© 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. 127996
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on quoted string literals in code (e.g. attribute values like "_token" or JavaScript strings like 'custom').

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

    name="attribute" tags

    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
    ALLES
    -1.02
    NotBe
    -0.96
    
    -0.94
    つは
    -0.93
    Estás
    -0.92
    Público
    -0.92
    ksamkeit
    -0.88
    ledad
    -0.88
     широкий
    -0.88
     with
    -0.86
    POSITIVE LOGITS
     برگزار
    0.96
     λό
    0.93
    原始内容存档于
    0.92
     city
    0.88
    ője
    0.86
     odc
    0.85
    ","
    0.83
    ORPG
    0.82
    RY
    0.82
    ]=="
    0.82
    Activations Density 0.002%

    No Known Activations