My word... samwho is doing some of the best technical explainers on the internet right now.

polotics 4 hours ago | parent | next [-]

Leading to my question: Ok keeping a zero and a minus-zero does make sense for some limits calculations... But when all you have is 4 bits, is this not quite wasteful? Would using the bits for eg. a 2.5 not improve the model?

	▲	polotics 4 hours ago \| parent [-]
		Oh well that's a rabbit hole: NVIDIA Blackwell has this, also GGUFs sidestep this with Qi_j / Qi_K... Great article, spikes curiosity!

▲

seabass 3 hours ago | parent | prev [-]

Heartily second that! It was cool to see a combination of DOM, SVG, and canvas visualization all in use for this post.