© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 36422
    Prev
    Next
    INDEX
    Explanations

    questions and exclamations

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

    The neuron spikes on content words whenever the text is asking a question (especially rhetorical or surprised questions).

    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
     qualquer
    -0.94
    stdc
    -0.92
     podla
    -0.88
     reprezenta
    -0.87
     alunos
    -0.87
    ”—
    -0.85
    Istanbul
    -0.83
     nieco
    -0.82
     فنا
    -0.82
    ANTLR
    -0.82
    POSITIVE LOGITS
    ?
    1.68
    ?!
    1.19
     still
    1.09
     does
    1.02
     Apparently
    1.02
     what
    1.01
     that
    1.00
    !!!
    1.00
     massively
    0.98
    ?"
    0.97
    Activations Density 0.010%

    No Known Activations