The tested LLMs fared much worse, though, when the Apple researchers modified the GSM-Symbolic benchmark by adding "seemingly relevant but ultimately inconsequential statements" to the questions
Good thing they're being trained on random posts and comments on the internet, which are known for being succinct and accurate.
No wonder Trump killed the bipartisan immigration reform bill. I didn't realize the anti-immigrant sentiment had gotten to this level in so much of the population. It's frankly horrifying.