© 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. 73005
    Prev
    Next
    INDEX
    Explanations

    code and structured text delimiters

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

    The neuron detects code block terminators (e.g. “End” in VB or loop/conditional closing keywords like “endfor”).

    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
     all
    -1.59
     every
    -1.24
    grafías
    -1.20
     as
    -1.16
    atisch
    -1.16
     {
    -1.14
    ！（
    -1.14
    INION
    -1.13
     Drž
    -1.13
     pener
    -1.11
    POSITIVE LOGITS
    1.31
    u
    1.27
    na
    1.23
     takich
    1.20
     dogged
    1.19
          
    1.18
    這裡
    1.17
    k
    1.14
    te
    1.13
    ani
    1.13
    Activations Density 0.001%

    No Known Activations