First, I think you're referring to SET IDENTITY_INSERT ON (underscore), which --as far as I remember allows you to provide your own unique ID values (MSSQL Server).
Second, provided I understand your question correctly, you shouldn't have to worry about that in SQLite3, since what you're trying to do is its 'default' behaviour.
Example (in SQLite):
Let's create a new table. The table consists of an ID attribute and some Name attribute. The ID field is defined as an integer PRIMARY KEY.
Now, let's add some values to this new table. At first we'll let SQLite take care of the ID value, then we'll manually provide a (unique) ID value, and lastly we'll have SQLite generate the ID value once again automatically...
BEGIN TRANSACTION;
CREATE TABLE IF NOT EXISTS MyColorsTable(ColorID integer PRIMARY KEY, ColorName text);
--create a record and let SQLite create the ID value
INSERT INTO MyColorsTable(ColorName) VALUES('Red');
INSERT INTO MyColorsTable(ColorName) VALUES('Green');
--manually provide an ID value now
INSERT INTO MyColorsTable(ColorID,ColorName) VALUES(456,'Blue');
INSERT INTO MyColorsTable(ColorID,ColorName) VALUES(789,'Purple');
--and lastly let SQLite do the job of providing an ID...
INSERT INTO MyColorsTable(ColorName) VALUES('Brown');
INSERT INTO MyColorsTable(ColorName) VALUES('Black');
COMMIT;
The resulting table values are shown below and I believe is something what you're trying to accomplish...
SELECT * FROM MyColorsTable
ColorID ColorName
1 Red
2 Green
456 Blue
789 Purple
790 Brown
791 Black
Again, not sure if that's what you were asking, but that's the closest I could think of.