Recall that a Grothendieck topos is a category of sheaves on some site .
If is the category of open subsets of a topological space (or some manifold or the like) then one calls a petit topos.
If on the other hand is a category of all test spaces in some sense, such as
the category Top of all topological spaces;
etc. one call a gros topos .
Objects in a gros topos may be thought of as spaces modeled on in the sense described at motivation for sheaves, cohomology and higher stacks and at space.
Also the objects in a petit topos – a category of sheaves on the category of open subsets of a topological space – are a kind of generalized spaces, but generalized spaces over on which the rigid structure of morphisms in (only inclusions of subsets, no more general maps) induces a correspondingly rigid structure so that they are not all that general. In fact, is equivalent to etale spaces over .
There is another notion of ‘large’ topos associated to a space (or more generally an object in a site), namely the topos of sheaves on the slice category , with the obvious notion of covering family; a family is covering if it is so under the functor . This site is often referred to as the large site of , as compared to the small site, which is as above. The topos can be viewed as spaces modelled Top (or more generally some site), but parameterised by the representable sheaf .
The petit topos of an object in a site is the category of sheaves on the small site of .
For a topological space, the petit topos that it defines is the category of sheaves on the category of open subsets of . A general object in this topos is an etale space over . The space itself is incarnated as the terminal object .
A gros topos in which is incarnated is a category of sheaves on a site of test spaces with which may be probed. For instance for Top, or Diff or CartSp with their standard coverages, is such a gros topos.
In good cases the intrinsic properties of do not depend on whether one regards it as an object of a petit or a gros topos. For instance at cohomology in the section Nonabelian sheaf cohomology with constant coefficients it is discussed how the nonabelian cohomology of a paracompact manifold with constant coefficients gives the same answer in each case.
Some aspects of an axiomatic characterization of petit vs. gros toposes is in
There is also something relevant in this article:
Mathhieu Anel?, Grothendieck topologies from unique factorization systems (arXiv:0902.1130)
Mamuka Jibladze, Homotopy types for “gros” toposes, thesis, pdf