The IR filter is located behind the optical feedback and its mount(circular front of the laser, held in place by three bolts, with two wires comming out of it). and is just glued on the housing.
In your first picture you shifted the OF to the side, now there is another plate that you must remove to get to the filter. (two more bolts)
Removing this unit will reveal the IR filter, for some reason mine was dirty and upon removing the filter all together the output cleared itself up to a nicer profile. (before doing that my "dot" looked like some firework effect)
keeping the optical feedback off, will result in no regulation, which seems to make the laser output about 2X its running power (although more unstable, ive gotten ~30mw of mine and nothing failed.. so try running without it at your own risk)... but it can always be bolted back on after removal if you so desire.
Very good point, IIRC Dont lase me bro figured that out originally, but thats what I did to my unit also to correct the output after the profle was cleaned up by ridding of the dirty filter. (divergance wise, there is no collimators of any kind after the crystals)
http://laserpointerforums.com/f48/473nm-lab-pictures-65908.html
I got some pictures in that thread with the original output, and that after the removal of the filter and addition of a collimator.