Simplified steps from

1. Say you have a column A with A1 to A100 rows with duplicate data that you want to eliminate
2. Go to Data > Advanced Filter. Under the input tab, choose list range as A1 to A100.
3. I don't understand why Gnumeric needs a criteria range for everything but it does... Even for duplicate-value-finding that involves no mathematical formulas. So choose a criteria range. Say B1 to B2. Make sure that the first cell of the criteria range has the same column header as the list range. So if, say your cell A1 was named "Numbers" then B1 should also be "Numbers". Leave B2 blank.
4. Check unique records only
5. Go to the Output tab. Choose an output range (longer than the input range, just to prevent any complaints from gnumeric). Click OK.
You should have the original list with no duplicates in the output range.
This was a major doorstopper because I completely subbed the butter with applesauce, subbed half the flour with wheat bran, subbed the fishy ammonia thingy with regular baking soda, forgot the vanilla extract and used 100% brown sugar. And of course, I have no measures, everything was eyeballed. I narrowly escaped breaking my teeth as I bit into one without cares. They are rather good (in a wow, this is so healthy for me way) dipped in hot tea.