Probably not. The dye molecule has its own absorbance spectrum. If it absorbs at the same wavelength you are using to measure the small molecule, then your standard curve will give an incorrect result. I would guess you are measuring the steroid absorbance in the ultraviolet. All dyes are aromatic compounds, so they all absorb in the UV. Therefore, it is likely that the dye's absorbance spectrum overlaps with the small molecule's. It is also possible that the absorbance spectrum of the small molecule will be different after labeling because of the chemical change.

NO

Better to draw a standard curve after the labeling. Because the molar absorptivity may differ.

Should be possible. Spectra will always overlap, but in a known way. Providers of dye labels give a correction factor for their labeling agents, here is an example:

https://info.gbiosciences.com/blog/how-to-determine-degree-of-protein-labeling

I argee that you should prepare a new calibration curve. This will not take much time, but it will give you confidence that this factor has been taken into account in the validation of the methodology.

This is not directly related to your question, but though the spectra in solvent will probably not change much (unless you have a steroid that is chromophoric in itself), the spectra in complex environment may change, if you have a solvatochromic water soluble fluorophore, since after that it will be amphiphilic and reside in lipid/water or lipid/protein/water interface where the solvation is different. This may be issue if your steroid-dye conjugate concentration is above CMC and you are using water for solvation, but then again, you probably are not.

A separate issue is that the physicochemical properties of the resulting steroid-dye conjugate will be very different from the original steroid. The steroids and sterols are quite small molecules, about 300-350 g/mol, whereas the fluorescent dyes tend to be at least as large and often much larger, up to 1500 g/mol, so you can think it as much or more as a dye labelled with a steroid than vice versa; the situation is very different from adding a label on 20 kDa protein. So, unless the particular steroid-dye conjugate you are using has been carefully characterized to resemble the behaviour of the original steroid with respect to property you are interested in, then the a priori assumption is that there will be little similarity in the behaviour of the dye-steroid conjugate and the steroid. This is probably clear to you, but I have seen people use dyes with acidic/basic groups and twice as large as the small molecule, and then use these for evaluating the intracellular localization (that is largely determined by pKa, acidic/basic groups, lipid bilayer solubility, and DNA binding for small molecules) of the small molecule, though this of course will be mostly determined by the larger and chemically more complex dye. If you have protein binding experiment (such as a steroid processing enzyme) and you see binding with relatively similar Kd for the original (e.g. radioactive) steroid, and the dye-steroid conjugate, then it is relatively simple, but if you are going to use it cells or complex environment, the first paper or two should likely be about checking that the behaviour of your dye-steroid conjugate reflects the behaviour of the steroid itself.