Research map / Large language models / Retrieval and knowledge
LUMINA: Detecting Hallucinations in RAG System with Context–Knowledge Signals
Samuel Yeh, Sharon Li, Tanwi Mallick
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use LUMINA instead of ReDeEP | ReDeEP · HalluRAG, Llama2-13B | AUROC: 0.7645 -> 0.9166 | Improved |
| use LUMINA instead of ReDeEP | ReDeEP · RAGTruth, Llama2-13B | AUROC: 0.8055 -> 0.8569 | Improved |
| use LUMINA instead of ReDeEP | ReDeEP · RAGTruth, Llama3-8B | AUROC: 0.7495 -> 0.7446 | Hurt |
| add entropy normalization to layer weighting in information processing rate (ablation) | LUMINA · RAGTruth, Mistral-7B | AUROC: 0.6111 -> 0.7679 | Improved |
| add layer weighting to entropy normalization (ablation) | LUMINA · RAGTruth, Llama2-7B | AUROC: 0.7164 -> 0.7652 | Improved |
| combine external context and internal knowledge scores (ablation) | LUMINA · RAGTruth, Llama2-13B | AUPRC | Improved |
| use RBF kernel instead of cosine kernel in MMD (ablation) | LUMINA external score · RAGTruth | AUPRC | No clear effect |
| vary top-k tokens for MMD approximation (50, 100, 500) (ablation) | LUMINA · RAGTruth | AUROC: <0.02% difference | No clear effect |
| vary random documents used for d' (ablation) | LUMINA · RAGTruth | AUROC: std <0.0025 | No clear effect |
| set lambda to 0.5 instead of tuned 0.2 (ablation) | LUMINA · RAGTruth, Llama2-7B and Llama3-8B | AUPRC: drop <0.025 | No clear effect |
| inject noise in retrieved and random documents (ablation) | LUMINA · RAGTruth, Llama2-13B | AUPRC: >0.1 drop | Hurt |
| inject noise in retrieved and random documents (ablation) | LUMINA · RAGTruth, other LLMs | AUPRC | No clear effect |
Closest papers by meaning
- ReDeEP: Detecting Hallucination in Retrieval-Augmented Generation via Mechanistic Interpretability (ICLR 2025 · similarity 0.95)
- LLM-Check: Investigating Detection of Hallucinations in Large Language Models (NeurIPS 2024 · similarity 0.93)
- HaDeMiF: Hallucination Detection and Mitigation in Large Language Models (ICLR 2025 · similarity 0.93)
- Scalable Token-Level Hallucination Detection in Large Language Models (NeurIPS 2026 · similarity 0.92)
- In-Context Sharpness as Alerts: An Inner Representation Perspective for Hallucination Mitigation (ICML 2024 · similarity 0.91)
- OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation (CVPR 2024 · similarity 0.91)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
