English

LLM-Oriented Retrieval Tuner

Computation and Language 2024-03-05 v1

Abstract

Dense Retrieval (DR) is now considered as a promising tool to enhance the memorization capacity of Large Language Models (LLM) such as GPT3 and GPT-4 by incorporating external memories. However, due to the paradigm discrepancy between text generation of LLM and DR, it is still an open challenge to integrate the retrieval and generation tasks in a shared LLM. In this paper, we propose an efficient LLM-Oriented Retrieval Tuner, namely LMORT, which decouples DR capacity from base LLM and non-invasively coordinates the optimally aligned and uniform layers of the LLM towards a unified DR space, achieving an efficient and effective DR without tuning the LLM itself. The extensive experiments on six BEIR datasets show that our approach could achieve competitive zero-shot retrieval performance compared to a range of strong DR models while maintaining the generation ability of LLM.

Keywords

Cite

@article{arxiv.2403.01999,
  title  = {LLM-Oriented Retrieval Tuner},
  author = {Si Sun and Hanqing Zhang and Zhiyuan Liu and Jie Bao and Dawei Song},
  journal= {arXiv preprint arXiv:2403.01999},
  year   = {2024}
}

Comments

16 pages, 8 figures, 5 tables

R2 v1 2026-06-28T15:08:19.118Z