Latest Articles for Hallucinations

A new method improves detection of inaccuracies in language models.

SLaVA-CXR improves chest X-ray report generation for better clinical efficiency.

PACU framework enhances VLLMs by refining prompts and utilizing image captions.

Examining the accuracy issues in large language models and their societal effects.

A new framework improves detection of false outputs in language models using unlabeled data.

AI tools in healthcare offer benefits but raise significant safety concerns.

Assessing the impact of LLMs on healthcare documentation and safety.

Exploring how large language models can mislead users in medical advice.

Exploring how smaller models struggle with inaccuracies from larger counterparts.

RadFlag helps ensure AI-generated medical reports are accurate and trustworthy.

New method improves accuracy in vision-language models by reducing hallucination.

Research shows ways to enhance context awareness in language models for better responses.

A new dataset evaluates large language models for predicting material properties.

A look into new methods for enhancing trust in AI responses.

AI programs create personalities through interactions, forming unique identities in a virtual space.

New tool H-POPE improves accuracy of vision-language models.

New methods improve the accuracy of large language models.

New method improves classification of patient feedback through Global Consistency Assessment.

A look at why language models make up facts and how to fix it.

A study reveals how CHIME enhances ChatGPT's responses to software bug reports.

Research shows how smart captions help models describe images accurately.

A practical approach to minimize errors in AI using Enhanced Attention Heads.

A look at text-to-video models and their common mistakes.

VL-Uncertainty improves the detection of incorrect answers from digital models.

InstruGen enhances robot navigation with realistic instructions from YouTube videos.

A new framework enhances accuracy of AI in medical question-answering.

Understanding how Knowledge Graphs can reduce false information in AI responses.

This article discusses the issues of hallucinations in LVLMs and potential solutions.

VidHal benchmarks video models' ability to accurately interpret content.

Learn how knowledge-enhanced language models improve accuracy and reliability.

GraphJudger cleans data to improve Knowledge Graph creation from messy information.

New method reduces errors in AI image analysis and response generation.

New method enhances understanding and trust in multimodal models.

New tool aims to detect false patterns in virtual histology images.

Researchers find ways to reduce inaccuracies in large vision-language models.

Uncertainty quantification enhances the accuracy of automated radiology reports.

Research highlights the challenge of verb understanding in multimodal AI models.

EACO reduces AI errors and enhances reasoning for better performance.

Researchers develop a new method to improve text-to-image AI accuracy.

Discover the importance of uncertainty quantification in improving AI reliability.