>We show that a variety of modern deep learning tasks exhibit a "double-descent" phenomenon where, as we increase model size, performance first gets worse and then gets better.

▲

smallerize 5 hours ago | parent | next [-]

Does this mean that if your model is "overfitting", the solution is to train for even more epochs?

▲

ForceBru 6 hours ago | parent | prev [-]

Right, isn't double descent one of the reasons why modern Extremely Large Language Models work at all? I think I heard somewhere that basically all today's "smart" (reasoning, solving math problems, etc) LLMs are trained in the "double descent" territory (whatever this means, I'm not entirely sure).

	▲	mxwsn 3 hours ago \| parent \| next [-]
		No, there are more training tokens than parameters in LLMs. They are in the classical first descent setting.
	▲	SiempreViernes 5 hours ago \| parent \| prev [-]
		No, double descent is a symptom of whatever it is that makes the deep models work at all. It's just the name for something you see happen when it works. The reason it works has something to do with how all those extra dimensions work as a regularisation term in the fit.