The Oath

Gary Marcus

Rank 12 of 47
|
Score 83
Gary Marcus
@GaryMarcus
--
@Mbounge_ You don’t think maybe turbo was supposed to be 5 and didn’t work out as planned?
4/12/2024, 3:33:17 AM
X
In reply to:
Mbongeni Ndlovu
@Mbounge_
·
660d
@GaryMarcus When 5 drops - please plot this again
Gary Marcus
@GaryMarcus
·
660d
What happens when you plot GPT-2, 3, 4, and Turbo side-by-side?

Below I have plotted one common measure, MMLU, where there are easy to find data going back to GPT-2. (There may be others with data going back that far; this is just a first quick attempt.)


What I see is an…
Gary Marcus
@GaryMarcus
·
660d
Could we see GPT 3 and 3.5 and GPT 4 on the same plot? And Gemini Pro 1.5 and Claude Opus?

The statement is a speculative comment within a technical conversation about AI model performance, lacking broader societal implications.

FacebookInstagramTwitterYouTube

© 2023-2024 The Oath, All rights reserved.