IPCC thinks that “tipping points” leading to positive feedback loops are highly unlikely.
In general, you should rather expect negative feedback loops. For example, the warming response to increased CO2 emission is logarithmic, meaning that the more CO2 you emit, the weaker effect additional effect has, and quite significantly so. Of course, our problem is that we are emitting so much, and our emissions grow so fast, that this diminishing effect is of little consolation. The point is, negative feedback loops are much more likely than positive ones.
There are some pretty clearly known positive feedback loops as well - mainly that ice has much higher albedo than water, and that ice covered areas that today are net emitters of heat will become net absorbers if they lose more ice. Specifically Greenland is known to already have become a net absorber, meaning that there is no way nkw to prevent the eventual permanent melting of all of Greenland's ice.
Those are already accounted for in the 2.7C etc estimates, though. The CO2 forcing just accounts for 0.5C IIRC of the warming, with known positive feedbacks (most notably water vapor, but also albedo etc) driving the rest.
The question becomes, are unknown feedbacks more likely to be positive or negative? I'd lean toward negative, which is what most experts think, but definitely an unknown.