© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129751
    Prev
    Next
    INDEX
    Explanations

    The neuron detects question phrasing—especially interrogative words (who, which, etc.) and polite request markers like “please.”

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

    examples of questions and explanations

    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
    accum
    -0.91
     női
    -0.87
     taraf
    -0.84
    Reconnect
    -0.83
    intermediate
    -0.83
    рова
    -0.82
     maîtres
    -0.82
     šaty
    -0.82
     πρ
    -0.82
    )}_
    -0.81
    POSITIVE LOGITS
    而非
    0.91
     tables
    0.84
     ::=
    0.83
    qty
    0.82
     officers
    0.81
    lijden
    0.81
    entet
    0.80
    ({"
    0.80
    ?,?,
    0.80
     their
    0.79
    Activations Density 0.125%

    No Known Activations