Gamma controls how far the influence of a single training example reaches, which in turn effects how tightly the decision boundaries end up surrounding points in the input space

Small gamma: larger similarity radius, points farther apart are considered similar, more points grouped together, smoother decision values, more regularization

Large gamma: kernel value decays more quickly, points have to be very close to be considered similar, results in more complex tightly constrained decision boundaries

Increase gamma → sharpening kernel