Let me start by saying that publications are usually evaluated not only according to their journal, but also according to the times they have been cited and to the impact they have had, which is a bit harder to gauge and a less objective metric. There are some cases where (at least where I work) the committee requires candidates to select a small-ish number of publications that they deem as the most important (5-12), and these are then read and carefully evaluated individually.
Regarding journals, the answer, as often happens, pretty much depends on the field. In many communities, the impact factor (IF) is surely important, but there are also other (often intangible) factors that come into play.
If many fields, the journals at the very top are few and well known (often Science, Nature and a few others). Papers published in these journals are usually considered to be the top of the cream. For all the other journals, the way they are considered by single scientists or committees is a combination of IF and of these people's own background.
For instance, I work in a pretty multidisciplinary field (in between physics, chemistry and materials science). Some journals are more chemistry- and materials-science-related, and these tend to have higher impact factors than their physics counterparts. As a result, I often find myself publishing in journals that, according to their IF, would be very good from a physicist perspective (>10). However, in the physics community these are regarded way lower than other physics-only journals that have a long history of publishing exceptionally important papers. A notable example is Physical Review Letters, which has published many famous papers like the original papers on the Higgs' Boson or the first detection of gravitational waves and has an IF of ~ 8.8 or PNAS, which is widely known and has an IF of ~ 9.5. I can tell you that, for many physicists, a paper published in PRL is valued higher than a paper published in a materials science journal with an IF of, say, 15 (everything else, such as the number of citations, being the same). I believe that in other communities the attitude is probably very similar: journals that are well known and have a long history will be ranked high, regardless of their IF.