HITL-RAG Faithfulness Evaluation Dataset for Cultural Tourism Information (Version 1.1)
Abstract
This dataset supports the evaluation of faithfulness in Retrieval-Augmented Generation (RAG) systems applied to cultural tourism information, with an explicit focus on Human-in-the-Loop (HITL) governance. It is designed as an evaluation artifact rather than a performance benchmark, emphasizing traceability, grounding fidelity, and decision transparency. The dataset consists of structured query–response evaluations, including retrieved text chunks, candidate answers generated by a RAG-based language model, reference explanations produced by an LLM-based judge, automatic similarity metrics, and final human evaluation decisions. Faithfulness is assessed with respect to curated cultural knowledge chunks, while system-level fallback and failure-trapping mechanisms are explicitly separated from content-level grounding. A structured HITL error taxonomy (E0–E10) is applied to distinguish faithful responses, controlled abstentions, content errors, and system failures. Repeated queries are intentionally included to observe consistency across evaluation cycles, allowing analysis of generative variation while maintaining stable classification outcomes under human judgment. Versioning follows a snapshot-based policy to ensure reproducibility and cross-version comparison. This release (Version 1.1) is deposited under embargo and will be made publicly accessible following completion of the associated journal publication.
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.