"Local RAG Refusal Calibration Benchmark and Results"
Abstract
"This dataset provides the reproducibility artifact for the IEEE Access manuscript \"When Good Retrieval Is Not Enough: Prompt Design Outperforms Retrieval-Score Gating in Local RAG Refusal.\" It contains Python code, tests, derived evaluation instances, cached local model outputs, summary result files, statistical outputs, and verification scripts for a controlled local retrieval-augmented generation evaluation. The evaluation covers SQuAD v2, HotpotQA, and a bounded Natural Questions stress test. The artifact supports verification of coverage, unsupported-answer risk, utility, latency, and paired statistical results reported in the manuscript. The included scripts allow users to verify the provided results without rerunning generation, or rerun the full experiment with a local Ollama server."
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.