© 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. 128366
    Prev
    Next
    INDEX
    Explanations

    The neuron fires primarily on numeral tokens (i.e. numeric literals and standalone numbers).

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

    Java DAO and UI screen

    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
    čnosti
    -1.14
     now
    -1.13
    Nhưng
    -1.07
     alimente
    -1.05
    recen
    -1.03
     impresionantes
    -1.02
     étui
    -1.02
    膩
    -1.02
    camisetas
    -1.02
     völ
    -1.02
    POSITIVE LOGITS
     such
    1.67
     такую
    1.35
     this
    1.11
     impli
    1.03
     AHEAD
    0.95
     extrem
    0.94
     certi
    0.92
    '])){
    0.91
     такой
    0.91
     ACTIV
    0.90
    Activations Density 0.013%

    No Known Activations