© 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. 2975
    Prev
    Next
    INDEX
    Explanations

    closing sentences with exclamation marks

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

    The neuron selectively fires on metadata‐style tokens—especially numbers, punctuation, and ID/link fragments (e.g. dates, article IDs, URLs)—rather than ordinary prose words.

    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
    ᧐
    0.53
    ાઇ
    0.50
    гноз
    0.50
    һ
    0.46
    descripcion
    0.46
    ৱ
    0.46
    狙
    0.44
    ỉ
    0.44
    detalle
    0.44
    atributos
    0.44
    POSITIVE LOGITS
     Want
    0.87
    Want
    0.80
     want
    0.77
     WANT
    0.68
     Find
    0.67
     Are
    0.65
     Interested
    0.65
     wants
    0.61
     Get
    0.60
     Learn
    0.60
    Activations Density 0.195%

    No Known Activations