2 Comments
User's avatar
Andrew E Scott's avatar

I like the insight that a Q&A-trained LLM is naturally a poor tutor, and needs to be cajoled into performing better. You may be interested in how AERO locally is framing-up their thinking about Gen AI:

https://www.edresearch.edu.au/research/discussion-papers/strengthening-evidence-how-genai-can-improve-teaching-and-learning

Expand full comment
Kendra Vant's avatar

Super interesting eh? Obvious one it gets pointed out but easy to overlook.

Thank for the link, I take a read

Expand full comment