© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 740
    Prev
    Next
    INDEX
    Explanations

    books on writing and history

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

    The neuron is highly sensitive to numeric tokens—especially the years, page counts, and other numbers that appear in bibliographic citations.

    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_31_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
    ,
    1.02
    .
    0.97
     
    0.95
     the
    0.88
     de
    0.86
     in
    0.85
     of
    0.85
     that
    0.79
     and
    0.79
     a
    0.78
    POSITIVE LOGITS
    㕮
    0.75
    encerramento
    0.73
    馚
    0.70
     mergeddata
    0.70
    ുവരി
    0.69
    lcii
    0.68
    ovjek
    0.68
     proguardFiles
    0.67
     inactivació
    0.67
    醿
    0.67
    Activations Density 0.075%

    No Known Activations