© 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. 3739
    Prev
    Next
    INDEX
    Explanations

    begin or question markers

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

    This neuron strongly activates on standalone numeric tokens (numbers, especially with digits and decimal points).

    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
     deadlock
    0.59
    ：（
    0.54
    ,}$
    0.53
    ARCHIVO
    0.52
    .]
    0.52
    ENCY
    0.52
     adulter
    0.51
    ,",
    0.51
    :
    0.50
     độ
    0.50
    POSITIVE LOGITS
    zelfde
    0.64
    Similarly
    0.60
    as
    0.59
     Similarly
    0.57
    Regards
    0.56
     Tige
    0.56
     Looks
    0.55
     Lobkovic
    0.55
     Treated
    0.55
    in
    0.54
    Activations Density 0.078%

    No Known Activations