© 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. 3980
    Prev
    Next
    INDEX
    Explanations

    The neuron is acting as a crude positional marker, firing most strongly on tokens at the very start of the input.

    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
    тур
    0.71
    }$.
    0.70
    Gruß
    0.68
    ur
    0.68
    ណ៍
    0.67
     meaning
    0.66
    hal
    0.64
    的就是
    0.64
    ș
    0.64
    à
    0.63
    POSITIVE LOGITS
     Temos
    0.95
    सँग
    0.91
     ahuv
    0.89
     проблемы
    0.86
     ідэ
    0.85
    aparikkh
    0.85
     яе
    0.85
     дзяржа
    0.82
     прадстаў
    0.82
    ependence
    0.82
    Activations Density 0.000%

    No Known Activations