I assume that you're in computer science, since otherwise you probably shouldn't list any conference papers as publications. So I'll answer as a computer scientist.
No, you should not list such papers as publications in your CV, because those papers are not actually published. Moreover, there is likely an expectation that the same paper can be published at a different conference. You must not list the "same" paper at more than one conference.*
I know of several conferences/workshops like the one your describe in computational geometry (my home field), including EuroCG, the Fall Workshops, and the Young Researchers Forum at SOCG. At all three venues, submissions are lightly peer-reviewed, only a subset of submissions are accepted, and a booklet of abstracts is distributed to participants and/or on the web. But no formal proceedings is issued at these events, because it is expected that accepted papers will later appear in more polished form at a more formally reviewed conference. (Some early iterations of EuroCG did have formal proceedings, despite the expectation of later publication, but other conferences were unwilling to accept papers that appeared in those proceedings.)
As others have said, it's perfectly fine to list those talks under "Unpublished Workshop Talks", especially early in your career. You might even include the acceptance rate if you want to emphasize that the venue carries some prestige. Alternatively, if you did publish the paper elsewhere, you might include the phrase "Also presented at ..." after the publication info.
*...but journals are different. In most subfields of computer science, conferences papers can be published later in refereed journals, usually in a more expanded/complete form. Even so, I recommend listing each paper only once in your CV, including all publication venues for each paper, rather than listing the same paper once under "conference papers" and again under "journal papers".