© 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
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11203
    Prev
    Next
    INDEX
    Explanations

    function return types

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

    The neuron activates on tokens representing formal legal or identification terms and sequences—e.g. “legal entity,” “individual,” and long numeric ID strings (tax/VAT/ABN numbers).

    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
     private
    0.73
    private
    0.63
    私
    0.62
     public
    0.61
     Private
    0.56
    至今
    0.55
     national
    0.55
     readonly
    0.55
     *)(
    0.55
     internal
    0.54
    POSITIVE LOGITS
    void
    1.83
     void
    1.75
    Void
    1.35
     voids
    1.22
     Void
    1.17
    Voids
    1.13
    VOID
    1.03
    IBAction
    0.73
    apses
    0.68
     cách
    0.68
    Activations Density 0.005%

    No Known Activations