© 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. 25516
    Prev
    Next
    INDEX
    Explanations

    mathematical variables and indices

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

    The neuron selectively activates on tokens containing digits (e.g. numeric subscripts, exponents, or standalone numbers) within mathematical 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
     not
    -1.30
     additional
    -1.13
    および
    -1.12
     barely
    -1.11
     will
    -1.10
     would
    -1.08
     almost
    -1.07
     even
    -1.06
     in
    -1.05
    ﾛ
    -1.05
    POSITIVE LOGITS
     nélk
    1.36
     körül
    1.34
     indivíduos
    1.29
    の準備
    1.27
     superbes
    1.25
     différences
    1.24
     magnifiques
    1.24
     decisões
    1.23
     incroy
    1.23
     of
    1.23
    Activations Density 0.023%

    No Known Activations