© 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. 14638
    Prev
    Next
    INDEX
    Explanations

    wrote: and email structures

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

    The neuron is selectively activating on mentions of writing actions—e.g. the verbs “wrote” (in email/forum quotes) and “Tweet.”

    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
     to
    -2.38
    providedIn
    -1.86
     ות
    -1.84
    цією
    -1.82
     febrero
    -1.70
     bonos
    -1.70
    webtoken
    -1.69
    واده
    -1.69
     gardien
    -1.65
     Saltar
    -1.63
    POSITIVE LOGITS
    ↵
    2.30
    .
    1.81
    lo
    1.73
     Wirkungen
    1.71
     incidencia
    1.71
    на
    1.69
    的確
    1.67
    がありますが
    1.63
     geomé
    1.62
     claramente
    1.61
    Activations Density 0.000%

    No Known Activations