© 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. 5804
    Prev
    Next
    INDEX
    Explanations

    actions and studies

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

    The neuron responds strongly to numerals and numeric measurements (e.g. statistics, years, decimal values).

    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
     wasn
    0.66
    isn
    0.61
     isn
    0.58
     shouldn
    0.58
     Wasn
    0.58
     Shouldn
    0.57
     usually
    0.57
    :
    0.57
    并没有
    0.55
     आमतौर
    0.55
    POSITIVE LOGITS
     across
    0.55
     throughout
    0.55
     from
    0.54
     през
    0.50
     undertakes
    0.50
     favours
    0.48
     during
    0.48
     decomposes
    0.47
     aspires
    0.46
     meets
    0.46
    Activations Density 0.236%

    No Known Activations