© 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. 2394
    Prev
    Next
    INDEX
    Explanations

    punctuation or code structure

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

    The neuron consistently fires on numeric tokens (digits, measurements, and quantities).

    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
    :
    0.74
     Castel
    0.73
    <strong>
    0.71
     Kail
    0.69
     Floren
    0.69
     Jorge
    0.68
     Tenerife
    0.64
     Remedies
    0.63
     Dysfunction
    0.63
     Custom
    0.62
    POSITIVE LOGITS
     толькі
    0.66
    ্যাশ
    0.60
     മാത്രം
    0.57
    ູ
    0.57
    washed
    0.56
    cleaned
    0.55
    કો
    0.51
    এত
    0.50
     προϊόν
    0.50
    خواه
    0.50
    Activations Density 0.000%

    No Known Activations