The cockpit crew consisted of 4 members: two pilots, navigator and engineer. The airforce chief watched in the back of the cockpit. The navigator had read height aloud. The numbers he called (recorded by the voice recorder) match radio height recoded by the FDR. The baro altitude recorded by the FDR has too coarse granularity (max 256 possible digital values), but we know that the terrain under the glideslope is seriously uneven (hills and ravines), unlike more flat terrain Poles used to at home. Therefore, the numbers the navigator called couldn't be from baro, no matter what pressure might be set (everybody agrees that most baro altimeters had correct pressure set, just one of several PIC's baro altimeters - the one connected to TAWS - was set by PIC with a quick button press to the standard pressure much higher than QFE 4.5 seconds after the first TAWS alarm in order to shut up annoying TAWS alarms, because the PIC knew that false TAWS alarms were guaranteed at that airfield because the airport was absent from TAWS database). Baro altimeters were set to QFE, i.e. showed height relative to runway elevation (Soviet SOP, still used in Russia). In the last ravine baro height was negative (the runway is on a hill). So, we can be sure that unfortunately nobody looked at the baro altimeters (except the airforce chief earlier, but he said nothing in the ravine). The navigator called the radio height since height 300 meters, perhaps as he used to in the Yak-40 aircraft (smaller and able to climb steeper). I'm not sure whether to call radio height since 300 meters is described as SOP in the Yak-40 manual, but it's definitely not so in the Tu-154M manual.
The pilots (and their commanders) punished themselves with the capital punishment.
I think that they didn't press the (unarmed) automatic go-around button because the PIC intended to land and intentionally descended under the glideslope ("scud running"). He pulled the yoke when he at last unexpectedly saw trees ahead in the thick fog - too late because of too steep ascending slope of the last ravine.