looking like noise, not oscillation. does it disappear when you shorten the input? (this could be just the white noise from the input resistance, amplified)
Also, if I remember correctly, the input bias should be adjusted without input connection (or maybe 10k to ground), so that even a small current gives a meaningful reading. Then the input is shortened and the servo should do its magic. The DC resistance of a MC cartridge is close enough to zero to do the same.
the gain of a phono stage is highest at lowest frequencies, and the paradise is no exception. so even a small input bias current offset may cause the servo to no longer be able to compensate, as it is designed to be relatively "weak". R43a,b lower values will make it stronger. And, if you have a MC cartridge connected and look at the output DC voltage, then pull the input cable out and see a jump in output voltage, it is an indication that the input bias needs adjusting.
just my two cents....