A mode-locked laser is mostly appreciated for its peak power. When we go from ps to fs, the peak power will dramatically increase even if the pulse energy remains the same. However, a laser cavity is very sensitive to losses. When we add losses inside a cavity, the laser threshold will slightly increase to overcome this loss, consequently, we expect a slightly lower efficiency, which means slightly low output energy for a certain pump power. I guess a low loss SA would definitely output energy slightly higher than a high loss SA. But, in a mode-locked laser, we all care about the pulse duration of the phase-locked pulses that will increase the peak power. I might be wrong but it might help.
You said "high energy" I suspect you didn't mean energy. More loss in the cavity will always cause the pulse energy to go down. I think you probably want to talk about the pulse intensity i.e. the instantaneous peak power. A mode-locked laser isn't more powerful than CW laser, and it won't have the pulse energy of a Q-switched laser. However, mode-locking squeezes the energy into a very short time resulting in very high instantaneous peak power. So when mode-locking goodness is measured in shorter pulses and higher peak power.
A saturable absorber is one of the ways to achieve passive mode locking. So, to the degree that the saturable absorber is the non-linearity which is causing the laser to mode lock, yes, the saturable absorber causes very short very intense pulses.
However, that isn't quite what you asked. You seem to be asking what if you have a laser which is ALREADY mode locked and you add MORE saturable absorber. In that case the intensity will probably decrease. There is an optimum amount of saturable absorber. You want just enough so that the non-linearity of the loss vs intensity drives the gain competition to favor the mode-locked condition. However, once you are getting mode locked transform limited pulses, more non-linearity does not help, and more loss just robs the cavity of energy. This is a pretty sharp function. Enough saturable absorber to result in any mode-locking is basically enough to result in transform limited mode locking. So more is essentially always bad. You want a little more to be stable, but the pulses won't get shorter or brighter.
Depending on the type of laser you are working on. For example, in fiber laser systems where there are plenty of gain, low-loss saturable absorber does not give much advantages, and instead a moderate loss (> 1.5dB) is better if you want to have stronger saturable absorption depth. If you are working on free space laser systems, such gain media such a Ti-Sapphire, thin-disk, etc with a low single-pass gain, a low-loss saturable absorber will be needed, so that you will not compromise your lasing condition.
The total insertion loss of an SA consists of saturable absorption, background non-saturable absorption, coupling loss. What you want is a low non-saturable absorption loss, but a large saturable absorption loss.
A few other things you have to consider include: saturable absorber's response time (e.g. SESAM ~15ps, CNT-SA ~500fs, Graphene-SA ~300fs) and dispersion/nonlinear-effect balance in you laser cavity.