Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jan 9, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

A hallucination detection and mitigation framework for faithful text summarization using LLMs.

Shenling Liu¹, Yang Gao², ShaSha Li³

¹Education School, National University of Defense Technology, Deyalu Street, ChangSha, 410073, HuNan Province, China. liushenling@nudt.edu.cn.

Scientific Reports

|December 3, 2025

Summary

Related Concept Videos

Higher Mental Functions of the Brain: Language

Higher Mental Functions of the Brain: Language

Language is a system of communication that allows the expression of thoughts, ideas, and feelings. The brain processes language in both hemispheres.
Language formation and comprehension take place in the dominant hemisphere. The dominant hemisphere is responsible for understanding the meaning of spoken, written, or sign language, as well as the ability to communicate. For most people, the left hemisphere is the dominant one. The right hemisphere, then, gives tone and emotional context to the...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

A Lu<sub>3</sub>Sc<sub>2</sub>Ga<sub>3</sub>O<sub>12</sub>: Eu<sup>3+</sup> red-emitting phosphor with excellent optical properties and thermal stability was obtained by partially replacing Lu<sup>3+</sup> with Gd<sup>3+</sup> / Y<sup>3</sup>.

Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy·2024

Same author

Adaptative machine vision with microsecond-level accurate perception beyond human retina.

Nature communications·2024

Same author

Case report: A rare case of hereditary hemochromatosis caused by a mutation in the HAMP gene in Fuyang, China.

Frontiers in medicine·2024

Same author

Ficus hirta Vahl. ameliorates liver fibrosis by triggering hepatic stellate cell ferroptosis through GSH/GPX4 pathway.

Journal of ethnopharmacology·2024

Same author

Progestogen-driven B7-H4 contributes to onco-fetal immune tolerance.

Cell·2024

Same author

Alteration in lipid metabolism is involved in nitrogen deficiency response in wheat seedlings.

Plant physiology and biochemistry : PPB·2024

Same journal

Peripheral B-cell receptor repertoire predicts immune-related adverse events following immune checkpoint inhibitor therapy in advanced renal cell carcinoma.

Scientific reports·2026

Same journal

Effects of black soldier fly (Hermetia illucens L.) larvae zoocompost on the mineral element content of blue honeysuckle berries.

Scientific reports·2026

Same journal

Investigation on absorption refrigeration performance of R1243zf with imidazolium ionic liquid as the working pairs.

Scientific reports·2026

Same journal

DeepTriage-CN: integrating clinical text with vital signs for emergency department admission prediction in an aging population.

Scientific reports·2026

Same journal

Gold nanoparticles as dual-action antiviral agents: disruption of SARS-CoV-2 viral envelopes and RNA integrity.

Scientific reports·2026

Same journal

Comparison of capillary microsampling and venous blood for multi-pathogen serosurveillance.

Scientific reports·2026

See all related articles

This summary is machine-generated.

This study introduces a new framework to detect and reduce hallucinations in automatic text summarization using large language models (LLMs). The Question-Answer Generation, Sorting, and Evaluation (Q-S-E) method improves summary accuracy and user trust.

Area of Science:

Natural Language Processing
Artificial Intelligence
Machine Learning

Background:

Large language models (LLMs) have advanced automatic text summarization.
Hallucinations in summaries (information not in source text) reduce accuracy and user satisfaction.
Current methods struggle to detect and mitigate hallucinations effectively and transparently.

Purpose of the Study:

To introduce a novel framework for hallucination detection and mitigation in text summarization.
To quantitatively detect hallucinations using a Question-Answer Generation, Sorting, and Evaluation (Q-S-E) methodology.
To enhance the transparency and faithfulness of LLM-based summarization.

Main Methods:

Developed a hallucination detection and mitigation framework utilizing a Question-Answer Generation, Sorting, and Evaluation (Q-S-E) methodology.

Keywords:

Factual accuracy Hallucination detection Q-S-E methodology Text summarization

Related Experiment Videos

Last Updated: Jan 9, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Integrated an iterative hallucination resolution mechanism powered by LLMs.

Employed three benchmark datasets: CNN/Daily Mail, PubMed, and ArXiv for evaluation.

Main Results:

The proposed framework significantly improves the factual consistency of generated summaries.
The approach effectively preserves the informational completeness of the source text.
Experimental results demonstrate marked improvements across benchmark datasets.

Conclusions:

The Q-S-E methodology provides a quantitative approach to hallucination detection.
The iterative resolution mechanism enhances transparency and faithfulness in summarization.
This framework offers a promising solution to improve the reliability of automatic text summarization.