© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25244
    Prev
    Next
    INDEX
    Explanations

    preposition followed by verb or noun

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

    The neuron is activated by action‐oriented verbs describing procedural steps or changes (e.g., “deduct,” “transferred,” “processed,” “written,” “prove,” “hide,” “closed”).

    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
     bloc
    -0.94
     verre
    -0.92
     artistico
    -0.88
     ready
    -0.87
    唠
    -0.86
     rute
    -0.84
     ¤
    -0.84
    zembro
    -0.83
     agrico
    -0.82
    ならず
    -0.82
    POSITIVE LOGITS
     Getränke
    0.95
     pisze
    0.91
    ['
    0.90
    我已经
    0.87
     fueran
    0.87
    īg
    0.86
    Norway
    0.86
    vestidos
    0.86
    hlás
    0.86
     Су
    0.85
    Activations Density 0.070%

    No Known Activations