© 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. 12997
    Prev
    Next
    INDEX
    Explanations

    requires('module')

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

    The neuron detects programming code keywords—e.g. tokens used to declare variables or define functions.

    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
    ส่
    0.55
    这是一
    0.55
     *(
    0.53
     прису
    0.52
    🫂
    0.52
    Н
    0.52
     Using
    0.50
    แห่ง
    0.50
    С
    0.50
    У
    0.50
    POSITIVE LOGITS
    cales
    0.77
     sequelize
    0.76
    mongoose
    0.75
     mongoose
    0.68
    alarının
    0.66
     Nodo
    0.63
    oli
    0.63
    érir
    0.61
    oses
    0.61
     mongo
    0.61
    Activations Density 0.003%

    No Known Activations