# InterpDetect: Interpretable Signals for Detecting Hallucinations in Retrieval-Augmented Generation

arXiv: https://arxiv.org/abs/2510.21538 · HTML: https://mlsanity.com/paper/2510.21538/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Hallucination detection in RAG, one of 15 directions in Retrieval-augmented generation.
- Builds on 19 of its 25 references that are in the map.
- First posted on arXiv 24 Oct 2025.

## Placed in

- Retrieval-augmented generation / [Hallucination detection in RAG](https://mlsanity.com/rag/hallucination-detection-in-rag/)

## Builds on (most cited first, 19 of 19)

- [Retrieval-Augmented Generation for Large Language Models: A Survey](https://mlsanity.com/paper/2312.10997/) (2023)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [A Survey on Large Language Models for Code Generation](https://mlsanity.com/paper/2406.00515/) (2024)
- [Ragas: Automated Evaluation of Retrieval Augmented Generation](https://mlsanity.com/paper/2309.15217/) (2023)
- [RAGTruth: A Hallucination Corpus for Developing Trustworthy Retrieval-Augmented Language Models](https://mlsanity.com/paper/2401.00396/) (2023)
- [RAGBench: Explainable Benchmark for Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2407.11005/) (2024)
- [ReDeEP: Detecting Hallucination in Retrieval-Augmented Generation via Mechanistic Interpretability](https://mlsanity.com/paper/2410.11414/) (2024)
- [Neuron-Level Knowledge Attribution in Large Language Models](https://mlsanity.com/paper/2312.12141/) (2023)
- [LettuceDetect: A Hallucination Detection Framework for RAG Applications](https://mlsanity.com/paper/2502.17125/) (2025)
- [SemEval-2025 Task 3: Mu-SHROOM, the Multilingual Shared Task on Hallucinations and Related Observable Overgeneration Mistakes](https://mlsanity.com/paper/2504.11975/) (2025)
- [Retrieval Augmented Generation Evaluation in the Era of Large Language Models: A Comprehensive Survey](https://mlsanity.com/paper/2504.14891/) (2025)
- [LRP4RAG: Detecting Hallucinations in Retrieval-Augmented Generation via Layer-wise Relevance Propagation](https://mlsanity.com/paper/2408.15533/) (2024)
- [Understanding the Interplay between Parametric and Contextual Knowledge for Large Language Models](https://mlsanity.com/paper/2410.08414/) (2024)
- [InfiBench: Evaluating the Question-Answering Capabilities of Code Large Language Models](https://mlsanity.com/paper/2404.07940/) (2024)
- [Quantifying reliance on external information over parametric knowledge during Retrieval Augmented Generation (RAG) using mechanistic analysis](https://mlsanity.com/paper/2410.00857/) (2024)
- [FRED: Financial Retrieval-Enhanced Detection and Editing of Hallucinations in Language Models](https://mlsanity.com/paper/2507.20930/) (2025)
- [Large Language Models for Code Summarization](https://mlsanity.com/paper/2405.19032/) (2024)
- [MechIR: A Mechanistic Interpretability Framework for Information Retrieval](https://mlsanity.com/paper/2501.10165/) (2025)
- [HalluSearch at SemEval-2025 Task 3: A Search-Enhanced RAG Pipeline for Hallucination Detection](https://mlsanity.com/paper/2504.10168/) (2025)

## Built on by (most cited first, 1 of 1)

- [INTRYGUE: Induction-Aware Entropy Gating for Reliable RAG Uncertainty Estimation](https://mlsanity.com/paper/2603.21607/) (2026)

## Abstract (arXiv)

Retrieval-Augmented Generation (RAG) integrates external knowledge to mitigate hallucinations, yet models often generate outputs inconsistent with retrieved content. Accurate hallucination detection requires disentangling the contributions of external context and parametric knowledge, which prior methods typically conflate. We investigate the mechanisms underlying RAG hallucinations and find they arise when later-layer FFN modules disproportionately inject parametric knowledge into the residual stream. To address this, we explore a mechanistic detection approach based on external context scores and parametric knowledge scores. Using Qwen3-0.6b, we compute these scores across layers and attention heads and train regression-based classifiers to predict hallucinations. Our method is evaluated against state-of-the-art LLMs (GPT-5, GPT-4.1) and detection baselines (RAGAS, TruLens, RefChecker). Furthermore, classifiers trained on Qwen3-0.6b signals generalize to GPT-4.1-mini responses, demonstrating the potential of proxy-model evaluation. Our results highlight mechanistic signals as efficient, generalizable predictors for hallucination detection in RAG systems.