In that case, you can just index t:
out = t[x[:, None], y]
print(out)
> tensor([[ 1, 3, 5, 7, 9, 11, 13, 15],
[ 33, 35, 37, 39, 41, 43, 45, 47],
[ 65, 67, 69, 71, 73, 75, 77, 79],
[ 97, 99, 101, 103, 105, 107, 109, 111],
[129, 131, 133, 135, 137, 139, 141, 143],
[161, 163, 165, 167, 169, 171, 173, 175],
[193, 195, 197, 199, 201, 203, 205, 207],
[225, 227, 229, 231, 233, 235, 237, 239]])
The last rows will be different, since in your example you’ve indexed row15 and row13.