© 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. 3135
    Prev
    Next
    INDEX
    Explanations

    language detection

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

    This neuron responds to paragraph or section breaks (i.e. blank‐line/formatting boundaries) in the text.

    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
     gør
    0.84
     maakt
    0.81
     (|
    0.75
     tonel
    0.75
     brinda
    0.73
     varios
    0.73
     funciona
    0.73
     equates
    0.73
     dahil
    0.73
     across
    0.72
    POSITIVE LOGITS
    गा
    0.64
    http
    0.63
    ACO
    0.61
    PASSWORD
    0.61
    hasht
    0.61
    U
    0.60
    ENU
    0.60
    ﺍ
    0.60
    жде
    0.59
    ttps
    0.58
    Activations Density 0.431%

    No Known Activations