3 ms·

Here is a study: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0236153 https://journals.plos.org/plosone/article?id=10.1371/journal... And,

Here is a study: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0236153 https://journals.plos.org/plosone/article?id=10.1371/journal... And, since LLMs are so bad at math currently, we may find that, by improving their math ability with gobs of synthetic data, we get improvements in general reasoning.

conformist 3y ago
Thanks, this is a nice reference! Interesting that the abstract and introduction mention the lack of existing evidence - seems like an under-studied question? One aspect that's quite easy to criticise about this study is that it uses existing groups of students with different levels of maths training. This means that there is possibly self-selection etc, and one may argue there might also be a causal effect in the opposite direction (e.g. folks that are good at reasoning like to do maths.)