With real tubes, there are some leakage current around cutoff.
One reason for this is that some electron from a filament don't go through the control grid but go directly to the plate.
The other reason is that 300B's characteristic curves are not so correct.
In your plot, the curve for Eb=350 is more straight than that for Eb=300 and
the curve for Eb=450 is more straight than that for Eb=400.
Don't you think that this is strange?
My tube model is good at low Eb region but not so good at high Eb region or cutoff region.