Splitting Array Elements into Multiple Elements

Yes that is correct, the "small" is in reference to the max string length allowed.

To help you understand and work through the second here is a breakdown:

=UNIQUE(
       DROP(
           REDUCE("",GPO[about:config setting equivalent],
                LAMBDA(p,q,
                        VSTACK(p,TEXTSPLIT(q,,";")))),1))

so line 1 is the UNIQUE, which you already know

line 2 is DROP which will drop any given number of elements of an array.  You can drop rows or columns or both and you can use negative numbers to drop from the end instead of the beginning.  In this case we are just dropping the 1st element to get rid of the "" initialization being used in the REDUCE

line 3 is REDUCE which is a helper function to the LAMBDA in line 4.  REDUCE basically has an initial value and an array and will call the LAMBDA for each element in the array until the final answer is provided.  so simple example is initial value 0 and an array of expenses and then in LAMBDA you have a formula to add the tax rate for each expense so then each call of that LAMBDA will pass it a new expense value and that LAMBDA will return the prior value/sum with the new calculated tax on the newly passed expense added to it so in the end you get a sum of the tax on all the items.

line 4 is LAMBDA which takes in some number of variables and then applies a formula to those variable and returns a value (much like a user defined function).  In this case it is used with the helper function REDUCE so it must have an accumulator (p) and a value (q).  In the tax sample I said above it might look like LAMBDA(p, q, p+0.07*q) so each call would have the accumulator (i.e. prior value returned from the LAMBDA) added to the newly passed value time 7%.

line 5 is VSTACK of each TEXTSPLIT.  So I believe you understand the TEXTSPLIT and this line just stackes each row from the textsplit on top of each other.

So the reason for the REDUCE - LAMBDA is that this will apply the TEXTSPLIT to each value individually and then build up a 'single' 2d table of the results.  There is major limitation with respect to array of arrays with LAMBDA functions but using REDUCE to return a 'single' table is a known 'work around' but since we initialized the REDUCE with "" the first row will be "" so we use DROP to get rid of that row.

I recommend you continue to learn more about LAMBDA as it is extrememly useful and powerful.

Finally w/r to UNIQUE and returning only cases where it occurs exactly 1 time, you gave an excellent application example but I wanted to mention another that I find it very useful.  If you have List A and you want to fill in List B using values from List A but not repeat any, then you can have a helper column tell you which unique values from List A are still left by using 

=UNIQUE( VSTACK( ListA, ListB ), , 1)

so then you can add Data Validation on the elements in List B and point that Data Valadation list to that helper column so they fill in only unique values from List A.