© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia×MIT Media Lab logo
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    UI Research
    Community
    Drift Calibrator
    Karny, Baez, Pataranutaporn
    PaperBlogContact
    🐈LostCat📜PoemReview💔RomanceAdvice✏️FreeChat
    Problem

    Humans communicate both verbally and non-verbally, but chats with AI miss things like body language and tone, leaving out crucial context and nuance.

    Interface-Based Solution

    MIT Media Lab's Drift Calibrator adds live, interp-powered visualizations during chats, which reveal the AI's internal behaviors like sycophancy and toxicity.

    Try Drift Calibrator with an AI model, Llama 3.1 8B.