I don't think transfer functions like HLG and PQ are used with the internal processing of HDRO in ACR. My understanding is that it's all done in 32 bit floating point. On the other hand, when you export an HDRO file say in avif format, either HLG or PQ is used. I think ACR uses PQ.
There is a standard under development, see here.
I'm not sure what you mean with your comments on exposure. The best you can do is expose to the right without any highlight clipping. As I see it, ACR expands the highlights above SDR standard white when the HDR button is pressed. This undoes the compression in this area caused by the tone curve used. The tone curve still improves the contrast in the SDR part of the image.
Dave
Yes, true on many points. Sorry for my late reply.
Yes, I imagine it's done internally with floating 32.
By "target exposure", I meaned same brightness for mid-grey.
However, and if I compare with Davinci resolve for example, my guess is that there is a conversion for displaing the datas, and this conversion need to be function of the choosen display curve of the display (HLG or PQ).
The way it is done with SDR worflow (AFIK) is that the curve (linear light -> rec709alikee) is mainly handle in the dcp profile, so I'm pretty confused by this one click "hdr" button in the user interface.