ISTR some researchers did a test replacing the reasoning tokens with ".", and reasoning still improved the results because dummy tokens are extra passes through the same network, allowing for more computation.

▲

devmor an hour ago | parent [-]

If you can find this research I would love to see it. That’s extremely interesting.

	▲	arthurcolle 29 minutes ago \| parent [-]
		"Let's Think Dot by Dot: Hidden Computation in Transformer Language Models" by Jacob Pfau, William Merrill, and Samuel R. Bowman https://arxiv.org/pdf/2404.15758