© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 9822
    Prev
    Next
    INDEX
    Explanations

    what if, explore options

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

    This neuron activates on the introductory phrases of structured list items or section headings—especially when they begin with enumerators (like “The next thing noted…,” “For a major violation…,” “Milk intended…,” or transition cue words such as “Conversely,” “However,” “This channel deploys,” etc.).

    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
    e
    0.64
    POLY
    0.61
    ז
    0.60
    コン
    0.57
     и
    0.55
    fl
    0.55
    ன்ஸ்
    0.55
     and
    0.54
     ८
    0.54
    errorCode
    0.54
    POSITIVE LOGITS
     الأرض
    0.61
     .......
    0.60
     fiancée
    0.60
     情况
    0.60
     Yor
    0.59
     学校
    0.58
     défendre
    0.57
     Trường
    0.57
     organización
    0.56
     frère
    0.56
    Activations Density 0.422%

    No Known Activations