© 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. 34-GEMMASCOPE-RES-131K
    7. 47374
    Prev
    Next
    INDEX
    Explanations

    initialization, code structure, or introductions

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

    This neuron detects Java code declarations, especially the “public” access modifier and class or method definition keywords (e.g. public, class, extends).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    ઞ
    -0.89
     delitos
    -0.89
    hObject
    -0.84
     accep
    -0.84
    リュー
    -0.83
    Selain
    -0.82
    Office
    -0.82
    İş
    -0.81
    setScene
    -0.81
     Ampli
    -0.81
    POSITIVE LOGITS
     that
    0.97
     inflation
    0.86
     some
    0.85
     through
    0.85
    .......
    0.84
    isIn
    0.82
     those
    0.82
     measure
    0.81
     orientation
    0.81
    ......
    0.80
    Activations Density 0.015%

    No Known Activations