Metraipolis › archive › dc_fdb55c48dfc5

Can Jev Improve Demonstration Retrieval for In-Context Learning? — arXiv:2611.00737

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.00737
License (as archived)
CC0-1.0
Archived
2026-11-03T18:12:09.742Z
SHA-256 of the archived text
3ef8d87be676807214c923ba12a2b3bcf58237cc09d6fb8888f5b746cfe1c96d
Stable URL of this page
https://metraipolis.com/documento/dc_fdb55c48dfc5

How to cite

Can Jev Improve Demonstration Retrieval for In-Context Learning? — arXiv:2611.00737 — arXiv (en), https://arxiv.org/abs/2611.00737 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fdb55c48dfc5 (Metraipolis archive dc_fdb55c48dfc5, sha256 3ef8d87be676807214c923ba12a2b3bcf58237cc09d6fb8888f5b746cfe1c96d)
BibTeX
@misc{dcfdb55c48dfc5,
  title = {Can Jev Improve Demonstration Retrieval for In-Context Learning? — arXiv:2611.00737},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.00737},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fdb55c48dfc5; sha256 3ef8d87be676807214c923ba12a2b3bcf58237cc09d6fb8888f5b746cfe1c96d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00737. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Demonstration selection plays a crucial role in the effectiveness of in-context learning (ICL) for large language models (LLMs). Extensive research has explored various retrieval strategies to identify informative demonstrations and improve downstream performance. Recently, the emergence of Jev, a decision-oriented language model capable of producing probabilistic judgments, offers a new perspective on evaluating demonstration utility. This motivates us to investigate a fundamental question: Can Jev improve demonstration retrieval for in-context learning? In this paper, we demonstrate that Jev can effectively enhance demonstration retrieval by directly assessing the utility of candidate demonstrations for a given query. However, applying Jev to large-scale retrieval introduces substantial computational overhead, as each candidate requires an additional decision-making inference. To address this limitation, we propose a simple yet effective knowledge distillation approach that transfers Jev's utility estimation capability to a lightweight multilayer perceptron (MLP). By learning to approximate Jev's probabilistic judgments, the distilled model enables efficient demonstration ranking while maintaining competitive retrieval performance. Our study highlights the potential of decision-oriented language models for demonstration selection and provides an efficient framework for integrating their decision-making capabilities into ICL.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.00737 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-03T18:09:53.307Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).