© 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. 22-GEMMASCOPE-RES-131K
    7. 13996
    Prev
    Next
    INDEX
    Explanations

    surprise and expressions

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

    The neuron fires on numeric tokens—i.e. constants or standalone numbers in formulas, code, or math expressions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     if
    -1.97
     for
    -1.92
     with
    -1.73
     when
    -1.68
     as
    -1.66
    ,
    -1.64
     on
    -1.62
     because
    -1.61
     during
    -1.59
     by
    -1.58
    POSITIVE LOGITS
    擽
    2.13
    ビックリ
    2.02
    びっくり
    1.99
    感じが
    1.95
    の方も
    1.90
    マジで
    1.82
    様な
    1.80
    きち
    1.79
    色々と
    1.72
    すっ
    1.72
    Activations Density 0.648%

    No Known Activations