Tag
Ai2 introduces TutorMoments, a replay-based evaluation framework and dataset for measuring whether LLMs can balance when to help and when to hold back in one-on-one math tutoring. Preliminary results show models tend to over-help, and prompt engineering only partially closes the gap to human tutors.
Phil Birchenall created a custom GPT powered by ChatGPT to help his 12-year-old daughter Daisy with math struggles by giving the AI tutor the personality of their dog Izzy, resulting in an engaging and effective learning tool.