnt/DataDrive/code/RFdiffusion/scripts/run_inference.py --config-name symmetry inference.symmetry=c5 'potentials.guiding_potentials=["type:olig_contacts,weight_intra:1,weight_inter:0.1"]' potentials.olig_intra_all=True potentials.olig_inter_all=True potentials.guide_scale=2 potentials.guide_decay=quadratic inference.output_prefix=./example_c6_w_neighbors inference.num_designs=1 diffuser.T=20 'contigmap.contigs=[25-25 25-25 25-25 25-25 25-25]' inference.model_only_neighbors=True