© 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-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 26112
    Prev
    Next
    INDEX
    Explanations

    scores with parentheses or commas

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

    This neuron fires on numeric score values in tennis match results (e.g. set scores like “6-4,” tiebreak points in parentheses, etc.).

    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
     like
    -1.67
     that
    -1.62
     who
    -1.59
     from
    -1.48
     it
    -1.45
     of
    -1.39
     their
    -1.38
     on
    -1.38
     which
    -1.31
    さんに
    -1.25
    POSITIVE LOGITS
     作り方
    1.48
    COMO
    1.33
    Kamu
    1.32
    がありますが
    1.30
    Gön
    1.27
    Karakter
    1.26
    Relig
    1.24
    TÜ
    1.23
     humeur
    1.23
    Warna
    1.23
    Activations Density 0.003%

    No Known Activations