© 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. 64775
    Prev
    Next
    INDEX
    Explanations

    Eureka Village On a The Atlanta

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

    The neuron fires on code‐style structural tokens—especially open‐parenthesis characters (and related tokens like “new” and import/module identifiers)—i.e. it detects function‐call and similar syntax in source code.

    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
     Kiedy
    -1.00
     váll
    -1.00
     znajdziesz
    -0.98
    みます
    -0.98
     három
    -0.96
     aparentemente
    -0.96
     Jeśli
    -0.95
    僰
    -0.91
     napisał
    -0.90
    aried
    -0.90
    POSITIVE LOGITS
     of
    1.16
     from
    1.15
     to
    1.04
    其他的
    0.93
     februar
    0.92
    TRIBUTES
    0.92
     literatura
    0.91
     хотите
    0.91
    bbean
    0.90
    เตอร์
    0.90
    Activations Density 0.000%

    No Known Activations