How does λ affect Generalized Advantage Estimation?

Holding the value-function approximation and discount factor fixed, what is the usual effect of increasing λ in Generalized Advantage Estimation?