© 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. 26443
    Prev
    Next
    INDEX
    Explanations

    delimiting punctuation

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

    This neuron fires on the opening double‐quote that begins a case name in a legal “Recommended Citation” (i.e. the quoted case‐title in judicial citations).

    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
     have
    -1.12
    tuigen
    -1.06
     it
    -1.00
     which
    -0.97
    复制代码
    -0.96
     their
    -0.95
     there
    -0.90
    Kamera
    -0.89
     stund
    -0.89
     at
    -0.89
    POSITIVE LOGITS
     warmen
    0.98
     абсолютно
    0.94
     geda
    0.92
    IENTS
    0.92
    tember
    0.91
    amples
    0.91
     weichen
    0.90
     pescoço
    0.90
     extiende
    0.90
    يي
    0.89
    Activations Density 0.001%

    No Known Activations