Recent Comments

    LLM tagged posts

    Size doesn’t matter: Just a small number of malicious files can corrupt LLMs of any size

    October 13, 2025, Categories  Physics, Technology/Electronics

    Size doesn't matter: just a small number of malicious files can corrupt LLMs of any size
    Overview of our experiments, including examples of clean and poisoned samples, as well as benign and malicious behavior at inference time. (a)DoS pretraining backdoor experiments. Credit: arXiv (2025). DOI: 10.48550/arxiv.2510.07192

    Large language models (LLMs), which power sophisticated AI chatbots, are more vulnerable than previously thought. According to research by Anthropic, the UK AI Security Institute and the Alan Turing Institute, it only takes 250 malicious documents to compromise even the largest models.

    The vast majority of data used to train LLMs is scraped from the public internet. While this helps them to build knowledge and generate natural responses, it also puts them at risk from data poisoning attacks...

    Read More

    Cyber defense innovation could significantly boost 5G network security

    October 10, 2025, Categories  Physics, Technology/Electronics

    Breakthrough development could significantly boost 5G network security
    Proposed FedLLMGuard Architecture. Credit: University of Portsmouth

    A framework for building tighter security into 5G wireless communications has been created by a Ph.D. student working with the University of Portsmouth’s Artificial Intelligence and Data Center.

    With its greater network capacity and ability to rapidly transmit huge amounts of information from one device to another, 5G is a critical component of intelligent systems and services—including those for health care and financial services.

    However, the dynamic nature of 5G networks, the high volumes of data shared and the ever changing types of information transmitted means that these networks are extremely vulnerable to cyber threats and increasing risks of attack.

    Hadiseh Rezaei, a Ph.D...

    Read More

    Algorithm based on LLMs doubles lossless data compression rates

    May 16, 2025, Categories  Technology/Electronics

    A powerful lossless data compression algorithm based on LLMs
    Image comparing the lossless compression rates of LMCompress with the traditional state-of-the-art methods and the large-model-based method that was proposed independently by a DeepMind-Meta&INRIA team. The comparison is done on four types of data: image, video, audio, and text. It shows that LMCompress consistently outperforms the others on all data types. Note that the DeepMind result on video is not available. Credit: Li et al.

    People store large quantities of data in their electronic devices and transfer some of this data to others, whether for professional or personal reasons. Data compression methods are thus of the utmost importance, as they can boost the efficiency of devices and communications, making users less reliant on cloud data services and external storage devices.

    R...

    Read More

    Why GPT cannot think like us

    February 24, 2025, Categories  Physics, Technology/Electronics

    ChatGPT
    Credit: Unsplash/CC0 Public Domain

    Artificial Intelligence (AI), particularly large language models like GPT-4, has shown impressive performance on reasoning tasks. But does AI truly understand abstract concepts, or is it just mimicking patterns? A new study from the University of Amsterdam and the Santa Fe Institute reveals that while GPT models perform well on some analogy tasks, they fall short when the problems are altered, highlighting key weaknesses in AI’s reasoning capabilities. The work is published in Transactions on Machine Learning Research.

    Analogical reasoning is the ability to draw a comparison between two different things based on their similarities in certain aspects...

    Read More