The function takes any real input and returns a output that is approximately linear with a negative slope. The output magnitude is slightly larger than the input magnitude.

