© 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. 33550
    Prev
    Next
    INDEX
    Explanations

    names of places and things

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

    The neuron fires strongly on the initial capitalized tokens of multi-word article or section titles (i.e. the first few words of a proper-name heading).

    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
     цвят
    -1.27
     държа
    -1.08
     bienvenida
    -1.04
     salonu
    -1.00
    IPMENT
    -0.99
     dourado
    -0.99
     перец
    -0.99
     кръ
    -0.98
    Gia
    -0.98
    niei
    -0.98
    POSITIVE LOGITS
     setTotal
    1.06
    1.02
    no
    1.00
     $("<
    0.97
     iota
    0.97
    ara
    0.95
     prosedur
    0.92
    うわ
    0.92
    ad
    0.91
    Enum
    0.91
    Activations Density 0.359%

    No Known Activations