In the Bellman$-$Ford algorithm, what does the relaxation of an edge $(u,v)$ with weight $w$ mean?

It updates the weight of the vertex $v$ if a shorter path from $u$ is found.

It finds the shortest path between $u$ and $v.$

It updates the weight of the vertex $u$ if a shorter path from $v$ is found.

It deletes the edge $(u,v)$ if the weight is too large.