Why RAGs Hallucinate: Penalty-Aware Evaluation of Retrieval-Augmented Generation Systems with Knowledge-Gap Canaries

arXiv:2608.26385 · cs.CL, cs.AI · Submitted 2026-08-26 · Read on arXiv

cs.CL, cs.AI

Submitted: 2026-08-26

Updated: 2026-08-26

Comments: 13 pages, 1 figure, 5 tables. Code, full per-request logs, and all judge votes: https://github.com/adorosario/why-rags-hallucinate

Code: https://github.com/adorosario/why-rags-hallucinate

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• RAG — 82× in this paper · explained in 3 episodes
• Abstention — 32× in this paper · explained in 1 episode
• Accuracy — 26× in this paper · explained in 1 episode
• Benchmark — 14× in this paper · explained in 2 episodes
• Grounding — 14× in this paper · explained in 3 episodes
• Hallucination — 11× in this paper · explained in 3 episodes
• Consistency — 3× in this paper · explained in 3 episodes
• Retrieval-Augmented Generation (RAG) — 3× in this paper · explained in 16 episodes
• LLM judge — 2× in this paper · explained in 2 episodes
• Adam — 1× in this paper · explained in 2 episodes
• Calibration — 1× in this paper · explained in 7 episodes
• Cooperation — 1× in this paper · explained in 2 episodes
• Failure Taxonomy — 1× in this paper · explained in 2 episodes
• Generative Model — 1× in this paper · explained in 2 episodes
• Generative Models — 1× in this paper · explained in 2 episodes
• LLM-as-a-Judge — 1× in this paper · explained in 3 episodes
• Verification — 1× in this paper · explained in 2 episodes

Sources

• Ragas: Automated Evaluation of Retrieval Augmented Generation
• SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge
• Language Models (Mostly) Know What They Know
• Why Language Models Hallucinate
• HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models
• FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation
• LLM Evaluators Recognize and Favor Their Own Generations
• ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems
• Measuring short-form factuality in large language models
• Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering