© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 3926
    Prev
    Next
    INDEX
    Explanations

    or conjunction

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

    The neuron specializes in spotting formally defined legal terms or proper‐name headings—i.e. the capitalized multiword phrases (like “User Generated Content,” “Carolina Custom Contractors,” “Services,” etc.) that mark defined terms in a terms‐of‐service or contract.

    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
    だったので
    0.78
    もあった
    0.75
     perché
    0.74
     எப்படி
    0.71
     защото
    0.71
    だった
    0.70
    でした
    0.70
     Еще
    0.68
     кілька
    0.68
     ಏಕೆ
    0.68
    POSITIVE LOGITS
     किंवा
    0.95
     અથવા
    0.94
     or
    0.92
    하거나
    0.92
    又は
    0.91
     أو
    0.89
     hoặc
    0.89
     или
    0.87
     veya
    0.87
     eller
    0.84
    Activations Density 0.019%

    No Known Activations