© 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. 31-GEMMASCOPE-2-RES-16K
    7. 13396
    Prev
    Next
    INDEX
    Explanations

    dark matter, DCF analysis, dark fantasy, dark humor

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

    The neuron selectively responds to tokens that begin with a capital “D” (e.g. Douglas, Decades, DC, Dakari, etc.).

    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
    WELL
    0.65
    крови
    0.64
    ้น
    0.62
    <unused83>
    0.61
    χος
    0.61
     काला
    0.59
     এবারের
    0.59
     பின்பற்ற
    0.57
    <unused53>
    0.57
    んは
    0.56
    POSITIVE LOGITS
     Dec
    1.13
    dec
    1.09
     dec
    1.03
    Dec
    1.03
     DEC
    1.02
    deco
    1.00
    Dek
    0.96
     deco
    0.95
    dek
    0.94
     Dek
    0.90
    Activations Density 0.269%

    No Known Activations