© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 3639
    Prev
    Next
    INDEX
    Explanations

    explaining breakdown or outlining approach

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron strongly activates on rare or nonstandard whitespace characters (e.g. figure spaces, thin spaces) in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ۰
    0.94
    0
    0.94
     ﺍﻟ
    0.88
    uuuu
    0.85
    BUILD
    0.84
    NOWLEDGE
    0.84
    tif
    0.83
    <unused2221>
    0.82
    FFER
    0.82
     ,
    0.81
    POSITIVE LOGITS
     Chances
    0.88
    我又
    0.87
     Admittedly
    0.87
     It
    0.86
    ائب
    0.82
    ून
    0.82
     There
    0.81
    ल
    0.81
    কিন্তু
    0.79
     Humble
    0.79
    Activations Density 0.024%

    No Known Activations