© 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. 10-GEMMASCOPE-RES-131K
    7. 85449
    Prev
    Next
    INDEX
    Explanations

    grace, graceful, or gracefulness

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

    The neuron strongly activates on the substring “Grace” (in names like Grace Kelly or Grace Wilson, and in words like graceful), i.e. anything containing “Grace.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    紕
    -2.92
     dostup
    -2.80
    柺
    -2.67
     Menschheit
    -2.66
     stavu
    -2.56
    塒
    -2.50
    riezmann
    -2.48
    </strong>
    -2.47
     muestran
    -2.45
    CUSSION
    -2.41
    POSITIVE LOGITS
    A
    3.22
    2
    3.00
    This
    2.92
    的
    2.77
     ihre
    2.61
    If
    2.59
    ]
    2.58
    ：
    2.58
     haar
    2.56
     hennes
    2.56
    Activations Density 0.005%

    No Known Activations