Least-Squares means are predictions from a model over a regular grid, possibly averaged over other dimensions. Also use this tag for the R packages emmeans and lsmeans.
Least-Squares means are the means of predictions from a model over a regular grid, averaged over zero or more additional dimensions. In some contexts these are called "predicted marginal means", "covariate adjusted means", or "estimated marginal means", among other names.
The term originated with the work of Walt Harvey in the 1970s, and was adopted by SAS and later by other software. See this post for several references.