Enregistré dans:
Détails bibliographiques
Auteurs principaux: Goyal, Agam, Mukherjee, Koyel, Saxena, Apoorv, Phukan, Anirudh, Chandrasekharan, Eshwar, Sundaram, Hari
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2604.06097
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908977542987776
author Goyal, Agam
Mukherjee, Koyel
Saxena, Apoorv
Phukan, Anirudh
Chandrasekharan, Eshwar
Sundaram, Hari
author_facet Goyal, Agam
Mukherjee, Koyel
Saxena, Apoorv
Phukan, Anirudh
Chandrasekharan, Eshwar
Sundaram, Hari
contents Dense retrievers in retrieval-augmented generation (RAG) systems exhibit systematic biases -- including brevity, position, literal matching, and repetition biases -- that can compromise retrieval quality. Query rewriting techniques are now standard in RAG pipelines, yet their impact on these biases remains unexplored. We present the first systematic study of how query enhancement techniques affect dense retrieval biases, evaluating five methods across six retrievers. Our findings reveal that simple LLM-based rewriting achieves the strongest aggregate bias reduction (54\%), yet fails under adversarial conditions where multiple biases combine. Mechanistic analysis uncovers two distinct mechanisms: simple rewriting reduces bias through increased score variance, while pseudo-document generation methods achieve reduction through genuine decorrelation from bias-inducing features. However, no technique uniformly addresses all biases, and effects vary substantially across retrievers. Our results provide practical guidance for selecting query enhancement strategies based on specific bias vulnerabilities. More broadly, we establish a taxonomy distinguishing query-document interaction biases from document encoding biases, clarifying the limits of query-side interventions for debiasing RAG systems.
format Preprint
id arxiv_https___arxiv_org_abs_2604_06097
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Masking or Mitigating? Deconstructing the Impact of Query Rewriting on Retriever Biases in RAG
Goyal, Agam
Mukherjee, Koyel
Saxena, Apoorv
Phukan, Anirudh
Chandrasekharan, Eshwar
Sundaram, Hari
Information Retrieval
Dense retrievers in retrieval-augmented generation (RAG) systems exhibit systematic biases -- including brevity, position, literal matching, and repetition biases -- that can compromise retrieval quality. Query rewriting techniques are now standard in RAG pipelines, yet their impact on these biases remains unexplored. We present the first systematic study of how query enhancement techniques affect dense retrieval biases, evaluating five methods across six retrievers. Our findings reveal that simple LLM-based rewriting achieves the strongest aggregate bias reduction (54\%), yet fails under adversarial conditions where multiple biases combine. Mechanistic analysis uncovers two distinct mechanisms: simple rewriting reduces bias through increased score variance, while pseudo-document generation methods achieve reduction through genuine decorrelation from bias-inducing features. However, no technique uniformly addresses all biases, and effects vary substantially across retrievers. Our results provide practical guidance for selecting query enhancement strategies based on specific bias vulnerabilities. More broadly, we establish a taxonomy distinguishing query-document interaction biases from document encoding biases, clarifying the limits of query-side interventions for debiasing RAG systems.
title Masking or Mitigating? Deconstructing the Impact of Query Rewriting on Retriever Biases in RAG
topic Information Retrieval
url https://arxiv.org/abs/2604.06097