I would expect an anchor to be permanently deformed from a trial test with its full intended system design load applied (IIRC that is acceptable as long as there is not anchor failure under BS 7883), but not for a proof test of the service load (anticipated real world use load). 6kN is a pretty standard periodical proof test load for industry, and any anchor made for that use should not be affected by that kind of force.
It is worth bearing in mind that peak forces seen by breaking ropes vary from 6 to 16 kN. The standards usually link back to a value of 6 kN based on an estimate / guess that that is what most people will walk away from. (Though not if you already have a 'bad back'.) Parachutes are reckon to open with a force of around 12 kN but it should be remembered that the parachute harness is a hell of a lot better at spreading the load into the whole of the body than most SRT harnesses that cavers use. The late Paul Seddon wrote a decent report on the topic entitled "Harness suspension: review and evaluation of existing information" in 2002. It looks like the HSE web site has lost the page so if anyone wants a copy, PM me.
I have not specifically checked if IC anchors flex at 6 kN but some FMEA calculations I had done indicated they would. My memory is that this flexing in BP anchors becomes a permanent set at around 10 kN, that is the head is left permanently distorted. (Though I may not have saved the evidence of this.)
For the uninitiated, EN 795 is for anchors being used with PPE and BS 7883 is the UK code of practice on them. The anchors used in caves under the BCA set up, conform to EN 959 which is for mountaineering anchors. A key feature of EN 795 anchors is that they are removable where as EN 959 are placed by gluing into the rock and thus are not. Further complicating matters EN 795 anchors fall into various types but EN 795 Type A anchors (which are similar to EN 895 anchors) use fixing methods to attach them to part of the structure (such as a cast in threaded sleeve). EN 795 Type anchors are required to hold a 12 kN static axial load. BS 7883 goes onto propose that each EN 795 Type A anchor after installation should be subject to a 6 kN axial load. Where as EN 895 anchors are only required to hold a 15 kN axial load and no comment is made about testing installed anchors. Hence the push to adopt routine pull testing.