© 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. 128323
    Prev
    Next
    INDEX
    Explanations

    The neuron selectively fires on numeric tokens (multi‐digit numbers and codes).

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

    this weekend

    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
     jūs
    -1.10
    *}[
    -1.04
    MatInputModule
    -0.98
    írit
    -0.97
    waukee
    -0.93
    失败
    -0.93
    '@
    -0.93
    "=>$
    -0.93
    の作成
    -0.90
    ]#
    -0.90
    POSITIVE LOGITS
     Karya
    0.97
    прочем
    0.94
     algas
    0.93
     mascarilla
    0.93
    相关文章
    0.91
    üks
    0.90
     caña
    0.88
    źć
    0.88
     Tempat
    0.87
     annonces
    0.87
    Activations Density 0.002%

    No Known Activations