© 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. 22-GEMMASCOPE-RES-131K
    7. 129102
    Prev
    Next
    INDEX
    Explanations

    The main thing this neuron does is detect numerical tokens (digits or numeric literals) in the text.

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

    objects and related entities

    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
    3
    -1.03
     serupa
    -0.93
    нт
    -0.90
     Worldwide
    -0.90
     intermédiaire
    -0.88
     Enhanced
    -0.88
    Worldwide
    -0.87
    装置
    -0.87
    Episodes
    -0.87
    Ngay
    -0.85
    POSITIVE LOGITS
    ༼
    1.06
     нарушения
    1.02
     who
    1.01
    !=-
    0.98
     pohy
    0.98
    ==-
    0.94
     comien
    0.94
    駐車
    0.93
     аллер
    0.93
     there
    0.92
    Activations Density 0.042%

    No Known Activations