The Jess manual documents <> as returning TRUE if the first argument is not equal to all subsequent arguments. It would appear, however, that the function returns TRUE if any two adjacent arguments are not equal. (I can envisage the implementation being a copy of the implementations of <, <=, >, >= and =, running the test for all neighbours.)
Either the documentation or the implementation should be fixed. It's difficult to see the general usefulness of the implementation for more than 2 arguments. The described semantics seem to be a little more useful. Kind regards Wolfgang