© 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. 128909
    Prev
    Next
    INDEX
    Explanations

    This neuron detects emphatic, excited punctuation and interjections (e.g. repeated exclamation marks, emoticons, “Wow,” “Amazing,” all-caps emphasis).

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

    interjections and transitions

    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
     dziew
    -1.01
    ном
    -0.99
     additional
    -0.94
     сентября
    -0.92
    ржа
    -0.91
    ='',
    -0.87
     alternatives
    -0.86
     up
    -0.85
    phem
    -0.84
    modific
    -0.84
    POSITIVE LOGITS
     digress
    1.63
     Anyway
    1.45
    Anyway
    1.39
     Anyways
    1.30
    さて
    1.05
    Anyways
    1.03
    anyway
    0.90
     comunque
    0.88
     anyways
    0.87
     winch
    0.83
    Activations Density 0.056%

    No Known Activations