Often reference voltages are generated using normal LED/diode, not zener. A reference from simple resistor divider will not provide any regulation of input voltage, a zener will provide regulation and a cleaner reference... so how does the LED/diode reference work and why is it used instead of those?
This is a Current vs Voltage plot of a typical diode:
Note that when the current is high enough the mode of operation will be "Forward mode". In that mode the current can change significantly but the voltage (the Knee voltage) stays more or less constant. For a typical silicon diode this voltage can be between 0.5 V and 1 V depending on the model of diode and the forward current.
The behavior of an LED is very similar but the voltage is different and can be as low as 1.5 V (IR LED) or as high as 4 V (Blue LED). The value of the Knee voltage of an LED depends on the materials it is made of and that is determined by what colour of light is emits.
That behavior is similar to how zener diodes behave. The voltage stays more or less constant over a wide range of current. Zener diodes operate in the region called "Break down" which is "Reverse mode" btw.
Compare that to a resistor which does not work as a voltage reference. When you increase the current through a resistor by for example 10% then you get a 10% higher voltage across the resistor.
For diodes, LEDs and zeners you would get for example only 1% (just a random number I made up) more voltage for the same 10% current increase. For an ideal voltage reference component (these do not exist, some ICs come close though), the voltage would not change at all.
The forward voltage of an LED changes only a little when its current or its temperature changes. BUT an LED is not rated for its voltage which is a range of voltages. A white LED might be 2.8V to 3.6V or any voltage in between. Buy thousands of LEDs and test them all. You might find one that has the voltage you need.