© 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. 2077
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on various inflected forms of the verb “take” (e.g. “take,” “taking,” “took”).

    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
     blazing
    0.57
     lingered
    0.56
     حامل
    0.55
     lành
    0.55
    悭
    0.55
     barcodes
    0.55
     usher
    0.54
    $'
    0.54
    lains
    0.54
     tapped
    0.52
    POSITIVE LOGITS
     advantage
    1.18
    advantage
    1.08
     Seriously
    1.00
     aback
    0.98
     Turns
    0.96
     seriously
    0.93
     care
    0.92
     shortcuts
    0.92
     Advantage
    0.92
     ventaja
    0.91
    Activations Density 0.319%

    No Known Activations