According to the article in Athletics Coach (spring 1995, by Norman Poole) , Cram ran 10-12 x 200m (30sec recovery) in June in 26-27sec (~800m pace). Later this might become (in a racing week, though not tapering)1x600m (1.17) and 6x200m(30sec rec)at a faster pace of 25-26sec, or 8x200m (60sec rec) in 25-26. Like Coe the last few were accelerated towards the end - which could mean tempo-change or faster (23-24). As claimed for Coe, these guys were fit enough to run very close (95%+) to their best 200m even at the end of these sets. That is, in part, what separates these guys from sprinters.
If we look at Coe (21.7 claimed 200m PR) and his legal best 400m (46.87) and relay split (45.5), that is a relationship of 2.10-2.15x 200PR for relay/open (45.6-46.7). Then for Cram, if we give him a best of 22.5 (his lack of relative speed may be due not so much to ability as that he didn't seem to do the specific weight-training that we know Coe did under George Gandy)that equates to a range of 47.3-48.4 using the same multiplier. While it is possible that Cram was faster than that, nothing indictates that this was in fact so, and we know that he trained more for speed-endurance than speed, anyway.
He did do sessions like 8-10x150 (250 jog recovery), which was a lot of recovery for him (apparently). These might have been equally progressive, starting at 800m GP (about 19sec) and going down to 17sec (or about 97% of his presumed PR of 16.5sec for 150m, based on a 200m PR of 22.5).