© 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. 128192
    Prev
    Next
    INDEX
    Explanations

    The neuron is primarily detecting sentence-terminal punctuation (periods and question marks).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    punctuation followed by word

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     and
    -0.96
    wählen
    -0.94
    llingwood
    -0.87
     BEAUTIFUL
    -0.84
    licitations
    -0.83
     ultimatum
    -0.82
     schließlich
    -0.81
    موس
    -0.81
    — 
    -0.81
     техники
    -0.81
    POSITIVE LOGITS
     يمكن
    0.87
    %
    0.86
    !
    0.82
    urtstag
    0.80
     Už
    0.78
    amata
    0.78
    çadas
    0.78
     Nordrhein
    0.76
     herzlich
    0.76
     might
    0.76
    Activations Density 0.424%

    No Known Activations