What I have seen we have no standards in reporting ancient admixtures of recent populations. Commercial services give sometimes 70% Western Hunter Gatherer, sometimes 15%. Methods are poorly described, if at all. Obviously there are admixture overlaps and the results depend on how admixture tools identify them. For instance Yamnaya Samara overlaps with several hunter gatherer groups. I am sure that also European early farmers (LBK_EN) overlap with later Near East admixture in Europe and later Near Eastern admixture exists in Mediterranean. In other words admixture runs without Near Eastern samples assume EEF, but it doesn't give best fit. My test uses following qpAdm files:
Levant_Neolithic (sample I0867)
BedouinB (selected most distant samples from Europeans)
The results looks reasonable and qpAdm gave positive best coefficients without errors with only one exception, Swedes. I don't know why modern Swedes failed, because ancient Gotlanders made a perfect fit.