© 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. 12346
    Prev
    Next
    INDEX
    Explanations

    secret version code block

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

    This neuron is tuned to detect numeric tokens—especially parts of floating-point or otherwise unusually large numbers—embedded in the text.

    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
    tection
    0.78
     wavelet
    0.70
     xay
    0.70
    ら
    0.70
     diel
    0.68
     matriline
    0.68
     diodes
    0.68
     Purification
    0.68
     semigroup
    0.67
     quantification
    0.67
    POSITIVE LOGITS
    헴
    0.82
    합니다
    0.73
    $).
    0.70
    ੀ
    0.69
    Қ
    0.69
    )\}$
    0.68
    κό
    0.67
    하게
    0.66
     참석
    0.66
    cómo
    0.66
    Activations Density 0.000%

    No Known Activations