# DeCoRe: Decoding by Contrasting Retrieval Heads to Mitigate Hallucinations

arXiv: https://arxiv.org/abs/2410.18860 · HTML: https://mlsanity.com/paper/2410.18860/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Builds on 50 of its 73 references that are in the map.
- First posted on arXiv 24 Oct 2024.

## Builds on (most cited first, 30 of 50)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [Mistral 7B](https://mlsanity.com/paper/2310.06825/) (2023)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [Survey of Hallucination in Natural Language Generation](https://mlsanity.com/paper/2202.03629/) (2022)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization](https://mlsanity.com/paper/1808.08745/) (2018)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [MuSiQue: Multihop Questions via Single-hop Question Composition](https://mlsanity.com/paper/2108.00573/) (2021)
- [Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Models](https://mlsanity.com/paper/2309.01219/) (2023)
- [Latent Retrieval for Weakly Supervised Open Domain Question Answering](https://mlsanity.com/paper/1906.00300/) (2019)
- [Contrastive Decoding: Open-ended Text Generation as Optimization](https://mlsanity.com/paper/2210.15097/) (2022)
- [Knowledge Neurons in Pretrained Transformers](https://mlsanity.com/paper/2104.08696/) (2021)
- [BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks](https://mlsanity.com/paper/1709.01686/) (2017)
- [Dissecting Recall of Factual Associations in Auto-Regressive Language Models](https://mlsanity.com/paper/2304.14767/) (2023)
- [Large Language Models Struggle to Learn Long-Tail Knowledge](https://mlsanity.com/paper/2211.08411/) (2022)

## Built on by (most cited first, 25 of 25)

- [Steering off Course: Reliability Challenges in Steering Language Models](https://mlsanity.com/paper/2504.04635/) (2025)
- [PiCSAR: Probabilistic Confidence Selection And Ranking for Reasoning Chains](https://mlsanity.com/paper/2508.21787/) (2025)
- [HICD: Hallucination-Inducing via Attention Dispersion for Contrastive Decoding to Mitigate Hallucinations in Large Language Models](https://mlsanity.com/paper/2503.12908/) (2025)
- [Loki's Dance of Illusions: A Comprehensive Survey of Hallucination in Large Language Models](https://mlsanity.com/paper/2507.02870/) (2025)
- [Revisiting Long-context Modeling from Context Denoising Perspective](https://mlsanity.com/paper/2510.05862/) (2025)
- [CAFE: Retrieval Head-based Coarse-to-Fine Information Seeking to Enhance Multi-Document QA Capability](https://mlsanity.com/paper/2505.10063/) (2025)
- [Representation-based Broad Hallucination Detectors Fail to Generalize Out of Distribution](https://mlsanity.com/paper/2509.19372/) (2025)
- [Understanding and Mitigating Hallucinations in Multimodal Chain-of-Thought Models](https://mlsanity.com/paper/2603.27201/) (2026)
- [An Analysis of Decoding Methods for LLM-based Agents for Faithful Multi-Hop Question Answering](https://mlsanity.com/paper/2503.23415/) (2025)
- [Preference Heads in Large Language Models: A Mechanistic Framework for Interpretable Personalization](https://mlsanity.com/paper/2604.22345/) (2026)
- [Noise Augmented Fine Tuning for Mitigating Hallucinations in Large Language Models](https://mlsanity.com/paper/2504.03302/) (2025)
- [Think Less, Know More: State-Aware Reasoning Compression with Knowledge Guidance for Efficient Reasoning](https://mlsanity.com/paper/2604.09150/) (2026)
- [Context-Fidelity Boosting: Enhancing Faithful Generation through Watermark-Inspired Decoding](https://mlsanity.com/paper/2604.22335/) (2026)
- [Hallucinations as Orthogonal Noise: Inference-Time Manifold Alignment via Dynamic Contextual Orthogonalization](https://mlsanity.com/paper/2606.03022/) (2026)
- [SSFO: Self-Supervised Faithfulness Optimization for Retrieval-Augmented Generation](https://mlsanity.com/paper/2508.17225/) (2025)
- [Temporal Guidance for Large Language Models](https://mlsanity.com/paper/2601.21744/) (2026)
- [Listen to the Layers: Mitigating Hallucinations with Inter-Layer Disagreement](https://mlsanity.com/paper/2602.09486/) (2026)
- [A Systematic Study of Training-Free Methods for Trustworthy Large Language Models](https://mlsanity.com/paper/2604.15789/) (2026)
- [APCD: Adaptive Path-Contrastive Decoding for Reliable Large Language Model Generation](https://mlsanity.com/paper/2605.09492/) (2026)
- [FIDES: Faithful Inference via Deep Evidence Signals for Retrieval-Memory Conflict in RAG](https://mlsanity.com/paper/2606.05644/) (2026)
- [Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads](https://mlsanity.com/paper/2607.01002/) (2026)
- [Beyond Tokens: A Survey on Decoding Methods for Large Language and Vision-Language Models](https://mlsanity.com/paper/2608.14797/) (2026)
- [Information-Guided Frontier Decoding: Contextual Utility-Driven Commitment in dMLLMs](https://mlsanity.com/paper/2608.26641/) (2026)
- [Invisible Ink, Visible Lies: How Production Watermarking Causes LLMs to Hallucinate](https://mlsanity.com/paper/2610.04860/) (2026)
- [MAPLE: A Multi-Axis Parametric Knowledge Extraction Lens for LLM Evaluation](https://mlsanity.com/paper/2610.19827/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) often hallucinate, producing unfaithful or factually incorrect outputs by misrepresenting the provided context or incorrectly recalling internal knowledge. Recent studies have identified specific attention heads within the Transformer architecture, known as retrieval heads, responsible for extracting relevant contextual information. We hypothesise that masking these retrieval heads can induce hallucinations and that contrasting the outputs of the base LLM and the masked LLM can reduce hallucinations. To this end, we propose Decoding by Contrasting Retrieval Heads (DeCoRe), a novel training-free decoding strategy that amplifies information found in the context and model parameters. DeCoRe mitigates potentially hallucinated responses by dynamically contrasting the outputs of the base LLM and the masked LLM, using conditional entropy as a guide. Our extensive experiments confirm that DeCoRe significantly improves performance on tasks requiring high contextual faithfulness, such as summarisation (XSum by 18.6%), instruction following (MemoTrap by 10.9%), and open-book question answering (NQ-Open by 2.4% and NQ-Swap by 5.5%).