© 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. 10-GEMMASCOPE-RES-131K
    7. 84130
    Prev
    Next
    INDEX
    Explanations

    HTML label attributes

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

    This neuron detects colon tokens used as key‐value or section‐header separators (i.e. the “:” marking the start of a new field or block).

    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
    "
    -3.80
    $
    -3.59
    *
    -3.41
    what
    -3.14
    le
    -2.97
    When
    -2.97
    \
    -2.88
    For
    -2.86
    ers
    -2.84
    I
    -2.83
    POSITIVE LOGITS
    荩
    3.16
     masterful
    2.94
    archiwizowane
    2.89
     terrific
    2.86
    
    2.83
     grotesque
    2.81
    哙
    2.81
     marvelous
    2.78
    嵛
    2.73
    鈇
    2.72
    Activations Density 0.001%

    No Known Activations