© 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. 12907
    Prev
    Next
    INDEX
    Explanations

    ```javascript`

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

    The neuron primarily fires on numeric tokens—especially larger numbers or decimal‐style measurements—spotting when the text contains explicit numeric quantities.

    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
     {
    0.55
    AIM
    0.54
     menor
    0.54
    كام
    0.53
    Hoist
    0.53
     HOP
    0.52
     Pg
    0.51
    HOP
    0.51
    ]),
    0.50
    Hop
    0.50
    POSITIVE LOGITS
     ки
    0.59
     кі
    0.59
    മാണ
    0.58
    verständ
    0.57
    ㄹ
    0.57
     ezért
    0.56
    能不能
    0.56
    collectibles
    0.56
     പക്ഷ
    0.56
     क़
    0.54
    Activations Density 0.000%

    No Known Activations