A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions

By Lei Huang, Weijiang Yu, Weitao Ma, Weihong Zhong, Zhangyin Feng, Haotian Wang, Qianglong Chen, Weihua Peng, Xiaocheng Feng, Bing Qin and 1 more. ACM Trans. Inf. Syst., 2023

3,329 citations recorded by the source corpus.

The emergence of large language models (LLMs) has marked a significant breakthrough in natural language processing (NLP), fueling a paradigm shift in information acquisition. Nevertheless, LLMs are prone to hallucination, generating plausible yet nonfactual content. This phenomenon raises significant concerns over the reliability of LLMs in real-world information retrieval (IR) systems and has attracted intensive research to detect and mitigate such hallucinations. Given the open-ended general-purpose attributes inherent to LLMs, LLM hallucinations present distinct challenges that diverge from prior task-specific models. This divergence highlights the urgency for a nuanced understanding and comprehensive overview of recent advances in LLM hallucinations. In this survey, we begin with an innovative taxonomy of hallucination in the era of LLM and then delve into the factors contributing to hallucinations. Subsequently, we present a thorough overview of hallucination detection methods and benchmarks. Our discussion then transfers to representative methodologies for mitigating LLM hallucinations. Additionally, we delve into the current limitations faced by retrieval-augmented LLMs in combating hallucinations, offering insights for developing more robust IR systems. Finally, we highlight the promising research directions on LLM hallucinations, including hallucination in large vision-language models and understanding of knowledge boundaries in LLM hallucinations.

Subfields: Artificial Intelligence, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/1e909e2a8cdacdcdff125ebcc566f37cb869a1c8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Electrical and Electronic Engineering

Citations in the corpus

A Survey on Hallucination in Large Vision-Language Models (2024, background): “A growing body of literature has explored various factors contributing to hallucinations in LLMs [Ji et al. , 2023; Zhang et al. , 2023b; Huang et al. , 2023a].”
Hallucination is Inevitable: An Innate Limitation of Large Language Models (2024, background): “For an extensive review of hallucination in LLMs, readers are directed to recent surveys [24, 29, 37, 46, 66].”
Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction (2024, neutral): “These models exhibit promising intelligence in generality and versatility that, despite issues like hallucinations [28], are still considered to take a solid step toward the general artificial intelligence (AGI).”
A Survey on the Memory Mechanism of Large Language Model-based Agents (2024, neutral): “One of the most prominent studies is Reflexion [138], which proposes verbal reinforcement learning for LLM-based agents.”
A framework to assess clinical safety and hallucination rates of LLMs for medical text summarisation (2025, neutral): “…these differ from traditional natural language processing (NLP) taxonomies 16 , which have separated hallucination types into distinct categories, for example, into “ intrinsic ” and “ extrinsic ” 24 , “ factuality ” and “ faithfulness ” 16 , “ factual mirage ” and “ silver lining ” 25 errors.”
AlphaEvolve: A coding agent for scientific and algorithmic discovery (2025, positive): “This evaluation mechanism allows AlphaEvolve to avoid any incorrect suggestions from the base LLM [44].”
Evaluation of Retrieval-Augmented Generation: A Survey (2024, background): “By retrieving relevant information from external sources, RAG significantly reduces the incidence of hallucinations [20] or factually incorrect generative outputs, thereby improving the content’s reliability and richness.”
MARS: Multi-hop Adaptive Retrieval and SPARQL Generation for KGQA (2026, unknown): “Large language models (LLMs) have achieved strong performance across diverse language and knowledge-intensive tasks [19,53,54], but remain prone to hallucinations [18,21], factual inconsistency [15], and multi-hop reasoning failures [25].”
ReMindRAG: Low-Cost LLM-Guided Knowledge Graph Traversal for Efficient RAG (2025, unknown): “Retrieval-Augmented Generation (RAG) [23] has become a prominent method for augmenting Large Language Models (LLMs) with external knowledge resources, enabling them to generate more accurate outputs and thereby expanding their practical utility [17, 32].”
Information Retrieval Using Large Language Models on a Restricted Access Environment (2026, unknown): “When unaware of the subject, it hallucinates [6].”
MG2-RAG: Multi-Granularity Graph for Multimodal Retrieval-Augmented Generation (2026, unknown): “In particular, MLLMs may produce multimodal hallucinations and often lack access to private or domain-specific knowledge, since their parameters are primarily trained on large-scale public corpora [33, 68, 79].”
Higher-Order Knowledge Representations for Agentic Scientific Reasoning (2026, unknown): “This architectural constraint manifests in well-documented failure modes including factual hallucination, temporal knowledge degradation, and diminished performance on long-tail or domain-specific queries [15–18].”
Mycroft - Retrieval Augmented Generation for SDK Documentation (2025, unknown): “When unaware of the subject, it hallucinates [5].”
ReservoirChat: Interactive Documentation Enhanced with LLM and Knowledge Graph for ReservoirPy (2025, unknown): “This phenomenon is called hallucination [10].”