On: Teaching GPT-3 to Identify Nonsense

[317070]

I have seen the figure 0.4 kWh for 100 pages of text come about in the paper for generating text with the model. So that goes up to $0.002/page of text when you take into account hardware cost too.

Source: https://arxiv.org/pdf/2005.14165.pdf section 6.3. Or for a more monetary overview: https://www.lesswrong.com/posts/N6vZEnCn6A95Xn39p/are-we-in-...

[hhh → 317070]

Cheaper than my workplace pays to print a colour page of paper.