mlsanityData to 14 Oct 2026

Quantifying Prior Dominance in RAG Systems

arXiv 2606.23695 ↗First posted 29 Apr 2026cs.CL cs.AIcited by 0

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

Abstract (arXiv)

Retrieval-Augmented Generation (RAG) grounds Large Language Models in external knowledge, yet current evaluations rely on discrete heuristics that suffer from ''epistemic blindness'' - failing to distinguish genuine contextual information extraction from parametric memory recall. To address this, we introduce the Normalized Context Utilization (NCU) metric, leveraging continuous token log-probabilities across zero-shot, oracle, and adversarial conditions to strictly quantify contextual information gain. Evaluating architectures ranging from 1.5B to 72B parameters alongside a proprietary commercial API reveals that for strict factual extraction (without Chain-of-Thought reasoning), traditional scaling laws exhibit extreme diminishing returns: highly efficient Small Language Models (SLMs) match or outperform high-capacity architectures. Furthermore, we demonstrate that ``Prior Dominance'' correlates with model scale and proprietary alignments. The evaluated commercial API not only overrode explicit external evidence in nearly half of adversarial conflicts, but also frequently suffered from systemic confidence collapse (Negative Transfer) when its parametric priors were contradicted. Our findings highlight the structural epistemic advantage and superior contextual adherence of SLMs in strict extraction workflows.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in RAG robustness and adaptive retrieval, one of 15 directions in Retrieval-augmented generation. Builds on 37 of its 46 references that are in the map. First posted on arXiv 29 Apr 2026.

Builds on

37 of 46 references.

Closest ancestors · 3

Same direction, newest first

  1. 2026Exploring Knowledge Conflicts for Faithful LLM Reasoning: Benchmark and Method13 Apr
  2. 2026Seeing through the Conflict: Transparent Knowledge Conflict Handling in Retrieval-Augmented Generation11 Jan
  3. 2025RAG-RL: Advancing Retrieval-Augmented Generation via RL and Curriculum Learning17 Mar

Other references · 34

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  3. 2025Qwen3 Technical Reportcited by 8,820
  4. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  5. 2020Scaling Laws for Neural Language Modelscited by 6,586
  6. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  7. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  8. 2022Training Compute-Optimal Large Language Modelscited by 3,019
  9. 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,212
  10. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,141
  11. 2022Survey of Hallucination in Natural Language Generationcited by 2,075
  12. 2021Improving language models by retrieving from trillions of tokenscited by 1,293
22 more
  1. 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,282
  2. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,128
  3. 2023In-Context Retrieval-Augmented Language Modelscited by 702
  4. 2023Textbooks Are All You Needcited by 569
  5. 2023Active Retrieval Augmented Generationcited by 516
  6. 2023Ragas: Automated Evaluation of Retrieval Augmented Generationcited by 418
  7. 2021Entity-Based Knowledge Conflicts in Question Answeringcited by 356
  8. 2023Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflictscited by 341
  9. 2023FreshLLMs: Refreshing Large Language Models with Search Engine Augmentationcited by 318
  10. 2023ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systemscited by 265
  11. 2022Rich Knowledge Sources Bring Complex Knowledge Conflicts: Recalibrating Models to Reflect Conflicting Evidencecited by 144
  12. 2025Large Language Models Hallucination: A Comprehensive Surveycited by 49
  13. 2024RAG-Star: Enhancing Deliberative Reasoning with Retrieval Augmented Verification and Refinementcited by 40
  14. 2024RAGEval: Scenario Specific RAG Evaluation Dataset Generation Frameworkcited by 39
  15. 2023Small Language Models Improve Giants by Rewriting Their Outputscited by 18
  16. 2025A Survey on Collaborative Mechanisms Between Large and Small Language Modelscited by 14
  17. 2025A Survey of Scaling in Large Language Model Reasoningcited by 10
  18. 2025Fine-Tune an SLM or Prompt an LLM? The Case of Generating Low-Code Workflowscited by 6
  19. 2025Redefining Retrieval Evaluation in the Era of LLMscited by 3
  20. 2026Revisiting RAG Retrievers: An Information Theoretic Benchmarkcited by 3
  21. 2026Mitigating Hallucinations in Large Vision-Language Models without Performance Degradationcited by 3
  22. 2026Kalman-Inspired Runtime Stability and Recovery in Hybrid Reasoning Systemscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Grounding Long-Context Reasoning with Contextual Normalization for Retrieval-Augmented Generation
  2. 2025Retrieval-Augmented Generation: A Comprehensive Survey of Architectures, Enhancements, and Robustness Frontiers
  3. 2025Principled Context Engineering for RAG: Statistical Guarantees via Conformal Prediction
  4. 2024Inference Scaling for Long-Context Retrieval Augmented Generation
  5. 2024From RAGs to rich parameters: Probing how language models utilize external knowledge over parametric information for factual queries
  6. 2025Parameters vs. Context: Fine-Grained Control of Knowledge Reliance in Language Models
  7. 2025Conflict-Aware Soft Prompting for Retrieval-Augmented Generation
  8. 2024Quantifying reliance on external information over parametric knowledge during Retrieval Augmented Generation (RAG) using mechanistic analysis