What do people do when they don't know when to stop, and what is the digital/simulated/artificial equivalent anyway?
In electronics, rotary positional encoders are often used to replace traditional analog potentiometers (variable resistors) for things like audio volume controls or frequency equalizers.
Like the title says "Round and Round We Go!" this encoder type of physical control knob has no maximum stopping point, unlike the potentiometer where you have the tactile feedback and you can easily feel when the old-school control is all the way up.
Once you reach maximum volume or gain with an encoder knob, you can keep turning it clockwise another 360 degrees forever, it gives the impression that you're "cranking it up to 11" but nothing ever changes beyond that point.
Maybe some of the same differences apply when it comes to the virtual rotary position encoders used in the LLMs.
What do people do when they don't know when to stop, and what is the digital/simulated/artificial equivalent anyway?
In electronics, rotary positional encoders are often used to replace traditional analog potentiometers (variable resistors) for things like audio volume controls or frequency equalizers.
Like the title says "Round and Round We Go!" this encoder type of physical control knob has no maximum stopping point, unlike the potentiometer where you have the tactile feedback and you can easily feel when the old-school control is all the way up.
Once you reach maximum volume or gain with an encoder knob, you can keep turning it clockwise another 360 degrees forever, it gives the impression that you're "cranking it up to 11" but nothing ever changes beyond that point.
Maybe some of the same differences apply when it comes to the virtual rotary position encoders used in the LLMs.