Yang Zhang

Examining the rise and risks of jailbreak prompts in language models.

A new method for lowering expenses in machine learning language services.

Analyzing how step count data can expose personal information and privacy risks.

New methods aim to tackle online toxicity using prompt learning techniques.

Study reveals vulnerabilities in updated language models against adversarial attacks.

Test-time adaptation methods face vulnerabilities from poisoning attacks, challenging their effectiveness.

Evaluating feature attribution methods through soundness and completeness for better AI predictions.

Researchers reveal La Ni O, a new superconductor that operates under high pressure.

Addressing privacy threats during digital twin migrations in smart transportation.

Enhancing vehicle interactions through optimized migration and resource management.

New model improves speech recognition in noisy environments by focusing on a single speaker.

Exploring golden cutting points to enhance quantum circuit efficiency.

Modern language models enhance text normalization accuracy and quality.

Research explores how language models enhance planning for multiple robots in complex tasks.

This article discusses the privacy concerns of using GPT models in cloud settings.

Recent observations reveal new findings about Pluto's atmosphere and pressure changes.

Examination of jailbreak attacks shows weaknesses in language model safety.

Custom LLMs raise safety concerns, particularly with instruction backdoor attacks.

Exploring the security challenges of prompt engineering with LLMs.

A new method enhances clarity in ore image segmentation for better processing.

Diverse samples enhance the effectiveness of machine learning model theft.

Examining the necessity of hydrogen in superconducting nickelates and its implications.

A new framework assesses the effectiveness of image safety classifiers against harmful content.

Inductive GNNs face privacy threats from link stealing attacks.

ReAd enhances robot teamwork by refining planning and feedback mechanisms.

A method to reduce the size of large language models while maintaining their performance.

A new framework enhances train fault detection using advanced deep learning techniques.

Seed-TTS creates lifelike speech from text for various applications.

A new method improves Large Language Models' ability to forget sensitive information.

A new model enhances depth estimation accuracy using self-supervised learning techniques.

Examining qubit interactions through correlation functions in integrable quantum circuits.

Forward Gradient Unrolling offers efficient solutions for optimizing large machine learning models.

A new framework enhances learning efficiency in multi-agent systems through innovative world modeling.

Research reveals language models struggle with false reasoning, raising safety concerns.

A new model improves accuracy in speech-to-text capabilities across multiple languages.

A new framework controls in-context learning to prevent misuse in AI models.

A study introduces SeqMIA to improve privacy against membership inference attacks.

PandaX-4T aims to detect solar B neutrinos and explore dark matter.

Examining the threats posed by autonomous language model agents and their weaknesses.

Examining the risks and strategies of model hijacking in federated learning systems.

Exploring privacy risks in masked image modeling and their implications.

HiFuse enhances the speed and efficiency of heterogeneous graph neural networks.

PandaX-4T searches for new particles to explain cosmic mysteries.

Examining the effects of updates on safety, bias, and authenticity in image generation.

A new dataset aims to improve human motion tracking in virtual environments.

Examining how important data points attract more security risks in machine learning.

PandaX-4T experiment searches for axion-like particles and dark photons.

A system making music creation easy and accessible for all skill levels.

Researchers probe boson production with photons, revealing crucial interactions in particle physics.

This article examines how different layers affect LLM performance.

A novel approach improves the detection of genuine signatures against forgeries.

A novel method for training agents to adapt across various tasks using diverse experiences.

TRIDENT enhances machine recognition of new object combinations, addressing key challenges in AI learning.

Some materials become messier when cooled, showcasing inverse melting.

A new method for intuitive control of robotic arms using Augmented Reality.

Scientists study rare neutrino processes to reveal cosmic mysteries.

A new method for efficiently pruning image-generating AI models while preserving quality.

Scientists investigate the Higgs boson's interactions with bottom quarks at the LHC.

Connecting social media insights to drought impacts for better understanding.

Scientists search for rare neutrino decay to unravel cosmic secrets.

Scientists investigate dark matter particles energized by the Sun.