© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 40550
    Prev
    Next
    INDEX
    Explanations

    code control structures

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

    The neuron fires on string literals (especially all-caps or mixed-case words in quotes) used as constants or comparison targets in code (e.g. “contact”, “Screen print”, “EMAIL_ADDRESS”, “Embossing”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     enorme
    -1.47
     الجيش
    -1.27
    ∈
    -1.27
    new
    -1.25
    name
    -1.24
    cluding
    -1.24
     sinds
    -1.23
    beta
    -1.23
     вернуть
    -1.20
    ter
    -1.20
    POSITIVE LOGITS
     this
    2.41
     there
    2.11
     or
    2.09
     этого
    1.99
     that
    1.95
     when
    1.84
     aufmerksam
    1.55
     Fußballspieler
    1.50
     químico
    1.49
     permitirá
    1.46
    Activations Density 0.014%

    No Known Activations