© 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. 64655
    Prev
    Next
    INDEX
    Explanations

    foreign characters and short words

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

    This neuron activates on opening parentheses that introduce alternate names, numbers, or non-Latin script forms (i.e. the “(” marking a parenthetical name, date, or original-script term).

    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
     paillettes
    -1.11
     czuj
    -1.07
     entweder
    -1.01
     oleju
    -0.98
     tył
    -0.95
    预计
    -0.94
     Même
    -0.94
     remplace
    -0.94
     potong
    -0.92
    卝
    -0.91
    POSITIVE LOGITS
     from
    1.38
     after
    1.23
     in
    1.18
     without
    1.12
     to
    1.05
     or
    1.05
     on
    1.04
     through
    1.01
     because
    1.00
     by
    1.00
    Activations Density 0.022%

    No Known Activations