© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26
    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 127550
    Prev
    Next
    INDEX
    Explanations

    location or position

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

    This neuron activates on very common function words (e.g. “in,” “the,” “and,” “local,” etc.), effectively detecting high-frequency stopwords.

    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
     with
    -1.79
     on
    -1.66
     from
    -1.57
     began
    -1.34
     for
    -1.34
     was
    -1.33
     to
    -1.30
     or
    -1.30
     of
    -1.29
     including
    -1.29
    POSITIVE LOGITS
    何より
    1.56
     любой
    1.47
     любого
    1.40
     />
    1.39
     jusqu
    1.37
    family
    1.35
     ("%
    1.33
    こういった
    1.32
    このように
    1.30
    ここの
    1.30
    Activations Density 0.187%

    No Known Activations