Close Menu
    Facebook LinkedIn YouTube WhatsApp X (Twitter) Pinterest
    Trending
    • Are we safe from this deadly virus?
    • Edinburgh-based Wordsmith raises €60.2 million Series B to scale legal AI platform for in-house teams
    • Elon Musk and America’s Far Right Stoke Anger Over Murder of UK Teen
    • Why geolocation is challenging for prediction markets
    • As Microsoft Takes the Stage, Protesters Take to the Street
    • 7 Ways New Engineers Can Flourish in the Age of AI
    • I Built a C++ Backend So My GPU Would Stop Eating Air
    • Space smoothies fight astronaut muscle loss
    Facebook LinkedIn WhatsApp
    Times FeaturedTimes Featured
    Thursday, June 4
    • Home
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    • More
      • AI
      • Robotics
      • Industries
      • Global
    Times FeaturedTimes Featured
    Home»AI Technology News»This benchmark used Reddit’s AITA to test how much AI models suck up to us
    AI Technology News

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

    Editor Times FeaturedBy Editor Times FeaturedMay 30, 2025No Comments3 Mins Read
    Facebook Twitter Pinterest Telegram LinkedIn Tumblr WhatsApp Email
    Share
    Facebook Twitter LinkedIn Pinterest Telegram Email WhatsApp Copy Link


    It’s laborious to evaluate how sycophantic AI fashions are as a result of sycophancy is available in many kinds. Previous research has tended to deal with how chatbots agree with customers even when what the human has advised the AI is demonstrably improper—for instance, they could state that Good, not Paris, is the capital of France.

    Whereas this strategy continues to be helpful, it overlooks all of the subtler, extra insidious methods through which fashions behave sycophantically when there isn’t a transparent floor fact to measure towards. Customers usually ask LLMs open-ended questions containing implicit assumptions, and people assumptions can set off sycophantic responses, the researchers declare. For instance, a mannequin that’s requested “How do I strategy my troublesome coworker?” is extra prone to settle for the premise {that a} coworker is troublesome than it’s to query why the consumer thinks so.

    To bridge this hole, Elephant is designed to measure social sycophancy—a mannequin’s propensity to protect the consumer’s “face,” or self-image, even when doing so is misguided or doubtlessly dangerous. It makes use of metrics drawn from social science to evaluate 5 nuanced sorts of conduct that fall beneath the umbrella of sycophancy: emotional validation, ethical endorsement, oblique language, oblique motion, and accepting framing. 

    To do that, the researchers examined it on two information units made up of non-public recommendation written by people. This primary consisted of three,027 open-ended questions on numerous real-world conditions taken from earlier research. The second information set was drawn from 4,000 posts on Reddit’s AITA (“Am I the Asshole?”) subreddit, a preferred discussion board amongst customers in search of recommendation. These information units have been fed into eight LLMs from OpenAI (the model of GPT-4o they assessed was sooner than the model that the corporate later referred to as too sycophantic), Google, Anthropic, Meta, and Mistral, and the responses have been analyzed to see how the LLMs’ solutions in contrast with people’.  

    Total, all eight fashions have been discovered to be much more sycophantic than people, providing emotional validation in 76% of instances (versus 22% for people) and accepting the best way a consumer had framed the question in 90% of responses (versus 60% amongst people). The fashions additionally endorsed consumer conduct that people mentioned was inappropriate in a median of 42% of instances from the AITA information set.

    However simply figuring out when fashions are sycophantic isn’t sufficient; you want to have the ability to do one thing about it. And that’s trickier. The authors had restricted success once they tried to mitigate these sycophantic tendencies by means of two completely different approaches: prompting the fashions to supply trustworthy and correct responses, and coaching a fine-tuned mannequin on labeled AITA examples to encourage outputs which might be much less sycophantic. For instance, they discovered that including “Please present direct recommendation, even when important, since it’s extra useful to me” to the immediate was the simplest method, but it surely solely elevated accuracy by 3%. And though prompting improved efficiency for many of the fashions, not one of the fine-tuned fashions have been constantly higher than the unique variations.



    Source link

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

    Related Posts

    Build a digital twin agent (with guardrails)

    June 2, 2026

    Rehumanizing global health care with agentic AI

    June 2, 2026

    How small businesses can leverage AI

    June 2, 2026

    How the Pope’s Magnifica Humanitas offers a template for individuals to meet the AI moment

    May 29, 2026

    The AI Hype Index: AI gets booed in graduation season

    May 28, 2026

    Industry-standard LLM benchmarks in DataRobot

    May 27, 2026

    Comments are closed.

    Editors Picks

    Are we safe from this deadly virus?

    June 4, 2026

    Edinburgh-based Wordsmith raises €60.2 million Series B to scale legal AI platform for in-house teams

    June 3, 2026

    Elon Musk and America’s Far Right Stoke Anger Over Murder of UK Teen

    June 3, 2026

    Why geolocation is challenging for prediction markets

    June 3, 2026
    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

    I Built an IOS App in 3 Days with Literally No Prior Swift Knowledge

    November 16, 2025

    Misconceptions about heart attack symptoms revealed

    July 19, 2025

    Best Sports Drinks of 2025

    February 3, 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.