It sounds as if rather than ACES they used a legacy workflow with a simple Rec.709 1D LUT as a display transform.
The major problem with this is that no data above 1.0 in the EXRs would have been visible to the animators, as they had no tone mapping. It is therefore likely that they would have reduced the lighting levels to non physically plausible values, in order to keep most of the rendered image data below 1.0. This may limit your ability to create an HDR deliverable, and force you to fake it in the grade.
Good luck!