EDIT: I'm not sure I realized it was 50 points we were talking about here, or just let my attention wander for a bit! Suspended solids can make a difference (see the comments), but that's notyou'd almost have to be measuring the case here.SG of slurry for it to make that much of a difference!
If the original recipe called for 2 cans of extract, and you used 3 then that's just under a 50% increase (the recipe also included grains.). 1.070 plus an extra 50% is 1.105, which puts you in the ball park for your reading of 1.120. Depending upon how much suspended sediment was in the jar, an additional 10-15 points is not improbable.
PS: 1.12 is not the same as 1.120 - the additional 0 shows you have measured it to 1/1000 precision rather than 1/100 precision.
My original answer follows.
It's most likely the sediment that caused the high reading, if the hydrometer is otherwise calibrated and you have corrected for temperature. It's widely believed that the hydrometer is not affected by suspended sediment - but in fact it is:
Hydrometer analysis begins after thoroughly mixing the sediment and water, after which particles settle out of the water column according to Stokes’s law. The density of a sediment-water suspension depends on the concentration and specific gravity of the sediments present in the mixture. If the suspension is allowed to stand, particles will settle out of the suspension and the density of the sediment-water suspension will decrease.
Although this is discussing drilling, the same principles apply. While the solids are in suspension, the density of the liquid/solid mixture is the average of the liquid and solid densities. When the solids fall out of suspension, then it's just the liquid density that is measured.
It's not possible to determine what the original gravity was without knowing the average density of the sediment in the sample jar.
Just chalk this one up to experience - no real harm done. Next time leave the gravity sample to settle and also to settle to room temperature, so the hydrometer reading is accurate.