© 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. 129252
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on function/method definition lines (e.g. “private void…”, “def …”, “func …”), i.e. it detects code that declares a function.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    code continuation keywords

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     will
    -1.23
     can
    -1.09
     or
    -1.04
     are
    -1.04
    }$,
    -1.02
     would
    -1.01
     otras
    -1.00
    return
    -0.99
     is
    -0.99
    而成
    -0.97
    POSITIVE LOGITS
    hockey
    1.03
     agrí
    1.01
     Comunidad
    1.00
     Política
    0.97
     catedral
    0.95
     profesora
    0.95
    stattung
    0.94
     Estación
    0.94
     Línea
    0.94
    ሬ
    0.93
    Activations Density 0.004%

    No Known Activations