Keywords
retrieval self-check
distractor density
paired simulation
reproducibility
Abstract
We evaluated retrieval self-check for retrieval-augmented answering under 26% distractor density. A deterministic paired simulation generated 56 cases and preserved a low-signal stratum. Mean evidence faithfulness changed from 0.578 to 0.619; the paired difference was +0.041 (95% interval +0.039 to +0.043). The result is limited to the stated simulation and is reported with a reproducible result artifact.
References
Sang, Y. (2025). Robustness of Fine-Tuned Llms Under Noisy Retrieval Inputs. 2025 6th International Conference on Artificial Intelligence and Electromechanical Automation (AIEA), 417-420. https://doi.org/10.1109/aiea66061.2025.11160531
Takamura, T., & Umezawa, A. (2025). Privacy-Preserving Retrieval-Augmented Generation on Local Devices for Regenerative Medicine Applications. https://doi.org/10.1101/2025.10.20.25337146
Kim, M. H. (2026). Pool-Gated Retrieval: Beyond Retrieval-Augmented Generation Toward Accountable Evidential Admission. https://doi.org/10.20944/preprints202606.0414.v1
Gai, Z., & Kuang, L. (2026). Adaptive Optimization for Retrieval-Augmented Generation via Diagnostic-Driven Strategy Generation. https://doi.org/10.36227/techrxiv.177155637.75354706/v1
Fu, E. L. (2025). Industrial AI Guardrails: Solving Field Service Reliability with Retrieval-Augmented Generation. Global Business & Economics Journal. https://doi.org/10.70924/f83n6wqz/g91xh08t
Wankhade, N. W. (2026). Explainable Retrieval-Augmented Generation Framework for Evidence-Aligned and Faithful Legal Reasoning. https://doi.org/10.21203/rs.3.rs-8947587/v1
Maharana, P. R., & Joshi, K. (2024). Retrieval augmented generation for building datasets from scientific literature. https://doi.org/10.26434/chemrxiv-2024-qjx32
Zandigohar, M., & Dai, Y. (2026). RAGulate: Retrieval-Augmented Generation for Post-hoc Literature-Grounded Regulatory Assessment. https://doi.org/10.64898/2026.01.20.700704
Ibiloye, A. C. (2025). Long Text Language Ensemble Models: Extension of Retrieval-Augmented Generation (RAG). https://doi.org/10.31219/osf.io/b94ng_v1
Joseph, A. (2025). Reducing Hallucinations in Large Language Models Through Integrated Self-Verification and Retrieval-Augmented Generation. Volume 2B: 45th Computers and Information in Engineering Conference (CIE), V02BT02A032. https://doi.org/10.1115/detc2025-169730