Just a few experiments making GPT-4 remedy math issues in 16 completely different languages
It’s mentioned that arithmetic is a common language — mathematical ideas, theorems, and definitions could be expressed as symbols which might be comprehensible no matter language.
On this article, I take a look at the mathematical capabilities of GPT-4 in sixteen completely different languages.
Early experiments confirmed GPT-4 scoring extremely on the SAT Math and AP Calculus assessments and on undergraduate-level arithmetic. Nonetheless, nearly all of these experiments take a look at GPT-4’s mathematical capabilities solely in English. To higher perceive GPT-4’s mathematical capabilities past English, I immediate it on the identical math issues in fifteen different languages.
So, how good is GPT-4 at math in numerous languages? In principle, it needs to be equally good (or unhealthy) throughout all languages, however sadly (as you might need guessed), this isn’t the case. GPT-4 is significantly better at fixing math issues in English. Relying on the language, GPT-4 might remedy among the issues. For historically under-resourced languages, nevertheless, similar to Burmese and Amharic, GPT-4 was unable to unravel the issues I gave it.
I take advantage of mathematical issues from the Venture Euler web site to check GPT-4. (That is additionally a throwback to one in every of my one in every of my earlier articles from this yr, the place I used immediate engineering utilizing ChatGPT to unravel a number of Venture Euler issues). Venture Euler, named for the eponymous mathematician, is an internet site with a whole lot of mathematical and pc programming issues ranging in issue. Began in 2001, they boast over 850 issues (as of October 2023) and launch a brand new query roughly each week.
The beauty of Venture Euler questions is that every downside has a numerically “right” reply — this makes it straightforward to verify if GPT-4’s reply is objectively right or not. In addition they are usually much more difficult than high-school or college-level math issues. At present, there isn’t a large-scale complete understanding of GPT-4’s (or different massive language fashions, for that matter) math…
