© 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. 24-GEMMASCOPE-2-RES-16K
    7. 1393
    Prev
    Next
    INDEX
    Explanations

    patriotic quotations and historical figures

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

    The neuron fires strongly on the very first substantive word of each new comment or document segment, effectively marking segment beginnings.

    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_24_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
    荐
    0.68
    anée
    0.67
    gini
    0.67
    effic
    0.66
    neos
    0.66
    支撑
    0.65
     кори
    0.63
    ācijas
    0.63
    ી
    0.63
    U
    0.63
    POSITIVE LOGITS
     imitating
    1.02
     überprüft
    0.98
     burrito
    0.94
     Zidane
    0.93
     Aladdin
    0.93
    EventRecorder
    0.93
     rafting
    0.91
     limpar
    0.91
     interrupts
    0.90
     wholeheartedly
    0.90
    Activations Density 0.000%

    No Known Activations