[seqfan] A082446 Thue-Morse replications

[Kevin Ryde]

A082446 is Thue-Morse parity with certain replications of the values.
The OFFSET shown is 0, but the name description reads like it's for 1.
Which would be intended?

I quite like 0 since it gives a formula like, in gp,

    A082446(n) = (hammingweight(n\3) + (n%3)) % 2;

but maybe the description with 3*2^k is for some block copying.

Also I think the sample values are missing a 0 after the first 47 values,
ie. insert a 0 there.  Could that be right?  I tried by formula above
and by following the description.