In simple words: reducing the light emitted by the Nixies a bit.

Well ... why don't we reduce the anode voltage a bit then - that should do fine ? Yes. Should. But doesn't. Or at least not very good.

Fact is that the Nixies need a particular cathode current to entirely light up the cathode at all. If you just reduce the anode voltage the current through the anode resistor drops and so does the cathode current. Up to a certain point where the cathode is only partially covered with the discharge glow - or doesn't come on at all. Both is not what we wanted, right ?

So here is my approach that is a bit more complicated but seems to work fine. At least for my taste better than just reducing the anode voltage. Your individual mileage may vary however.

The circuit shown above is a simple oscillator with a 555 chip. The only uncommon part of the layout is the 1M Ohm Potentiometer and the two diodes. This part of the circuit allows to vary the duty-cycle of the output signal between 5 and 95 percent. The transistor in the upper part is wired into the high voltage supply line between power supply and the nixie anode resistors. The 555 is driving another transistor that simply cuts off the voltage on the base of that "main" transistor.

Result: the lower the duty cycle of the 555 output signal, the longer is the conductive period of the main transistor - during which the full current flows through the anode resistors. The trick is, that the high voltage is switched and not regulated.

The component values in the schematic allow a frequency variation between 3 - 8.5 ms (about 117 - 330 Hz) and the best results with a set of ZM1080s is at about 3.5 ms (285 Hz). Here the regulation of the duty cycle works from 5 to 95 %, where 95% is pretty dim, but still readable and the cathodes fully covered.

The transistors must be any of the high voltage types, since it is either attached to the HV-supply or has to sink the base voltage of the main transistor. For my prototype I picked a BU-407 as main transistor, because I had it at hand. You might come away with an even smaller one like the BF-458 or even one MPSA-42, if you don't have 6 x ZM1040 or bigger tubes running.

This circuit has its limitations of course, but it works far better than just reducing the anode voltage and current, which cannot be reduced below a certain point.
Unfortunately I cannot offer pictures from the scope and the Nixies at one time ...

Advantages:
You can use a simple potentiometer with all pins at low voltage levels to set the intensity. No need to pick something isolated or something of very uncommon value.

Disadvantages:

Apart from the "high complexity" you cannot use e.g. a photo-transistor for automatic regulation (dark outside = low intensity) or such. This will require some more parts. I thought about dual FETs and an inverter stage probably. I'm open for ideas here. (Grin !)

Known Problems so far:

If you have a "flashing dot" neon and reduced its intensity with a higher anode resistor already it may fail to operate below a certain dimming. Or flash erratically.