I'm not well studied in this field, but I suspect that global average temperature by proxy for a single year is difficult to measure accurately. The 1991 cooling from Mt. Pinatubo happened recently enough that accurate measurements were made, though cooling is still somewhat difficult to pin down exactly due to other variation factors, sunspots being a small variation and the Pacific current being a big variation.
1990, at the time, was the hottest year on record, driven in part by the El Nino that year. (El Nino warms the Earth and usually lasts for a year, maybe 18 months or 2 years sometimes).
The El Nino continued into 1991, though it weakened greatly by the end of 1991 (Peak El Nino being about 12 months is about right). And the Mt. Pinatubo eruption happened on June 15, so it didn't affect climate in 1991 until nearly half way through the year. 1991 might have been warmer than 1990 prior to the eruption (I looked for a source on the first half of 1991 but couldn't find it).
Overall, 1991 was colder than 1990 but only a little, and probably would have been warmer without Pinatubo. 1992 was considerably colder than 1993 and 1994 showed gradual warming and 1995 set a new hottest year on record at the time. The fall 1990 to 1991 was likely volcano driven but the bigger fall in 1992 was probably partially driven by the weakening El Nino.
Chart used as source if you scroll down
The article above - see quote says that Pinatubo lowered temperatures less than your source suggests.
in the year to follow, volcanic particles in the atmosphere would
lower global temperatures by an average of 0.9 degrees Fahrenheit (0.5
But, something to keep in mind about a 0.5 degree drop is that oceans are very slow to react, and unlike land, oceans don't warm up well under direct sunlight due to evaporation and the cooling effect of photon-generated evaporation. Because Earth is about 70% covered by ocean nearly all that 0.5 degree variation is over the land, land surface temperatures drop by nearly 1.5 degrees C and Oceans change very little.
Finally, and this is important, volcanic "winters" aren't just about temperature drop but also about sunlight blocking. Less sunlight means less crop growth and it only takes one frost at the wrong time to kill a lot of crops, so even if a 1 or 2 degree Celsius drop or, lets say 2-3 degree drop over land that might not sound devastating, but temperature drops like that can be associated with a considerably higher chance of frost and crop-death. The 1991 eruption wasn't associated with crop failures, but the 1883 and 1815 eruptions were. Sunlight blocking may have been as big or an even bigger factor as temperature drop, but late frosts may have played a key role as well. (I lack the knowledge to give precise data on this).
I realize I've not answered your question with a number like you asked. You can find estimates of temperature drop in various articles, usually about 1-2 degrees from big volcanic eruptions like the one in 1815, but neither the 1883 nor the 1815 were accurately measured. From what I gather (but I haven't found it said specifically), the 1883 cold spell lead to greater interest in global temperature, which is why many global average temperature charts begin in 1884, after Krakatoa. We may not have an accurate estimate on the temperature drop from a super-volcano. Narrowing down an accurate estimate based on geological evidence for one year, even supported with some local temperatures and historical records is tricky. I don't think anyone knows the specifics of temperature drop by super-volcano with accuracy. The numbers that get reported are probably somewhere between guesses and estimates.
Also, not all eruptions are alike. A higher SO2 content should create a longer cooling. Volcanoes vary in sulfur content.