Turbidity currents generated from sediment-carrying freshwater discharges into the sea contain a fluid that is less dense than ambient seawater. From experiments it is known that such currents will eventually lift from their substrate either in part or as a whole through buoyancy reversal. This ascent will happen when their density is lowered below that of seawater through settling of suspended sediment from the top or deposition from the bottom of the flows. Evidence for large-scale lofting of suspended sediment from the top of giant sand- and gravel-carrying turbidity currents in the Labrador Sea comes from two independent lines of observations: (1) The first is a distinct sedimentary facies consisting of stacked, centimeter-thick graded mud layers that contain grains of ice-rafted debris (IRD) supported by the mud. Deposition of these unusual layers requires a graded- layer–forming process that is slow enough to allow the incorporation of IRD; this is not possible with normal mud-carrying turbidity currents. (2) The second observation is the presence of a huge abyssal sand and gravel plain in the central Labrador Sea that received its sediment from bed-load–rich meltwater discharges from the Hudson Strait outlet of the Pleistocene Laurentide Ice Sheet. These discharges turned into turbidity currents that released rising columns of freshwater that carried fine-grained suspended sediment and spread out at a water level where their density equaled that of ambient seawater. Deposition from these slow turbid interflows would allow the incorporation of IRD in the accumulating graded mud deposits. The IRD-spiked graded mud facies is restricted to Heinrich layers within 300 km radius of the Hudson Strait ice-stream terminus, tying the sand-carrying turbidity currents via fine-grained sediment lofting to Heinrich events. Estimated total discharge volumes of individual currents are on the order of 103 km3, supporting the notion that Heinrich ice-rafting events were times of maximum meltwater generation.