© 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. 34-GEMMASCOPE-RES-131K
    7. 31119
    Prev
    Next
    INDEX
    Explanations

    Crosse Alerts

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

    The neuron is primarily activated by capitalized tokens in headings and named entities—e.g. names of people, places, organizations, dates—essentially picking out proper-noun or title-style tokens.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Mop
    -0.85
    restes
    -0.83
     крестьян
    -0.78
     dreapta
    -0.77
    scrapy
    -0.75
     Harz
    -0.75
    centerline
    -0.75
    agliari
    -0.74
    いただける
    -0.72
     Wordle
    -0.70
    POSITIVE LOGITS
    дови
    0.76
    焼け
    0.74
     Cou
    0.71
     Amish
    0.71
    >`;
    0.69
    stā
    0.69
    DataOutputStream
    0.68
    wk
    0.68
    目指
    0.67
     drifting
    0.67
    Activations Density 0.009%

    No Known Activations