I was actually looking for Rufus Stephens, but ran across Mifs Stephens instead.
When I saw this search result on Ancestry.com, I was pretty certain I knew what was going on. I was really doubtful there was a "Mifs" Stephens.
Seeing the occupation of milliner in the search results made me more convinced that "Mifs" was not really "Mifs."
The actual image from the census leads me to conclude the actual entry is "Miss Stephens." During the time period in question what looks like an "fs," isn't. It is a double "s." I realize that there are people who transcribe what it "looks" like, but it is important to realize what was meant. In old German script, an "L" looks quite a bit like an English "B." If I encounter that handwriting in the records of an American church, do I transcribe it was it "looks" or what was meant? I'd go with the intent.
In this case, the census enumerator has already set the precedent for not listing women by their names. The "Widow Graham" is listed right above "Miss Stephens." The enumerator is simply following his apparent pattern by using women's first names.
Another good instance of where searching without the first name may help you locate the person of interest.
And another good instance of looking at an entry in the context of other entries on the page or record.