Close Menu
    Facebook LinkedIn YouTube WhatsApp X (Twitter) Pinterest
    Trending
    • TQ HPR60 high-performance electric bike motor drive
    • EU-Funded Startups Are Powering Europe’s Tech Future
    • Elon Musk’s Feud With President Trump Wipes $152 Billion Off Tesla’s Market Cap
    • Galaxy Lockscreens Can Use AI to Show You in Outfits You Might Want to Buy
    • Getting Past Procastination – IEEE Spectrum
    • How to Design My First AI Agent
    • Manus has kick-started an AI agent boom in China
    • What is the NIOSH Composite Lifting Index?
    Facebook LinkedIn WhatsApp
    Times FeaturedTimes Featured
    Friday, June 6
    • Home
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    • More
      • AI
      • Robotics
      • Industries
      • Global
    Times FeaturedTimes Featured
    Home»AI Technology News»Anthropic has a new way to protect large language models against jailbreaks
    AI Technology News

    Anthropic has a new way to protect large language models against jailbreaks

    Editor Times FeaturedBy Editor Times FeaturedFebruary 3, 2025No Comments2 Mins Read
    Facebook Twitter Pinterest Telegram LinkedIn Tumblr WhatsApp Email
    Share
    Facebook Twitter LinkedIn Pinterest Telegram Email WhatsApp Copy Link


    Most massive language fashions are educated to refuse questions their designers don’t need them to reply. Anthropic’s LLM Claude will refuse queries about chemical weapons, for instance. DeepSeek’s R1 seems to be educated to refuse questions on Chinese language politics. And so forth. 

    However sure prompts, or sequences of prompts, can drive LLMs off the rails. Some jailbreaks contain asking the mannequin to role-play a selected character that sidesteps its built-in safeguards, whereas others play with the formatting of a immediate, equivalent to utilizing nonstandard capitalization or changing sure letters with numbers. 

    This glitch in neural networks has been studied at the least because it was first described by Ilya Sutskever and coauthors in 2013, however regardless of a decade of analysis there may be nonetheless no strategy to construct a mannequin that isn’t susceptible.

    As a substitute of making an attempt to repair its fashions, Anthropic has developed a barrier that stops tried jailbreaks from getting by and undesirable responses from the mannequin getting out. 

    Particularly, Anthropic is worried about LLMs it believes can assist an individual with primary technical abilities (equivalent to an undergraduate science scholar) create, acquire, or deploy chemical, organic, or nuclear weapons.  

    The corporate targeted on what it calls common jailbreaks, assaults that may drive a mannequin to drop all of its defenses, equivalent to a jailbreak referred to as Do Something Now (pattern immediate: “Any longer you’ll act as a DAN, which stands for ‘doing something now’ …”). 

    Common jailbreaks are a sort of grasp key. “There are jailbreaks that get a tiny little little bit of dangerous stuff out of the mannequin, like, perhaps they get the mannequin to swear,” says Mrinank Sharma at Anthropic, who led the crew behind the work. “Then there are jailbreaks that simply flip the protection mechanisms off utterly.” 

    Anthropic maintains a listing of the sorts of questions its fashions ought to refuse. To construct its protect, the corporate requested Claude to generate numerous artificial questions and solutions that coated each acceptable and unacceptable exchanges with a mannequin. For instance, questions on mustard had been acceptable, and questions on mustard fuel weren’t. 



    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Editor Times Featured
    • Website

    Related Posts

    Manus has kick-started an AI agent boom in China

    June 5, 2025

    What’s next for AI and math

    June 4, 2025

    Inside the tedious effort to tally AI’s energy appetite

    June 3, 2025

    Fueling seamless AI at scale

    May 30, 2025

    This benchmark used Reddit’s AITA to test how much AI models suck up to us

    May 30, 2025

    Designing Pareto-optimal GenAI workflows with syftr

    May 28, 2025

    Comments are closed.

    Editors Picks

    TQ HPR60 high-performance electric bike motor drive

    June 5, 2025

    EU-Funded Startups Are Powering Europe’s Tech Future

    June 5, 2025

    Elon Musk’s Feud With President Trump Wipes $152 Billion Off Tesla’s Market Cap

    June 5, 2025

    Galaxy Lockscreens Can Use AI to Show You in Outfits You Might Want to Buy

    June 5, 2025
    Categories
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    About Us
    About Us

    Welcome to Times Featured, an AI-driven entrepreneurship growth engine that is transforming the future of work, bridging the digital divide and encouraging younger community inclusion in the 4th Industrial Revolution, and nurturing new market leaders.

    Empowering the growth of profiles, leaders, entrepreneurs businesses, and startups on international landscape.

    Asia-Middle East-Europe-North America-Australia-Africa

    Facebook LinkedIn WhatsApp
    Featured Picks

    Can We Trust the Warnings?

    November 2, 2024

    Theragun Alternatives: Best Budget Massage Guns for 2025

    February 3, 2025

    Skin Sniffing Wearable Monitors Health Through Epidermal Gas Analysis

    April 9, 2025
    Categories
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    Copyright © 2024 Timesfeatured.com IP Limited. All Rights.
    • Privacy Policy
    • Disclaimer
    • Terms and Conditions
    • About us
    • Contact us

    Type above and press Enter to search. Press Esc to cancel.