3 ms·

Here is an even easier one, ask llms to take the integral from 0 to 3 of 1/(x-1)^3. It fails to notice it's an improper integral and just gives an answer.

• floam 2y ago
  ChatGPT definitely noticed: o1, o3-mini, o3-mini-high. Maybe 4o will get it wrong? I wouldn’t try it for math.
  • vmg12 2y ago
    I tried 4.5 which i thought was the best model, seems like the reasoning models do get it.