PrismRAG: Boosting RAG Factuality with Distractor Resilience and Strategized Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kachuee, Mohammad, Gollapudi, Teja, Kim, Minseok, Huang, Yin, Sun, Kai, Yang, Xiao, Wang, Jiaqi, Shah, Nirav, Liu, Yue, Colak, Aaron, Kumar, Anuj, Yih, Wen-tau, Dong, Xin Luna
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915408882171904
author Kachuee, Mohammad
Gollapudi, Teja
Kim, Minseok
Huang, Yin
Sun, Kai
Yang, Xiao
Wang, Jiaqi
Shah, Nirav
Liu, Yue
Colak, Aaron
Kumar, Anuj
Yih, Wen-tau
Dong, Xin Luna
author_facet Kachuee, Mohammad
Gollapudi, Teja
Kim, Minseok
Huang, Yin
Sun, Kai
Yang, Xiao
Wang, Jiaqi
Shah, Nirav
Liu, Yue
Colak, Aaron
Kumar, Anuj
Yih, Wen-tau
Dong, Xin Luna
contents Retrieval-augmented generation (RAG) often falls short when retrieved context includes confusing semi-relevant passages, or when answering questions require deep contextual understanding and reasoning. We propose an efficient fine-tuning framework, called PrismRAG, that (i) trains the model with distractor-aware QA pairs mixing gold evidence with subtle distractor passages, and (ii) instills reasoning-centric habits that make the LLM plan, rationalize, and synthesize without relying on extensive human engineered instructions. Evaluated across 12 open-book RAG QA benchmarks spanning diverse application domains and scenarios, PrismRAG improves average factuality by 5.4%, outperforming state-of-the-art solutions.
format Preprint
id arxiv_https___arxiv_org_abs_2507_18857
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PrismRAG: Boosting RAG Factuality with Distractor Resilience and Strategized Reasoning
Kachuee, Mohammad
Gollapudi, Teja
Kim, Minseok
Huang, Yin
Sun, Kai
Yang, Xiao
Wang, Jiaqi
Shah, Nirav
Liu, Yue
Colak, Aaron
Kumar, Anuj
Yih, Wen-tau
Dong, Xin Luna
Computation and Language
Artificial Intelligence
Machine Learning
Retrieval-augmented generation (RAG) often falls short when retrieved context includes confusing semi-relevant passages, or when answering questions require deep contextual understanding and reasoning. We propose an efficient fine-tuning framework, called PrismRAG, that (i) trains the model with distractor-aware QA pairs mixing gold evidence with subtle distractor passages, and (ii) instills reasoning-centric habits that make the LLM plan, rationalize, and synthesize without relying on extensive human engineered instructions. Evaluated across 12 open-book RAG QA benchmarks spanning diverse application domains and scenarios, PrismRAG improves average factuality by 5.4%, outperforming state-of-the-art solutions.
title PrismRAG: Boosting RAG Factuality with Distractor Resilience and Strategized Reasoning
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2507.18857