So my area of expertise, the three kingdoms' civil war of 190-280 CE, is often mentioned in such memes (that I have seen anyway) and only a 10th of China survived or worse death roll than world wars seemed to be an increasingly common belief. Sometimes another, the Yellow Turban rebellion of 184, pops up depending on how long the list is. I suspect a lot of the memes are based on someone (either directly or via a meme) seeing Wikipedia numbers.

I'm going to go with the Turbans first because it plays into the problem with more often seen three kingdoms numbers. Wikipedia says the Turban wars cost between three million and seven million lives, citing a non-academic internet article on worst wars and armed conflict that is now defunct with the link now going to a page on the curse of Kosovo instead. In the wiki article about the Yellow Turban revolt, it admits numbers are unknown while giving questionable numbers for the armed forces.

The Han forces are said to be 350,000 with no sources but when Huangfu Song and Zhu Jun combined their armies early in the campaign against the larger force of Bo Cai they could muster just over 40,000 soldiers. So unless Lu Zhi and some local forces had an absurd amount of troops, I'm going with unlikely. The Turbans are said to have 2,000 million (with an initial revolt under Zhang Jue of 360,000)which it claims to get from Paul Robb's "China in World History" but having used google preview, he doesn't seem to use the larger number. He does use the 360,000 under Zhang Jue which is from the Emperor Ling entry in the Book of Later Han.

While the infobox often refers to the 184 revolt, the years given in the war list and the article do cover till the destruction of Sima Ju and Xu He in the Turban stronghold of Qing in 206. However, it is unclear where the numbers of soldiers come from when extended to include the post 184 years. Even if one took those numbers blindly: two million and three hundred and sixty thousand soldiers who did not all die (even with the Han army's brutal policies towards the Turbans) going up to 3-7 million casualty war is hard to reconcile.

Not a historian but I don't recall any historian grouping the Turban revolts after 184 into one category. I understand why Wikipedia does it for their article but I would question the casualties as if being one war spread over a series of time. One considerable problem is who gets defined as a Turban? Some like Yang Feng and Ma Xiang seem to have tried it on for size, there is the possibility of wrong attribution by the records given the poor understanding of different groups. Which has led to uncertainty, with opinions swaying over time, over if the Five Pecks of Rice group of Hanzhong were ever connected to the Turbans.

We don't often know the numbers both of the "Turbans" and the forces that battled them, nor the casualties involved. When we do have numbers, there is always the danger of exaggeration with the Qingzhou Turbans as an example. This was a group that missed the 184 due to staying to farm due to the revolt having been earlier than planned then when the land collapsed into civil war, wandered about the north getting some heavy defeats before negotiating a deal to join Yan warlord Cao Cao. He is said to have gained three hundred thousand soldiers and seven hundred dependants but while the influx was sizeable enough to form a (not always reliable) core in his army and help populate his lands, that number is not entirely reliable either. Other sources give lesser numbers ranging from 100,000-300,000 for those joining Cao Cao. On top of that, having a quarter of Qing province move, plus still having such numbers after heavy defeats like the ones against Gongsun Zan (30,000 killed then in the next battle another tens of thousands killed plus 70'000 captured at Pan River) has indeed been questioned.

Now it isn't the say the Han forces, be they Han government or local warlords post 190 were not brutal, the Han commander Huangfu Song is said to have killed so many he could see the capital from Ye if he piled all the heads on top of each other. Surrenders were sometimes refused, the devastation of Turbans leaving their homes on mass and living off the land did not help authorities deal with famines. There were a lot of deaths but even if we took the numbers of deaths as straight, that the typical exaggerations didn't happen, there is no justification for those numbers.

What about the Three Kingdoms? The 36-40 million who died during the epidemics, the famines, the slaughters between various factions. It was a dreadful time as the Han collapsed for nearly a century of civil war and it left China much weaker. But 36 to 40 million dead?

Wikipedia puts the war at a too-early date, 184 is a narrative starting point for fiction rather than the historical starting point of the civil war (190) and memes that include both wars ignore some of those "40 million" who have to have died before 190. But wiki does make a caveat about the three kingdoms' figures "Also, note that the death range provided is actually the amount the population declined according to the census data and is likely an overestimation of actual combat fatalities. "

Depending on how the meme phrases it, that should be a red flag. 36-40 million were not killed in battle, they were not troops who were killed in the field. Epidemics that had been going on since the late 160s did not suddenly stop in 190, the collapse of governance led to widespread famine for a time and we have plenty of people who notably did not die in battle.

Now if the meme is careful and more talking of a general death toll, it might be a little high but still, war's impact on other things led to a general death toll of extraordinary proportions then the meme is still wrong. Wiki's caveat doesn't cover the extent of the problem with those figures.

The census for the before figures are not from the 180s or the 170s or even the 160s. The census from around 140 is often used and that puts the total population just under fifty million but that is several decades before the civil war, decades which saw natural disasters, war, loss of ground in the north, let alone the wave after wave of what was probably the Antonine Plague. This might have altered the population statistics downwards in the decades since before the civil war.

The problem with the "after" figures shortly after unity, which suggest a population of only sixteen million, is they weren't a census, they were tax records. It might be the best we have but it was not like for like. While the civil war was no doubt devastating to China's populace, the decline of the Later Han, the war itself and the manner of Jin's seizure of power had left a weakened central state. While only being able to have a tax and levy base of sixteen million people was a major problem for Jin, it doesn't tell us how many people were alive at that time. Treating it as an equivalent figure so we can draw "this is how many people died during the bloody civil war" means losing the context of Jin's statistics and what it tells us about government authority. Perhaps less exciting for some than "they all died" and big numbered decimation but the loss of authority and power at the centre was an important part of the era that gets missed by the memes.

I would suggest if they were so clumsy with the three kingdom numbers in the memes about the bloody civil wars, the other numbers may not have also been carefully scrutinised for accuracy or for putting into correct context.