© 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. 26166
    Prev
    Next
    INDEX
    Explanations

    international friendly vs. El

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

    The neuron fires on tokens that name formal entities or contexts—like dates, organizations, places, or other proper nouns and official‐sounding terms.

    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
     店舗
    -0.89
    López
    -0.88
     sardines
    -0.87
     $/
    -0.85
     arabes
    -0.84
    rát
    -0.84
    ومات
    -0.82
     optimize
    -0.82
    وبا
    -0.81
    🍧
    -0.81
    POSITIVE LOGITS
     his
    1.07
     when
    0.83
    ッピー
    0.82
    bunga
    0.81
    privateKey
    0.79
    高峰
    0.79
    ± 
    0.77
     nalazi
    0.77
     fellow
    0.77
     solo
    0.76
    Activations Density 0.139%

    No Known Activations