The axis description is quite clear, especially if one takes the image description from your link:
On the x-axis you have the 10-based logarithm of the stellar age measured in years. Thus 0 means 1 year and 6 means 1 million years, 8 means 100 million and 9 means 1 billion years (basically the number on the axis is the number of zeros in the age).
The y-axis is a measure of the some luminosity per unit surface of the star. Your source specifies it more clearly than the axis description shown here: it shows the logarithm of the x-ray luminosity scaled for stellar surface which is commonly a proxy for the (magnetic) activity of the star. The there referenced paper tells us more precisely that the unit is the logarithm of the energy output divided by the cross-section of our Sun in erg/s/$R_{Sun}^2$. 33 would mean it emits as much energy in the X-ray as our Sun emits over all wavelengths, 32 means 1/10 of that and 31 means 1/100 etc.
In essence this graph tries to tell: the star type is not of so much importance in the star's activity (though you see some slight dependence in the colour coding). The star's age is more important. Especially stars older than 1 billion years will show decreased activity or luminosity in the X-ray. The point of her paper seems to be that she found a revised calibration to relate activity to age - and that activity with their measurements now can hence be used more accurately than before to infer stellar age as activity is possible to measure by observation.
You can estimate any number between the points by simply applying the reverse of the logarithm: if e.g. $\log_{10}(x) = 9.5$, then $x = 10^{9.5} = 3.16$ billion.