The SENSITIVITY LABELS: section specifies the words that make up a human-readable representation of a sensitivity label, as well as the required combinations and combination constraints on these words. This section is used by the system to convert a human-readable representation of non-classification sensitivity label words into the internal bit-string form, and to translate the internal form to a human-readable representation.
In all cases, the SENSITIVITY LABELS: section must associate words with exactly the same compartment bits as the INFORMATION LABELS: section. Moreover, for every word in the INFORMATION LABELS: section with associated normal (non-inverse) compartment bits, there can be no word in the SENSITIVITY LABELS: section whose associated normal compartment bits dominate but do not equal the compartment bits of the information label word, unless the sensitivity label word is an alias. Also, for each inverse compartment word in the SENSITIVITY LABELS: section, there must be a corresponding inverse compartment word in the INFORMATION LABELS: section whose compartment bits are dominated by the sensitivity label word's compartment bits, and whose markings contain no normal bits.
The SENSITIVITY LABELS: section has a structure identical to the INFORMATION LABELS: section with the following exceptions:
All other keywords described above for the information labels section work in the sensitivity labels section, with the same purpose, rules, restrictions, and caveats.
Another difference between the INFORMATION LABELS: section and the SENSITIVITY LABELS: section is that the conventional order within the encodings file of sensitivity label words in the intelligence community is with the least sensitive words first.
Chapter 7, General Considerations for Specifying Encodings discusses some very important considerations concerning the specification of sensitivity label encodings.