AI RESEARCH
Demonstrations, CoT, and Prompting: A Theoretical Analysis of ICL
arXiv CS.LG
•
ArXi:2603.19611v1 Announce Type: new In-Context Learning (ICL) enables pretrained LLMs to adapt to downstream tasks by conditioning on a small set of input-output nstrations, without any parameter updates. Although there have been many theoretical efforts to explain how ICL works, most either rely on strong architectural or data assumptions, or fail to capture the impact of key practical factors such as nstration selection, Chain-of-Thought (CoT) prompting, the number of nstrations, and prompt templates.