There should be a tr(...) around the entire N = 4 SYM action below:
G_5 ~ G_10 should be DIVIDED BY Area[S_5] in Planck units
(So the calculations in purple marker directly below are wrong too.)
Since (e.g. for scalar) ∆(∆-4) = m^2, dimension of nonchiral irreps should be ∆^2 ~ m_string^2 ~ 1/(l_s)^2 (in units where L = 1)
