You are correct that the allocation factor column from the MABLE/Geocorr12 Results represents the proportion of the county in the PUMA. It should be noted that while PUMAs were drawn to match well with county boundaries there are some counties that cannot be identified perfectly
You may find the 2000-2010 PUMA Crosswalk spreadsheet (available here) useful for identifying counties to PUMAs. The 2010 PUMAs (used by the ACS starting with the 2012 sample) were named based on contained counties or other major geographic entities and those names are listed in the crosswalk. These names are also present in the MABLE/Geocorr12 Results. There is also an interactive map (as well as a larger map) available on the IPUMS-USA geography page that you can use to visualize how counties fit inside the two PUMA boundary definitions.
I hope this helps.