© 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. 10-GEMMASCOPE-RES-131K
    7. 40378
    Prev
    Next
    INDEX
    Explanations

    professions and group identifiers

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

    The neuron is tuned to technical, domain-specific keywords—especially programming and code identifiers (like jQuery, CSS, input, toggle, Meteor, PHP)—marking code snippets and technical jargon.

    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
    Breite
    -1.57
     идея
    -1.56
     when
    -1.55
    嗶
    -1.50
    Anmeldung
    -1.45
     लोगों
    -1.44
    ⼈
    -1.43
    ケモン
    -1.41
    条款
    -1.41
    Viernes
    -1.40
    POSITIVE LOGITS
    {
    1.52
     प्रोड
    1.50
    둠
    1.46
     Anfitrión
    1.44
     This
    1.39
    といけない
    1.38
    ğlık
    1.34
     fournisse
    1.30
    同样
    1.30
     DOLLARS
    1.29
    Activations Density 36.135%

    No Known Activations