Numerous studies have demonstrated that the Transformer architecture possesses the capability for in-context learning (ICL). In scenarios involving function approximation, context can serve as a control parameter for the model, endowing it with the universal approximation property (UAP). In practice, context is represented by tokens from a finite set, referred to as a vocabulary, which is the case considered in this paper, \emph{i.e.}, vocabulary in-context learning (VICL). We demonstrate that VICL in single-layer Transformers, without positional encoding, does not possess the UAP; however, it is possible to achieve the UAP when positional encoding is included. Several sufficient conditions for the positional encoding are provided. Our findings reveal the benefits of positional encoding from an approximation theory perspective in the context of ICL.
@article{arxiv.2511.06376,
title = {Vocabulary In-Context Learning in Transformers: Benefits of Positional Encoding},
author = {Qian Ma and Ruoxiang Xu and Yongqiang Cai},
journal= {arXiv preprint arXiv:2511.06376},
year = {2025}
}