xie-2021-icl-error-decreases-example-length
IN premise — summaries/2026/08/24/xie-2021-icl-bayesian-s1-introduction.md
Created 2026-08-25T02:58:55+00:00
In Xie et al. (2021), ICL error decreases with the length of each prompt example, indicating that information in the input tokens (not just the input→output label) contributes to concept inference.
Summary
Longer examples in a prompt help the model make fewer mistakes, and this isn't just because there's more label data to learn from. It means the model is actually pulling conceptual understanding out of the input content itself, not merely memorizing which answer goes with which question, so the structure of the example matters as much as the correct answer it demonstrates.