© 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. 101339
    Prev
    Next
    INDEX
    Explanations

    Herbert calls

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

    The neuron is primarily detecting numeric tokens or short alphanumeric identifiers (e.g. reference numbers, class or module IDs) in the text.

    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
     colocación
    -1.49
     not
    -1.48
     Developmental
    -1.41
     same
    -1.37
     transparencia
    -1.36
     now
    -1.35
     recomendación
    -1.34
     uzav
    -1.33
     agujeros
    -1.30
     actualización
    -1.28
    POSITIVE LOGITS
    ……」
    1.48
    ículo
    1.45
    чки
    1.42
    ulfate
    1.41
     lamentable
    1.41
     Wirken
    1.40
     metic
    1.36
    EReference
    1.35
    ーズン
    1.33
    TintMode
    1.33
    Activations Density 0.001%

    No Known Activations