© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 15094
    Prev
    Next
    INDEX
    Explanations

    none0AU[]

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

    The neuron fires on numeric literals or digits (i.e. any standalone numbers/version numbers) in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     straff
    -0.97
     география
    -0.97
    гови
    -0.92
    zionata
    -0.91
    <bos>
    -0.90
    Composição
    -0.89
     of
    -0.89
     manuelle
    -0.88
    _{{\
    -0.87
    瑣
    -0.87
    POSITIVE LOGITS
     none
    2.02
     no
    1.85
     nothing
    1.76
     neither
    1.58
    ไม่มี
    1.51
     هیچ
    1.50
    NONE
    1.42
     None
    1.41
     nenhum
    1.34
     żad
    1.30
    Activations Density 0.019%

    No Known Activations