Confident and Wrong: We Tested 17 AI Models on Questions a Middle Schooler Could Answer

Confident and Wrong: We Tested 17 AI Models on Questions a Middle Schooler Could Answer We tested 17 open-source large language models on 6 elementary questions. Basic multiplication. A train word problem. A logic syllogism your kid could solve. 6 of 17 models failed at least one question. 2 models scored 0 out of 6. And here's the part that should worry you: the wrong answers look exactly like the right ones. The Setup 6 questions. One correct answer each. No ambiguity. What is 7 times 8? A tra