However, it seems to me its more selecting for its own particular configuration of genes, that essentially select for a signature of themself.
Within a species the configuration between individuals is generally different enough to not ellicit altruistic behaviour.
In most animals, the 50% kinship gene relatedness threshold is definitely not enough to instigate self sacrificial behaviour. The children of a usocial species have to be like 11/16ths clones of eachother. And the runt suicide mechanism that triggers in rabbits requires a brood of true clones.
The reason I think it seems the genes select for individuals is because most circumstances will recquire that the current reproductive unit at the individual level needs to stay alive long enough for a net positive replication contribution to have been made.