I read at Wikipedia that
In his notebooks, Ramanujan generalized the Euler product for the zeta function as $\prod_p(x-p^{-s})\approx\frac{1}{Li_s(x)}$
I can also see that Wikipedia mentions at the references Ramanujan's Lost Notebook: Part I, but this seems irrelevant, as Ramanujan deals with continued fractions mostly there. Also after I searched notebook part I, for "zeta function" I found three or four results which do not mention at all $\prod_p(x-p^{-s})\approx\frac{1}{Li_s(x)}$
(or anything similar).
I was wondering if anybody could help me find this "generalization". Is it in another notebook?