© 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. 6663
    Prev
    Next
    INDEX
    Explanations

    twenty years

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

    The neuron consistently activates on spelled-out numbers (e.g. “seven,” “twenty,” “four,” “thirty-three,” 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_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
    ução
    0.71
    sz
    0.70
    sdf
    0.70
     工具
    0.67
    서
    0.67
    ش
    0.66
    ok
    0.66
    Nich
    0.66
    \|=
    0.66
    POSITIVE LOGITS
     milliamperes
    0.85
     assorted
    0.81
     extortion
    0.79
    сот
    0.77
     interspersed
    0.76
     underprivileged
    0.75
     oversee
    0.75
    летний
    0.74
     eight
    0.73
     arthritis
    0.73
    Activations Density 0.015%

    No Known Activations