© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 8833
    Prev
    Next
    INDEX
    Explanations

    def questions or handle

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

    The neuron is a “segment‐start” detector—it fires most strongly on the very first substantive token of each new text block (i.e. at the start of each example/section).

    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
     судь
    0.92
     года
    0.88
    льных
    0.82
     кислоты
    0.82
    महारा
    0.81
     परम
    0.80
    ጴ
    0.80
     такой
    0.80
    arquía
    0.79
    льность
    0.79
    POSITIVE LOGITS
     trasfer
    0.76
     Place
    0.75
     Business
    0.75
     luogo
    0.75
    Place
    0.70
     Tropical
    0.69
    ਮ
    0.68
     μέρος
    0.68
     Lease
    0.67
     Regione
    0.67
    Activations Density 0.001%

    No Known Activations