1 paper
Adhemar de Senneville, Xavier Bou, Jérémy Anger +2
There has been significant work on understanding the In-Context Learning capabilities of Large Language Models, especially on the induction circuit. For a few-shot classification t…