© 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. 72879
    Prev
    Next
    INDEX
    Explanations

    controller code

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

    This neuron is strongly triggered by code tokens that are PascalCase or otherwise start with an uppercase letter—e.g. class‐ or type‐style identifiers.

    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
    の説明
    -1.05
    assertArray
    -1.02
     quelques
    -0.97
     beschrieben
    -0.96
    隔壁
    -0.96
    Airplane
    -0.96
     mukaan
    -0.95
    utado
    -0.94
    Faj
    -0.93
     réessayer
    -0.93
    POSITIVE LOGITS
    Main
    1.12
     in
    1.07
    controller
    1.06
     which
    1.02
     doing
    0.98
     salmón
    0.98
    ♂
    0.97
     directly
    0.97
    alamu
    0.96
     controller
    0.96
    Activations Density 0.018%

    No Known Activations