© 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. 73110
    Prev
    Next
    INDEX
    Explanations

    input/output operations

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

    The neuron consistently fires on the token “io” (e.g. I/O package or module names), so it is detecting occurrences of the substring “io.”

    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
     as
    -2.03
    and
    -2.00
    of
    -1.92
    ’
    -1.88
     when
    -1.84
    S
    -1.84
    ’,
    -1.84
     (‘
    -1.80
     where
    -1.73
    $,
    -1.73
    POSITIVE LOGITS
    
    2.16
     Anſ
    1.95
     AVAILABLE
    1.92
    荸
    1.91
    翃
    1.87
    сети
    1.84
     Inſ
    1.80
    对付
    1.80
     Afrique
    1.77
     MÉXICO
    1.76
    Activations Density 0.053%

    No Known Activations