© 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. 33052
    Prev
    Next
    INDEX
    Explanations

    code generation and identifiers

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

    The neuron consistently fires on capitalized identifiers—proper names, class/type tokens, and other PascalCase or uppercase-starting words.

    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
     Psic
    -0.99
     without
    -0.90
     постепенно
    -0.87
    鏄
    -0.85
     ständig
    -0.83
    kontrol
    -0.82
     mesta
    -0.81
    Đăng
    -0.81
    ówek
    -0.81
    地名
    -0.80
    POSITIVE LOGITS
    点头
    1.00
    GroupLayout
    0.96
    js
    0.96
    TokenNameDOT
    0.96
    Id
    0.95
    mrow
    0.93
    GeneratedMessage
    0.92
    還要
    0.91
     tendrán
    0.91
     is
    0.91
    Activations Density 0.021%

    No Known Activations