I've never heard of liquid-liquid extractions which use DMF (or DMSO) instead of water for the polar layer. I tried it out using hexane as the organic layer and it worked to an extent. The hexane layer changed colour, so something went in there. The DMF layer stayed green. When I added water, everything went into the hexane. Since DMF dissolves all kinds of things, seems like this might be a good way to selectively extract the less polar compounds out of the DMF. Then you could add brine, and do a water-chloroform extraction to get the other compounds out.