mlsanityData to 7 Oct 2026

Quantifying and Improving the Robustness of Retrieval-Augmented Language Models Against Spurious Features in Grounding Data

arXiv 2503.05587 ↗First posted 7 Mar 2025cs.CL cs.AI cs.LGcited by 10

On the map: Retrieval-augmented generation → RAG robustness and adaptive retrieval

Abstract (arXiv)

Robustness has become a critical attribute for the deployment of RAG systems in real-world applications. Existing research focuses on robustness to explicit noise (e.g., document semantics) but overlooks implicit noise (spurious features). Moreover, previous studies on spurious features in LLMs are limited to specific types (e.g., formats) and narrow scenarios (e.g., ICL). In this work, we identify and study spurious features in the RAG paradigm, a robustness issue caused by the sensitivity of LLMs to semantic-agnostic features. We then propose a novel framework, SURE, to empirically quantify the robustness of RALMs against spurious features. Beyond providing a comprehensive taxonomy and metrics for evaluation, the framework's data synthesis pipeline facilitates training-based strategies to improve robustness. Further analysis suggests that spurious features are a widespread and challenging problem in the field of RAG. Our code is available at https://github.com/maybenotime/RAG-SpuriousFeatures .

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Placed in RAG robustness and adaptive retrieval, one of 15 directions in Retrieval-augmented generation. Builds on 28 of its 43 references that are in the map. First posted on arXiv 7 Mar 2025.

Builds on

28 of 43 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024Trustworthiness in Retrieval-Augmented Generation Systems: A Survey16 Sep
  2. 2024ClashEval: Quantifying the tug-of-war between an LLM's internal prior and external evidence16 Apr

Other references · 26

  1. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  2. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,383
  3. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,895
  4. 2021Unsupervised Dense Information Retrieval with Contrastive Learningcited by 1,203
  5. 2022Atlas: Few-shot Learning with Retrieval Augmented Language Modelscited by 940
  6. 2023Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formattingcited by 778
  7. 2019Latent Retrieval for Weakly Supervised Open Domain Question Answeringcited by 691
  8. 2024From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judgecited by 480
  9. 2024The Power of Noise: Redefining Retrieval for RAG Systemscited by 275
  10. 2023Better Zero-Shot Reasoning with Role-Play Promptingcited by 220
  11. 2024Large Language Models for Data Annotation and Synthesis: A Surveycited by 201
  12. 2023Chain-of-Note: Enhancing Robustness in Retrieval-Augmented Language Modelscited by 180
14 more
  1. 2024ProSA: Assessing and Understanding the Prompt Sensitivity of LLMscited by 145
  2. 2024Does Prompt Formatting Have Any Impact on LLM Performance?cited by 132
  3. 2023Human-like Summarization Evaluation with ChatGPTcited by 126
  4. 2024Typos that Broke the RAG's Back: Genetic Attack on RAG Pipeline by Simulating Documents in the Wild via Low-level Perturbationscited by 64
  5. 2023RECALL: A Benchmark for LLMs Robustness against External Counterfactual Knowledgecited by 51
  6. 2024DALK: Dynamic Co-Augmentation of LLMs and KG to answer Alzheimer's Disease Questions with Scientific Literaturecited by 50
  7. 2023A New Benchmark and Reverse Validation Method for Passage-level Hallucination Detectioncited by 42
  8. 2022Spurious Features Everywhere -- Large-Scale Detection of Harmful Spurious Features in ImageNetcited by 41
  9. 2023Neural Retrievers are Biased Towards LLM-Generated Contentcited by 37
  10. 2024HtmlRAG: HTML is Better Than Plain Text for Modeling Retrieved Knowledge in RAG Systemscited by 37
  11. 2023Influence of External Information on Large Language Models Mirrors Social Cognitive Patternscited by 16
  12. 2024RAGged Edges: The Double-Edged Sword of Retrieval-Augmented Chatbotscited by 11
  13. 2023A New Dataset and Empirical Study for Sentence Simplification in Chinesecited by 4
  14. 2025MuDAF: Long-Context Multi-Document Attention Focusing through Contrastive Learning on Attention Headscited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 7; 2026: 3.

10 papers.

  1. 2025TreeReview: A Dynamic Tree of Questions Framework for Deep and Efficient LLM-based Scientific Peer Reviewcited by 31
  2. 2026ToolPRMBench: Evaluating and Advancing Process Reward Models for Tool-using Agentscited by 13
  3. 2026Bridging Reasoning Trajectories in On-Policy Distillation via Near-Future Guidancecited by 12
  4. 2025Out of Style: RAG's Fragility to Linguistic Variationcited by 10
  5. 2025The Quest for Efficient Reasoning: A Data-Centric Benchmark to CoT Distillationcited by 7
  6. 2025RARE: Retrieval-Aware Robustness Evaluation for Retrieval-Augmented Generation Systemscited by 7
  7. 2025GuiLoMo: Allocating Expert Number and Rank for LoRA-MoE via Bilevel Optimization with GuidedSelection Vectorscited by 6
  8. 2025Who's Your Judge? On the Detectability of LLM-Generated Judgmentscited by 6
  9. 2026Retrieval-Augmented LLM Agents: Learning to Learn from Experiencecited by 6
  10. 2025Exposing the Cracks: Vulnerabilities of Retrieval-Augmented LLM-based Machine Translationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Evaluating the Retrieval Robustness of Large Language Models
  2. 2025RAG Makes Guardrails Unsafe? Investigating Robustness of Guardrails under RAG-style Contexts
  3. 2025On the Diminishing Returns of Complex Robust RAG Training in the Era of Powerful LLMs
  4. 2026How You Ask Matters! Adaptive RAG Robustness to Query Variations
  5. 2025QE-RAG: A Robust Retrieval-Augmented Generation Benchmark for Query Entry Errors
  6. 2024RAGGED: Towards Informed Design of Scalable and Stable RAG Systems
  7. 2024Astute RAG: Overcoming Imperfect Retrieval Augmentation and Knowledge Conflicts for Large Language Models
  8. 2026Retrieved But Not Reliable: A Survey on Attacks, and Defenses in Retrieval-Augmented Generation